Bladder cancer (BC) is the most common malignancy of the urinary tract. According to the American Cancer Society, 79,030 new cases of bladder cancer and 18,540 deaths were estimated to occur in the United States alone in 2017 (Siegel et al., 2017). Predominantly of urothelial histology, invasive BC arises from non-invasive papillary or flat precursors, and many BC patients suffer multiple relapses prior to progression, providing ample lead-time for early detection and treatment prior to metastasis (Netto, 2013).

Although most urothelial carcinomas arise in the bladder, 5–10% originate in the renal pelvis and/or ureter (Rouprêt et al., 2015; Soria et al., 2017). The annual incidence of these upper tract urothelial carcinomas (UTUC) in Western countries is 1–2 cases per 100,000 (Rouprêt et al., 2015; Soria et al., 2017), but the disease occurs at a much higher rate in populations exposed to aristolochic acid (AA) (Chen et al., 2012; Grollman, 2013; Lai et al., 2010; Taiwan Cancer Registry, 2017). AA is a carcinogenic and nephrotoxic nitrophenanthrene carboxylic acid produced by Aristolochia plants (National Toxicology Program, 2011). An etiological link between AA exposure and UTUC has been established in several populations (Grollman, 2013; Grollman et al., 2009; Jelakovic et al., 2012; National Toxicology Program, 2011). The profound public health threat posed by the medicinal use of Aristolochia herbs is illustrated in Taiwan, which has the highest incidence of UTUC in the world (Chen et al., 2012; Yang et al., 2002). In recent years, more than one-third of the adult population in Taiwan has been prescribed herbal remedies containing AA (Hsieh et al., 2008), resulting in an unusually high (37%) proportion of UTUC cases relative to urothelial cancers worldwide (Taiwan Cancer Registry, 2017).

Tumors of the upper and lower urinary tracts differ in important ways, including etiology, but they have many common features (Green et al., 2013), such as the somatic alterations that drive their growth (Lee et al., 2018). High rates of activating mutations in the upstream promoter of the TERT gene are found in the majority of urothelial neoplasms of both upper and lower tracts (Huang et al., 2013; Killela et al., 2013; Scott et al., 2014; Yuan et al., 2016). TERT promoter mutations predominantly affect two hot spots, g.1295228 C > T and g.1295250 C > T. These mutations generate CCGGAA/T or GGAA/T motifs that alter the binding site for ETS transcription factors and subsequently stimulate increased TERT promoter activity (Horn et al., 2013; Huang et al., 2013). TERT promoter mutations occur in up to 80% of invasive urothelial carcinomas of the bladder and in several of their histologic variants (Allory et al., 2014; Cowan et al., 2016; Killela et al., 2013; Kinde et al., 2013; Nguyen et al., 2016). Moreover, TERT promoter mutations occur in 60–80% of BC precursors, including papillary urothelial neoplasms of low malignant potential (Rodriguez Pena et al., 2017), non-invasive low-grade papillary urothelial carcinoma, non-invasive high-grade papillary urothelial carcinoma and ‘flat’ carcinoma in situ (CIS), as well as in urinary cells from a subset of these patients (Kinde et al., 2013). TERT promoter mutations have thus been established as a common genetic alteration in urothelial neoplasms (Cheng et al., 2017; Killela et al., 2013; Kinde et al., 2013; Yuan et al., 2016).

Other important oncogene-activating mutations include those in FGFR3, RAS and PIK3CA genes, which occur in a high fraction of non-muscle invasive bladder cancers (Humphrey et al., 2016; Netto, 2011). In muscle-invasive bladder cancers, mutations in the TP53, CDKN2A, MLL and ERBB2 genes are also frequently found (Cancer Genome Atlas Research Network, 2014; Lin et al., 2010; Mo et al., 2007; Netto, 2011; Sarkis et al., 1995; Sarkis et al., 1994; Sarkis et al., 1993; Wu, 2005). Mutations in these genes are also present in UTUC (Hoang et al., 2013; Lee et al., 2018; Moss et al., 2017; Sfakianos et al., 2015).

Urine cytology is a non-invasive method for the detection of BC. Although it has value for the detection of high-grade BC, the test is unable to detect the vast majority of low-grade tumors (Lotan and Roehrborn, 2003; Netto and Tafe, 2016; Zhang et al., 2016). Urine cytology also fails to detect the majority of UTUCs (Baard et al., 2017). These facts, together with the high cost and invasive nature of repeated endoscopy and follow-up biopsy procedures, have led to many attempts to develop alternative minimally invasive methods to detect urothelial cancers. Strategies to identify BC include urine- or serum-based genetic and protein assays for screening and surveillance (Allory et al., 2014; Bansal et al., 2014; Ellinger et al., 2015; Fradet and Lockhard, 1997; Goodison et al., 2012; Hurst et al., 2014; Kawauchi et al., 2009; Kinde et al., 2013; Krüger et al., 2003; Moonen et al., 2007; Ralla et al., 2014; Sarosdy et al., 2006; Serizawa et al., 2011; Skacel et al., 2003; Wang et al., 2014; Yafi et al., 2015). Currently available U.S. Food and Drug Administration (FDA) approved assays for BC diagnosis include the ImmunoCyt test (Scimedx Corp), the nuclear matrix protein 22 (NMP22) immunoassay test (Matritech), and multitarget FISH (UroVysion) (Fradet and Lockhard, 1997; Kawauchi et al., 2009; Krüger et al., 2003; Moonen et al., 2007; Sarosdy et al., 2006; Skacel et al., 2003; Yafi et al., 2015). Sensitivities between 62% and 69% and specificities between 79% and 89% have been reported for some of these tests. However, due to assay performance inconsistencies, cost or requirements for technical expertise, integration of such assays into routine clinical practice has not yet occurred. Furthermore, none of the FDA-approved BC assays have yet been validated for clinical use in detection of UTUC. Therefore, a non-invasive test that predicts which patients are most likely to develop urothelial cancer could be medically and economically important.

As urothelial cells from the upper and lower urinary tracts are in direct contact with urine, we hypothesized that genetic analyses of exfoliated urinary cells could be used to detect neoplasia in these organs in a non-invasive fashion. The current study assesses the performance of a massively parallel sequencing-based assay, termed UroSEEK, for the detection of BC and UTUC through the genetic analysis of urinary cell DNA. UroSEEK has three components: detection of intragenic mutations in regions of ten genes (FGFR3, TP53, CDKN2A, ERBB2, HRAS, KRAS, PIK3CA, MET, VHL and MLL) frequently mutated in urothelial tumors (Cancer Genome Atlas Research Network, 2014; Hoang et al., 2013; Lee et al., 2018; Lin et al., 2010; Mo et al., 2007; Moss et al., 2017; Netto, 2011; Sarkis et al., 1995; Sarkis et al., 1994; Sarkis et al., 1993; Sfakianos et al., 2015; Wu, 2005); detection of mutations in the TERT promoter (Huang et al., 2013; Killela et al., 2013; Scott et al., 2014; Yuan et al., 2016); and detection of aneuploidy (Kinde et al., 2012; Kinde et al., 2011). The exons and specific genes chosen for inclusion in UroSEEK were chosen on the basis of BC mutations recorded in the COSMIC database (Supplementary file 1). Selected amplicons from VHL and MET were also included in the hope that renal cell carcinomas might also shed cells into urine, although urine samples from patients with these cancers were not included in our study.

UroSEEK was applied to three independent cohorts of patients. The first (called the BC early detection cohort) exhibited microscopic hematuria or dysuria and supplied urine samples prior to any surgical procedures. A small fraction (4% to 5%) of patients with microscopic hematuria later develops urothelial malignancy (Wein et al., 2012; Mishriki et al., 2008), so the decision as to which patients should undergo cystoscopy is often difficult. The second cohort (called the UTUC cohort) consisted of Taiwanese patients with UTUC who supplied a urine sample prior to nephroureterectomy. Such patients might benefit from a non-invasive test that could be used to screen individuals at increased risk for UTUC, such as those exposed to herbal remedies containing the carcinogen AA. The third cohort (called the BC surveillance cohort) included patients who had already been diagnosed with BC and were therefore at high risk for recurrence (Wein et al., 2012). Because urine cytology is relatively insensitive for the detection of recurrence, cystoscopies are performed as often as every three months in such patients in the U.S. In fact, the cost of managing these patients is in aggregate higher than the cost of managing any other type of cancer, amounting to 3 billion dollars annually (Netto and Epstein, 2010).

A schematic of the approach used in this study is provided in Figure 1. A flow diagram indicating the number of patients evaluated in this study and the major results are presented in Figure 2.

Figure 1

Download asset Open asset

Figure 2

Download asset Open asset

BC early detection cohort

Cohort characteristics

A total of 570 patients were included in the early detection cohort, each with one urine sample analyzed. 90% of the patients had hematuria, 3% had lower urinary tract symptoms (LUTS), and 9% had other indications suggesting they were at risk for BC. The median age of the participants was 58 years (range 5 to 89 years; Table 1a). As expected from prior studies of patients at risk for BC, 70% of the patients were male (Siegel et al., 2017; Wein et al., 2012). Patients (n = 175; 31%) developed BC after a median follow-up period of 18 months (range 0 to 40 months). For each patient who developed BC, we selected two other patients who presented with similar symptoms but did not develop BC during the follow-up period. By design, the fraction of cases in this cohort developing BC was higher than the fraction (5%) of patients with similar presentations who would have developed BC in standard clinical practice. The characteristics of the tumors developing in the 570 patients are summarized in Table 1a and detailed in Supplementary file 2.

Table 1

Gender	n	%	Ten-gene multiplex positive	TERT positive	Aneuploidy positive	UroSEEK positive	Cytology positive*	Uroseek or cytology positive*
Table 1a. Demographic, clinical and genetic features of the early detection cohort
Males without recurrence	172	59%	3 (2%)	10 (6%)	2 (1%)	13 (8%)	0 (0%)	13 (8%)
Males with recurrence	32	11%	26 (81%)	21 (66%)	19 (59%)	29 (91%)	16 (50%)	30 (94%)
Females without recurrence	81	28%	2 (2%)	2 (2%)	1 (1%)	5 (6%)	0 (0%)	5 (6%)
Females with recurrence	9	3%	4 (44%)	4 (44%)	3 (33%)	6 (67%)	1 (11%)	6 (67%)
Indication
Hematuria without recurrence	346	61%	6 (2%)	15 (4%)	5 (1%)	22 (6%)	0 (0%)	17 (5%)
Hematuria with recurrence	163	29%	108 (66%)	90 (55%)	76 (47%)	134 (82%)	18 (11%)	32 (2%)
LUTS without recurrence	11	2%	0 (0%)	2 (18%)	0 (0%)	2 (18%)	0 (0%)	2 (18%)
LUTS with recurrence	3	1%	2 (67%)	1 (33%)	0 (0%)	2 (67%)	1 (33%)	2 (67%)
Other without recurrence	38	7%	1 (3%)	0 (0%)	1 (3%)	2 (5%)	0 (0%)	2 (5%)
Other with recurrence	9	2%	9 (100%)	8 (89%)	5 (56%)	9 (100%)	2 (22%)	9 (100%)
Detected Tumor Diagnosis
PUNLMP	2	1%	0 (0%)	1 (50%)	0 (0%)	1 (50%)	0 (0%)	0 (0%)
CIS	7	5%	4 (57%)	4 (57%)	1 (14%)	6 (86%)	3 (43%)	6 (86%)
LGTCC	31	21%	15 (48%)	18 (58%)	9 (29%)	22 (71%)	0 (0%)	4 (13%)
HGTCC	49	33%	34 (69%)	28 (57%)	26 (53%)	40 (82%)	4 (8%)	11 (22%)
INTCC	61	41%	48 (79%)	36 (59%)	35 (57%)	57 (93%)	9 (15%)	16 (26%)
Cytology diagnosis*
Positive	21	6%	16 (76%)	12 (57%)	16 (76%)	20 (95%)	N/A	N/A
Atypical	105	30%	21 (20%)	21 (30%)	12 (11%)	30 (29%)	N/A	N/A
Negative	221	64%	4 (2%)	9 (4%)	1 (0.4%)	12 (5%)	N/A	N/A
Table 1b. Demographic, clinical and genetic features of the Surveillance cohort.
Males without recurrence	59	30%	3 (5%)	8 (14%)	3 (5%)	10 (17%)	0 (0%)	8 (14%)
Males with recurrence	90	45%	45 (50%)	53 (59%)	20 (22%)	59 (66%)	20 (22%)	53 (59%)
Females without recurrence	17	9%	5 (29%)	3 (18%)	0 (0%)	6 (35%)	0 (0%)	6 (35%)
Females with recurrence	33	17%	15 (45%)	19 (58%)	11 (33%)	33 (100%)	6 (18%)	19 (58%)
Original Tumor Diagnosis
PUNLMP	12	4%	5 (42%)	2 (17%)	1 (8%)	6 (50%)	0 (0%)	2 (17%)
CIS	25	8%	11 (44%)	13 (52%)	6 (24%)	14 (56%)	5 (20%)	10 (40%)
LGTCC	107	35%	27 (25%)	34 (32%)	8 (7%)	41 (38%)	0 (0%)	59 (55%)
HGTCC	62	20%	22 (36%)	24 (39%)	10 (16%)	30 (49%)	4 (7%)	16 (26%)
INTCC	104	34%	39 (38%)	47 (45%)	29 (28%)	54 (52%)	20 (19%)	34 (33%)
Original Tumor Stage
pTis	25	8%	11 (44%)	13 (52%)	6 (24%)	14 (56%)	5 (20%)	10 (40%)
pTa	181	58%	54 (30%)	60 (33%)	19 (19%)	77 (43%)	4 (2%)	77 (43%)
pT1	71	23%	28 (39%)	35 (49%)	22 (31%)	39 (55%)	14 (20%)	23 (32%)
pT2	23	7%	9 (9%)	9 (39%)	7 (30%)	12 (52%)	5 (22%)	10 (43%)
pT3	9	3%	1 (11%)	2 (22%)	0	2 (22%)	1 (11%)	1 (11%)
pT4	1	0.3%	1 (100%)	1 (100%)	0	1 (100%)	N/A	N/A
Routine cytology diagnosis*
Positive	30	15%	21 (21%)	25 (83%)	20 (67%)	27 (90%)	N/A	N/A
Atypical	95	48%	38 (40%)	43 (45%)	18 (19%)	50 (53%)	N/A	N/A
Negative	71	36%	12 (17%)	13 (18%)	3 (4%)	19 (27%)	N/A	N/A

1. *Cytology was available on only a subset of cases.

