QLC-Net: A Hybrid Quantum Support Vector Machine infused Convolutional Neural Network Approach to Detect Bronchogenic Carcinoma (Lung Cancer) in Pathological Slides

Overview

Recently, medical professionals have been in significant decline due to lack of interest, yet the rising population is increasing this disparity of medical treatment ratio to patients. Especially for periphery regions, there is a lack of equipment and prof essionals ready to make proper diagnoses, oftentimes pathological slides are left for late diagnosis or are even never examined. This raises the question regarding the feasibility of utilizing artificial intelligence for the purpose of pathological slide b iopsy diagnosis. In particular this study was focused on the disease Bronchogenic Carcinoma (Lung Cancer) and the utilization of a Quantum Support vector machine infused Convolutional Neural Network (QC -CNN). The alternative hypothesis was that the QC - CNN would have at least an 80% accuracy in identifying healthy from cancerous, and would be able to identify the subcategory of the cancer. This was proven by the experiment after a multitude of trials with various epoch numbers and learning rates that resulted in a final model and trial with the highest variables consisting of a sample of 200 pathological slides and a raw accuracy of 90% and overall machine accuracy of 93%. The high accuracy of these results help support the advancement of technology, and alth ough it wouldn’t be used immediately for patients, this can still be a valuable tool for educators and students alike who are studying pathology. Furthermore proving its accuracy assists in debunking some ethical concerns of AI, rather than replacing medical professionals this will only be assistance.