1. Cancer Biology
  2. Computational and Systems Biology

Radiomics: Rethinking the role of clinical imaging

Radiomics has the potential to improve the management of cancer patients, but further research is required before it can be adopted into routine clinical practice.
https://doi.org/10.7554/eLife.30563
Share this article
Cite this article
  1. James PB O'Connor
(2017)
Radiomics: Rethinking the role of clinical imaging
eLife 6:e30563.
https://doi.org/10.7554/eLife.30563
  2. Download BibTeX
  3. Download .RIS
  1. James PB O'Connor  Is a corresponding author
  1. University of Manchester, United Kingdom
  2. The Christie Hospital, United Kingdom

The pressing need to improve cancer survival rates continuously motivates scientists to uncover the genetic and molecular composition of tumors, and use this information to develop more effective cancer therapies. Since tumors differ in their biological makeup, treatments can now often be tailored towards individual patients, in a strategy termed ‘personalized medicine’ (Schilsky, 2010). However, personalized medicine can only improve patient outcomes if there are reliable tests available to identify and analyze the biology of individual tumors, and detect which patients benefit from a given therapy (La Thangue and Kerr, 2011).

Medical imaging techniques underpin much decision-making in cancer medicine and can be used to determine, among other things, the size of a tumor and if it has spread into the surrounding tissue or other parts of the body. These and other tumor characteristics act as prognostic indicators and can help clinicians assign treatments for patients and monitor their response to therapy. However, only a few imaging techniques have been approved for use as companion diagnostic tools or as predictive biomarkers that can identify which patients are likely to benefit from a specific therapy (O'Connor et al., 2017). Now, in eLife, a team of researchers led by Hugo Aerts, Robert Gillies and Philippe Lambin – including Patrick Grossmann as first author – report on how an image analysis approach called radiomics could be used to enhance personalized medicine (Grossmann et al., 2017).

Radiomics uses computer algorithms to process the data collected by different medical imaging techniques and is becoming increasingly popular in cancer imaging research. One of its key characteristics is that radiomics recognizes that digital medical images are not only pictures – they are also are complex data. Radiomic analyses extract and measure an array of ‘features’ that describe the image texture and distributions of individual voxel values – the units that make up a 3D image – within a tumor. Each voxel represents a tiny amount of tissue and contains around 105 to 107 neoplastic and stromal cells, depending on tumor type and voxel dimensions.

Features can be descriptive terms derived from radiology reports or mathematical quantities. Feature extraction requires several complex steps in a defined pipeline: defining the regions of interest and segmenting three-dimensional images, extracting features and converting images into ‘mineable’ data. Finally, data from imaging techniques, and from biofluids and tumor tissue, are combined into models to predict patient prognosis or benefit from a specific therapy (Brady et al., 2016; Gillies et al., 2016).

Grossmann et al. – who are based at various insitiutes in the USA, Candada and The Netherlands – used radiomics to extract over 600 image features in computer tomography images from 262 patients with non-small cell lung cancer. They next examined how these features relate to the molecular characteristics of the tumors (such as the modulation of immune pathways and tumor suppressor proteins) and whether radiomic features predicted if the patients recovered after treatment. The findings were validated on an independent set of data from 89 other patients. Grossmann et al. discovered a connection between the radiomic phenotype of a tumor, the signaling pathways inside cells that drive how cancer develops, and clinical treatment outcomes. Moreover, by combining the data from radiomic, molecular and clinical studies, the overall chance of a patient surviving the cancer could be better predicted than when using clinical data alone.

The study by Grossmann et al. demonstrates the potential strength of radiomics, especially when combined with genomic and clinical information. One of the main benefits is that medical imaging data acquired in routine healthcare can be used in a new way to inform clinicians about the biology of a tumor and provide potential prognostic or predictive information. Imaging data from a large number of patients can be investigated rapidly. Further, radiomic image analysis is a non-invasive procedure that provides three-dimensional reconstructions of a tumor and avoids the problems associated with samples obtained through invasive biopsies (O'Connor et al., 2015).

However, several challenges need to be overcome before radiomics can be used as a companion diagnostic or predictive tool. Significant statistical problems can arise from analysis of such ‘big data’, with overfitting leading to relationships being seen were there are none (false discovery; Limkin et al., 2017). Once relationships are discovered, these findings need to be replicated in prospective studies.

Moreover, the terminology used for the derived image features needs to be standardized. At present, different investigators use their own names for the features, which makes it difficult to compare data between studies. In addition, the number of features incorporated into radiomic analyzes appears to vary even within individual research groups, leading to a phenomenon termed 'biomarker drift'. Biomarkers only become fit for guiding clinical decision-making when they become fixed in their acquisition, analysis and quality assurance processes, all defined clearly in standardized operating procedures (O'Connor et al., 2015).

Finally, ‘Picture Archiving and Communications System’ platforms must be adapted to integrate with radiomics algorithms in order to enable widespread clinical use. This will require coordinated efforts from clinical researchers and imaging scientists, along with additional computational power and infrastructure.

While considerable effort is required to rise to these challenges, the potential rewards are great. If enough evidence of the predictive power of radiomics can be produced, then radiomics may become a viable solution to help deliver personalized medicine in the clinic.