   N/A Not Available.

Genetic analysis

We performed three separate tests for genetic abnormalities that might be found in urinary cells derived from BC (Figure 1). First, we evaluated mutations in selected regions of ten genes that have been shown to be frequently altered in urothelial tumors (Figure 3 and Supplementary file 3). For this purpose, we designed a specific set of primers that allowed us to detect mutations in as few as 0.03% of urinary cells (Supplementary file 4). The capacity to detect such low-mutant fractions was a result of the incorporation of molecular barcodes in each of the primers, thereby substantially reducing the artifacts associated with massively parallel sequencing (Kinde et al., 2011). Second, we evaluated TERT promoter mutations. A singleplex PCR was used for this analysis because the unusually high GC-content of the TERT promoter precluded its inclusion in the multiplex PCR design. Third, we evaluated the extent of aneuploidy using a technique in which a single PCR is used to co-amplify ~38,000 members of a subfamily of long interspersed nucleotide element-1 retrotransposons (L1 retrotransposons, also called LINEs). L1 retrotransposons, like other human repeats, have spread throughout the genome via retrotransposition and are found on all 39 non-acrocentric autosomal arms (Kinde et al., 2012).

Figure 3

Download asset Open asset

The multiplex assay detected mutations in 68% of the 175 urinary cell samples from the individuals who developed BC during the course of this study (95% CI, 61% to 75%) (Table 1a and Supplementary file 5). A total of 246 mutations were detected in eight of the ten target genes (Figure 3 and Supplementary file 5). The median mutant allele frequency (MAF) in the urinary cells with detectable mutations was 8% (8.14%. The most commonly altered genes were TP53 (45% of the total mutations) and FGFR3 (20% of the total mutations; Figure 3). The distribution of mutant genes was roughly consistent with expectations based on previous exome-wide sequencing studies of BCs (Cancer Genome Atlas Research Network, 2014). At the thresholds used, 1.7% of the 395 patients in the early detection cohort who did not develop BC during the course of the study had a detectable mutation in any of the ten genes. At the same thresholds, none of the 188 urinary cell samples from healthy individuals had a mutation in any of the ten genes assayed (100% specificity, 95% CI, 98% to 100%).

Mutations in the TERT promoter were detected in 57% of the 175 urinary cell samples from the patients who developed cancer during the study interval (95% CI, 49% to 64%; Table 1a and Supplementary file 6). The median TERT MAF in the urinary cells was 6% (5.76%). Mutations were detected in three positions; 98% of the mutations were at TERT:g.1295228 (79%) and TERT:g.1295250 (19%), which are 66 and 88 bp upstream of the transcription start site, respectively. The remaining 2% of mutations were found at TERT:g.1295242. The first two of these positions have been previously shown to be critical for the appropriate transcriptional regulation of TERT. In particular, the mutant alleles recruit the GABPA/B1 transcription factor, resulting in the H3K4me2/3 mark of active chromatin and reversing the epigenetic silencing present in normal cells (Stern et al., 2015). Of the 395 patients in this cohort who did not develop BC during the course of the study, only 4% had a detectable mutation in the TERT promoter. Finally, only one of the 188 urinary samples from healthy individuals harbored a TERT promoter mutation.

Aneuploidy was detected in 46% (95% CI, 39% to 54%) of the 175 urinary cell samples from the patients who developed BC during the course of the study (Table 1a and Supplementary file 7). The most commonly altered chromosome arms were 5q, 8q, and 9p. These three chromosome arms harbor well-known oncogenes and tumor suppressor genes that have been shown to undergo copy number alterations in many cancers, including BC (Vogelstein et al., 2013). Aneuploidy was detected in 1.5% of the urinary cell samples from the 395 patients who did not develop BC during the course of the study, but it was not detected in any of the 188 urinary samples from healthy individuals.

Comparison with primary tumors

DNAs from resected or biopsied tumor samples from 102 of the patients enrolled in the BC early detection cohort were available for comparison and were examined with the same three assays used to probe the urinary cell samples. In 91 (89%) of these 102 cancers, at least one mutation in the eleven genes studied was present (in the 10-gene panel or the TERT promoter). Moreover, at least one of the mutations identified in the urine samples from these 102 patients was also identified in 83% of the corresponding primary BC samples (Supplementary files 5 and 6).

Analysis of these tumors also shed light on the basis for ‘false negatives,’ the urine samples with no detectable mutations from patients who ultimately developed BC. We attributed false negatives to the possibility that the corresponding BC either did not harbor a mutation in any of these 11 genes or the fraction of neoplastic cells in the urine sample was insufficient to allow detection with the assays used. We identified a mutation in at least one of the 11 genes in 62% of the primary tumors from patients with false negative urine tests (Supplementary file 3 and 8). We concluded that 62% of the 29 false negative tests were due to insufficient cancer cells in the urine while the remaining 38% were due to the absence of any of the queried mutations in the primary tumor tissue.

UroSEEK: biomarkers in combination

The ten-gene multiplex assay, the TERT singleplex assay, and the aneuploidy assays yielded 68%, 57%, and 46% sensitivities, respectively, when used separately (Table 1a and Supplementary files 5, 6, and 7). Sensitivity was increased when the three assays were performed on each urine cell sample. In samples without TERT promoter mutations (n = 45), mutations in one of the other ten genes were detected (Figure 4 and Supplementary file 5). Conversely, 35 samples negative for mutations in the multiplex assay were detected by virtue of TERT promoter mutations (Figure 4 and Supplementary file 6). Finally, ten of the urinary cell samples without any detectable mutations in the 11 genes were positive for aneuploidy (Figure 4 and Supplementary file 7). Thus, when the three assays were used together (test termed ‘UroSEEK’), and a positive result in any one of the assays was sufficient to score a sample as positive, the sensitivity rose to 83% (95% CI, 76% to 88%). Only one of the 188 samples from healthy individuals was scored positive by UroSEEK (specificity 99.5%, CI 97% to 100%). Twenty-six (6.5%) of the 395 patients in the BC early detection cohort who did not develop BC during the course of the study scored positive by the UroSEEK test (specificity 93%, CI 91% to 96%). On average, UroSEEK positivity preceded the diagnosis of BC by 2.3 months, and in eight cases, by >one year (Figure 5 and Supplementary file 2).

Figure 4

Download asset Open asset

Figure 5

Download asset Open asset

UroSEEK plus cytology

As both cytology and UroSEEK are non-invasive tests and can be performed on the same urine sample, we assessed their performance in combination. Cytology was available for 347 patients in the BC early detection cohort (Table 1a and Supplementary file 2). Among the 40 patients who developed biopsy-proven cancer in this cohort, cytology was positive in 17 cases (43% sensitivity), and UroSEEK was positive in 100% of these cancer patients. UroSEEK was also positive in 95% of 23 cancer patients whose urines were negative by cytology. Thus, in combination, UroSEEK plus cytology achieved 95% (95% CI, 83% to 99%) sensitivity, a 12% increase over UroSEEK and a 52% increase over cytology. Finally, none of the 299 patients who did not develop cancer over the course of the study were positive by cytology (100% specificity), but 20 (6.6%) were positive by UroSEEK, giving the combination of UroSEEK and cytology a specificity of 93% (95% CI, 90% to 96%).

UTUC cohort

Cohort characteristics

The gender distribution of this cohort, 32 females and 24 males, is atypical of UTUC patients in Western countries where males predominate (Shariat et al., 2011), but is consistent with earlier epidemiologic studies of Taiwanese individuals with known exposures to AA (summary in Table 2; individual data in Supplementary file 9)(Chen et al., 2012). Tobacco use was reported by 18% in this cohort and were all males. Based on estimated glomerular filtration rate (eGFR) values, renal function was unimpaired (chronic kidney disease (CKD) stage 0–2) in 45% of the subjects, while mild-to-moderate renal disease (CKD stage 3) or severe disease (CKD stages 4–5) was noted for 43% and 12% of the cohort, respectively (Table 2). Tumors were confined to a single site along the upper urinary tract in the majority of cases (38% renal pelvis; 39% ureter), while multifocal tumors affecting both renal pelvis and ureter occurred in 23% of the patients. Synchronous bladder cancer (diagnosed within 3 months prior to nephroureterectomy) was present in 38% of patients. Tumors were classified as high grade in 89% of the cases, with the majority categorized as muscle-invasive (T2-T4, 66%; Table 2).

Table 2

	N	%	Ten-gene multiplex positive	TERT positive	Aneuploidy positive	UroSEEK positive
All subjects	56	100%	64%	29%	39%	75%
Gender
Males	24	43%	71%	33%	54%	83%
Females	32	57%	59%	25%	28%	69%
CKD stage
0–2	25	45%	68%	36%	44%	76%
3A	14	25%	50%	21%	43%	71%
3B	10	18%	80%	20%	40%	80%
4	4	7%	25%	50%	0%	50%
5	3	5%	100%	0%	33%	100%
Tumor grade
Low	6	11%	67%	50%	17%	67%
High	50	89%	64%	26%	42%	76%
Tumor stage
Ta	11	20%	73%	55%	45%	82%
T1	8	14%	50%	0%	38%	75%
T2	10	18%	80%	20%	10%	80%
T3	24	43%	67%	33%	54%	79%
T4	3	5%	0%	0%	0%	0%
Upper urinary tract tumor site
Lower ureter	17	30%	76%	18%	35%	76%
Upper ureter	1	2%	100%	0%	0%	100%
Ureterovesical junction	2	4%	0%	0%	0%	0%
Lower ureter and upper ureter	2	4%	100%	50%	50%	100%
Renal pelvis	21	38%	57%	38%	38%	76%
Renal pelvis and lower ureter	4	7%	75%	25%	50%	100%
Renal pelvis and upper ureter	5	9%	40%	40%	60%	60%
Renal pelvis, lower ureter, upper ureter	4	7%	75%	25%	50%	75%
Synchronous bladder cancer
Present	21	38%	52%	29%	33%	62%
Absent	35	63%	71%	29%	43%	83%
UTUC risk factors
Aristolactam-DNA adducts present	54	96%	65%	30%	39%	74%
Smoking history	10	18%	70%	30%	60%	70%
CKD, chronic kidney disease.

Genetic analysis

The multiplex assay detected at least one mutation in 36 of the 56 urinary cell samples from UTUC patients (64%, 95% CI, 51% to 76%; Table 2 and Supplementary file 10). A total of 57 mutations were detected in nine of the ten target genes (Figure 3). The median MAF in the urinary cells was 5.6% and ranged from 0.3% to 80%. The most commonly altered genes were TP53 (n = 33, 58% of the 57 mutations) and FGFR3 (n = 9, 16% of the 57 mutations; Figure 3).

Mutations in the TERT promoter were detected in 16 of the 56 urinary cell samples from UTUC patients (29%, 95% CI, 18% to 42%; Table 2 and Supplementary file 11). The median TERT MAF in the urinary cells was 2.22% and ranged from 0.59% to 46.3%. One of the 188 urinary samples from healthy individuals harbored a mutation (TERT g.1295250C > T with a MAF of 0.39%). In the UTUC urinary cell samples, most of the TERT mutations (94%) were at one of two positions, TERT:g.1295228 (67%) and TERT:g.1295250 (28%), which are 66 and 88 bp upstream of the transcription start site, respectively. A third position, TERT:g.1295242, was also involved in the remaining 6% of cases.

Aneuploidy was detected in 22 of the 56 urinary cell samples from UTUC patients (39%, 95% CI, 28% to 52%, Supplementary file 12 and 13). The most commonly altered chromosome arms were 1q, 7q, 8q, 17 p, and 18q.

Comparison with primary tumors

The distribution of mutant genes in primary tumors (Supplementary file 14) was consistent with findings from some (Hoang et al., 2013; Lee et al., 2018; Yuan et al., 2016) but not all (Moss et al., 2017; Sfakianos et al., 2015), exome-wide and targeted sequencing studies of UTUCs. In the present study, TP53 mutations were found only in high-grade UTUCs, while FGFR3 mutations dominated in low-grade tumors (present in 5/6). Such mutational patterns have been previously reported by others (Sfakianos et al., 2015). However, the overall frequency of FGFR3 mutations in our UTUC cohort (21%) was relatively low compared to values reported by Moss et al. (2017) (74%) and Sfakianos et al. (2015) (54%), but was comparable to values reported by Hoang et al. (2013) (8%) and Lee et al. (2018) (13%). We attribute this difference to the race/ethnicity profile of the cohorts under comparison, as FGFR3 mutation levels are relatively low in UTUCs from Han Chinese patients (3–9%) compared to Western patients (36–60%), as reported by Yuan et al. (2016). Our cohort was Taiwanese and principally of Han Chinese descent, as were the Hoang et al. (2013) (Taiwanese) and Lee et al. (2018) (Korean) cohorts, whereas Western patients were examined in the Sfakianos et al. (2015) and Moss et al. (2017) studies.

