Adewole Adamson, a dermatologist and researcher, believes AI-assisted screening will lead to more overdiagnosis as the goal should be to find cancers that will ultimately kill people, not just find more cancer. However, AI could potentially help address the problem of overdiagnosis by using information embedded in medical records to examine the trajectories of different patients’ cancers over time. Adamson suggests adding a third category to the data that the algorithms learn from: “Maybe cancer”, which would encompass slides or images that provoke disagreement among experts.
Key takeaways:
- Microsoft has partnered with a digital pathology company, Paige, to build the world’s largest image-based AI model for identifying cancer, using a training data set of 4 million images.
- AI-supported cancer screening models have shown promising results, reducing workload by 44% and detecting 20% more cancers in a recent clinical trial.
- However, there are concerns that AI-assisted screening could lead to overdiagnosis, as it may not be able to distinguish between lethal and nonlethal cases of cancer.
- Adewole Adamson, a dermatologist and researcher, suggests that AI could potentially help address the problem of overdiagnosis by adding a third category to the data that the algorithms learn from: 'Maybe cancer'.