Utilizing Deep Learning Methods for Cancer Detection through Analysis of MicroRNA Expression Profiles

Abstract

Integration of cutting-edge computational methods and genomic data analysis has become crucial in the quest for early cancer diagnosis and enhanced diagnostic accuracy. The genomic sequences of microRNAs (miRNAs), which are important cancer biomarkers, provide important information for this. In this study, we propose a novel deep learning-based framework for cancer detection with a focus on FNNs and a hybrid DNN model with an accuracy of over 90.7%. Our method aims to identify detailed genomic patterns and features that improve the sensitivity and specificity of cancer detection by painstakingly curating and preprocessing large miRNA datasets gathered from various patient cohorts. This research sets the stage for further exploration of deep learning methodologies within the context of miRNA-based cancer detection, promising advancements in personalized diagnosis and prognosis. Our method aims to identify detailed genomic patterns and features that improve the sensitivity and specificity of cancer detection by painstakingly curating and preprocessing large miRNA datasets gathered from various patient cohorts. Our approach seeks to improve sensitivity and specificity by deciphering complex genetic patterns. By utilizing these datasets, we show off the effectiveness of our model and its clinical potential, giving an accuracy of 90.7% for our Hybrid Feedforward and Dense Neural Network model as compared to current state of the art machine learning models. This research promises revolutionary advances in customized oncology, providing a route towards improved diagnostic accuracy and early intervention. It also proves that miRNA expressions values are not sequential in nature. It also lays the groundwork for the development of deep learning in miRNA-based cancer detection. © 2024 IEEE.