Tumor samples from all 56 patients with UTUC were available for comparison and were subjected to the same three assays used to analyze the urinary cell samples. At least one mutation could be identified in the urinary cells from 39 UTUC cases. In 35 (90%) of these 39 cases, at least one of the mutations identified in the urine sample (Supplementary file 10 and 11) was also identified in the corresponding tumor DNA sample (Supplementary file 14 and 15). When all 80 mutations identified in the urinary cells were considered, 63 (79%) were identified in the corresponding tumor sample. The discrepancies between urine and tumor samples in any of the three assays might be explained by the fact that we had access to only one tumor per patient, even though more than one anatomically distinct tumor was often evident clinically (Table 2). In addition, DNA was extracted from only one location in each tumor; thus, intratumoral heterogeneity (McGranahan et al., 2015) could have been responsible for some of the discrepancies.

The tumor data also helped to establish why 17 of the 56 urinary cell samples from UTUC patients did not contain detectable mutations. From the evaluation of the primary tumor samples, we found that only four (24%) of these 17 urine samples were from patients whose tumors did not contain any of the queried mutations (Supplementary file 14 and 15). Thus, we concluded that the main reason for failure of the mutation test was an insufficient number of cancer cells in the urine, which accounted for 13 (76%) of the 17 failures.

Aneuploidy was observed in 22 of the urinary cell samples (Supplementary file 12). Overall, 96% of the chromosomal gains or losses observed in the urinary cells were also observed in the primary tumors (Supplementary file 13). Conversely, aneuploidy was not observed in 34 of the urinary cell samples. Evaluation of the corresponding 56 tumors with the same assay demonstrated that all but three were aneuploid. Therefore, as with mutations, the main reason for failure of the aneuploidy assay was insufficient neoplastic DNA in the urinary cells.

UroSEEK: biomarkers in combination

The ten-gene multiplex assay, the TERT singleplex assay, and the aneuploidy assays yielded 64%, 29%, and 39% sensitivities, respectively, when used separately in the UTUC cohort (Table 2). Mutations in one of the other ten genes were detected in 23 samples without TERT promoter mutations (Figure 4). Conversely, three samples without detectable mutations in the multiplex assay scored positive for TERT promoter mutations (Figure 4). Furthermore, three of the urinary cell samples without any detectable mutations were positive for aneuploidy (Figure 4). Thus, when the three assays were used together, and a positive result in any one assay was sufficient to score a sample as positive, the sensitivity rose to 75% (95% CI 62.2% to 84.6%).

To determine the basis for the increased sensitivity afforded by the combination assays, we evaluated data from the primary tumors of the three patients whose urinary cell samples exhibited aneuploidy but did not harbor detectable mutations. We found that these three tumors did not contain any mutations in the 11 queried genes, which explained why these same assays were negative when applied to the urinary cell DNAs. These three tumors were aneuploid, thus enabling their detection through copy number variations in the urinary cell samples.

Correlation with clinical features

The most clinically desirable biomarkers are those associated with early stage tumor development as they enable surgical removal of lesions before widespread metastasis. In UTUC, ten-year cancer specific survival rates show that 91% of patients with stage T1 malignancies are expected to be cured by surgery, compared to only 78%, 34% and 0% of patients with stage T2, T3, or T4 tumors, respectively (Li et al., 2008). In our cohort, UroSEEK was equally sensitive for detecting early and late UTUCs. The test was positive in 15 (79%) of 19 patients with stage Ta/T1 tumors and 27 (73%) of 37 patients with stage T2-T4 tumors (Table 2). Sensitivity was comparable across gender, tumor grade, tumor location and risk factors for developing UTUC (Table 2), indicating that the assay was suitable for evaluation of diverse patient populations. UroSEEK performance was also comparable in UTUC cohorts with and without synchronous BC (Table 2).

UroSEEK was also considerably more sensitive than urine cytology in the UTUC cohort (Figures 2 and 6). Cytology was available in 42 cases, and of these, four (9.5%) were positive on cytology. UroSEEK detected all four of these cases. In addition, UroSEEK was positive in 5/7 cases that had an equivocal cytology diagnosis of suspicious for malignancy, and 22/31 samples that were negative on cytology.

Figure 6

Download asset Open asset

Aristolochic acid exposure

The activated metabolites of AA bind covalently to the exocyclic amino groups in purine bases, with a preference for dA, leading to characteristic A > T transversions (Hollstein et al., 2013; Moriya et al., 2011). To determine whether individuals in the UTUC cohort had been exposed to AA, we quantified renal cortical DNA adducts using mass spectrometry (Yun et al., 2012). All but two of the 56 patients had detectable aristolactam (AL)-DNA adducts (Table 2) with levels ranging from 0.4 to 68 dA-AL-I adducts per 10⁸ nucleotides. Moreover, the A > T signature mutation (Hoang et al., 2013) associated with AA was highly represented in the mutational spectra of TP53 (18/32 A > T) and HRAS (2/2 A > T) found in urinary cells (Supplementary file 10).

Cytology is a non-invasive test that is highly specific, and in expert hands nearly always indicates the presence of urothelial malignancy when positive. This specificity was verified in our study: all 51 patients in the BC early detection cohort whose urine samples were positive by cytology developed biopsy-proven BC. However, cytology is not particularly sensitive. UroSEEK adds considerably to sensitivity, as it raised the sensitivity of cytology from 43% to 95% in the BC early detection cohort, from 25% to 71% in the BC Surveillance cohort, and from 10% to 75% in the UTUC cohort. The increased sensitivity was further highlighted by the fact that UroSEEK-positive results preceded clinical diagnosis or positive cytology by months to years in the BC surveillance cohort.

The advantage of using UroSEEK in addition to cytology was particularly evident for low-grade tumors. Cytology was negative in all 49 cases in the BC early detection cohort, while 2/3 of these patients were positive with UroSEEK. Similarly, UroSEEK correctly identified 80% of low-grade UTUC while none were detected by cytology. Another example of the utility of the combination of UroSEEK plus cytology was evident in patients with an equivocal cytology reading. A relatively large number of urine samples receive such an equivocal cytologic reading, even in the hands of a sub-specialized, board-certified cytopathology expert such as employed in the current study (Barkan et al., 2016). In the BC early detection cohort, for example, 105 urine samples were scored as ‘atypical’, and of these cases, 19% developed recurrence while the other 81% did not. UroSEEK was positive in 95% of the atypical cases that developed BC, but only in 13% of the atypical cases that did not develop cancer. These results demonstrate that UroSEEK can be used to more confidently interpret atypical cytology results.

Although UroSEEK is more sensitive than cytology, it is less specific. In this study, we assessed specificity in several independent ways. The first, and in some ways, most straightforward, was in a collection of urine samples from healthy individuals. In these 188 individuals, only one sample was positive, yielding a specificity of 99.5% (CI 97% to 100%). Such high specificity can be considered the technical specificity of the test, but biological specificities are also important. In the BC early detection cohort, 26 of the 395 patients who did not develop BC scored positive, yielding a specificity of 93% (CI 90.50% to 96%), or a false positive rate of 6.5%. These ‘false positives’ detected by UroSEEK could result from several factors. First, we cannot be certain that the patients whose urinary cells harbored genetic alterations did not have cancer. The follow-up period for many of patients was only one year, and cystoscopy was not generally performed. Second, it is possible that there are clonal proliferations in the bladder epithelium that increase with age. The patients in the BC early detection cohort were on average older than the 188 healthy individuals used as controls (40 years vs. 58 years). Although this explanation is speculative, clonal proliferations that are not considered neoplastic have been described in the bone, skin, and other tissues (Risques and Kennedy, 2018). Clonal proliferations may also be the basis for any discordance between mutations in urinary cells and in the primary tumors of the same patients. Although in the majority of cases, at least one of the mutations identified in the urine was also present in the primary tumor, this was not true in 22% of the cases in the BC early detection cohort. In these cases, UroSEEK could be detecting clonal proliferations in the bladder epithelium that did not progress to cancer, and such proliferations may be more common in patients with BC than in the general population Risques and Kennedy, 2018Because only one biopsy from the primary tumor was available for comparison, it is also possible that intratumoral heterogeneity explains some of the discrepancies. False positives in the BC surveillance cohort could be explained in similar ways. False positives are not unique to our study; they have been observed in all other molecular assays for BC, including FDA-approved tests (Dimashkieh et al., 2013; Gopalakrishna et al., 2017; Hajdinjak, 2008). Whether the false positives in these other assays have the same biological basis is an important area for future research.

There are two factors that limit sensitivity for genetically-based biomarkers. First, a sample can only be scored as positive for the biomarker if it contains DNA from a sufficient number of neoplastic cells to be detected by the assay. Second, the tumor from which the neoplastic cells were derived must harbor the genetic alteration that is queried. Combination assays can increase sensitivity by assessing more genetic alterations, and are thereby more likely to detect at least one genetic alteration present in the tumor. However, mutations in clinical samples are often present at low allele frequencies (Supplementary files 5, 6, 10 and 11), requiring high coverage of every base queried. It would be prohibitively expensive to perform whole exome sequencing at 10,000x coverage, for example, so some compromise is needed. In our study, we evaluated carefully selected regions of 11 genes, including TERT, together with copy number analysis of 39 chromosome arms. Even if a tumor does not contain a genetic alteration in one of the 11 genes assessed, it might still be detectable by the urinary cell assay for aneuploidy. The sensitivity of aneuploidy detection however is less than that of the mutation assays. Simulations demonstrated that DNA containing a minimum of 1% neoplastic cells is required for reliable aneuploidy detection, while mutations present in as few as 0.03% of the DNA templates can be detected by the mutation assays used in our study (Bettegowda et al., 2014; Kinde et al., 2013; Wang et al., 2016). Nevertheless, urinary cell samples that had relatively high fractions of neoplastic cells but did not contain a detectable mutation in the 11 queried genes should still be detectable due to their aneuploidy. In addition, some of the mutations in the 11 genes queried, such as large insertions or deletions or complex changes, might be undetectable by mutation-based assays, but such samples might still score positive in a test for aneuploidy.

For UTUC, although the approach described here has significant potential for screening purposes, we emphasize that the current study demonstrates proof-of-principle rather than clinical applicability given the small number of patients evaluated. Another caveat is that our assays cannot distinguish between UTUC and BC. We consider this a strength of the assay since the detection of BC is equally important given that patients exposed to AA are at risk for BC as well as for UTUC (Poon et al., 2015). It has been estimated that 100 million people in China are at risk for UTUC as a result of exposure to this carcinogen (Grollman, 2013; Hu et al., 2004). Non-invasive, sensitive methods to screen the large numbers of at-risk individuals for UTUC in such populations are thus clearly desirable. UroSEEK could also be used to monitor for UTUC recurrence in bladder, which occurs in 22% to 47% of cases, or in the contralateral tract affecting 2% to 6% of patients (Rouprêt et al., 2015) and up to 30% of AA-related UTUC patients (Chen et al., 2013). Currently, no such screening methods are available, as illustrated in the current study where urine cytology failed to detect 90% of UTUC cases. Radiologic tests, such as MRI or CT-scans, are not well suited for screening, and the latter confers significant radiation exposure. Ureteroscopy is often definitive, but in addition to being invasive, requires highly skilled clinicians and is also ill-suited as a screening tool (Golan et al., 2015).

Liquid biopsy has recently gained attention as a non-invasive approach to screen for cancer. Although this concept often refers to blood samples, it can be applied to other body fluids, such as urine (Patel et al., 2017; Sidransky et al., 1991; Togneri et al., 2016). Urine contains DNA from several sources, including (i) glomerular filtration of circulating free DNA (Botezatu et al., 2000) released by normal and tumor cells from sites throughout the body; (ii) DNA released directly into urine by normal and tumor cells of the urinary tract; and (iii) intact normal or malignant cells of the urinary tract exfoliated into urine (Bettegowda et al., 2014; Dawson et al., 2013; Dressman et al., 2003; Forshew et al., 2012; Haber and Velculescu, 2014; Kinde et al., 2013; Springer et al., 2015; Vogelstein and Kinzler, 1999; Wang et al., 2015a; Wang et al., 2015b; Wang et al., 2016). We chose the latter option for the current study to increase sensitivity and specificity.

While optimizing conditions for the current study, we compared the relative performance of mutation assays in matched plasma and urine samples obtained from 14 UTUC patients. In each case, a TERT or TP53 mutation was first identified in the primary tumor. That particular mutation was subsequently queried in DNA from the urine and plasma using a singleplex assay. Mutations were detected in 93% (13/14) of the urinary cell DNA samples compared to 36% (5/14) of the plasma samples. Importantly, the plasma test failed to identify any of the six non-muscle-invasive cancers (Ta/T1), while all six (100%) were identified in the matched urinary cell DNA samples. The superior performance in urinary cells was likely due to a substantial enrichment for mutated DNA in these cells compared to plasma; the median MAF in plasma when a mutation was detectable was only 0.3% compared to 15% in the urinary cells.

Our study lays the conceptual and practical framework for a novel test that could be used in the management of patients with urothelial cancer. Large prospective trials will be required to demonstrate the ability of UroSEEK to improve the management of patients with hematuria or dysuria or patients at risk for urothelial cancer recurrence. Before carrying out large-scale trials to evaluate such clinical utility, it is informative to predict what the performance characteristics of such a test might be. As one example, consider the use of UroSEEK plus cytology in patients presenting to their physician with microscopic hematuria or dysuria, a commonly encountered situation. In large population-based studies involving over 80,000 individuals participating in health screening, the fraction of individuals with micro-hematuria ranged from 2.4% to 31.1% (Davis et al., 2012; Wein et al., 2012). It has been estimated that 5% of such patients actually have urothelial cancer (Khadra et al., 2000). In the current study, UroSEEK plus cytology had a sensitivity of 95% and a specificity of 93% for BC in patients with this presentation. These results extrapolate to a positive predictive value (PPV) of 66% (95% CI, 55% to 74%) and a negative predictive value (NPV) of 99.3% (95% CI, 97.3% to 99.8%). These values are well above those generally considered to be diagnostically helpful and are considerably higher than achieved in FDA-approved tests for this indication (Dimashkieh et al., 2013; Hajdinjak, 2008).

