Histological Grading of Breast Carcinoma: Variability between Breast Subspecialty Pathologists and General Pathologists
K Woolf, J Wang, X Wang, LM Schiffhauer, DG Hicks, P Tang. University of Rochester Medical Center, Rochester, NY; RTI Health Solutions, Research Triangle Park, NC
Background: The Nottingham grading system is the most commonly used grading scheme for breast carcinomas and is based on a combined score for tubule formation, nuclear cytology, and mitotic rate. While guidelines of the grading system attempt to define criteria for the score within each category, the assignment of these scores by pathologists remains subjective. With subspecialty practice has become more popular; we sought to investigate the agreement in histologic grading between general pathologists and breast pathologists.
Design: Ninety-seven breast carcinomas were pulled from the files at our departmental file and reviewed first independently by five pathologists in the breast subspecialty group and then re-examined for a consensus if there was a disagreement. The Nottingham score (3-9) and grade (1-3) were compared between the originally reported grade given by general pathologists and the consensus grade from the breast pathologists for each case.
Results: Among the 97 cases, the complete agreement was 71 (73.2%) for histologic grade, 76 (78.6%) for tubule formation, 55 (56.7%) for nuclear grade, and 69 (71.1%) for mitotic count. There was good agreement between the consensus grade and the originally reported grade (kappa 0.65), as well as the consensus total score and the originally reported score (kappa 0.62). Comparing the individual scores for tubules, nuclear cytology, and mitotic figures showed very good agreement for the tubule score (kappa 0.73), but less agreement for the nuclear score (kappa 0.39) and mitotic rate (0.35). The agreement for histologic grading among 5 breast pathologists were fairly good, with 49 cases with complete agreement, 18 cases with agreement on 4 pathologists, 29 cases with 3 pathologists. Only one case had discrepancy among all 5 pathologists (G1 by 2, G2 by 2, and G3 by 1), which might due to multiple slides with tumor from this case.
Conclusions: While it is reassuring that there is very good agreement between the final grade and the score that grade is based upon, there remains a lack of reproducibility especially in the nuclear and mitotic scores. Subspecialty practice may not have an effect on grading consistency.
Monday, March 22, 2010 1:00 PM
Poster Session II # 46, Monday Afternoon