Comparison among different classification systems regarding the pathological response of preoperative chemotherapy in relation to the long-term outcome

Neoadjuvant chemotherapy (NAC) is increasingly used for operable disease. However there are several pathological response classification systems and the correlation between the pathological response to NAC according to each system and the patient outcome is still under debate. From 1998 to 2006, 370 primary breast cancer patients underwent curative surgical treatment after NAC containing both anthracycline and taxane at the National Cancer Center Hospital. We retrospectively evaluated the clinical and pathological response using the cTMN, Fisher's, Chevailler's, and the Japanese Breast Cancer Society classification systems (JBCS) respectively, and analyzed the correlation between each pathological response and disease free survival (DFS). Ninety-five (26%) patients had tumor recurrence. The five-year DFS according to Fisher's system was pCR, 80% and pINV, 63%. The five-year DFS according to Chevallier's system was Grade 1, 83%, Grade 2, 85%, Grade 3, 62%, and Grade 4, 65%. The five-year DFS according to the JBSC system was Grade 3, 77%, Grade 2, 68%, Grade 1a, 68%, Grade 1b, 58%, and Grade 0, 52%. None of the pathological response systems reached a statistically significant difference. In the classification by the post-treatment number of metastatic axillary lymph nodes, the 5-year DFS was n = 0, 86%; n = 1-3, 64%; n = 4-9, 44%; and n > 10 positive: 25% (P < .0001). In pathologically node negative patients, there were no significant differences in the DFS among all the classification systems. All three classifications analyzed were considered inadequate as the prognostic marker of the long-term outcome after NAC and further studies are warranted to optimize the prediction.