Utilizing Interpretable AI to Distinguish prePMF From ET

Andrew Srisuwananukorn, MD, assistant professor at The Ohio State University Comprehensive Cancer Center, discusses the main findings from research of an artificial intelligence (AI)-powered machine learning algorithm which evaluated digital whole-slide images of diagnostic bone marrow biopsies and accurately differentiated prefibrotic primary myelofibrosis (pre-PMF) from essential thrombocythemia (ET). According to findings that were presented by Srisuwananukorn at the 2023 American Society of Hematology Annual Meeting, the AI distinguished between differentiated pre-PMF and ET with a 92.3% accuracy rate.

According to this research, the algorithm focused on areas of high biological interest and specifically examined bone marrow cellularity compared with areas of fat, bone, or background tissue.

Transcription:

0:09 | [There was an] area under the receiver operator curve of 0.9, a sensitivity of 66.6% specificity of 100%, and an accuracy of 92.3% in diagnosing prefibrotic myelofibrosis. In addition, we did a qualitative analysis to try to understand what is being used in these AI algorithms to make those predictions. With this qualitative interpretation of quote unquote, opening the black box of our AI algorithms, we were able to see that preferentially areas of bone marrow cellularity were chosen for the prediction of 1 vs the other disease. Reassuringly, the algorithm was not using nonsensical portions of the image, such as fat or cortical bone or even background artifacts.

0:54 | We believe this AI algorithm is using biological reasons. I view this type of tool as a companion diagnostic tool, but I do not believe [that] AI tools can replace the judgment of a human physician. I think it is really up to us as pathologists and clinicians to say when an AI algorithm tool is not working appropriately. What I hope is that this can be used for better information for the patient to understand their disease.