MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641071448 A) filed by Chaitanya Bharathi Institute Of Technology on June 09, 2026, for Federated Explainable Transformer-Based Multimodal Analytics Framework For Privacy- Preserving Early Blood Cancer Prognosis Using Distributed Clinical Data.
Inventors include Dr. Y. Rama Devi; Dr. K. Mary Sudha Rani; Dr. T. Sridevi; Dr. Y. C. A. Padmanabha Reddy; Dr. Garlapati Narayana; Dr. B. Sateesh; Smt. K. Swathi; Ms. G. Kavita; Dr. Ch. Indirapriyadarsini; and Dr. D. Naga Jyothi.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: A federated explainable transformer-based multimodal analytics framework for privacy-preserving early blood cancer prognosis using distributed clinical data is disclosed. The framework comprises a plurality of healthcare institution nodes, multimodal data acquisition and preprocessing modules, transformer-based modality encoders, a cross-modal attention fusion engine, a federated learning engine, a privacy-preservation module, a federated aggregation server, an explainability engine, and a prognostic analytics engine. The system processes heterogeneous clinical information including laboratory results, pathology images, genomic data, biomarker profiles, electronic health records, radiological scans, and longitudinal patient monitoring data to generate prognostic predictions without transferring raw patient information outside institutional boundaries. Federated learning enables collaborative model training across distributed healthcare organizations, while privacy- preserving mechanisms ensure secure data utilization. The framework further provides interpretable clinical explanations, risk stratification outputs, disease progression forecasts, survival probability estimations, and treatment-response predictions to support informed clinical decision-making and precision oncology applications.
Disclaimer: Curated by HT Syndication.