We envision that the first application of UroSEEK would be patients such as those described here in the cohorts used for early detection of BC and UTUC. Patients with hematuria who might otherwise be referred to cystoscopy would be tested by UroSEEK plus cytology. Such tests could be ordered by general practitioners and do not require consultation with a urologist. Only if a test was positive would cystoscopy be required. The sensitivities, specificities, and PPV and NPV of UroSEEK plus cytology suggest that this strategy is well within the boundaries of currently accepted medical guidelines. Optimistically, 95% of patients would be spared the discomfort and inconvenience of cystoscopy as well as its unintended consequences. Only patients who have positive UroSEEK, persistent symptoms, or hematuria would undergo cystoscopy. The savings in this approach would be considerable, as we estimate the cost of UroSEEK plus cytology to be less than 1/3 of the cost of cystoscopy.

1. 1. Allory Y
   2. Beukers W
   3. Sagrera A
   4. Flández M
   5. Marqués M
   6. Márquez M
   7. van der Keur KA
   8. Dyrskjot L
   9. Lurkin I
   10. Vermeij M
   11. Carrato A
   12. Lloreta J
   13. Lorente JA
   14. Carrillo-de Santa Pau E
   15. Masius RG
   16. Kogevinas M
   17. Steyerberg EW
   18. van Tilborg AA
   19. Abas C
   20. Orntoft TF
   21. Zuiverloon TC
   22. Malats N
   23. Zwarthoff EC
   24. Real FX

   (2014) Telomerase reverse transcriptase promoter mutations in bladder cancer: high frequency across stages, detection in urine, and lack of association with outcome

   European Urology 65:360–366.

   https://doi.org/10.1016/j.eururo.2013.08.052

   • PubMed
   • Google Scholar
2. 1. Baard J
   2. de Bruin DM
   3. Zondervan PJ
   4. Kamphuis G
   5. de la Rosette J
   6. Laguna MP

   (2017) Diagnostic dilemmas in patients with upper tract urothelial carcinoma

   Nature Reviews Urology 14:181–191.

   https://doi.org/10.1038/nrurol.2016.252

   • PubMed
   • Google Scholar
3. 1. Bansal N
   2. Gupta A
   3. Sankhwar SN
   4. Mahdi AA

   (2014) Low- and high-grade bladder cancer appraisal via serum-based proteomics approach

   Clinica Chimica Acta 436:97–103.

   https://doi.org/10.1016/j.cca.2014.05.012

   • PubMed
   • Google Scholar
4. 1. Barkan GA
   2. Wojcik EM
   3. Nayar R
   4. Savic-Prince S
   5. Quek ML
   6. Kurtycz DF
   7. Rosenthal DL

   (2016) The paris system for reporting urinary cytology: The quest to develop a standardized terminology

   Advances in Anatomic Pathology 23:193–201.

   https://doi.org/10.1097/PAP.0000000000000118

   • PubMed
   • Google Scholar
5. 1. Bettegowda C
   2. Sausen M
   3. Leary RJ
   4. Kinde I
   5. Wang Y
   6. Agrawal N
   7. Bartlett BR
   8. Wang H
   9. Luber B
   10. Alani RM
   11. Antonarakis ES
   12. Azad NS
   13. Bardelli A
   14. Brem H
   15. Cameron JL
   16. Lee CC
   17. Fecher LA
   18. Gallia GL
   19. Gibbs P
   20. Le D
   21. Giuntoli RL
   22. Goggins M
   23. Hogarty MD
   24. Holdhoff M
   25. Hong SM
   26. Jiao Y
   27. Juhl HH
   28. Kim JJ
   29. Siravegna G
   30. Laheru DA
   31. Lauricella C
   32. Lim M
   33. Lipson EJ
   34. Marie SK
   35. Netto GJ
   36. Oliner KS
   37. Olivi A
   38. Olsson L
   39. Riggins GJ
   40. Sartore-Bianchi A
   41. Schmidt K
   42. Shih lM
   43. Oba-Shinjo SM
   44. Siena S
   45. Theodorescu D
   46. Tie J
   47. Harkins TT
   48. Veronese S
   49. Wang TL
   50. Weingart JD
   51. Wolfgang CL
   52. Wood LD
   53. Xing D
   54. Hruban RH
   55. Wu J
   56. Allen PJ
   57. Schmidt CM
   58. Choti MA
   59. Velculescu VE
   60. Kinzler KW
   61. Vogelstein B
   62. Papadopoulos N
   63. Diaz LA

   (2014) Detection of circulating tumor DNA in early- and late-stage human malignancies

   Science Translational Medicine 6:224ra24.

   https://doi.org/10.1126/scitranslmed.3007094

   • PubMed
   • Google Scholar
6. 1. Botezatu I
   2. Serdyuk O
   3. Potapova G
   4. Shelepov V
   5. Alechina R
   6. Molyaka Y
   7. Ananév V
   8. Bazin I
   9. Garin A
   10. Narimanov M
   11. Knysh V
   12. Melkonyan H
   13. Umansky S
   14. Lichtenstein A

   (2000)

   Genetic analysis of DNA excreted in urine: a new approach for detecting specific genomic DNA sequences from cells dying in an organism

   Clinical chemistry 46:1078–1084.

   • PubMed
   • Google Scholar
7. 1. Cancer Genome Atlas Research Network

   (2014) Comprehensive molecular characterization of urothelial bladder carcinoma

   Nature 507:315–322.

   https://doi.org/10.1038/nature12965

   • PubMed
   • Google Scholar
8. 1. Chai H
   2. Brown RE

   (2009)

   Field effect in cancer-an update

   Annals of Clinical and Laboratory Science 39:331–337.

   • PubMed
   • Google Scholar
9. 1. Chen CH
   2. Dickman KG
   3. Huang CY
   4. Moriya M
   5. Shun CT
   6. Tai HC
   7. Huang KH
   8. Wang SM
   9. Lee YJ
   10. Grollman AP
   11. Pu YS

   (2013) Aristolochic acid-induced upper tract urothelial carcinoma in Taiwan: clinical characteristics and outcomes

   International Journal of Cancer 133:14–20.

   https://doi.org/10.1002/ijc.28013

   • PubMed
   • Google Scholar
10. 1. Chen CH
    2. Dickman KG
    3. Moriya M
    4. Zavadil J
    5. Sidorenko VS
    6. Edwards KL
    7. Gnatenko DV
    8. Wu L
    9. Turesky RJ
    10. Wu XR
    11. Pu YS
    12. Grollman AP

    (2012) Aristolochic acid-associated urothelial cancer in Taiwan

    PNAS 109:8241–8246.

    https://doi.org/10.1073/pnas.1119920109

    • PubMed
    • Google Scholar
11. 1. Cheng L
    2. Montironi R
    3. Lopez-Beltran A

    (2017) TERT Promoter mutations occur frequently in urothelial papilloma and papillary urothelial neoplasm of low malignant potential

    European Urology 71:497–498.

    https://doi.org/10.1016/j.eururo.2016.12.008

    • PubMed
    • Google Scholar
12. 1. Cowan ML
    2. Springer S
    3. Nguyen D
    4. Taheri D
    5. Guner G
    6. Mendoza Rodriguez MA
    7. Wang Y
    8. Kinde I
    9. Del Carmen Rodriguez Pena M
    10. VandenBussche CJ
    11. Olson MT
    12. Cunha I
    13. Fujita K
    14. Ertoy D
    15. Kinzler K
    16. Bivalacqua T
    17. Papadopoulos N
    18. Vogelstein B
    19. Netto GJ

    (2016) Detection of TERT promoter mutations in primary adenocarcinoma of the urinary bladder

    Human Pathology 53:8–13.

    https://doi.org/10.1016/j.humpath.2016.02.009

    • PubMed
    • Google Scholar
13. 1. Davis R
    2. Jones JS
    3. Barocas DA
    4. Castle EP
    5. Lang EK
    6. Leveillee RJ
    7. Messing EM
    8. Miller SD
    9. Peterson AC
    10. Turk TM
    11. Weitzel W
    12. American Urological A
    13. American Urological Association

    (2012) Diagnosis, evaluation and follow-up of asymptomatic microhematuria (AMH) in adults: AUA guideline

    The Journal of Urology 188:2473–2481.

    https://doi.org/10.1016/j.juro.2012.09.078

    • PubMed
    • Google Scholar
14. 1. Dawson SJ
    2. Tsui DW
    3. Murtaza M
    4. Biggs H
    5. Rueda OM
    6. Chin SF
    7. Dunning MJ
    8. Gale D
    9. Forshew T
    10. Mahler-Araujo B
    11. Rajan S
    12. Humphray S
    13. Becq J
    14. Halsall D
    15. Wallis M
    16. Bentley D
    17. Caldas C
    18. Rosenfeld N

    (2013) Analysis of circulating tumor DNA to monitor metastatic breast cancer

    New England Journal of Medicine 368:1199–1209.

    https://doi.org/10.1056/NEJMoa1213261

    • PubMed
    • Google Scholar
15. 1. Dimashkieh H
    2. Wolff DJ
    3. Smith TM
    4. Houser PM
    5. Nietert PJ
    6. Yang J

    (2013) Evaluation of urovysion and cytology for bladder cancer detection: a study of 1835 paired urine samples with clinical and histologic correlation

    Cancer Cytopathology 121:591–597.

    https://doi.org/10.1002/cncy.21327

    • PubMed
    • Google Scholar
16. 1. Douville C
    2. Springer S
    3. Kinde I
    4. Cohen JD
    5. Hruban RH
    6. Lennon AM
    7. Papadopoulos N
    8. Kinzler KW
    9. Vogelstein B
    10. Karchin R

    (2018) Detection of aneuploidy in patients with cancer through amplification of long interspersed nucleotide elements (LINEs)

    PNAS 115:1871–1876.

    https://doi.org/10.1073/pnas.1717846115

    • PubMed
    • Google Scholar
17. 1. Dressman D
    2. Yan H
    3. Traverso G
    4. Kinzler KW
    5. Vogelstein B

    (2003) Transforming single DNA molecules into fluorescent magnetic particles for detection and enumeration of genetic variations

    PNAS 100:8817–8822.

    https://doi.org/10.1073/pnas.1133470100

    • PubMed
    • Google Scholar
18. 1. Ellinger J
    2. Müller SC
    3. Dietrich D

    (2015) Epigenetic biomarkers in the blood of patients with urological malignancies

    Expert Review of Molecular Diagnostics 15:505–516.

    https://doi.org/10.1586/14737159.2015.1019477

    • PubMed
    • Google Scholar
19. 1. Forshew T
    2. Murtaza M
    3. Parkinson C
    4. Gale D
    5. Tsui DW
    6. Kaper F
    7. Dawson SJ
    8. Piskorz AM
    9. Jimenez-Linan M
    10. Bentley D
    11. Hadfield J
    12. May AP
    13. Caldas C
    14. Brenton JD
    15. Rosenfeld N

    (2012) Noninvasive identification and monitoring of cancer mutations by targeted deep sequencing of plasma DNA

    Science Translational Medicine 4:ra168.

    https://doi.org/10.1126/scitranslmed.3003726

    • PubMed
    • Google Scholar
20. 1. Fradet Y
    2. Lockhard C

    (1997)

    Performance characteristics of a new monoclonal antibody test for bladder cancer: ImmunoCyt trade mark

    The Canadian Journal of Urology 4:400–405.

    • PubMed
    • Google Scholar
21. 1. Golan S
    2. Nadu A
    3. Lifshitz D

    (2015) The role of diagnostic ureteroscopy in the era of computed tomography urography

    BMC Urology 15:74.

    https://doi.org/10.1186/s12894-015-0068-8

    • PubMed
    • Google Scholar
22. 1. Goodison S
    2. Chang M
    3. Dai Y
    4. Urquidi V
    5. Rosser CJ

    (2012) A multi-analyte assay for the non-invasive detection of bladder cancer

    PLoS One 7:e47469.

    https://doi.org/10.1371/journal.pone.0047469

    • PubMed
    • Google Scholar
23. 1. Gopalakrishna A
    2. Fantony JJ
    3. Longo TA
    4. Owusu R
    5. Foo WC
    6. Dash R
    7. Denton BT
    8. Inman BA

    (2017) Anticipatory Positive Urine Tests for Bladder Cancer

    Annals of Surgical Oncology 24:1747–1753.

    https://doi.org/10.1245/s10434-016-5763-5

    • PubMed
    • Google Scholar
24. 1. Green DA
    2. Rink M
    3. Xylinas E
    4. Matin SF
    5. Stenzl A
    6. Roupret M
    7. Karakiewicz PI
    8. Scherr DS
    9. Shariat SF

    (2013) Urothelial carcinoma of the bladder and the upper tract: disparate twins

    The Journal of Urology 189:1214–1221.

    https://doi.org/10.1016/j.juro.2012.05.079

    • PubMed
    • Google Scholar
25. Book

    1. Grollman AP
    2. Scarborough J
    3. Jelakovic B

    (2009)

    Aristolochic Acid Nephropathy: An Environmental and Iatrogenic Disease

    Fishbein J, editors. Elsevier.

