Biology subjects

Sukhadia, S. S.

Publications and source records attributed to Sukhadia, S. S..

6 recordsLinked to original sources

bioRxiv2026

Interpreting and Validating a Deep Learning Model Predictive of Spatial Morphologic-Molecular Patterns in Lung Adenocarcinoma, Using Ground Truth Immunohistochemistry

Lung adenocarcinoma (LUAD), the most common subtype of non-small cell lung cancer, exhibits profound histological and molecular heterogeneity. While genomic profiling has identified key oncogenic drivers and immune signatures, its use is limited by cost, technical demands and tissue availability. In addition, spatial transcriptomics provides spatially resolved molecular insights but remains challenging and time-consuming. To address this gap, we developed XpressO-Lung, an explanatory deep learning model that predicts gene expression heterogeneity spatially in tumor and its microenvironment on hematoxylin and eosin based diagnostic (Dx) whole-slide images (WSIs) by learning associations between tissue morphology and the corresponding bulk-transcriptomic data. Utilizing 200 LUAD cases from The Cancer Genome Atlas, XpressO-Lung predicted spatial expression patterns of NAPSA, TP53I3, CD8A, TTF1, KRT7, CDKN2A, FOXO1, KEAP1, RB1 and TP53 on Dx-WSIs with AUCs ranging from 0.64 to 0.92. The predicted spatial gene expression patterns aligned with the known morphologic interactions of the tumor and its microenvironment, capturing biological events directly on Dx-WSIs. These spatio-morpho-molecular associations were further validated using immunohistochemistry on an external set of clinical samples at Dartmouth Health, demonstrating concordance between model-predicted spatial patterns and observed histomorphologic features. By coupling predictive performance with spatial interpretability of gene expression on Dx-WSIs, the XpressO-Lung model bridges histopathology and bulk-transcriptomics, enabling explainable spatio-morpho-genomic analyses to advance biomarker discovery, therapeutic stratification and precision oncology in LUAD.

pathology↗

bioRxiv2026

X-SPATIO: An Explanatory Deep Learning Pipeline for the Prediction and Visualization of Spatially Resolved Biomarker Expression in Triple-Negative Breast Cancer

Histopathologic evaluation remains central to cancer diagnosis and treatment planning, yet the molecular programs underlying distinct tissue morphologies are not routinely accessible in clinical workflows. Spatial transcriptomic/proteomic platforms provide region-specific molecular measurements but are limited by cost, throughput, and scalability. Most computational pathology models rely on either bulk tissue-based gene expression or a focused gene/protein expression-panel prediction, thereby obscuring subregion-specific morpho-molecular relationships and limiting spatial interpretation of a wider gene/protein expression network. This limitation is particularly significant in triple-negative breast cancer (TNBC), which exhibits pronounced spatial heterogeneity across tumor, stroma, and immune compartments. We developed X-SPATIO, a spatially compatible computational pipeline designed to directly link hematoxylin and eosin (H&E) morphology with region-matched mRNA and protein expression. The model was trained on H&E-defined regions of interest paired with spatially-resolved transcriptomic and proteomic data obtained from GeoMx Digital Spatial Profiler. Using a multiple-instance learning approach, X-SPATIO captures morpho-molecular associations, generating spatio-morphologic attention maps that indicate predictive tissue regions. X-SPATIO demonstrated strong performance across biologically relevant spatial biomarkers, achieving area under the curve values ranging [0.79, 0.97]. Attention maps revealed spatial patterns consistent with known biology, indicating alignment between learned features and tissue organization. By integrating spatial molecular ground truth with routine histopathology, X-SPATIO enables cost-effective inference of spatial biomarker expression and establishes a foundation for biologically grounded discovery and precision oncology in TNBC.

pathology↗

bioRxiv2025

Interpretable Deep Learning Reveals Biologically Relevant Spatial Gene Expression Patterns in Lung Tumors and their Microenvironment

Lung adenocarcinoma (LUAD), the most common subtype of non-small cell lung cancer (NSCLC) exhibits profound histological and molecular heterogeneity, hindering accurate prognosis and effective treatment. Current approaches to assess this heterogeneity, such as histopathology, molecular profiling, and spatial transcriptomics are constrained by high costs, long turnaround times, and limited tissue availability, making them challenging for widespread prognostic use. To address this gap, we developed XpressO-Lung, an explanatory deep learning model that predicts gene expression heterogeneity, spatially, in tumor and its microenvironment, on hematoxylin and eosin (H&E) based diagnostic (Dx) whole-slide images (WSIs) by learning associations between tissue morphology and the corresponding bulk-transcriptomic data. Utilizing 200 LUAD cases from The Cancer Genome Atlas (TCGA), XpressO-Lung predicted spatial expression patterns of NAPSA, SLC47A1, TP53I3, KLRB1, FAM189A1, TICAM1, CD8A, CXCL13, TTF, CDH3, KRT7 and CDKN2A genes (biomarkers) on the respective Dx-WSIs with AUCs ranging up to 0.92. More importantly, the predicted spatial gene expression patterns aligned with the known morphologic interactions of the tumor and its microenvironment, capturing biological events directly on Dx-WSIs. By coupling predictive performance with spatial interpretability of gene expression on Dx-WSIs, XpressO-Lung bridges histopathology and bulk-transcriptomics, enabling explainable morpho-genomic analyses to advance biomarker discovery and offer prognostic insights to inform precision oncology in LUAD, especially in low-resource settings.

