Background. The role of tumor-infiltrating lymphocytes (TILs) in breast cancer (BC) is still an issue for clinical research.Toward this end, a sensitivity analysis of neoadjuvant and adjuvant randomized clinical trials was performed according to disease subtypes. Methods. Pathological complete responses (pCRs) after neoadjuvant treatment according to the presence or absence of lymphocyte-predominant BC (LPBC) were extracted and cumulated as odds ratios (ORs) by adopting a random-effects model by subtype. Overall survival hazard ratios as a function of 10% incremental values of stromal TILs (sTILs) in adjuvant trials were extracted. The interaction test was adopted to determine the differential effect according to the subtype. Results. Eight trials (5,514 patients) were identified. With regard to neoadjuvant setting (4 studies), a significant interaction (p , .0001) according to LPBC was found. The