Intellectual Property India Publishes Patent Application for Artificial Intelligence-Based Two-Stage Hybrid Diagnostic System For Early Detection Of Pancreatic Ductal Adenocarcinoma Using Sequential Biomarker Screening And Multimodal Imaging Analysis

MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641071780 A) filed by Sr University on June 09, 2026, for Artificial Intelligence-Based Two-Stage Hybrid Diagnostic System For Early Detection Of Pancreatic Ductal Adenocarcinoma Using Sequential Biomarker Screening And Multimodal Imaging Analysis.

Inventors include Dr. H S Saraswathi; Dr. Subhashis Das; Kotaramma T S; Navya K G; Kaveri C; and Madhuri Deekshith S.

The application for the patent was published on June 19, 2026, under issue no. 25/2026.

Abstract: ARTIFICIAL INTELLIGENCE-BASED TWO-STAGE HYBRID DIAGNOSTIC SYSTEM FOR EARLY DETECTION OF PANCREATIC DUCTAL ADENOCARCINOMA USING SEQUENTIAL BIOMARKER SCREENING AND MULTIMODAL IMAGING ANALYSIS The present invention discloses an artificial intelligence-based two-stage hybrid diagnostic system for the early detection of Pancreatic Ductal Adenocarcinoma (PDAC). The system operates through a sequential pipeline that first evaluates patient biochemical biomarkers including CA 19 9, CEA, CA 125, CRP, and IL 6 using a machine learning screening module to generate a probabilistic risk score. Patients classified as high-risk are forwarded to the second stage, where multimodal imaging data such as CT, MRCP, or EUS is acquired and preprocessed for deep learning feature extraction using convolutional neural networks. The biomarker-derived risk representation and imaging-derived features are fused in a hybrid classification module to produce a composite PDAC likelihood score. An explainability layer generates SHAP plots for biomarker contributions and Grad CAM overlays for imaging features, thereby enhancing interpretability and clinician trust. The final output is a structured clinical report comprising risk scores, annotated imaging heatmaps, and recommended clinical actions. This invention significantly improves sensitivity and specificity for early PDAC detection compared to unimodal approaches, while reducing unnecessary imaging procedures and healthcare costs.

Disclaimer: Curated by HT Syndication.