    • Google Scholar
26. 1. Grollman AP

    (2013) Aristolochic acid nephropathy: Harbinger of a global iatrogenic disease

    Environmental and Molecular Mutagenesis 54:1–7.

    https://doi.org/10.1002/em.21756

    • PubMed
    • Google Scholar
27. 1. Haber DA
    2. Velculescu VE

    (2014) Blood-based analyses of cancer: circulating tumor cells and circulating tumor DNA

    Cancer Discovery 4:650–661.

    https://doi.org/10.1158/2159-8290.CD-13-1014

    • PubMed
    • Google Scholar
28. 1. Hajdinjak T

    (2008) UroVysion FISH test for detecting urothelial cancers: meta-analysis of diagnostic accuracy and comparison with urinary cytology testing

    Urologic Oncology: Seminars and Original Investigations 26:646–651.

    https://doi.org/10.1016/j.urolonc.2007.06.002

    • PubMed
    • Google Scholar
29. 1. Hoang ML
    2. Chen CH
    3. Sidorenko VS
    4. He J
    5. Dickman KG
    6. Yun BH
    7. Moriya M
    8. Niknafs N
    9. Douville C
    10. Karchin R
    11. Turesky RJ
    12. Pu YS
    13. Vogelstein B
    14. Papadopoulos N
    15. Grollman AP
    16. Kinzler KW
    17. Rosenquist TA

    (2013) Mutational signature of aristolochic acid exposure as revealed by whole-exome sequencing

    Science Translational Medicine 5:ra102.

    https://doi.org/10.1126/scitranslmed.3006200

    • PubMed
    • Google Scholar
30. 1. Hollstein M
    2. Moriya M
    3. Grollman AP
    4. Olivier M

    (2013) Analysis of TP53 mutation spectra reveals the fingerprint of the potent environmental carcinogen, aristolochic acid

    Mutation Research/Reviews in Mutation Research 753:41–49.

    https://doi.org/10.1016/j.mrrev.2013.02.003

    • PubMed
    • Google Scholar
31. 1. Horn S
    2. Figl A
    3. Rachakonda PS
    4. Fischer C
    5. Sucker A
    6. Gast A
    7. Kadel S
    8. Moll I
    9. Nagore E
    10. Hemminki K
    11. Schadendorf D
    12. Kumar R

    (2013) TERT promoter mutations in familial and sporadic melanoma

    Science 339:959–961.

    https://doi.org/10.1126/science.1230062

    • PubMed
    • Google Scholar
32. 1. Hsieh SC
    2. Lin IH
    3. Tseng WL
    4. Lee CH
    5. Wang JD

    (2008) Prescription profile of potentially aristolochic acid containing Chinese herbal products: an analysis of National Health Insurance data in Taiwan between 1997 and 2003

    Chinese Medicine 3:13.

    https://doi.org/10.1186/1749-8546-3-13

    • PubMed
    • Google Scholar
33. 1. Hu SL
    2. Zhang HQ
    3. Chan K
    4. Mei QX

    (2004) Studies on the toxicity of Aristolochia manshuriensis (Guanmuton)

    Toxicology 198:195–201.

    https://doi.org/10.1016/j.tox.2004.01.026

    • PubMed
    • Google Scholar
34. 1. Huang FW
    2. Hodis E
    3. Xu MJ
    4. Kryukov GV
    5. Chin L
    6. Garraway LA

    (2013) Highly recurrent TERT promoter mutations in human melanoma

    Science 339:957–959.

    https://doi.org/10.1126/science.1229259

    • PubMed
    • Google Scholar
35. 1. Humphrey PA
    2. Moch H
    3. Cubilla AL
    4. Ulbright TM
    5. Reuter VE

    (2016) The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part B: Prostate and Bladder Tumours

    European Urology 70:106–119.

    https://doi.org/10.1016/j.eururo.2016.02.028

    • PubMed
    • Google Scholar
36. 1. Hurst CD
    2. Platt FM
    3. Knowles MA

    (2014) Comprehensive mutation analysis of the TERT promoter in bladder cancer and detection of mutations in voided urine

    European Urology 65:367–369.

    https://doi.org/10.1016/j.eururo.2013.08.057

    • PubMed
    • Google Scholar
37. 1. Jelaković B
    2. Karanović S
    3. Vuković-Lela I
    4. Miller F
    5. Edwards KL
    6. Nikolić J
    7. Tomić K
    8. Slade N
    9. Brdar B
    10. Turesky RJ
    11. Stipančić Ž
    12. Dittrich D
    13. Grollman AP
    14. Dickman KG

    (2012) Aristolactam-DNA adducts are a biomarker of environmental exposure to aristolochic acid

    Kidney International 81:559–567.

    https://doi.org/10.1038/ki.2011.371

    • PubMed
    • Google Scholar
38. 1. Kawauchi S
    2. Sakai H
    3. Ikemoto K
    4. Eguchi S
    5. Nakao M
    6. Takihara H
    7. Shimabukuro T
    8. Furuya T
    9. Oga A
    10. Matsuyama H
    11. Takahashi M
    12. Sasaki K

    (2009) 9p21 index as estimated by dual-color fluorescence in situ hybridization is useful to predict urothelial carcinoma recurrence in bladder washing cytology

    Human Pathology 40:1783–1789.

    https://doi.org/10.1016/j.humpath.2009.06.011

    • PubMed
    • Google Scholar
39. 1. Khadra MH
    2. Pickard RS
    3. Charlton M
    4. Powell PH
    5. Neal DE

    (2000) A prospective analysis of 1,930 patients with hematuria to evaluate current diagnostic practice

    The Journal of Urology 163:524–527.

    https://doi.org/10.1016/S0022-5347(05)67916-5

    • PubMed
    • Google Scholar
40. 1. Killela PJ
    2. Reitman ZJ
    3. Jiao Y
    4. Bettegowda C
    5. Agrawal N
    6. Diaz LA
    7. Friedman AH
    8. Friedman H
    9. Gallia GL
    10. Giovanella BC
    11. Grollman AP
    12. He TC
    13. He Y
    14. Hruban RH
    15. Jallo GI
    16. Mandahl N
    17. Meeker AK
    18. Mertens F
    19. Netto GJ
    20. Rasheed BA
    21. Riggins GJ
    22. Rosenquist TA
    23. Schiffman M
    24. Shih I
    25. Theodorescu D
    26. Torbenson MS
    27. Velculescu VE
    28. Wang TL
    29. Wentzensen N
    30. Wood LD
    31. Zhang M
    32. McLendon RE
    33. Bigner DD
    34. Kinzler KW
    35. Vogelstein B
    36. Papadopoulos N
    37. Yan H

    (2013) TERT promoter mutations occur frequently in gliomas and a subset of tumors derived from cells with low rates of self-renewal

    PNAS 110:6021–6026.

    https://doi.org/10.1073/pnas.1303607110

    • PubMed
    • Google Scholar
41. 1. Kinde I
    2. Munari E
    3. Faraj SF
    4. Hruban RH
    5. Schoenberg M
    6. Bivalacqua T
    7. Allaf M
    8. Springer S
    9. Wang Y
    10. Diaz LA
    11. Kinzler KW
    12. Vogelstein B
    13. Papadopoulos N
    14. Netto GJ

    (2013) TERT promoter mutations occur early in urothelial neoplasia and are biomarkers of early disease and disease recurrence in urine

    Cancer Research 73:7162–7167.

    https://doi.org/10.1158/0008-5472.CAN-13-2498

    • PubMed
    • Google Scholar
42. 1. Kinde I
    2. Papadopoulos N
    3. Kinzler KW
    4. Vogelstein B

    (2012) FAST-SeqS: a simple and efficient method for the detection of aneuploidy by massively parallel sequencing

    PLoS One 7:e41162.

    https://doi.org/10.1371/journal.pone.0041162

    • PubMed
    • Google Scholar
43. 1. Kinde I
    2. Wu J
    3. Papadopoulos N
    4. Kinzler KW
    5. Vogelstein B

    (2011) Detection and quantification of rare mutations with massively parallel sequencing

    PNAS 108:9530–9535.

    https://doi.org/10.1073/pnas.1105422108

    • PubMed
    • Google Scholar
44. 1. Krüger S
    2. Mess F
    3. Böhle A
    4. Feller AC

    (2003) Numerical aberrations of chromosome 17 and the 9p21 locus are independent predictors of tumor recurrence in non-invasive transitional cell carcinoma of the urinary bladder

    International Journal of Oncology 23:41–48.

    https://doi.org/10.3892/ijo.23.1.41

    • PubMed
    • Google Scholar
45. 1. Lai MN
    2. Wang SM
    3. Chen PC
    4. Chen YY
    5. Wang JD

    (2010) Population-based case-control study of Chinese herbal products containing aristolochic acid and urinary tract cancer risk

    JNCI Journal of the National Cancer Institute 102:179–186.

    https://doi.org/10.1093/jnci/djp467

    • PubMed
    • Google Scholar
46. 1. Lee JY
    2. Kim K
    3. Sung HH
    4. Jeon HG
    5. Jeong BC
    6. Seo SI
    7. Jeon SS
    8. Lee HM
    9. Choi HY
    10. Kwon GY
    11. Kim KM
    12. Lee J
    13. Lim HY
    14. Park SH

    (2018) Molecular characterization of urothelial carcinoma of the bladder and upper urinary tract

    Translational Oncology 11:37–42.

    https://doi.org/10.1016/j.tranon.2017.10.008

    • PubMed
    • Google Scholar
47. 1. Levey AS
    2. Coresh J
    3. Greene T
    4. Stevens LA
    5. Zhang YL
    6. Hendriksen S
    7. Kusek JW
    8. Van Lente F
    9. Chronic Kidney Disease Epidemiology Collaboration

    (2006) Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate

    Annals of Internal Medicine 145:247–254.

    https://doi.org/10.7326/0003-4819-145-4-200608150-00004

    • PubMed
    • Google Scholar
48. 1. Levey AS
    2. Eckardt KU
    3. Tsukamoto Y
    4. Levin A
    5. Coresh J
    6. Rossert J
    7. De Zeeuw D
    8. Hostetter TH
    9. Lameire N
    10. Eknoyan G

    (2005) Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO)

    Kidney International 67:2089–2100.

    https://doi.org/10.1111/j.1523-1755.2005.00365.x

    • PubMed
    • Google Scholar
49. 1. Li CC
    2. Chang TH
    3. Wu WJ
    4. Ke HL
    5. Huang SP
    6. Tsai PC
    7. Chang SJ
    8. Shen JT
    9. Chou YH
    10. Huang CH

    (2008) Significant predictive factors for prognosis of primary upper urinary tract cancer after radical nephroureterectomy in Taiwanese patients

    European Urology 54:1127–1135.

    https://doi.org/10.1016/j.eururo.2008.01.054

    • PubMed
    • Google Scholar
50. 1. Lin HH
    2. Ke HL
    3. Huang SP
    4. Wu WJ
    5. Chen YK
    6. Chang LL

    (2010) Increase sensitivity in detecting superficial, low grade bladder cancer by combination analysis of hypermethylation of E-cadherin, p16, p14, RASSF1A genes in urine

    Urologic Oncology: Seminars and Original Investigations 28:597–602.

    https://doi.org/10.1016/j.urolonc.2008.12.008

    • PubMed
    • Google Scholar
51. 1. Lotan Y
    2. Roehrborn CG

    (2003) Sensitivity and specificity of commonly available bladder tumor markers versus cytology: results of a comprehensive literature review and meta-analyses

    Urology 61:109–118.

    https://doi.org/10.1016/S0090-4295(02)02136-2

    • PubMed
    • Google Scholar
52. 1. McGranahan N
    2. Favero F
    3. de Bruin EC
    4. Birkbak NJ
    5. Szallasi Z
    6. Swanton C

    (2015) Clonal status of actionable driver events and the timing of mutational processes in cancer evolution

    Science Translational Medicine 7:283ra54.

    https://doi.org/10.1126/scitranslmed.aaa1408

    • PubMed
    • Google Scholar
53. 1. Mishriki SF
    2. Nabi G
    3. Cohen NP

    (2008) Diagnosis of urologic malignancies in patients with asymptomatic dipstick hematuria: prospective study with 13 years' follow-up

    Urology 71:13–16.

    https://doi.org/10.1016/j.urology.2007.08.031

    • PubMed
    • Google Scholar
54. 1. Mo L
    2. Zheng X
    3. Huang HY
    4. Shapiro E
    5. Lepor H
    6. Cordon-Cardo C
    7. Sun TT
    8. Wu XR

    (2007) Hyperactivation of Ha-ras oncogene, but not Ink4a/Arf deficiency, triggers bladder tumorigenesis

    Journal of Clinical Investigation 117:314–325.

    https://doi.org/10.1172/JCI30062

    • PubMed
    • Google Scholar
55. 1. Moonen PM
    2. Merkx GF
    3. Peelen P
    4. Karthaus HF
    5. Smeets DF
    6. Witjes JA

    (2007) UroVysion compared with cytology and quantitative cytology in the surveillance of non-muscle-invasive bladder cancer

    European Urology 51:1275–1280.

    https://doi.org/10.1016/j.eururo.2006.10.044

    • PubMed
    • Google Scholar
56. 1. Moriya M
    2. Slade N
    3. Brdar B
    4. Medverec Z
    5. Tomic K
    6. Jelaković B
    7. Wu L
    8. Truong S
    9. Fernandes A
    10. Grollman AP

    (2011) TP53 Mutational signature for aristolochic acid: an environmental carcinogen

    International Journal of Cancer 129:1532–1536.

    https://doi.org/10.1002/ijc.26077

    • PubMed
    • Google Scholar
57. 1. Moss TJ
    2. Qi Y
    3. Xi L
    4. Peng B
    5. Kim TB
    6. Ezzedine NE
    7. Mosqueda ME
    8. Guo CC
    9. Czerniak BA
    10. Ittmann M
    11. Wheeler DA
    12. Lerner SP
    13. Matin SF

    (2017) Comprehensive genomic characterization of upper tract urothelial carcinoma

    European Urology 72:641–649.

    https://doi.org/10.1016/j.eururo.2017.05.048

    • PubMed
    • Google Scholar
58. 1. National Toxicology Program

    (2011)

    Aristolochic acids

    Report on Carcinogens : Carcinogen Profiles 12:45–49.

    • PubMed
    • Google Scholar
59. 1. Netto GJ
    2. Epstein JI

    (2010) Theranostic and prognostic biomarkers: genomic applications in urological malignancies

    Pathology 42:384–394.

    https://doi.org/10.3109/00313021003779145

    • PubMed
    • Google Scholar
60. 1. Netto GJ
    2. Tafe LJ

    (2016) Emerging bladder cancer biomarkers and targets of therapy

    Urologic Clinics of North America 43:63–76.

    https://doi.org/10.1016/j.ucl.2015.08.006

    • PubMed
    • Google Scholar
61. 1. Netto GJ

    (2011) Molecular biomarkers in urothelial carcinoma of the bladder: are we there yet?

