MOMA: Advances in colonoscopy screening accuracy with Artificial Intelligence
By Nitish Bhamidipati
Colorectal Cancer is the second deadliest cancer in the world, according to the International Agency for Research on Cancer, a division of the United Nations World Health Organization. While there have been significant improvements in recent years, we’ve arrived at a breakthrough. The latest interest in medical research has been Artificial Intelligence, specifically in the context of screening for cancers, the development of AI tools that can detect tumor growth throughout the body. An article recently published by Harvard Medical School (HMS) outlines the development of a new AI tool that can detect colorectal tumors and assess their overall “deadliness” just from analyzing images of tumor samples (called biopsies). With over 150,000 new colorectal cancer cases estimated for 2023 according to reports by the American Cancer Society, the creation of new software that can make cancer detection more accurate and less expensive can have incredible benefits to the general public, and potentially save hundreds of thousands of lives.
This AI tool known as Multi-omics Multi-cohort Assessment (MOMA) has been trained on almost 450,000 patients in clinical trials. To put that number into context, the more patients the tool is trained on, the greater its accuracy, because artificial intelligence tools require large data sets to train and refine its algorithms to recognize and assess cancers quicker and more accurately. In order to test this tool in a real world setting, researchers used the tool on real tumor samples from patients with no background information given. In this setting, the tool was able to accurately predict and outperform human pathologists in 3 different aspects: (1) Overall patient survival, (2) Years of cancer-freeness, (3) Effectiveness of different therapies based on genetic makeup of patients.
In addition to MOMA’s revolutionary abilities, it’s also cheap, being free to all researchers and clinicians. A report by Harvard researchers estimates that AI tools can potentially save $360 billion annually if healthcare institutions actively adopt them. Tools like MOMA can provide accurate clinical assessments in rural and underdeveloped locations with the lack of infrastructure to support expensive hospital developments and sophisticated machines.
The extent of this tool’s accuracy and ability has crossed a new level, and shows a new transition in medical technology appearing. With the threat of AI tools replacing the job of human pathologists, it’s important to think about how us as medically inclined individuals fit in this changing world. How are we going to adapt to new technologies and stay relevant? What new skills beyond simply medicine will we have to learn? The road ahead is intimidating, but hopeful. As long as technologies like MOMA keep developing, more lives will be saved. That’s a guarantee. This is just one of many examples that strengthen the argument that biology, chemistry, and overall medical understanding is not enough anymore. We must embrace the new possibilities of AI. That starts with understanding what AI is and how it works.
Start learning about AI here: https://cloud.google.com/learn/what-is-artificial-intelligence
Enjoy time traveling to the future.
Citations: