Posters
Presenting Author Academic/Professional Position
Staff
Academic/Professional Position (Other)
Data Scientist
Academic Level (Author 1)
Staff
Discipline/Specialty (Author 1)
Population Health and Biostatistics
Academic Level (Author 2)
Staff
Academic Level (Author 3)
Staff
Academic Level (Author 4)
Staff
Academic Level (Author 5)
Staff
Presentation Type
Poster
Discipline Track
Biomedical ENGR/Technology/Computation
Abstract Type
Research/Clinical
Abstract
Background High-resolution tissue microarray (TMA) technology allows prompt molecular profiling of multiple tissue specimens, making it ideal for analyzing candidate biomarkers quickly and effectively [1, 2]. Integrating TMA with digital pathology and machine learning enhances high-throughput, cost-effective studies, offering advanced image analysis and improved diagnostic accuracy. MUC13 (Mucin 13) is a transmembrane glycoprotein frequently overexpressed in colorectal cancer (CRC) [3]. MUC13 contributes to colonic tumorigenesis, progression and metastasis [4, 5], making it an attractive target for antibody-guided radiotheranostics in CRC. This study investigates the expression pattern of MUC13 and its association with patients' clinical characteristics in primary and metastatic CRC using this integrated approach. The goal is to identify patient populations likely to benefit from a MUC13-targeted radiotheranostics approach.
Methods Colorectal cancer TMAs spotted with 32 cases/96 cores/per slide (primary CRC, n=98) and with 60 cases/120 cores/per slide (liver metastasis, n=120) were provided by MDACC Research Histology Core Laboratory. Immunohistochemical staining (IHC) for MUC13 protein expression in TMAs was performed by avidin-biotin complex (ABC) method. The digital pathology-based quantitative analysis of MUC13 immunoreactivity involves preprocessing TMA IHC images, detecting and classifying cells or stroma, and analyzing cells staining intensity. Cells were detected using an adjusted intensity threshold and a cell expansion parameter. The classification model for tumor, stroma, and background classes was trained using three classifiers: Deep Neural Network (DNN), Random Tree (RT), K-Nearest Neighbors (KNN ). The Random Tree was ultimately selected based on a comparison of histological structures. Staining intensity was categorized into four levels, negative, weak, moderate, and strong for calculating the H-score. We correlated the H-score data with patients corresponding survival outcomes to gain deeper insights into prognostic significance.
ResultsThe digital pathology analysis of a high-resolution TMAs images dataset identified 39 primary CRC and 48 liver metastasis patient samples. The H-score analysis revealed significant differences between liver metastases (110.9±35.68, Mean±SD) and primary CRC samples (65.91±56.49, P
Conclusions Digital pathology approach provided accurate and efficient analysis of a high-resolution TMA dataset comprised of over 160 samples. The heterogeneity in H-scores underscores the complexity of tumor biology and the need for personalized cancer management. MUC13 expression shows potential as a prognostic biomarker, with distinct implications for primary CRC and liver metastases.
Recommended Citation
Huang, Beibei; Yamaguchi, Aiko; Wang, Jianbo; Sharma, Shilpa; Liu, Zhiwen; and Manning, Henry Charles, "Machine Learning Quantification of High-Resolution Tissue Microarray (TMA) Image on MUC13 IHC Analysis" (2025). Research Symposium. 31.
https://scholarworks.utrgv.edu/somrs/2025/posters/31
Included in
Machine Learning Quantification of High-Resolution Tissue Microarray (TMA) Image on MUC13 IHC Analysis
Background High-resolution tissue microarray (TMA) technology allows prompt molecular profiling of multiple tissue specimens, making it ideal for analyzing candidate biomarkers quickly and effectively [1, 2]. Integrating TMA with digital pathology and machine learning enhances high-throughput, cost-effective studies, offering advanced image analysis and improved diagnostic accuracy. MUC13 (Mucin 13) is a transmembrane glycoprotein frequently overexpressed in colorectal cancer (CRC) [3]. MUC13 contributes to colonic tumorigenesis, progression and metastasis [4, 5], making it an attractive target for antibody-guided radiotheranostics in CRC. This study investigates the expression pattern of MUC13 and its association with patients' clinical characteristics in primary and metastatic CRC using this integrated approach. The goal is to identify patient populations likely to benefit from a MUC13-targeted radiotheranostics approach.
Methods Colorectal cancer TMAs spotted with 32 cases/96 cores/per slide (primary CRC, n=98) and with 60 cases/120 cores/per slide (liver metastasis, n=120) were provided by MDACC Research Histology Core Laboratory. Immunohistochemical staining (IHC) for MUC13 protein expression in TMAs was performed by avidin-biotin complex (ABC) method. The digital pathology-based quantitative analysis of MUC13 immunoreactivity involves preprocessing TMA IHC images, detecting and classifying cells or stroma, and analyzing cells staining intensity. Cells were detected using an adjusted intensity threshold and a cell expansion parameter. The classification model for tumor, stroma, and background classes was trained using three classifiers: Deep Neural Network (DNN), Random Tree (RT), K-Nearest Neighbors (KNN ). The Random Tree was ultimately selected based on a comparison of histological structures. Staining intensity was categorized into four levels, negative, weak, moderate, and strong for calculating the H-score. We correlated the H-score data with patients corresponding survival outcomes to gain deeper insights into prognostic significance.
ResultsThe digital pathology analysis of a high-resolution TMAs images dataset identified 39 primary CRC and 48 liver metastasis patient samples. The H-score analysis revealed significant differences between liver metastases (110.9±35.68, Mean±SD) and primary CRC samples (65.91±56.49, P
Conclusions Digital pathology approach provided accurate and efficient analysis of a high-resolution TMA dataset comprised of over 160 samples. The heterogeneity in H-scores underscores the complexity of tumor biology and the need for personalized cancer management. MUC13 expression shows potential as a prognostic biomarker, with distinct implications for primary CRC and liver metastases.
