Virtual mouse brain histology from multi-contrast MRI via deep learning

  1. Zifei Liang
  2. Choong H Lee
  3. Tanzil M Arefin
  4. Zijun Dong
  5. Piotr Walczak
  6. Song Hai Shi
  7. Florian Knoll
  8. Yulin Ge
  9. Leslie Ying
  10. Jiangyang Zhang  Is a corresponding author
  1. New York University School of Medicine, United States
  2. University of Maryland, United States
  3. Memorial Sloan Kettering Cancer Center
  4. University at Buffalo, State University of New York, United States

Abstract

1H MRI maps brain structure and function non-invasively through versatile contrasts that exploit inhomogeneity in tissue micro-environments. Inferring histopathological information from MRI findings, however, remains challenging due to absence of direct links between MRI signals and cellular structures. Here, we show that deep convolutional neural networks, developed using co-registered multi-contrast MRI and histological data of the mouse brain, can estimate histological staining intensity directly from MRI signals at each voxel. The results provide three-dimensional maps of axons and myelin with tissue contrasts that closely mimics target histology and enhanced sensitivity and specificity compared to conventional MRI markers. Furthermore, the relative contribution of each MRI contrast within the networks can be used to optimize multi-contrast MRI acquisition. We anticipate our method to be a starting point for translation of MRI results into easy-to-understand virtual histology for neurobiologists and provide resources for validating novel MRI techniques.

Data availability

All data and source codes used in this study are available at https://www.github.com/liangzifei/MRH-net/. The data can also be found at datadryad.org

The following data sets were generated
The following previously published data sets were used
    1. Lein ES et al
    (2006) Allen Mouse Brain Atlas
    The reference data at http://connectivity.brain-map.org/static/referencedata.

Article and author information

Author details

  1. Zifei Liang

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  2. Choong H Lee

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  3. Tanzil M Arefin

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  4. Zijun Dong

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  5. Piotr Walczak

    Department of Diagnostic Radiology and Nuclear Medicine, University of Maryland, Baltimore, United States
    Competing interests
    The authors declare that no competing interests exist.
  6. Song Hai Shi

    Developmental Biology Program, Memorial Sloan Kettering Cancer Center
    Competing interests
    The authors declare that no competing interests exist.
  7. Florian Knoll

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  8. Yulin Ge

    Department of Radiology, New York University School of Medicine, New York, United States
    Competing interests
    The authors declare that no competing interests exist.
  9. Leslie Ying

    Departments of Biomedical Engineering, University at Buffalo, State University of New York, Buffalo, United States
    Competing interests
    The authors declare that no competing interests exist.
  10. Jiangyang Zhang

    Department of Radiology, New York University School of Medicine, New York, United States
    For correspondence
    jiangyang.zhang@nyulangone.org
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0740-2662

Funding

Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD074593)

  • Jiangyang Zhang

National Institute of Neurological Disorders and Stroke (R01NS102904)

  • Jiangyang Zhang

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (s16-00145-133) of the New York University.

Reviewing Editor

  1. Saad Jbabdi, University of Oxford, United Kingdom

Publication history

  1. Preprint posted: May 3, 2020 (view preprint)
  2. Received: July 20, 2021
  3. Accepted: January 27, 2022
  4. Accepted Manuscript published: January 28, 2022 (version 1)
  5. Version of Record published: February 11, 2022 (version 2)

Copyright

© 2022, Liang et al.

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

Metrics

  • 1,493
    Page views
  • 233
    Downloads
  • 0
    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. Zifei Liang
  2. Choong H Lee
  3. Tanzil M Arefin
  4. Zijun Dong
  5. Piotr Walczak
  6. Song Hai Shi
  7. Florian Knoll
  8. Yulin Ge
  9. Leslie Ying
  10. Jiangyang Zhang
(2022)
Virtual mouse brain histology from multi-contrast MRI via deep learning
eLife 11:e72331.
https://doi.org/10.7554/eLife.72331

Further reading

    1. Genetics and Genomics
    2. Medicine
    Joshua K Park, Shantanu Bafna ... Ron Do
    Research Article

    Background: Causality between plasma triglyceride (TG) levels and atherosclerotic cardiovascular disease (ASCVD) risk remains controversial despite more than four decades of study and two recent landmark trials, STRENGTH and REDUCE-IT. Further unclear is the association between TG levels and non-atherosclerotic diseases across organ systems.

