Inter-observer agreement among pathologists in grading the pathological response to neoadjuvant chemotherapy in breast cancer.

BACKGROUND

The degree of pathological response to neoadjuvant chemotherapy (NAC) was correlated with the prognosis in breast cancer. There are few studies published on inter-observer variability in the assessment of pathological responses among pathologists.

METHODS

We collected 64 surgically resected specimens from patients who had received NAC. Three pathologists assessed the pathological responses and classified them into 7 grades according to grading system of the Japanese Breast Cancer Society. The levels of concordance among pathologists were categorized into 3 classes: full concordance (all pathologists gave the same grade), partial concordance (two of them gave the same grade), and discordance (all three gave different grades). The inter-observer agreement among pathologists was estimated using the percentage concordance and Cohen's kappa statistics.

RESULTS

Full concordance, partial concordance, and discordance were seen in 28 (43%), 33 (52%), and 3 (5%) specimens, respectively. In most of partial concordance specimens (30 out of 33), the pathological response grades differed by only one level. The kappa value was 0.59. The concordance rate with regard to pCR was 97%.

CONCLUSIONS

Most of the judgments among pathologists differed within one level, but there is room for improving harmonization in the assessment of pathological responses.