Proliferation in breast cancer

Pathology services in the Western Norway Health Region – a center for applied digitization: Measuring proliferation in breast cancer

Samarbeidspartnere

Prof. Håvard Danielsen; Institute of cancer genetics and informatics, Oslo University Hospital

Formål

To train, test, validate and implement a machine learning system for calculating the Mitotic Activity Index (MAI), Ki-67 and PHH3 index in breast cancer

Problemstilling

Breast cancer is the most frequent female malignancy in the western world. To improve therapeutic decision-making, guidelines often combine conventional predictors to estimate relapse ⁄ mortality risk, but inaccuracy will cause over- and undertreatment. Proliferation factors are stronger prognostic indicators than Adjuvant! or NBCG Guidelines. Ki67 is also prognostic, and has been included in the Norwegian and Sankt Gallen therapy guidelines, but lacks independent laboratory test validation. As there is no formal standard operating procedure for Ki67 assessment based on scientific observations existent so far, the Norwegian guidelines nowadays also includes the classical mitosis counting, a time-consuming manual method. Both Ki67/PHH3 and mitosis counting can be automated and improved by means of digital image analysis.