    Nature Reviews Urology 9:41–51.

    https://doi.org/10.1038/nrurol.2011.193

    • PubMed
    • Google Scholar
62. 1. Netto GJ

    (2013) Clinical applications of recent molecular advances in urologic malignancies: no longer chasing a "mirage"?

    Advances in Anatomic Pathology 20:175–203.

    https://doi.org/10.1097/PAP.0b013e3182863f80

    • PubMed
    • Google Scholar
63. 1. Nguyen D
    2. Taheri D
    3. Springer S
    4. Cowan M
    5. Guner G
    6. Mendoza Rodriguez MA
    7. Wang Y
    8. Kinde I
    9. VandenBussche CJ
    10. Olson MT
    11. Ricardo BF
    12. Cunha I
    13. Fujita K
    14. Ertoy D
    15. Kinzler KW
    16. Bivalacqua TJ
    17. Papadopoulos N
    18. Vogelstein B
    19. Netto GJ

    (2016) High prevalence of TERT promoter mutations in micropapillary urothelial carcinoma

    Virchows Archiv 469:427–434.

    https://doi.org/10.1007/s00428-016-2001-2

    • PubMed
    • Google Scholar
64. 1. Patel KM
    2. van der Vos KE
    3. Smith CG
    4. Mouliere F
    5. Tsui D
    6. Morris J
    7. Chandrananda D
    8. Marass F
    9. van den Broek D
    10. Neal DE
    11. Gnanapragasam VJ
    12. Forshew T
    13. van Rhijn BW
    14. Massie CE
    15. Rosenfeld N
    16. van der Heijden MS

    (2017) Association of plasma and urinary mutant dna with clinical outcomes in muscle invasive bladder cancer

    Scientific Reports 7:5554.

    https://doi.org/10.1038/s41598-017-05623-3

    • PubMed
    • Google Scholar
65. 1. Poon SL
    2. Huang MN
    3. Choo Y
    4. McPherson JR
    5. Yu W
    6. Heng HL
    7. Gan A
    8. Myint SS
    9. Siew EY
    10. Ler LD
    11. Ng LG
    12. Weng WH
    13. Chuang CK
    14. Yuen JS
    15. Pang ST
    16. Tan P
    17. Teh BT
    18. Rozen SG

    (2015) Mutation signatures implicate aristolochic acid in bladder cancer development

    Genome Medicine 7:38.

    https://doi.org/10.1186/s13073-015-0161-3

    • PubMed
    • Google Scholar
66. 1. Ralla B
    2. Stephan C
    3. Meller S
    4. Dietrich D
    5. Kristiansen G
    6. Jung K

    (2014) Nucleic acid-based biomarkers in body fluids of patients with urologic malignancies

    Critical Reviews in Clinical Laboratory Sciences 51:200–231.

    https://doi.org/10.3109/10408363.2014.914888

    • PubMed
    • Google Scholar
67. 1. Risques RA
    2. Kennedy SR

    (2018) Aging and the rise of somatic cancer-associated mutations in normal tissues

    PLOS Genetics 14:e1007108.

    https://doi.org/10.1371/journal.pgen.1007108

    • PubMed
    • Google Scholar
68. 1. Rodriguez Pena MDC
    2. Tregnago AC
    3. Eich ML
    4. Springer S
    5. Wang Y
    6. Taheri D
    7. Ertoy D
    8. Fujita K
    9. Bezerra SM
    10. Cunha IW
    11. Raspollini MR
    12. Yu L
    13. Bivalacqua TJ
    14. Papadopoulos N
    15. Kinzler KW
    16. Vogelstein B
    17. Netto GJ

    (2017) Spectrum of genetic mutations in de novo PUNLMP of the urinary bladder

    Virchows Archiv 471:761–767.

    https://doi.org/10.1007/s00428-017-2164-5

    • PubMed
    • Google Scholar
69. 1. Rouprêt M
    2. Babjuk M
    3. Compérat E
    4. Zigeuner R
    5. Sylvester RJ
    6. Burger M
    7. Cowan NC
    8. Böhle A
    9. Van Rhijn BW
    10. Kaasinen E
    11. Palou J
    12. Shariat SF

    (2015) European association of urology guidelines on upper urinary tract urothelial cell carcinoma: 2015 Update

    European Urology 68:868–879.

    https://doi.org/10.1016/j.eururo.2015.06.044

    • PubMed
    • Google Scholar
70. 1. Sarkis AS
    2. Bajorin DF
    3. Reuter VE
    4. Herr HW
    5. Netto G
    6. Zhang ZF
    7. Schultz PK
    8. Cordon-Cardo C
    9. Scher HI

    (1995) Prognostic value of p53 nuclear overexpression in patients with invasive bladder cancer treated with neoadjuvant MVAC

    Journal of Clinical Oncology 13:1384–1390.

    https://doi.org/10.1200/JCO.1995.13.6.1384

    • PubMed
    • Google Scholar
71. 1. Sarkis AS
    2. Dalbagni G
    3. Cordon-Cardo C
    4. Melamed J
    5. Zhang ZF
    6. Sheinfeld J
    7. Fair WR
    8. Herr HW
    9. Reuter VE

    (1994) Association of P53 nuclear overexpression and tumor progression in carcinoma in situ of the bladder

    The Journal of Urology 152:388–392.

    https://doi.org/10.1016/S0022-5347(17)32745-3

    • PubMed
    • Google Scholar
72. 1. Sarkis AS
    2. Dalbagni G
    3. Cordon-Cardo C
    4. Zhang ZF
    5. Sheinfeld J
    6. Fair WR
    7. Herr HW
    8. Reuter VE

    (1993) Nuclear overexpression of p53 protein in transitional cell bladder carcinoma: a marker for disease progression

    JNCI Journal of the National Cancer Institute 85:53–59.

    https://doi.org/10.1093/jnci/85.1.53

    • PubMed
    • Google Scholar
73. 1. Sarosdy MF
    2. Kahn PR
    3. Ziffer MD
    4. Love WR
    5. Barkin J
    6. Abara EO
    7. Jansz K
    8. Bridge JA
    9. Johansson SL
    10. Persons DL
    11. Gibson JS

    (2006) Use of a multitarget fluorescence in situ hybridization assay to diagnose bladder cancer in patients with hematuria

    The Journal of Urology 176:44–47.

    https://doi.org/10.1016/S0022-5347(06)00576-3

    • PubMed
    • Google Scholar
74. 1. Scott GA
    2. Laughlin TS
    3. Rothberg PG

    (2014) Mutations of the TERT promoter are common in basal cell carcinoma and squamous cell carcinoma

    Modern Pathology 27:516–523.

    https://doi.org/10.1038/modpathol.2013.167

    • PubMed
    • Google Scholar
75. 1. Serizawa RR
    2. Ralfkiaer U
    3. Steven K
    4. Lam GW
    5. Schmiedel S
    6. Schüz J
    7. Hansen AB
    8. Horn T
    9. Guldberg P

    (2011) Integrated genetic and epigenetic analysis of bladder cancer reveals an additive diagnostic value of FGFR3 mutations and hypermethylation events

    International Journal of Cancer 129:78–87.

    https://doi.org/10.1002/ijc.25651

    • PubMed
    • Google Scholar
76. 1. Sfakianos JP
    2. Cha EK
    3. Iyer G
    4. Scott SN
    5. Zabor EC
    6. Shah RH
    7. Ren Q
    8. Bagrodia A
    9. Kim PH
    10. Hakimi AA
    11. Ostrovnaya I
    12. Ramirez R
    13. Hanrahan AJ
    14. Desai NB
    15. Sun A
    16. Pinciroli P
    17. Rosenberg JE
    18. Dalbagni G
    19. Schultz N
    20. Bajorin DF
    21. Reuter VE
    22. Berger MF
    23. Bochner BH
    24. Al-Ahmadie HA
    25. Solit DB
    26. Coleman JA

    (2015) Genomic characterization of upper tract urothelial carcinoma

    European Urology 68:970–977.

    https://doi.org/10.1016/j.eururo.2015.07.039

    • PubMed
    • Google Scholar
77. 1. Shariat SF
    2. Favaretto RL
    3. Gupta A
    4. Fritsche HM
    5. Matsumoto K
    6. Kassouf W
    7. Walton TJ
    8. Tritschler S
    9. Baba S
    10. Matsushita K
    11. Bastian PJ
    12. Martínez-Salamanca JI
    13. Seitz C
    14. Pycha A
    15. Otto W
    16. Karakiewicz PI
    17. Ficarra V
    18. Novara G

    (2011) Gender differences in radical nephroureterectomy for upper tract urothelial carcinoma

    World Journal of Urology 29:481–486.

    https://doi.org/10.1007/s00345-010-0594-7

    • PubMed
    • Google Scholar
78. 1. Sidransky D
    2. Von Eschenbach A
    3. Tsai YC
    4. Jones P
    5. Summerhayes I
    6. Marshall F
    7. Paul M
    8. Green P
    9. Hamilton SR
    10. Frost P

    (1991) Identification of p53 gene mutations in bladder cancers and urine samples

    Science 252:706–709.

    https://doi.org/10.1126/science.2024123

    • PubMed
    • Google Scholar
79. 1. Siegel RL
    2. Miller KD
    3. Jemal A

    (2017) Cancer Statistics, 2017

    CA: A Cancer Journal for Clinicians 67:7–30.

    https://doi.org/10.3322/caac.21387

    • PubMed
    • Google Scholar
80. 1. Skacel M
    2. Fahmy M
    3. Brainard JA
    4. Pettay JD
    5. Biscotti CV
    6. Liou LS
    7. Procop GW
    8. Jones JS
    9. Ulchaker J
    10. Zippe CD
    11. Tubbs RR

    (2003) Multitarget fluorescence in situ hybridization assay detects transitional cell carcinoma in the majority of patients with bladder cancer and atypical or negative urine cytology

    The Journal of Urology 169:2101–2105.

    https://doi.org/10.1097/01.ju.0000066842.45464.cc

    • PubMed
    • Google Scholar
81. 1. Soria F
    2. Shariat SF
    3. Lerner SP
    4. Fritsche HM
    5. Rink M
    6. Kassouf W
    7. Spiess PE
    8. Lotan Y
    9. Ye D
    10. Fernández MI
    11. Kikuchi E
    12. Chade DC
    13. Babjuk M
    14. Grollman AP
    15. Thalmann GN

    (2017) Epidemiology, diagnosis, preoperative evaluation and prognostic assessment of upper-tract urothelial carcinoma (UTUC)

    World Journal of Urology 35:379–387.

    https://doi.org/10.1007/s00345-016-1928-x

    • PubMed
    • Google Scholar
82. 1. Springer S
    2. Wang Y
    3. Dal Molin M
    4. Masica DL
    5. Jiao Y
    6. Kinde I
    7. Blackford A
    8. Raman SP
    9. Wolfgang CL
    10. Tomita T
    11. Niknafs N
    12. Douville C
    13. Ptak J
    14. Dobbyn L
    15. Allen PJ
    16. Klimstra DS
    17. Schattner MA
    18. Schmidt CM
    19. Yip-Schneider M
    20. Cummings OW
    21. Brand RE
    22. Zeh HJ
    23. Singhi AD
    24. Scarpa A
    25. Salvia R
    26. Malleo G
    27. Zamboni G
    28. Falconi M
    29. Jang JY
    30. Kim SW
    31. Kwon W
    32. Hong SM
    33. Song KB
    34. Kim SC
    35. Swan N
    36. Murphy J
    37. Geoghegan J
    38. Brugge W
    39. Fernandez-Del Castillo C
    40. Mino-Kenudson M
    41. Schulick R
    42. Edil BH
    43. Adsay V
    44. Paulino J
    45. van Hooft J
    46. Yachida S
    47. Nara S
    48. Hiraoka N
    49. Yamao K
    50. Hijioka S
    51. van der Merwe S
    52. Goggins M
    53. Canto MI
    54. Ahuja N
    55. Hirose K
    56. Makary M
    57. Weiss MJ
    58. Cameron J
    59. Pittman M
    60. Eshleman JR
    61. Diaz LA
    62. Papadopoulos N
    63. Kinzler KW
    64. Karchin R
    65. Hruban RH
    66. Vogelstein B
    67. Lennon AM

    (2015) A combination of molecular markers and clinical features improve the classification of pancreatic cysts

    Gastroenterology 149:1501–1510.

    https://doi.org/10.1053/j.gastro.2015.07.041

    • PubMed
    • Google Scholar
83. 1. Stern JL
    2. Theodorescu D
    3. Vogelstein B
    4. Papadopoulos N
    5. Cech TR

    (2015) Mutation of the TERT promoter, switch to active chromatin, and monoallelic TERT expression in multiple cancers

    Genes & Development 29:2219–2224.

    https://doi.org/10.1101/gad.269498.115

    • PubMed
    • Google Scholar
84. Website

    1. Taiwan Cancer Registry

    (2017) Bureau of Health Promotion, Dept. of Health, Taiwan. The incidence of renal pelvic and ureteral tumor in Taiwan

    Taiwan Cancer Registry. Accessed August 14, 2017.

    https://cris.bhp.doh.gov.tw/pagepub/Home.aspx?itemNo=cr.q.10
85. 1. Togneri FS
    2. Ward DG
    3. Foster JM
    4. Devall AJ
    5. Wojtowicz P
    6. Alyas S
    7. Vasques FR
    8. Oumie A
    9. James ND
    10. Cheng KK
    11. Zeegers MP
    12. Deshmukh N
    13. O'Sullivan B
    14. Taniere P
    15. Spink KG
    16. McMullan DJ
    17. Griffiths M
    18. Bryan RT

    (2016) Genomic complexity of urothelial bladder cancer revealed in urinary cfDNA

    European Journal of Human Genetics 24:1167–1174.

