About

Ashwini Esnakula is not listed with a detailed biography or research focus on the provided page. The page primarily contains a list of team members, including faculty, research scientists, graduate researchers, and undergraduate researchers, with no specific biographical or research information about Ashwini Esnakula.

Research topics

• Internal medicine
• Pathology
• Medicine
• Biology
• Cancer research
• Computer Science
• Radiology
• Oncology
• Computer vision
• Molecular biology
• Immunology
• Genetics

Selected publications

• Ectopic Splenic Tissue: Diagnostic Challenges in Limited Samples and Intraoperative Consultations

  International Journal of Surgical Pathology · 2026-02-25

  article

  The histologic diagnosis of ectopic splenic tissue can be challenging in limited biopsy and intraoperative samples. Accurate recognition is crucial to avoid misdiagnosis and guide appropriate patient management. This multi-institutional study of 41 lesions describes the clinicopathologic features of ectopic splenic tissue, with emphasis on diagnostic pitfalls. Fourteen patients (36%) had a history of malignancy-10 solid tumors, 3 hematologic neoplasms, and 1 with both. Notably, 1 patient demonstrated a superimposed low-grade B-cell lymphoma arising within ectopic splenic tissue. Thirteen had prior splenectomy. Among 35 specimens with available imaging, all demonstrated mass lesions, yet ectopic splenic tissue was suspected in only 17%. Initial radiologic impressions often favored neoplastic processes, including neuroendocrine tumor (3 patients), metastasis (1), and hepatocellular adenoma or carcinoma (3). Histologically, ectopic splenic tissue exhibited a broad morphologic spectrum, often mimicking inflammatory or neoplastic conditions. Fourteen specimens showed white pulp-predominant patterns resembling lymphoid tissue. Seven demonstrated mixed neutrophilic and lymphohistiocytic infiltrates suggestive of abscess or other inflammatory processes. Five specimens with red pulp predominance closely resembled hemorrhagic or vascular tumors. Eight exhibited solid or nested growth of monotonous epithelioid cells, mimicking well-differentiated neuroendocrine tumors. One showed mixed inflammatory infiltrates with fibroblastic proliferation, resembling an inflammatory myofibroblastic tumor. CD8 immunohistochemical staining consistently highlighted sinusoidal endothelial cells, providing key diagnostic confirmation. In conclusion, ectopic splenic tissue presents a wide range of morphologic appearances that can pose diagnostic challenges, especially in limited samples. Recognition of these patterns and use of CD8 immunostaining are essential to avoid misdiagnosis and unnecessary interventions.

  PublisherDOI

• Inferring Clinically Relevant Molecular Subtypes of Pancreatic Cancer from Routine Histopathology Using Deep Learning

  Research Square · 2026-01-16

  preprintOpen access
  PublisherOA PDFDOI

• Abstract 2743: Deploying artificial intelligence driven digital pathology for real world clinical decision-making in pancreatic cancer.