    Methods: Here, we conducted a phenome-wide, two-sample Mendelian randomization (MR) analysis using inverse-variance weighted (IVW) regression to systematically infer the causal effects of plasma TG levels on 2,600 disease traits in the European ancestry population of UK Biobank. For replication, we externally tested 221 nominally significant associations (p < 0.05) in an independent cohort from FinnGen. To account for potential horizontal pleiotropy and the influence of invalid instrumental variables, we performed sensitivity analyses using MR-Egger regression, weighted median estimator, and MR-PRESSO. Finally, we used multivariable MR controlling for correlated lipid fractions to distinguish the independent effect of plasma TG levels.

    Results: Our results identified 7 disease traits reaching Bonferroni-corrected significance in both the discovery (p < 1.92 × 10-5) and replication analyses (p < 2.26 × 10-4), suggesting a causal relationship between plasma TG levels and ASCVDs, including coronary artery disease (OR 1.33, 95% CI 1.24-1.43, p = 2.47 × 10-13). We also identified 12 disease traits that were Bonferroni-significant in the discovery or replication analysis and at least nominally significant in the other analysis (p < 0.05), identifying plasma TG levels as a novel potential risk factor for 9 non-ASCVD diseases, including uterine leiomyoma (OR 1.19, 95% CI 1.10-1.29, p = 1.17 × 10-5).

    Conclusions: Taking a phenome-wide, two-sample MR approach, we identified causal associations between plasma TG levels and 19 disease traits across organ systems. Our findings suggest unrealized drug repurposing opportunities or adverse effects related to approved and emerging TG-lowering agents, as well as mechanistic insights for future studies.

    Funding: RD is supported by the National Institute of General Medical Sciences of the National Institutes of Health (NIH) (R35-GM124836) and the National Heart, Lung, and Blood Institute of the NIH (R01-HL139865 and R01-HL155915).

    1. Medicine
    Astha Thakkar, Kith Pradhan ... Balazs Halmos
    Research Article

    Background: Cancer patients show increased morbidity with COVID-19 and need effective immunization strategies. Many healthcare regulatory agencies recommend administering 'booster' doses of COVID-19 vaccines beyond the standard 2-dose series, for this group of patients. Therefore, studying the efficacy of these additional vaccine doses against SARS-CoV-2 and variants of concern is of utmost importance in this immunocompromised patient population.

    Methods: We conducted a prospective single arm clinical trial enrolling patients with cancer that had received two doses of mRNA or one dose of AD26.CoV2.S vaccine and administered a 3rd dose of mRNA vaccine. We further enrolled patients that had no or low responses to three mRNA COVID vaccines and assessed the efficacy of a 4th dose of mRNA vaccine. Efficacy was assessed by changes in anti-spike antibody, T-cell activity and neutralization activity were again assessed at baseline and 4 weeks.

    Results: We demonstrate that a 3rd dose of COVID-19 vaccine leads to seroconversion in 57% of patients that were seronegative after primary vaccination series. The immune response is durable as assessed by anti-S antibody titers, T-cell activity and neutralization activity against wild-type SARS-CoV2 and BA1.1.529 at 6 months of follow up. A subset of severely immunocompromised hematologic malignancy patients that were unable to mount an adequate immune response (titer <1000 AU/mL) after the 3rd dose and were treated with a 4th dose in a prospective clinical trial which led to adequate immune-boost in 67% of patients. Low baseline IgM levels and CD19 counts were associated with inadequate seroconversion. Booster doses induced limited neutralization activity against the Omicron variant.

    Conclusions: These results indicate that 3rd dose of COVID vaccine induces durable immunity in cancer patients and an additional dose can further stimulate immunity in a subset of patients with inadequate response.

    Funding: Leukemia lymphoma society, National Cancer Institute.

    Clinical trial identifier: NCT05016622.