MUMBAI, India, July 11 -- Intellectual Property India has published a patent application (202521059662 A) filed by Mrs. Swati Anand Powar; and Dr. Pallavi Jha, Pune, Maharashtra, on June 21, for 'hybrid fused survival prediction model using pan cancer gene expression data with kernel-based clustering process.'
Inventor(s) include Mrs. Swati Anand Powar; and Dr. Pallavi Jha.
The application for the patent was published on July 11, under issue no. 28/2025.
According to the abstract released by the Intellectual Property India: "Hybrid Fused Survival prediction model using PAN cancer Gene Expression data with Kernel-based Clustering Process Abstract The present invention discloses a Hybrid Fused Survival Prediction System designed to enhance the accuracy and reliability of survival prediction in cancer patients using PAN-cancer gene expression data. The system addresses critical limitations in conventional diagnostic frameworks, particularly the challenge of handling high-dimensional gene expression data with a relatively low number of samples, which often results in low predictive performance and inconsistent clinical outcomes. The proposed system integrates a modified kernel-based Fuzzy C-Means (FCM) clustering technique for optimal gene selection, significantly improving the identification of prognostically relevant gene subsets. An advanced data transformation module is incorporated during preprocessing to refine the quality and consistency of input data. Following gene selection, the invention utilizes a robust feature extraction mechanism that captures both statistical and technical indicators, including Exponential Moving Average (EMA), Relative Strength Index (RSI), and Average True Range, which are vital in capturing trends and variability in gene activity. The core predictive engine of the system is a hybrid deep learning architecture comprising a Parallel Convolutional Neural Network (PCNN) and a Bidirectional Long Short-Term Memory (Bi-LSTM) network, enabling effective learning of spatial and temporal patterns inherent in gene expression profiles. To further enhance prediction robustness, an improved score-level fusion technique is introduced to aggregate outputs from the dual models, yielding a final survival prediction score. This invention is implemented in a computational environment and evaluated using PAN-cancer datasets. The system demonstrates superior performance over traditional survival models, particularly in minimizing Type I and Type II errors. The invention has strong implications in the field of precision medicine, offering a scalable and intelligent solution for early cancer prognosis, personalized treatment planning, and improved patient survival outcomes."
Disclaimer: Curated by HT Syndication.
