Contact informationPlease contact Tarjei S. Hveem for more information
The proportion of cells undergoing division is a prognostic marker in several cancer types, including breast and gynaecological cancers.
Mitotic index, the number of cells undergoing mitotis in ten high-power fields as assessed by a pathologist is a longstanding method in the field of pathology. An increased number of mitotic cells is normally associated with a worse prognosis for the patient. However, it is difficult to accurately identify mitotic figures and the process is hampered by a significant lack of agreement between pathologists.
We have developed a deep learning method for the automatic identification of mitotic cells in scanned tissue sections based on a publicly available breast cancer dataset with manually annotated mitotic cells. We have evaluated the performance of our automatic method in a total population of uterine sarcomas and found the prognostic impact to be strong. We plan to apply the method to other cancer types where mitotic index may be relevant for prognosis or treatment response and to evaluate the spatial distribution of the mitotic cells.
This text was last modified: 12.02.2020