Correlation of Ki-67 between Biopsy and Resection Specimens: Does “Hot-Spot” Counting Matter?
Zuzana Kos, Denis H Gravel, Robertson J Susan. Ottawa Hospital and University of Ottawa, Ottawa, ON, Canada
Background: The neoadjuvant setting is being increasingly explored as a means for the evaluation of therapeutic strategies in women with breast cancer. This setting permits study of the biological characteristics of responsive versus resistant tumours in vivo, using surrogate markers of response. The Ki-67 proliferation index is emerging as an important predictive and prognostic marker, and changes in Ki-67 labelling in response to short-term challenge with anti-estrogens and chemotherapeutic agents have been reported to be predictive of the long-term clinical effectiveness of the drug. However, studies specifically addressing the correlation between Ki-67 scores on biopsies and resection specimens are lacking. In the past few years, there has been a shift towards hot-spot counting, suggesting a possibility of higher Ki-67 scores in resection specimens versus biopsies as a direct result of selective counting and biopsy sampling discrepancies.
Design: Biopsy and resection specimens from 20 patients with estrogen receptor-positive breast carcinomas were selected. Ki-67 scores were determined by image-assisted hot-spot counting of 1000 cells in the resection specimens and biopsies when possible; in smaller biopsy samples, all neoplastic cells were counted. The samples were analysed for correlation using a two-tailed paired t-test.
Results: The mean Ki-67 index for the biopsies was 28.3 + 24.3%, with a range of 0.4 to 99.1%. The mean labelling index for resection specimens was 32.2 + 26.0 %, with a range of 4.8 to 91%. The resection specimens had a mean Ki-67 score that was 3.9% higher than on biopsy, but this was not statistically significant (p = 0.2). Difference in individual counts, however, ranged from 13% lower to 43 % higher on the resection specimen as compared to the biopsy, including 3 cases in which the Ki-67 would give a score that would be different with respect to the values suggested as surrogate cut-offs (10 to 14 % range) for designating a tumour as lower risk luminal A versus higher risk Luminal B.
Conclusions: These data show that despite the practice of hot-spot counting there is no significant difference between the Ki-67 labelling index determined on biopsies and subsequent resection specimens. This study also predicts that for interventions in trials dealing with groups, changes in Ki-67 scores may be predictive in most cases; however, interpretation of results, on a case by case basis, does require some caution.
Tuesday, March 5, 2013 1:00 PM
Poster Session IV # 58, Tuesday Afternoon