References

Article and author information

Author details

  1. James PB O'Connor

    James PB O'Connor is in the Division of Cancer Sciences, University of Manchester, Manchester, United Kingdom and at The Christie Hospital, Manchester, United Kingdom
    For correspondence
    james.o'connor@manchester.ac.uk
    Competing interests
    No competing interests declared

Publication history

  1. Version of Record published: September 6, 2017 (version 1)

Copyright

© 2017, O'Connor

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 2,164
    Page views
  • 206
    Downloads
  • 11
    Citations

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

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. James PB O'Connor
(2017)
Radiomics: Rethinking the role of clinical imaging
eLife 6:e30563.
https://doi.org/10.7554/eLife.30563

Categories and tags

Research organism

    1. Related to

    Defining the biological basis of radiomic phenotypes in lung cancer

    Patrick Grossmann, Olya Stringfield ... Hugo JWL Aerts
  2. Further reading

Further reading

    1. Cancer Biology
    2. Computational and Systems Biology

    Defining the biological basis of radiomic phenotypes in lung cancer

    Patrick Grossmann, Olya Stringfield ... Hugo JWL Aerts
    Research Article Updated

    Medical imaging can visualize characteristics of human cancer noninvasively. Radiomics is an emerging field that translates these medical images into quantitative data to enable phenotypic profiling of tumors. While radiomics has been associated with several clinical endpoints, the complex relationships of radiomics, clinical factors, and tumor biology are largely unknown. To this end, we analyzed two independent cohorts of respectively 262 North American and 89 European patients with lung cancer, and consistently identified previously undescribed associations between radiomic imaging features, molecular pathways, and clinical factors. In particular, we found a relationship between imaging features, immune response, inflammation, and survival, which was further validated by immunohistochemical staining. Moreover, a number of imaging features showed predictive value for specific pathways; for example, intra-tumor heterogeneity features predicted activity of RNA polymerase transcription (AUC = 0.62, p=0.03) and intensity dispersion was predictive of the autodegration pathway of a ubiquitin ligase (AUC = 0.69, p<10-4). Finally, we observed that prognostic biomarkers performed highest when combining radiomic, genetic, and clinical information (CI = 0.73, p<10-9) indicating complementary value of these data. In conclusion, we demonstrate that radiomic approaches permit noninvasive assessment of both molecular and clinical characteristics of tumors, and therefore have the potential to advance clinical decision-making by systematically analyzing standard-of-care medical images.

    1. Cell Biology
    2. Cancer Biology

    Transferred mitochondria accumulate reactive oxygen species, promoting proliferation

    Chelsea U Kidwell, Joseph R Casalini ... Minna Roh-Johnson
    Research Article Updated

    Recent studies reveal that lateral mitochondrial transfer, the movement of mitochondria from one cell to another, can affect cellular and tissue homeostasis. Most of what we know about mitochondrial transfer stems from bulk cell studies and have led to the paradigm that functional transferred mitochondria restore bioenergetics and revitalize cellular functions to recipient cells with damaged or non-functional mitochondrial networks. However, we show that mitochondrial transfer also occurs between cells with functioning endogenous mitochondrial networks, but the mechanisms underlying how transferred mitochondria can promote such sustained behavioral reprogramming remain unclear. We report that unexpectedly, transferred macrophage mitochondria are dysfunctional and accumulate reactive oxygen species in recipient cancer cells. We further discovered that reactive oxygen species accumulation activates ERK signaling, promoting cancer cell proliferation. Pro-tumorigenic macrophages exhibit fragmented mitochondrial networks, leading to higher rates of mitochondrial transfer to cancer cells. Finally, we observe that macrophage mitochondrial transfer promotes tumor cell proliferation in vivo. Collectively these results indicate that transferred macrophage mitochondria activate downstream signaling pathways in a ROS-dependent manner in cancer cells, and provide a model of how sustained behavioral reprogramming can be mediated by a relatively small amount of transferred mitochondria in vitro and in vivo.

    1. Cancer Biology
    2. Computational and Systems Biology

    Interrogating the precancerous evolution of pathway dysfunction in lung squamous cell carcinoma using XTABLE

    Matthew Roberts, Julia Ogden ... Carlos Lopez-Garcia
    Tools and Resources Updated

    Lung squamous cell carcinoma (LUSC) is a type of lung cancer with a dismal prognosis that lacks adequate therapies and actionable targets. This disease is characterized by a sequence of low- and high-grade preinvasive stages with increasing probability of malignant progression. Increasing our knowledge about the biology of these premalignant lesions (PMLs) is necessary to design new methods of early detection and prevention, and to identify the molecular processes that are key for malignant progression. To facilitate this research, we have designed XTABLE (Exploring Transcriptomes of Bronchial Lesions), an open-source application that integrates the most extensive transcriptomic databases of PMLs published so far. With this tool, users can stratify samples using multiple parameters and interrogate PML biology in multiple manners, such as two- and multiple-group comparisons, interrogation of genes of interests, and transcriptional signatures. Using XTABLE, we have carried out a comparative study of the potential role of chromosomal instability scores as biomarkers of PML progression and mapped the onset of the most relevant LUSC pathways to the sequence of LUSC developmental stages. XTABLE will critically facilitate new research for the identification of early detection biomarkers and acquire a better understanding of the LUSC precancerous stages.