    https://doi.org/10.1038/ejhg.2015.281

    • PubMed
    • Google Scholar
86. 1. Vogelstein B
    2. Kinzler KW

    (1999) Digital PCR

    PNAS 96:9236–9241.

    https://doi.org/10.1073/pnas.96.16.9236

    • PubMed
    • Google Scholar
87. 1. Vogelstein B
    2. Papadopoulos N
    3. Velculescu VE
    4. Zhou S
    5. Diaz LA
    6. Kinzler KW

    (2013) Cancer genome landscapes

    Science 339:1546–1558.

    https://doi.org/10.1126/science.1235122

    • PubMed
    • Google Scholar
88. 1. Wang K
    2. Liu T
    3. Ge N
    4. Liu L
    5. Yuan X
    6. Liu J
    7. Kong F
    8. Wang C
    9. Ren H
    10. Yan K
    11. Hu S
    12. Xu Z
    13. Björkholm M
    14. Fan Y
    15. Zhao S
    16. Liu C
    17. Xu D

    (2014) TERT promoter mutations are associated with distant metastases in upper tract urothelial carcinomas and serve as urinary biomarkers detected by a sensitive castPCR

    Oncotarget 5:12428–12439.

    https://doi.org/10.18632/oncotarget.2660

    • PubMed
    • Google Scholar
89. 1. Wang Y
    2. Springer S
    3. Mulvey CL
    4. Silliman N
    5. Schaefer J
    6. Sausen M
    7. James N
    8. Rettig EM
    9. Guo T
    10. Pickering CR
    11. Bishop JA
    12. Chung CH
    13. Califano JA
    14. Eisele DW
    15. Fakhry C
    16. Gourin CG
    17. Ha PK
    18. Kang H
    19. Kiess A
    20. Koch WM
    21. Myers JN
    22. Quon H
    23. Richmon JD
    24. Sidransky D
    25. Tufano RP
    26. Westra WH
    27. Bettegowda C
    28. Diaz LA
    29. Papadopoulos N
    30. Kinzler KW
    31. Vogelstein B
    32. Agrawal N

    (2015a) Detection of somatic mutations and HPV in the saliva and plasma of patients with head and neck squamous cell carcinomas

    Science Translational Medicine 7:293ra104.

    https://doi.org/10.1126/scitranslmed.aaa8507

    • PubMed
    • Google Scholar
90. 1. Wang Y
    2. Springer S
    3. Zhang M
    4. McMahon KW
    5. Kinde I
    6. Dobbyn L
    7. Ptak J
    8. Brem H
    9. Chaichana K
    10. Gallia GL
    11. Gokaslan ZL
    12. Groves ML
    13. Jallo GI
    14. Lim M
    15. Olivi A
    16. Quinones-Hinojosa A
    17. Rigamonti D
    18. Riggins GJ
    19. Sciubba DM
    20. Weingart JD
    21. Wolinsky JP
    22. Ye X
    23. Oba-Shinjo SM
    24. Marie SK
    25. Holdhoff M
    26. Agrawal N
    27. Diaz LA
    28. Papadopoulos N
    29. Kinzler KW
    30. Vogelstein B
    31. Bettegowda C

    (2015b) Detection of tumor-derived DNA in cerebrospinal fluid of patients with primary tumors of the brain and spinal cord

    PNAS 112:9704–9709.

    https://doi.org/10.1073/pnas.1511694112

    • PubMed
    • Google Scholar
91. 1. Wang Y
    2. Sundfeldt K
    3. Mateoiu C
    4. Shih I
    5. Kurman RJ
    6. Schaefer J
    7. Silliman N
    8. Kinde I
    9. Springer S
    10. Foote M
    11. Kristjansdottir B
    12. James N
    13. Kinzler KW
    14. Papadopoulos N
    15. Diaz LA
    16. Vogelstein B

    (2016) Diagnostic potential of tumor DNA from ovarian cyst fluid

    eLife 5:e15175.

    https://doi.org/10.7554/eLife.15175

    • PubMed
    • Google Scholar
92. Book

    1. Wein AJ
    2. Kavoussi LR
    3. Novick AC
    4. Partin AW
    5. Peters CA

    (2012)

    Campbell-Walsh Urology (10th)

    Philadelphia, PA: Elsevier/Saunders. ISBN 978-1-4160-6911-9.

    • Google Scholar
93. 1. Wu XR

    (2005) Urothelial tumorigenesis: a tale of divergent pathways

    Nature Reviews Cancer 5:713–725.

    https://doi.org/10.1038/nrc1697

    • PubMed
    • Google Scholar
94. 1. Yafi FA
    2. Brimo F
    3. Steinberg J
    4. Aprikian AG
    5. Tanguay S
    6. Kassouf W

    (2015) Prospective analysis of sensitivity and specificity of urinary cytology and other urinary biomarkers for bladder cancer

    Urologic Oncology: Seminars and Original Investigations 33:66.e25–66.66.

    https://doi.org/10.1016/j.urolonc.2014.06.008

    • PubMed
    • Google Scholar
95. 1. Yang MH
    2. Chen KK
    3. Yen CC
    4. Wang WS
    5. Chang YH
    6. Huang WJ
    7. Fan FS
    8. Chiou TJ
    9. Liu JH
    10. Chen PM

    (2002) Unusually high incidence of upper urinary tract urothelial carcinoma in Taiwan

    Urology 59:681–687.

    https://doi.org/10.1016/S0090-4295(02)01529-7

    • PubMed
    • Google Scholar
96. 1. Yuan X
    2. Liu C
    3. Wang K
    4. Liu L
    5. Liu T
    6. Ge N
    7. Kong F
    8. Yang L
    9. Björkholm M
    10. Fan Y
    11. Zhao S
    12. Xu D

    (2016) The genetic difference between Western and Chinese urothelial cell carcinomas: infrequent FGFR3 mutation in Han Chinese patients

    Oncotarget 7:25826–25835.

    https://doi.org/10.18632/oncotarget.8404

    • PubMed
    • Google Scholar
97. 1. Yun BH
    2. Rosenquist TA
    3. Sidorenko V
    4. Iden CR
    5. Chen CH
    6. Pu YS
    7. Bonala R
    8. Johnson F
    9. Dickman KG
    10. Grollman AP
    11. Turesky RJ

    (2012) Biomonitoring of aristolactam-DNA adducts in human tissues using ultra-performance liquid chromatography/ion-trap mass spectrometry

    Chemical Research in Toxicology 25:1119–1131.

    https://doi.org/10.1021/tx3000889

    • PubMed
    • Google Scholar
98. 1. Zhang ML
    2. Rosenthal DL
    3. VandenBussche CJ

    (2016) The cytomorphological features of low-grade urothelial neoplasms vary by specimen type

    Cancer Cytopathology 124:552–564.

    https://doi.org/10.1002/cncy.21716

    • PubMed
    • Google Scholar

Thank you for submitting your article "Non-invasive detection of bladder cancer through the analysis of driver gene mutations and aneuploidy" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by Ross Levine as Reviewing Editor (and one of the reviewers) and Charles Sawyers as the Senior Editor.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

In summary, in the two manuscripts the authors develop a non-invasive molecular test called UroSEEK to analyze DNA from urinary CELLS shed into the urine. UroSEEK combines three assessments including mutations from exons in 11 genes, TERT mutations, and aneuploidy. UroSEEK has the potential to augment the sensitivity of routine cytology for the detection of cancer in patients with bladder cancer and upper tract urothelial carcinomas. For the bladder study, they looked at two cohorts, one with hematuria and the second with previous bladder cancer history at risk for new bladder cancers. For the urothelial study, this included 56 subjects with urothelial cancer. For each cohort, they used the same 188 control urine specimens from healthy subjects as negative controls. The main conclusion from these studies is the same – urinary cell mutation testing has potential to detect urothelial and bladder cancers. As such, we think it is best to combine these papers into a single report, which would be of broad interest to the field.

Essential Revisions:

1) There is significant overlap between these two manuscripts, and reading these separately does not add value to either one's conclusion, especially in the same journal issue of eLife. The clinical and diagnostics audience for urothelial and bladder cancer is the same. While the clinical cohorts are different, the genes, methods, controls, and disease (based on their common mutation assessment with URoSEEK) are identical. The written methods are identical. The message is the same – urinary cell mutation testing has potential to detect urothelial and bladder cancers.

2) The issue of how this approach compares to cystoscopy is a major one and is not thoughtfully discussed. If both tests are similar in cost, and cystoscopy is more invasive but definitive, will patients and payers want to use an equally costly non-invasive test vs the more definitive test? How would the authors propose to use this as part of the diagnostic workup and can the impact be modeled in terms of costs, # procedures, etc.? Would add value.

3) The lack of specificity, particularly with respect to positive genetic data in 1) patients without cancer and 2) in patients who previously had cancer and did not recur, is a substantive issue. Can this be improved? Are there ways to deal with this issue in terms of proposing a diagnostic approach? On this note, the methods/assay used have not been thoroughly vetted for sensitivity, specificity, limits of detection and reproducibility for a clinical assay, typically done with serial dilutions and replicates. The authors indicate that one reason for false negatives was due to discordant mutations on the gene panel and the tumor. Another reason could be the assay's limits of detection. While this has not been completed, nonetheless, it does not significantly diminish the conclusion from these studies that urinary cell mutation testing could augment cytology.

4) Have the authors modeled whether including more, or less than 11 genes + TERT/aneuploidy, would impact performance of the test? Are there ways to improve the design to improve performance?

5) Why did the authors separate these assays especially that the three assays that make up UroSEEK can be combined in one NGS platform (examples include assays developed by Foundation Medicine, MSK-IMPACT, Oncomine panel, etc.) that can provide more comprehensive genetic analysis on many more genes including mutations and copy number alterations. Some of the currently available NSG platforms can also utilize cfDNA.

The list of genes in the assay, while includes some of the very commonly mutated genes in bladder cancer, it leaves out other genes that are both relevant and common including for example ARID1A, KDM6A, CDKN1A, STAG2, MLL2/MLL3, etc. Including such genes would very likely improve the assay.

Essential Revisions:

1) There is significant overlap between these two manuscripts, and reading these separately does not add value to either one's conclusion, especially in the same journal issue of eLife. The clinical and diagnostics audience for urothelial and bladder cancer is the same. While the clinical cohorts are different, the genes, methods, controls, and disease (based on their common mutation assessment with URoSEEK) are identical. The written methods are identical. The message is the same – urinary cell mutation testing has potential to detect urothelial and bladder cancers.

The two manuscripts have been integrated into a single paper as requested.

2) The issue of how this approach compares to cystoscopy is a major one and is not thoughtfully discussed. If both tests are similar in cost, and cystoscopy is more invasive but definitive, will patients and payers want to use an equally costly non-invasive test vs the more definitive test? How would the authors propose to use this as part of the diagnostic workup and can the impact be modeled in terms of costs, # procedures, etc.? Would add value.

This is an important point, and we appreciate the reviewer's bringing it up. Cystoscopy is definitely definitive but is invasive and always associated with pain and inconvenience and occasionally associated with side effects such as infection, bleeding, urethral stenosis, etc. The procedure also requires skilled professionals to perform it. As the great majority of patients in some screening settings (e.g., hematuria) do not have cancer, these issues are precisely those that inspired our study. It is widely agreed that a non-invasive test that could be used to screen patients in such settings could minimize the need for cystoscopy and its unintended consequences.

The cost of cystoscopy is up to $3200, not including the cost of the pathology of biopsies. The cost of ureteroscopy is similar. We envision that the cost of UroSEEK plus cytology would be <$750 when implemented in a commercial setting, using savings afforded by economy-of scale. We anticipate that >90% of patients with hematuria who would otherwise undergo cystoscopy could be spared this procedure on the basis of a negative UroSEEK test The entire cost savings, integrated over all patients, would therefore be considerable: for every 100 patients evaluated, up to $200,000 would be saved (assuming UroSEEK plus cytology in all patients plus cystoscopy in 5% of the patients who test positive with UroSEEK). This translates to many millions of dollars in the U.S. annually, and does not include the cost of follow-ups for the side effects of cystoscopy when performed. These issues are now briefly discussed at the end of the revised Discussion.

3) The lack of specificity, particularly with respect to positive genetic data in 1) patients without cancer and 2) in patients who previously had cancer and did not recur, is a substantive issue. Can this be improved? Are there ways to deal with this issue in terms of proposing a diagnostic approach?

We agree with the reviewers that specificity is important, but there are two kinds of specificity. Technical specificity can be assessed from the evaluation of the test in DNA from totally normal individuals, such as DNA from normal WBCs. The technical specificity of UroSEEK is very high, over 99%, as reported in the manuscript. Biological specificity is different and reflects the cohorts studied. In urines from normal individuals (without any hematuria, for example), the specificity is as high as in as in normal WBCs (>99%). As noted in the manuscript, though, this specificity is not as high in other cohorts. In the Discussion, this issue is discussed at length. For example, the 6.5% false positives in BC early detection cohort could be related to a lead time ahead of “clinical” detection. The follow-up period for many of patients was only one year, and cystoscopy was not performed in every patient as per current standards. The same issue could lead to apparently false positive results in the surveillance cohort, in which patients are at much higher risk for cancer than in the early detection cohorts. We will attempt to address such possibilities in future prospective studies that we are now planning where a longer follow up (e.g. 3 years minimum) should be helpful. False positives are not unique to our study; they have been observed in all other molecular assays for bladder cancer, including those that are FDA- approved. Even assuming that none of the false positives are due to as yet undetected cancers, the positive predictive values and negative predictive values in the early detection cohorts is considerably above what has been required for FDA-approval of diagnostic tests in general.

On this note, the methods/assay used have not been thoroughly vetted for sensitivity, specificity, limits of detection and reproducibility for a clinical assay, typically done with serial dilutions and replicates. The authors indicate that one reason for false negatives was due to discordant mutations on the gene panel and the tumor. Another reason could be the assay's limits of detection. While this has not been completed, nonetheless, it does not significantly diminish the conclusion from these studies that urinary cell mutation testing could augment cytology.