  Cancer Research · 2026-04-03

  article

  Abstract Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest malignancies, driven by limited therapeutic options and the absence of widely implemented, clinically actionable biomarkers. Transcriptomic subtyping, particularly the classical (CL) versus basal-like (BL) Moffitt classification, offers prognostic and predictive value: CL tumors show improved outcomes and greater sensitivity to 5 fluorouracil based regimens (e.g., FOLFIRINOX). However, BL tumors exhibit poor outcomes across treatment regimens and may benefit from clinical trial prioritization or intensified oversight. But routine use of RNA-based subtyping is hindered by cost, turnaround time, and restricted access to commercial assays such as Purity Independent Subtyping of Tumors (PurIST). To overcome these barriers, we developed a deep learning model that infers PDAC molecular subtypes directly from hematoxylin and eosin (H&E) whole slide images (WSIs) by integrating with matched RNA-sequencing data for supervised training. In this initial iteration, WSIs from 126 Pancreatic Cancer Action Network (PANCAN) patients with high-quality slides and confirmed CL or BL transcriptomic profiles were curated. Bulk RNA-seq underwent standardized preprocessing, including quality control, alignment, normalization, and quantification of Moffitt-derived gene signatures to generate high-confidence molecular labels for supervised training. These annotations served as ground truth for supervised training. Using five-fold cross-validation, the model classified PDAC tumors into CL and BL subtypes with strong performance (area under curve, AUC: 0.83; accuracy: 77%; specificity: 80%; sensitivity: 72%), comparable to existing image-based PurIST subtyping literature (AUC 0.83-0.86). Our ongoing work with larger multi-institutional datasets aims to further enhance accuracy and generalizability. This proof of concept establishes the feasibility of AI-driven digital pathology for rapid, scalable Moffitt PDAC molecular subtyping directly from WSIs of routine H&E slides. By eliminating the need for RNA-based assays, this approach offers a cost-effective and scalable alternative, particularly valuable for real-world and resource-limited clinical settings. Prospective validation studies will be crucial for refining performance, assessing clinical utility, and enabling integration into precision oncology workflows for PDAC treatment. Citation Format: Ashish Manne, Alejandro Leya, Abdul Rehman Akbar, Upender Manne, Anne Noonan, Anup Kasi, Ashwini Esnakula, Ravi Paluri, Anil Vasdev Parwani, Muhammad Khalid Khan Niazi. Deploying artificial intelligence driven digital pathology for real world clinical decision-making in pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2743.

  PublisherDOI

• 1381 Hepatic Epstein-Barr Virus–Associated Smooth Muscle Tumors: Insights from a Multi-Institutional Cohort

  Laboratory Investigation · 2026-03-01

  article
  PublisherDOI

• Inferring Clinically Relevant Molecular Subtypes of Pancreatic Cancer from Routine Histopathology Using Deep Learning

  medRxiv · 2026-01-08 · 1 citations

  articleOpen access

  Abstract Background and aims Molecular subtyping of pancreatic ductal adenocarcinoma (PDAC) into basal-like and classical has established prognostic and predictive value. However, its use in clinical practice is limited by cost, turnaround time, and tissue requirements, thereby restricting its application in the management of PDAC. We introduce PanSubNet (PANcreatic SUBtyping NETwork), an interpretable deep learning framework that predicts therapy-relevant molecular subtypes directly from standard hematoxylin and eosin (H&E)-stained whole-slide images. Methods PanSubNet was developed using data from 1 , 055 patients across two multi-institutional cohorts (PANCAN, n=846; TCGA, n=209) with paired histology and RNA sequencing data. Ground-truth labels were derived using the validated Moffitt 50-gene signature refined by GATA6 expression. The model employs dual-scale architecture that fuses cellular-level morphology with tissue-level architecture, leveraging attention mechanisms for multi-scale representation learning and transparent feature attribution. Results On internal validation within PANCAN using five-fold cross-validation, PanSubNet achieved mean area under the receiver operating characteristic curve (AUC) of 88.5% in high-confidence cases, with balanced sensitivity and specificity. External validation on the independent TCGA cohort without fine-tuning demonstrated robust generalizability (AUC 84.0% ). PanSubNet preserved and, in metastatic disease, strengthened prognostic stratification compared to RNA-seq–based labels. Prediction uncertainty linked to intermediate transcriptional states, not classification noise. Model predictions are aligned with established transcriptomic programs, differentiation markers, and DNA damage repair signatures. Conclusions By enabling rapid, cost-effective molecular stratification from routine H&E-stained slides, PanSubNet offers a clinically deployable and interpretable tool for genetic subtyping. We are gathering data from two institutions to validate and assess real-world performance, supporting integration into digital pathology workflows and advancing precision oncology for PDAC.

