A reproducible approach for scoring TIL in residual tumors after neoadjuvant treatment of breast cancer patients

Abstract Neoadjuvant chemotherapy (NAC) is standard of care for patients with locally advanced breast cancer. TIL scoring is prognostic for response and has additional predictive value to the residual cancer burden after NAC. However, NAC induces various changes in the tumor bed and reliability of TIL scoring in post-NAC samples has not yet been studied. Pre- and post- NAC FFPE blocs from 60 patients were collected. H&E and dual CD3/CD20 chromogenic IHC-stained tissues were prepared and scored for global, stromal and intratumoral TIL by two experienced pathologists. Digital TIL scoring was preformed using the HALO® image analysis software. Comparison of the different scoring methods was done by calculation the concordance correlation coefficient (CCC), the Bland-Altman method and Passing-Bablok regression analysis. In patients with residual disease we demonstrate a good inter-pathologist correlation for global and stromal TIL on H&E stained tissue (CCC value 0.73, CI 0.53–0.85 for stromal TIL on H&E). Good correlation for scoring using the two staining methods (CCC 0.81, CI 0.67–0.90 for stromal TIL score) and with digital TIL scoring (CCC 0.77, CI 0.59–0.88) is demonstrated. There is only a poor to moderate reliability for intratumoral TIL scores. While overall concordance for TIL scoring in patients with a complete response is poor, there is a moderate inter-pathologist reliability for stromal TIL scoring on CD3/CD20 stained tissue. This study demonstrates a good reliability for TIL scoring in patients after NAC treatment, comparable to results in untreated breast cancer patients, in patients that have detectable residual tumor. Since there is a good concordance with digital TIL scoring, the development of a validated algorithm could be helpful in the future.