By Lynn McCain

 

Congratulations to our co-chief resident, Dr. Corey Post, for being selected by the United States and Canadian Academy of Pathology as the recipient of the 2024 F. Stephen Vogel Award.  This award is given to a resident who has published an outstanding paper in one of the USCAP’s affiliated journals, Laboratory Investigation or Modern Pathology, and will be presented to him at the USCAP Honors Session on Saturday, March 23, 2024.

Post worked with Drs. Jerome Cheng, Liron Pantanowitz, and Maria Westerhoff, to submit a manuscript entitled “Utility of Machine Learning to Detect Cytomegalovirus in Digital Hematoxylin and Eosin-Stained Slides” to Laboratory Investigation, which was published in October 2023. This project focused on training machine learning models to identify cytomegalovirus (CMV) inclusions on routine H&E stained tissue sections. The final result yielded multiple models with excellent performance, with the highest showing a test accuracy of 99.93%, sensitivity of 99.85%, and an area under the curve of 0.9998.

Use of artificial intelligence to identify CMV can significantly reduce the amount of time needed for pathologists to review these cases, as a search can be time consuming when there are not many inclusions present. The machine-learning algorithm identifies suspected CMV inclusions which can be presented to the pathologist for confirmation, eliminating the need for exhaustive examination of multiple slides. This could also reduce the need to perform CMV immunohistochemical stains, which can delay results by another day. The faster turnaround time on these cases is important as the rapid identification of CMV in immunosuppressed or immunocompromised patients has a direct impact on therapeutic management.

Post, who is currently in his fourth year as a House Officer, came to Michigan from the University of Arkansas for Medical Sciences in Little Rock, where he had completed his medical training. Prior to that, he had completed a master’s degree in bioinformatics at the University of Illinois at Urbana Champaign. He selected Michigan for residency due to the outstanding pathology informatics division combined with the modern facilities and friendly people. “During my visit to Ann Arbor, I really enjoyed both the city and the campus, as well as the camaraderie among the residents and faculty members I interviewed with. I was drawn to the strength of the training program, with an emphasis on both anatomic and clinical pathology. I was also interested in the wealth of available fellowship options and will be staying here for fellowships in both hematopathology and molecular genetic pathology.”

Post went on to reflect on experiences as a resident that engaged his passions. “One of my passions is medical education, and I was very fortunate to get involved in a project during my first year with Dr. Madelyn Lew. “Over my four years in training, we have restructured the medical student elective in pathology to be more transparent and engaging, and I was able to use the work as a part of the Community of Medical Educators in Training (CoMET) fellowship program.” He also enjoyed working with the informatics team on several projects, including the upcoming digital pathology transition, as well as his research which led to his award-winning publication. Finally, he has felt honored to serve as co-chief for the residency program and has learned much about leadership and administration.

“I am thankful for the substantial faculty support that I received with my work. Dr. Pantanowitz served as a mentor to me during his time here and was the person who introduced me to this project. Dr. Cheng handled most of the technical aspects of the image analysis and machine learning, and the excellent performance is thanks to his hard work. And Dr. Westerhoff served as a guide from beginning to end; indeed someone had to train me how to recognize CMV before I could train a computer how to do it. Having experts like these really helped me to do the work that was needed and to make this publication a success.”

“I am honored that the awards committee recognized the importance of this study, and I am very grateful for the support that I received from faculty and department. There is a lot of promise for the application of artificial intelligence to digital pathology, but this comes with many challenges and concerns. Hopefully proof-of-concept projects like this continue to generate excitement for the utility of these tools in daily practice.”

 

Citation:

Corey S. Post, Jerome Cheng, Liron Pantanowitz, Maria Westerhoff. Utility of Machine Learning to Detect Cytomegalovirus in Digital Hematoxylin and Eosin-Stained Slides. Lab Invest 2023, 103(10): 100225. https://doi.org/10.1016/j.labinv.2023.100225.