Genomic Mastery: CNN-Driven Prognostic Detection in Mantle Cell Lymphoma

Abstract

Deep learning techniques are crucial in biomedical research, particularly in analyzing genomic data. Our research aims to overcome limitations of existing prognostic models for Mantle Cell Lymphoma (MCL) by introducing a CNN -based model trained on the entire genomic dataset and clinical data. The model is aimed at tissue and mutation specific diagnosis of MCL cancer thus resulting into increased diagnostic accuracy of the prognostic estimate obtained, which also increases the volume of the data required for correct medical decision making. The work applies convolutional neural network to various patient populations and clinical settings by assessing robustness and generalization properties. The dialogue with the CNN-based model should be explained in a way that is easily understandable. Our work starts with the creation of a genomic database through which a CNN-type model is trained, gradually improving the prediction, assessing the metrics of model performance, comparing it with what is already in the genomics data field, testing the stability and applicability of our models, interpreting the results of model validation, and taking the MCL research to the next level by outperforming the previous works in accuracy.