pathology↗

bioRxiv2025

XpressO-Melanoma: An Explainable Deep Learning Model for the Prediction of BRAF V600 Mutation Status in Cutaneous Melanomas

BRAF V600E mutations are critical oncogenic drivers in cutaneous melanoma, influencing treatment decisions and outcomes. However, conventional molecular assays face limitations, including tissue availability, cost, and access. To address this, we present an explainable deep learning model that predicts BRAF V600E mutation status directly from diagnostic whole-slide images (WSIs) of skin cutaneous melanoma. Using histopathological WSIs from The Cancer Genome Atlas (TCGA) and their corresponding mutation labels (BRAF wildtype vs. BRAF V600E), we trained a weakly supervised deep learning pipeline, XpressO, to identify tumor regions of interest (ROIs) predictive of BRAF mutation status. The model outputs attention heatmaps highlighting spatially relevant diagnostic features and computes a combined probability score from the top ten attention regions per WSI. These regions are further reviewed by a pathologist for biological appropriateness. On an independent test set, the model achieved an AUC of 0.79 with balanced precision and recall, correctly identifying 7 of 8 BRAF V600E mutant cases. This demonstrates the models ability to capture phenotypic correlates of mutation status and highlights the potential of computational pathology in precision oncology. Our approach offers a scalable, interpretable, and cost-effective alternative to molecular testing, particularly in resource-limited settings.

cancer biology↗

bioRxiv2023

Predicting Distant Recurrences in Invasive Breast Carcinoma Patients Using Clinicopathological Data: A cross-institutional and AI-based study

Breast cancer ranks second in the most common cancer in women worldwide with 30% of cases resulting into recurrence of the disease at distant organs post the treatment. While clinicians have utilized several clinicopathological measurements for prediction of distant recurrences in invasive breast carcinoma (IBC), none of those studies have showcased the potential of combining clinicopathological evaluations of IBC tumors pre and post therapies using machine learning (ML) or artificially intelligent (AI) models to predict the distant recurrence of the disease in respective patients. The goal of our study was to determine whether classification-based ML/AI techniques can predict distant recurrences in IBC patients using key clinicopathological measurements that includes pathological staging of tumor and surrounding lymph nodes deemed both pre- and post-neoadjuvant therapy, imaging-based therapy responses, and the status of adjuvant therapy administered to patients. We trained and tested clinicopathological ML/AI model using dataset from Duke University and validated it using external dataset from Dartmouth Hitchcock Medical Center (DHMC). Random Forest (RF) model performed best compared to C-Support Vector Classifier (SVC) and Multi-Layer Perceptron (MLP) yielding AUC ranging 0.75-1.0 (p<0.002) across both the institutions, thereby demonstrating the cross-institutional portability and validity of ML/AI models in the field of clinical research in cancer.

bioinformatics↗

bioRxiv2021

ImaGene: A web-based software platform for tumor radiogenomic evaluation and reporting

The field of radiomics has undergone several advancements in approaches to uncovering hidden quantitative features from tumor imaging data for use in guiding clinical decision-making for cancer patients. Radiographic imaging techniques provide insight into the imaging features of tumor regions of interest (ROIs), while immunohistochemistry and sequencing techniques performed on biopsy samples yield omics data. Potential associations between tumor genotype and phenotype can be identified from imaging and omics data via traditional correlation analysis, as well as through artificial intelligence (AI) models. However, at present the radiogenomics community lacks a unified software platform for which to conduct such analyses in a reproducible manner. To address this gap, we propose ImaGene, a web-based platform that takes tumor omics and imaging data sets as input, performs correlation analysis between them, and constructs AI models (optionally using only those features found to exhibit statistically significant correlation with some element of the opposing dataset). ImaGene has several modifiable configuration parameters, providing users complete control over their analysis. For each run, ImaGene produces a comprehensive report displaying a number of intuitive model diagnostics. To demonstrate the utility of ImaGene, exploratory studies surrounding Invasive Breast Carcinoma (IBC) and Head and Neck Squamous Cell Carcinoma (HNSCC) on datasets acquired from public databases are conducted. Potential associations are identified between several imaging features and 6 genes: CRABP1, SMTNL2, FABP1, HAS1, FAM163A and DSG1 for IBC, and 4 genes: CEACAM6, NANOG, ACSM2B, and UPK2 for HNSCC. In summary, the software provides researchers with a transparent tool for which to begin radiogenomic analysis and explore possible further directions in their research. We anticipate that ImaGene will become the standard platform for tumor analyses in the field of radiogenomics due to its ease of use, flexibility, and reproducibility, and that it can serve as an enabling centrepoint for an emerging radiogenomic knowledge base.

bioinformatics↗