Counting Mitoses With Digital Pathology in Breast Phyllodes Tumors

Area of Science:

• Pathology
• Oncology
• Digital Pathology

Background:

• Mitotic count is a key histologic criterion for grading and prognostication of phyllodes tumors (PTs).
• Current methods of mitosis counting can lead to misclassification due to variations in microscopes and field selection.

Purpose of the Study:

• To compare the correlation of 10 high-power field (HPF) versus whole slide mitotic counts with clinicopathologic parameters in PTs using digital pathology (DP).
• To assess the potential for developing an artificial intelligence (AI) protocol for mitosis counting.

Main Methods:

• Digital pathology was used to scan and view slides from 93 PT cases.
• Mitotic counts were performed on whole slide images and compared with counts from 10 HPFs.
• Correlations between mitotic counts and clinicopathologic parameters (grade, stromal atypia, hypercellularity, patient age, tumor size) were analyzed.

Main Results:

• Both whole slide and 10 HPF mitotic counts demonstrated statistically significant correlations with PT grade, stromal atypia, and stromal hypercellularity.
• Neither method showed a statistically significant correlation with patient age or tumor size.

Conclusions:

• Accurate mitosis counting in breast PTs is vital for precise grading.
• Digital whole slide analysis offers a foundation for training and validating future AI algorithms for mitosis counting.