  PublisherOA PDFDOI

• Inferring Clinically Relevant Molecular Subtypes of Pancreatic Cancer from Routine Histopathology Using Deep Learning

  arXiv (Cornell University) · 2026-01-06

  preprintOpen access

  Molecular subtyping of PDAC into basal-like and classical has established prognostic and predictive value. However, its use in clinical practice is limited by cost, turnaround time, and tissue requirements, thereby restricting its application in the management of PDAC. We introduce PanSubNet, an interpretable deep learning framework that predicts therapy-relevant molecular subtypes directly from standard H&E-stained WSIs. PanSubNet was developed using data from 1,055 patients across two multi-institutional cohorts (PANCAN, n=846; TCGA, n=209) with paired histology and RNA-seq data. Ground-truth labels were derived using the validated Moffitt 50-gene signature refined by GATA6 expression. The model employs dual-scale architecture that fuses cellular-level morphology with tissue-level architecture, leveraging attention mechanisms for multi-scale representation learning and transparent feature attribution. On internal validation within PANCAN using five-fold cross-validation, PanSubNet achieved mean AUC of 88.5% with balanced sensitivity and specificity. External validation on the independent TCGA cohort without fine-tuning demonstrated robust generalizability (AUC 84.0%). PanSubNet preserved and, in metastatic disease, strengthened prognostic stratification compared to RNA-seq based labels. Prediction uncertainty linked to intermediate transcriptional states, not classification noise. Model predictions are aligned with established transcriptomic programs, differentiation markers, and DNA damage repair signatures. By enabling rapid, cost-effective molecular stratification from routine H&E-stained slides, PanSubNet offers a clinically deployable and interpretable tool for genetic subtyping. We are gathering data from two institutions to validate and assess real-world performance, supporting integration into digital pathology workflows and advancing precision oncology for PDAC.

  PublisherDOI

• Inferring Clinically Relevant Molecular Subtypes of Pancreatic Cancer from Routine Histopathology Using Deep Learning

  ArXiv.org · 2026-01-06

  articleOpen access

  Molecular subtyping of PDAC into basal-like and classical has established prognostic and predictive value. However, its use in clinical practice is limited by cost, turnaround time, and tissue requirements, thereby restricting its application in the management of PDAC. We introduce PanSubNet, an interpretable deep learning framework that predicts therapy-relevant molecular subtypes directly from standard H&E-stained WSIs. PanSubNet was developed using data from 1,055 patients across two multi-institutional cohorts (PANCAN, n=846; TCGA, n=209) with paired histology and RNA-seq data. Ground-truth labels were derived using the validated Moffitt 50-gene signature refined by GATA6 expression. The model employs dual-scale architecture that fuses cellular-level morphology with tissue-level architecture, leveraging attention mechanisms for multi-scale representation learning and transparent feature attribution. On internal validation within PANCAN using five-fold cross-validation, PanSubNet achieved mean AUC of 88.5% with balanced sensitivity and specificity. External validation on the independent TCGA cohort without fine-tuning demonstrated robust generalizability (AUC 84.0%). PanSubNet preserved and, in metastatic disease, strengthened prognostic stratification compared to RNA-seq based labels. Prediction uncertainty linked to intermediate transcriptional states, not classification noise. Model predictions are aligned with established transcriptomic programs, differentiation markers, and DNA damage repair signatures. By enabling rapid, cost-effective molecular stratification from routine H&E-stained slides, PanSubNet offers a clinically deployable and interpretable tool for genetic subtyping. We are gathering data from two institutions to validate and assess real-world performance, supporting integration into digital pathology workflows and advancing precision oncology for PDAC.

  PublisherOA PDF

• Tubular Adenoma With Squamoid Morules Versus Composite Adenoma-Microcarcinoid

  Elsevier eBooks · 2025-10-25

  book-chapterSenior author
  PublisherDOI

• DNA methylation signatures as predictive biomarkers for chemotherapy (CT) resistance and survival in pancreatic ductal adenocarcinoma (PDA).

