LAUSR

Background. Classification of phyllodes tumors is challenging due unclear diagnostic criteria, recently addressed by consensus review criteria. Herein, we reviewed all malignant phyllodes tumor resections and reclassified them based on the consensus guidelines, correlating with outcome. We hypothesize that application of criteria would result in a significant proportion being "down-graded" to either borderline or benign phyllodes tumor. Methods. Primary resections of malignant phyllodes tumor were reviewed by four AP board-certified, breast fellowship-trained pathologists. Morphologic variables delineated in consensus guidelines (ie stromal cellularity, cellular atypia, tumor border, presence of heterologous elements, presence of stromal overgrowth) were evaluated. Following review, cases were reclassified as benign, borderline, or malignant. Results. Upon reclassification, 20% (5/20) cases were "down-graded" to borderline phyllodes tumor while 80% (15/20) remained malignant phyllodes tumor. Two morphologic features were statistically significant including broadly infiltrating tumor border in 80% (12/15) of malignant phyllodes tumors compared to none in borderline phyllodes tumor (0/5) (p = 0.004) and stromal overgrowth in 67% (10/15) of malignant phyllodes tumor compared to none in borderline phyllodes tumors (0/5) (p = 0.03). Upon review of the pathology reports, 30% (6/20) contained all 5 histomorphologic variables delineated in the consensus review criteria. Malignant phyllodes tumor resulted in five cases with recurrence (33.3%, 5/15) and three cases with metastases (20.0%, 3/15) and borderline phyllodes tumor resulted in one case with recurrence (20.0%, 1/5) and no metastases (0/5). Conclusion. The consensus guidelines for phyllodes tumor are useful for subclassification. We hypothesize that standardize reporting of the histomorphologic variables may lead to better consensus.