MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202611071731 A) filed by Shrawan Kumar Sharma; Ravinder; Priyanka Parihar; Ashish Sen; Sonali Rahul Dhave; and Dr. Rahul Kumar on June 10, 2026, for An Intelligent Multi-Stage Framework For Automated Lung Cancer Diagnosis Using Digital Pathology And Class Dependency Modelling.
Inventors include Ravinder; Shrawan Kumar Sharma; Priyanka Parihar; Ashish Sen; Sonali Rahul Dhave; and Dr. Rahul Kumar.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: It is a computer-implemented device and procedure of automated classification and subtyping of lung cancer based on the whole-slides histopathology images. The system digests hairy tissue slides by challenging patches to a pathology-specific feature extractor and then combining features to a multi-branch multi-instance learning structure. Class dependency modelling module allows the sharing of information between more than one classifier of cancer -subtypes and in this way, the system will be able to establish clinically significant dependencies between the lung-cancer subtypes and the morphological patterns. Also, the system facilitates mixed supervision to provide diagnostically significant regions of tissue through limited region-level annotations, which are optional. The invention produces slide-level diagnostic results consisting of cancer subtype estimates, presence results concerning particular patterns, and confidence measures. The invention can improve the accuracy, consistency and interpretability of diagnostic results and alleviate the burden on annotation because it synchronizes the computational workflow with the current workflow in pathology. Clinical decision support, cancer screening programs and efforts on large scale pathology research are all suited to this system, which can be combined with other modalities of medical imaging to assist in multimodal diagnostic pipelines.
Disclaimer: Curated by HT Syndication.
