Partnering with AI Against Breast Cancer
/
For over 40 years, from 2001's HAL 9000 through The Terminator and The Matrix, we have been conditioned to fear and distrust artificial intelligence (AI). “Thinking machines are coming,” Hollywood has warned us, “and they're out to get you!” Even the news wires are filled with scary stories about AI's that are after our jobs, and our women.
So it's a delight (and a relief) to finally read about a real-life AI that is working for the good of mankind.
A research team from Beth Israel Deaconess Medical Center (BIDMC) and Harvard Medical School (HMS) recently developed AI methods aimed at training computers to interpret pathology images, with the long-term goal of building AI-powered systems to make pathologic diagnoses more accurate.
The new AI method is based upon an algorithm known as “deep learning,” and it is already showing promise as a way to help pathologists improve diagnosis of breast cancer from images. At the International Symposium of Biomedical Imaging (ISBI), it boosted human accuracy from 96 to 99.5 percent.
“Identifying the presence or absence of metastatic cancer in a patient’s lymph nodes is a routine and critically important task for pathologists,” explained Andrew Beck, an associate professor in pathology at HMS. “Peering into the microscope to sift through millions of normal cells to identify just a few malignant cells can prove extremely laborious using conventional methods. We thought this was a task that the computer could be quite good at – and that proved to be the case.”
The team trained the computer to distinguish between cancerous tumor regions and normal regions based on a deep multilayer convolutional network. They started with hundreds of training slides for which a pathologist had labeled regions of cancer and regions of normal cells, and then extracted millions of these small training examples and used deep learning to build a computational model to classify them.
The team then identified the specific training examples for which the computer is prone to making mistakes and re-trained the computer using greater numbers of the more difficult training examples. In this way, the computer’s performance continued to improve, and “learn.”
“There have been many reasons to think that digitizing images and using machine learning could help pathologists be faster, more accurate and make more accurate diagnoses for patients,” Beck added. “This has been a big mission in the field of pathology for more than 30 years. But it’s been only recently that improved scanning, storage, processing and algorithms have made it possible to pursue this mission effectively.”
The results in the ISBI competition show that what the HMS computer is doing is genuinely intelligent and that the combination of human and computer interpretations will result in more precise and more clinically valuable diagnoses to guide treatment decisions.
In other words, AI's may eventually try to enslave humanity or drive us underground, but in the meantime they are proving to be invaluable partners in the fight against cancer.