Predicting Breast Cancer Patient Prognosis Using Deep Learning

AJAS · 2022 Computer Science

Dheepthi Mohanraj

Thumbnail supplied by the source for Predicting Breast Cancer Patient Prognosis Using Deep Learning

Overview

Breast cancer is the leading cause of cancer deaths among women worldwide, but early detection and diagnosis can significantly improve a patient’s prognosis. Conventionally, biopsy breast tissue is graded based on morphological and genetic features; lesion subtype, mitotic count, and HER2 (IHC) score are three of the most significant prognostic markers. However, the manual examination of breast tissue slides is time-consuming and subject to inter- and intra-pathologist variability. Also, many developing nations lack the skilled medical professionals necessary to provide thorough and timely diagnoses. Artificial intelligence innovations in healthcare can address these issues and the repercussions of late or incorrect diagnosis. In this research, I developed BrCaVision, a state-of-the-art automated system for determining malignancies, classifying histological subtypes, detecting mitosis, and scoring HER2 in a whole slide image (WSI). Using transfer learning methods, I constructed sophisticated convolutional neural network models that accomplished 85% accuracy for subtype determination and a sensitivity of 0.98 for mitosis detection in a fraction of the time required for gold-standard methodologies. Also, I developed a web application that allows medical professionals to input a WSI and obtain a HER2 score and heatmaps that mark tissue subtypes and mitotic cell locations. By deploying the application on a public web hosting service, professionals can acquire timely and accurate breast cancer diagnosis and prognosis prediction across the world. Being easily accessible and fully automated, BrCaVision serves as a step forward in the field of bioinformatics and can save lives with reliable, accurate, and efficient breast cancer detection.

Video

If the video does not play, watch it on the original project page.

From the student

The word ‘cancer’ is as evasive and enigmatic as the diseases that it defines. As far as I can remember, I have been enraptured by cancer ever since I could pronounce the term. My first true encounter with cancer was in a now-battered copy of Siddhartha Mukherjee’s “The Emperor of All Maladies”, an enthralling inquiry into the greatest victories and most alarming setbacks in cancer research. As a middle schooler, my first instinct was to be scared of the disease and the often impulsive endeavors that have been made to vanquish it. Yet, I gradually became drawn to cancer research’s unpredictability and swift pace - and its victories. Sidney Farber and Min Chiu Li soon became my heroes, brave oncologists who devoted their careers and lives to the pursuit of a cure that they could only imagine. I soon decided that I would be just like them.

I’ve wanted to become an oncologist since I was little, but I didn’t want to wait until I was 30 to delve into the field of cancer research. I also aspire to make a lasting impact as a researcher even before entering the professional realm. I've always been enthusiastic about technological fields of study after being introduced to coding in elementary school. I conducted my first research project in 5th grade and won first place at the NC State Science and Engineering Fair for a project regarding parabolic reflectors and their impacts on WiFi signal strength. However, after taking courses in biology and chemistry and reading “The Emperor of All Maladies,” I was quickly drawn to research in bioinformatics, which lay at the intersection of computer science and biology. In 8th grade, I analyzed data from NCBI’s Gene Reference Sequence and AmaZonia! to identify soluble factors and integrin subunits that are crucial for the development of renal cells; my project (named "MicroArray Gene Expression Analysis for Renal Bioengineering") was nominated for BROADCOM Masters. Yet, after my uncle was diagnosed with aggressive and invasive carcinoma in the summer of my freshman year of high school, I decided to start my first endeavor into cancer research - but I wanted to utilize my passion for technology as well. In 10th grade, I developed a neural network model to differentiate between images of benign and malignant (cancerous) breast tissue and created a web application portal that would analyze uploaded tissue images. My subsequent project (named "AIConvPath: A Deep Learning Approach to the Automatic Classification of Histopathological Images of Malignant Cell Tissues to Detect Breast Cancer") advanced to the NC State Science and Engineering Fair and placed 1st place in the Advanced ‘Technology/Engineering’ category at the state NC Student Academy of Sciences (NCSAS) competition - I later presented this project at the 2021 American Association for the Advancement of Science Annual Meeting.

In 11th grade, I constructed a state-of-the-art automated system for determining malignancies, classifying histological subtypes, scoring HER2 expression, and detecting mitosis in whole slide images of biopsy breast tissue. In short, I’d tried to develop a precise and comprehensive system for predicting a breast cancer patient’s prognosis. My project received a Third Grand Award at the 2021 Regeneron International Science and Engineering Fair and a third-place Oral Presentation Award at the 2021 National Junior Science and Humanities Symposium. I am presenting this project at the 2022 American Association for the Advancement of Science Annual Meeting.