  Journal of Clinical Oncology · 2025-05-28

  article

  4151 Background: Predicting innate treatment resistance to traditional CT in PDA can optimize therapeutic strategies and improve patient outcomes. This study investigates methylation changes in genes encoding proteins implicated in preclinical models (PDA cell lines or mouse models) influencing drugs commonly used to treat PDA, including 5-fluorouracil (5FU), oxaliplatin, irinotecan, gemcitabine (Gem), nanoliposomal irinotecan, and nab-paclitaxel. Using a comprehensive literature review (1970–2024), we curated a panel of relevant genes and analyzed their methylation patterns in The Cancer Genome Atlas (TCGA) database. We hypothesized that DNA methylation changes affect gene expression and protein production, contributing to chemotherapy resistance. Methods: PDA patient methylation data were accessed from the TCGA database. Survival analyses were performed using elastic net multivariate regression to identify significant methylation signatures, followed by Kaplan-Meier analysis. Model parameters, including alpha (α) and lambda (λ), were optimized through 100 iterations to minimize error. Our curated panel consisted of 138 genes, predominantly Gem-specific or Gem + 5FU (n = 93). Results: The TCGA database provided methylation data for 184 PDA patients (106 had Gem or Gem-based therapy), with 133/138 genes in our analysis. Our analysis identified 23 cytosines followed by guanine residue (CpG) methylation signatures within the panel, ranging from 1 to 23 CpG sites. The best-performing signature, containing 21 CpG sites, stratified patients into significantly different survival groups (17 months (m) vs. not evaluable, p = 0.004). The second-best signature, with 8 CpG sites, stratified survival as 17m vs. 66.94m (p = 0.03). Interestingly, signatures with the most (n = 23) and least (n = 1) CpG sites also demonstrated strong stratification, with survival differences of 15.15m vs. 30.02m (p = 0.01) and 18.67m vs. 44.38m (p = 0.01), respectively. Many signatures included multiple CpG sites from single genes. A Gem or Gem + 5FU-specific panel (n = 93) applied to patients treated with Gem-based therapy identified an 8-CpG signature distinguishing high-risk patients (20.84m vs. 49.38m, p = 0.02). Conclusions: This study highlights the potential of CpG methylation signatures to predict treatment outcomes in PDA. These findings may guide the identification of high-risk patients and the optimization of CT regimens for improved survival. The identification of methylation signatures associated with genes implicated in chemotherapy resistance provides valuable insights into the underlying mechanisms of innate treatment resistance in PDA. Further validation of these methylation signatures could contribute to more effective and targeted therapeutic approaches in clinical practice.

  PublisherDOI

• Endoscopic Mucosal Resection and Endoscopic Submucosal Dissection Reporting of Adenocarcinoma in Polyp

  Elsevier eBooks · 2025-10-25

  book-chapter1st authorCorresponding
  PublisherDOI

Frequent coauthors

• Jinping Lai

  66 shared

• Xiuli Liu

  66 shared

• Dengfeng Cao

  Washington University in St. Louis

  65 shared

• Joeffrey Chahine

  Brigham and Women's Hospital

  65 shared

• Daniel García Sánchez

  Hospital General Universitario de Elche

  65 shared

• Ashwin Akki

  Cleveland Clinic

  53 shared

• Lina Abdul Karim

  Washington University in St. Louis

  50 shared

• Ian S. Hagemann

  50 shared

Labs

• AI4PathPI

  From Pixels to Prognosis: AI in Action!

Education

• Other

  Kasturba Medical College, Mangalore, India

  2000

• M.S., Tumor Biology

  Georgetown University

  2007

• Other, Anatomic and Clinical Pathology

  Howard University Hospital

  2009

• Other, Surgical Pathology

  Washington University/Barnes Jewish Hospital

  2013

• Other, Gastrointestinal, Hepatobiliary, and Pancreatic Pathology

  Cleveland Clinic

  2014

Awards & honors

• 2013 Anatomic and Clinical Pathology, American Board of Path…