ScholarWorksIndianapolis
  • Communities & Collections
  • Browse ScholarWorks
  • English
  • Català
  • Čeština
  • Deutsch
  • Español
  • Français
  • Gàidhlig
  • Italiano
  • Latviešu
  • Magyar
  • Nederlands
  • Polski
  • Português
  • Português do Brasil
  • Suomi
  • Svenska
  • Türkçe
  • Tiếng Việt
  • Қазақ
  • বাংলা
  • हिंदी
  • Ελληνικά
  • Yкраї́нська
  • Log In
    or
    New user? Click here to register.Have you forgotten your password?
  1. Home
  2. Browse by Author

Browsing by Author "Couetil, Justin"

Now showing 1 - 2 of 2
Results Per Page
Sort Options
  • Loading...
    Thumbnail Image
    Item
    302 Diagnostic Evidence Gauge of Spatial Transcriptomics (DEGAS-ST): Using transfer learning to map clinical data to spatial transcriptomics in prostate cancer
    (Cambridge University Press, 2024-04-03) Couetil, Justin; Alomari, Ahmed K.; Zhang, Jie; Huang, Kun; Johnson, Travis S.; Medical and Molecular Genetics, School of Medicine
    OBJECTIVES/GOALS: The 'field effect' is a concept in pathology that pre-malignant tissue changes forecast health. Spatial transcriptomics could detect these changes earlier than histopathology, suggesting new early cancer screening methods. Knowing how normal tissue damage relates to cancer’s origin and progression may improve long-term outcomes. METHODS/STUDY POPULATION: We trained DEGAS, our machine learning framework, with prostate cancer data, combining both general cancer patterns and in-depth genetic information from individual tumors. The Tumor Cancer Genome Atlas (TCGA) shows how gene patterns in tumors relate to patient outcomes, emphasizing the differences between tumors from different patients (intertumor). On the other hand, spatial transcriptomics (ST) shows the genetic variety within a single tumor (intratumor) but has limited samples, making it hard to know which genetic differences are important for treatment. DEGAS bridges these areas by finding tissue sections that resemble those in TCGA profiles and are key indicators of patient survival. DEGAS serves as a valuable tool for generating clinically-important hypotheses. RESULTS/ANTICIPATED RESULTS: DEGAS identified benign-appearing glands in a normal prostate as being highly associated with poor progression-free survival. These glands have transcriptional signatures similar to high-grade prostate cancer. We confirmed this finding in a separate prostate cancer ST dataset. By integrating single cell (SC) data we demonstrated that cells annotated as cancerous in the SC data map to regions of benign glands in the ST dataset. We pinpoint several genes, chiefly Microseminoprotein-β (MSMB, PSP94), where reduced expression is highly correlated with poor progression-free survival. Cell type specific differential expression analysis further revealed that loss of MSMB expression associated with poor outcomes occurs specifically in luminal epithelia, the putative progenitor of prostate cancer. DISCUSSION/SIGNIFICANCE: DEGAS reveals that normal-appearing tissue can be highly-associated with tumor progression and underscores the importance of the 'field effect' in cancer research. Traditional analysis may miss such nuance, hiding key transitional cell states. Validating gene markers could boost early cancer detection and understanding of metastasis.
  • Loading...
    Thumbnail Image
    Item
    Predicting melanoma survival and metastasis with interpretable histopathological features and machine learning models
    (Frontiers Media, 2023-01-06) Couetil, Justin; Liu, Ziyu; Huang, Kun; Zhang, Jie; Alomari, Ahmed K.; Medical and Molecular Genetics, School of Medicine
    Introduction: Melanoma is the fifth most common cancer in US, and the incidence is increasing 1.4% annually. The overall survival rate for early-stage disease is 99.4%. However, melanoma can recur years later (in the same region of the body or as distant metastasis), and results in a dramatically lower survival rate. Currently there is no reliable method to predict tumor recurrence and metastasis on early primary tumor histological images. Methods: To identify rapid, accurate, and cost-effective predictors of metastasis and survival, in this work, we applied various interpretable machine learning approaches to analyze melanoma histopathological H&E images. The result is a set of image features that can help clinicians identify high-risk-of-metastasis patients for increased clinical follow-up and precision treatment. We use simple models (i.e., logarithmic classification and KNN) and "human-interpretable" measures of cell morphology and tissue architecture (e.g., cell size, staining intensity, and cell density) to predict the melanoma survival on public and local Stage I-III cohorts as well as the metastasis risk on a local cohort. Results: We use penalized survival regression to limit features available to downstream classifiers and investigate the utility of convolutional neural networks in isolating tumor regions to focus morphology extraction on only the tumor region. This approach allows us to predict survival and metastasis with a maximum F1 score of 0.72 and 0.73, respectively, and to visualize several high-risk cell morphologies. Discussion: This lays the foundation for future work, which will focus on using our interpretable pipeline to predict metastasis in Stage I & II melanoma.
About IU Indianapolis ScholarWorks
  • Accessibility
  • Privacy Notice
  • Copyright © 2025 The Trustees of Indiana University