We appreciate the reviewers’ view that such detailed vetting is not critical for the conclusions made in this manuscript. But we would like to highlight several pieces of data that address this point. The reproducibility of the test has been exhaustively examined, because tests were scored positive if and only if an independent assay of DNA from the same urine sample was also positive. Moreover, the coefficients of variation of the duplicates were high, as noted in the table below (which can easily be recalculated by readers using the data recorded in the Supplementary files, wherein the results of each replicate are specified). We note that the CVs are relatively low even when the MAFs are low, meaning that there are only a few molecules present, helping to establish technical sensitivity (limit of sensitivity).

UroSEEK requires no complex molecular biological manipulations of DNA; it simply uses PCR of DNA purified from the urine. Thus, many of the issues that are faced by other next generation sequencing technologies, such as those involved in end-polishing of DNA, ligations, and captures, are avoided. The reproducibility and sensitivity of SafeSeqS, which is the basis for UroSEEK, has been previously documented in the original paper (Kinde et al., Detection and quantification of rare mutations with massively parallel sequencing. Proc Natl Acad Sci U S A, 108(23), 9530-9535).

Finally, we are in the process of licensing UroSEEK to a small start-up company that was founded by some of the senior authors of this study. As part of their due diligence and preparation for an upcoming clinical trial of UroSEEK based on the results reported in our manuscript, they have prepared a Clinical Laboratory Improvement Amendment (CLIA) report. Though the report cannot be made public, it provides additional evidence of the reproducibility of UroSEEK in a completely different laboratory.

4) Have the authors modeled whether including more, or less than 11 genes + TERT/aneuploidy, would impact performance of the test? Are there ways to improve the design to improve performance?

This is an excellent question. The genes chosen for the analysis were based on a comprehensive evaluation of the literature on bladder cancer (insufficient data existed for UTUC, but prior publications indicate that the genetic alterations in upper and lower urinary tract cancers are similar). We modeled how additional genes would impact the assay by looking at the inclusion of other amplicons from the top 20 genes contributing to bladder cancer development listed in the COSMIC database. As shown in the figure in Supplementary file 1, the 59 amplicons chosen are above the asymptote in this graph, meaning that more amplicons would not substantially to sensitivity. More amplicons would however, increase the cost and indirectly decrease the sensitivity by decreasing specificity. The reason for this decrease in sensitivity is that the probabilistic models used to interpret rare mutations are rooted in multiple hypothesis testing; the more amplicons studied, the greater the possibility for false positives. The Bonferroni correction is an early example of an approach for dealing with this issue. The new figure illustrates that our choice of amplicons was reasonable. Moreover, the point in the graph (at 59 amplicons) shows that the expected sensitivity was actually realized in the primary bladder cancers that we studied, and was not just theoretical based on expectations from the COSMIC database.

We have included this graph as a new figure in Supplementary file 1 to illustrate this important concept.

5) Why did the authors separate these assays especially that the three assays that make up UroSEEK can be combined in one NGS platform (examples include assays developed by Foundation Medicine, MSK-IMPACT, Oncomine panel, etc.) that can provide more comprehensive genetic analysis on many more genes including mutations and copy number alterations. Some of the currently available NSG platforms can also utilize cfDNA.

As mentioned in the manuscript, TERT promoter region is very CG rich and difficult to amplify. This forced us to exclude it from the rest of the panel and sequence it alone. The Aneuploidy test requires a completely different amplification strategy due the large number of total regions that are assessed (~30,000).

We are keenly aware of the other tests cited by the reviewers because they are largely based on the strategy we described earlier, using endogenous rather than exogenous molecular barcodes (Kinde reference above). Though it is beyond the scope of our article to discuss their strengths and weaknesses, we note that these other tests are designed to detect relatively high fractions of mutant DNA, much higher than those we observed in many urine samples, at a much larger number of genomic positions. Additionally, they are far more expensive than UroSEEK and not well-suited for screening purposes. UroSEEK is unique so far in that it can assess a relatively small (but adequate) number of amplicons at very high depth and sensitivity at minimal cost.

The list of genes in the assay, while includes some of the very commonly mutated genes in bladder cancer, it leaves out other genes that are both relevant and common including for example ARID1A, KDM6A, CDKN1A, STAG2, MLL2/MLL3, etc. Including such genes would very likely improve the assay.

We again thank the reviewers for this suggestion, but note that the inclusion of these other amplicons would not substantially increase the sensitivity of the assay, as documented by the figure in Supplementary file 1 (and now included in the revised manuscript). We also note that cancers that do not harbor mutations in the 59 individual amplicons evaluated can still be aneuploid and can therefore be detected with the third component of UroSEEK, which detects aneuploidy.

Author details

1. Simeon U Springer

   1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
   2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

   Contribution

   Resources, Formal analysis, Validation, Investigation, Writing—review and editing

   Contributed equally with

   Chung-Hsin Chen and Maria Del Carmen Rodriguez Pena

   For correspondence

   gnetto@uabmc.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2891-2111

2. Chung-Hsin Chen

   Department of Urology, National Taiwan University Hospital, Taipei, Taiwan

   Contribution

   Conceptualization, Investigation, Writing—review and editing, Project administration, Resources, Funding acquisition

   Contributed equally with

   Simeon U Springer and Maria Del Carmen Rodriguez Pena

   Competing interests

   No competing interests declared

3. Maria Del Carmen Rodriguez Pena

   1. Department of Pathology, Johns Hopkins University, Baltimore, United States
   2. Department of Pathology, University of Alabama at Birmingham, Birmingham, United States

   Contribution

   Resources, Formal analysis, Investigation, Writing—review and editing

   Contributed equally with

   Simeon U Springer and Chung-Hsin Chen

   For correspondence

   mdcrodriguez@uabmc.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-3439-7013

4. Lu Li

   Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1920-4965

5. Christopher Douville

   Department of Biomedical Engineering, Institute for Computational Medicine, Johns Hopkins University, Baltimore, United States

   Contribution

   Formal analysis, Investigation

   For correspondence

   cdouvil1@jhmi.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2510-4151

6. Yuxuan Wang

   1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
   2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2932-6042

7. Joshua David Cohen

   1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
   2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

   Contribution

   Formal analysis, Investigation

   For correspondence

   cohen@jhmi.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1158-5668

8. Diana Taheri

   1. Department of Pathology, Johns Hopkins University, Baltimore, United States
   2. Department of Pathology, Isfahan Kidney Diseases Research Center, Isfahan University of Medical Sciences, Isfahan, Iran

   Contribution

   Resources, Formal analysis, Investigation

   Competing interests

   No competing interests declared

9. Natalie Silliman

   1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
   2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

10. Joy Schaefer

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Formal analysis, Investigation

    For correspondence

    mjschaefer68@gmail.com

    Competing interests

    No competing interests declared

11. Janine Ptak

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Formal analysis, Investigation

    For correspondence

    jptak1@jhmi.edu

    Competing interests

    No competing interests declared

12. Lisa Dobbyn

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Formal analysis, Investigation

    For correspondence

    ldobbyn1@jhmi.edu

    Competing interests

    No competing interests declared

13. Maria Papoli

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Formal analysis, Investigation

    For correspondence

    mpopoli81@gmail.com

    Competing interests

    No competing interests declared

14. Isaac Kinde

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Conceptualization, Writing—review and editing

    For correspondence

    ik@jhmi.edu

    Competing interests

    No competing interests declared

15. Bahman Afsari

    1. Department of Oncology, Johns Hopkins University, Baltimore, United States
    2. Division of Biostatistics and Bioinformatics, Department of Oncology, Sidney Kimmel Cancer Center, Johns Hopkins School of Medicine, Baltimore, United States

    Contribution

    Formal analysis

    Competing interests

    No competing interests declared

16. Aline C Tregnago

    Department of Pathology, Johns Hopkins University, Baltimore, United States

    Contribution

    Resources, Formal analysis, Investigation

    For correspondence

    tregnago.aline@gmail.com

    Competing interests

    No competing interests declared

17. Stephania M Bezerra

    Department of Pathology, AC Camargo Cancer Center, Sao Paulo, Brazil

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

18. Christopher VandenBussche

    Department of Pathology, Johns Hopkins University, Baltimore, United States

    Contribution

    Resources, Investigation

    For correspondence

    cjvand@jhmi.edu

    Competing interests

    No competing interests declared

19. Kazutoshi Fujita

    Department of Pathology, Osaka University, Osaka, Japan

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

20. Dilek Ertoy

    Department of Pathology, Hacettepe University, Ankara, Turkey

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

21. Isabela W Cunha

    Department of Pathology, AC Camargo Cancer Center, Sao Paulo, Brazil

    Contribution

    Resources, Investigation

    For correspondence

    iwerneck0210@gmail.com

    Competing interests

    No competing interests declared

22. Lijia Yu

    Department of Pathology, University of Alabama at Birmingham, Birmingham, United States

    Contribution

    Formal analysis

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-6735-9569

23. Trinity J Bivalacqua

    Department of Urology, Johns Hopkins University, Baltimore, United States

    Contribution

    Investigation

    For correspondence

    tbivala1@jhmi.edu

    Competing interests

    No competing interests declared

24. Arthur P Grollman

    1. Department of Pharmacological Sciences, Stony Brook University, Stony Brook, United States
    2. Department of Medicine, Stony Brook University, Stony Brook, United States

    Contribution

    Conceptualization, Funding acquisition, Writing—review and editing

    Competing interests

    No competing interests declared

25. Luis A Diaz

    1. Department of Forensic Medicine and Pathology, National Taiwan University Hospital, Taipei, Taiwan
    2. Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, United States

    Contribution

    Conceptualization, Writing—review and editing

    Competing interests

    No competing interests declared

26. Rachel Karchin

    1. Department of Biomedical Engineering, Institute for Computational Medicine, Johns Hopkins University, Baltimore, United States
    2. Department of Oncology, Johns Hopkins University, Baltimore, United States

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

27. Ludmila Danilova

    1. Division of Biostatistics and Bioinformatics, Department of Oncology, Sidney Kimmel Cancer Center, Johns Hopkins School of Medicine, Baltimore, United States
    2. Department of Pathology, Hacettepe University, Ankara, Turkey

    Contribution

    Investigation

    Competing interests

    No competing interests declared

28. Chao-Yuan Huang

    Department of Urology, National Taiwan University Hospital, Taipei, Taiwan

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

29. Chia-Tung Shun

    Department of Forensic Medicine and Pathology, National Taiwan University Hospital, Taipei, Taiwan

    Contribution

    Investigation

    Competing interests

    No competing interests declared

30. Robert J Turesky

    1. Masonic Cancer Center, University of Minnesota, Minneapolis, United States
    2. Department of Medicinal Chemistry, University of Minnesota, Minneapolis, United States

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

31. Byeong Hwa Yun

    1. Masonic Cancer Center, University of Minnesota, Minneapolis, United States
    2. Department of Medicinal Chemistry, University of Minnesota, Minneapolis, United States

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

32. Thomas A Rosenquist

    Department of Pharmacological Sciences, Stony Brook University, Stony Brook, United States

    Contribution

    Writing—review and editing

    Competing interests

    No competing interests declared

33. Yeong-Shiau Pu

    Department of Urology, National Taiwan University Hospital, Taipei, Taiwan

    Contribution

    Resources, Supervision, Funding acquisition, Investigation, Writing—review and editing

    Competing interests

    No competing interests declared

34. Ralph H Hruban

    Department of Pathology, Johns Hopkins University, Baltimore, United States

    Contribution

    Conceptualization, Visualization

    For correspondence

    rhruban@jhmi.edu

    Competing interests

    No competing interests declared

35. Cristian Tomasetti

    1. Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States
    2. Division of Biostatistics and Bioinformatics, Department of Oncology, Sidney Kimmel Cancer Center, Johns Hopkins School of Medicine, Baltimore, United States

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

36. Nickolas Papadopoulos

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    Competing interests

    Founder of Personal Genome Diagnostics and PapGene and advises Sysmex-Inostics. These companies and others have licensed technologies from Johns Hopkins, of which BV, KK, and NP are inventors on a patent (U.S. 20140227705 A1) and receive royalties. The terms of these arrangements are managed by the university in accordance with its conflict of interest policies.

37. Ken W Kinzler

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    Competing interests

    Ken W Kinzler: Founder of Personal Genome Diagnostics and PapGene and advises Sysmex-Inostics. These companies and others have licensed technologies from Johns Hopkins, of which BV, KK, and NP are inventors on a patent (U.S. 20140227705 A1) and receive royalties. The terms of these arrangements are managed by the university in accordance with its conflict of interest policies.

38. Bert Vogelstein

    1. Howard Hughes Medical Institute, Ludwig Center for Cancer Genetics and Therapeutics, Baltimore, United States
    2. Sidney Kimmel Comprehensive Cancer Center, Baltimore, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    bertvog@gmail.com

    Competing interests

    Bert Vogelstein: Founder of Personal Genome Diagnostics and PapGene and advises Sysmex-Inostics. These companies and others have licensed technologies from Johns Hopkins, of which BV, KK, and NP are inventors on a patent (U.S. 20140227705 A1) and receive royalties. The terms of these arrangements are managed by the university in accordance with its conflict of interest policies.

    "This ORCID iD identifies the author of this article:" 0000-0003-0766-3854

39. Kathleen G Dickman

    1. Department of Pharmacological Sciences, Stony Brook University, Stony Brook, United States
    2. Department of Medicine, Stony Brook University, Stony Brook, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    kathleen.dickman@stonybrook.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-1308-2992

40. George J Netto

    1. Department of Pathology, Johns Hopkins University, Baltimore, United States
    2. Department of Pathology, University of Alabama at Birmingham, Birmingham, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Visualization, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    gnetto@uabmc.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-3915-9134

• 6,885

  Page views

• 887

  Downloads

• 122

  Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.