NACT has become an important treatment for patients with inoperable locally advanced breast cancer. However, the response to chemotherapy varies among different subtypes of breast cancer. Therefore, it is crucial to predict the probability of achieving pCR before receiving NACT, which will help in selecting an accurate treatment plan. Luminal-type breast cancer exhibits a variable response to NACT, with the majority of patients not achieving a complete response. Therefore, the early identification of the nonresponsive subgroup within this subset will shield them from the unnecessary toxicity of NACT and assist in selecting those patients who stand to benefit most from NACT[19]. Thus, identifying clinicopathological factors to predict pCR in luminal breast cancer continues to pose a challenge. Previous studies have employed a nomogram based on clinicopathological features to predict pCR.
One of the indications for NACT is to downstage locally advanced breast cancer to a more operable stage, specifically converting from mastectomy to BCS. In the EORTC 10902 trial, the breast conservation rate was 35% in the neoadjuvant arm versus 22% in the adjuvant arm [30]. For luminal breast cancer, the rate of downstaging is lower than that in HER2-positive cancer and TNBC (16–18% vs 41% and 42%, respectively) [31, 32]. Another study by Kim et al. [33] explored the potential benefits of NACT in clinically node-positive luminal subtype breast cancer. This study enrolled 244 patients with clinically node-positive breast cancer treated with NACT, categorizing them by molecular subtype. Results showed that 61.4% of patients classified as Luminal HER2 negative underwent total mastectomy, while 38.6% underwent BCS, with a conversion rate of 16.3%. Although our study did not evaluate the rate of conversion from mastectomy to BCS, the rate for BCS in the entire cohort was 46.4%. Notably, patients who achieved a complete response underwent more BCS compared to those with a non-complete response (56.9% vs 43.4% respectively), which is higher than the results from Kim et al. Thus, BCS remains a feasible and viable surgical option for patients with luminal-type breast cancer who received NACT.
In this study, we investigated the effect of NACT on complete response in luminal-type breast cancer. We developed a novel nomogram to individually predict the probability of pCR in patients with this subtype of breast cancer. The nomogram incorporated several covariates—clinical T stage, ER Allred score, PR Allred score, and Ki67 index—based on multivariate analysis. Furthermore, our findings indicated that patients with lower clinical T stage, and lower ER and PR Allred scores, along with a high Ki67 index, were associated with a higher pCR rate from NACT.
The VG of our nomogram differs from the TG. First, the VG tended to include older women compared to those in the TG. Additionally, there were more patients with aggressive forms of breast cancer in the VG than in the TG. Consequently, more patients underwent mastectomy rather than breast-conserving surgery. Furthermore, the VG exhibited an increased rate of complete response compared to the TG. We believe that the small sample size of the VG (n = 330) and changes in the indications for NACT over the years may contribute to the observed differences. Despite these differences, the AUC of the ROC curve in the VG was 0.63, indicating good predictive ability. Through this nomogram, we can establish the likelihood of pCR. However, further large external and prospective studies are needed to validate our results.
Ki67 is a nuclear antigen expressed in the active phases of the cell cycle: G1, S, G2, and M phase. It is recognized as a proliferative marker of cells[20]. Furthermore, Ki67 serves as a valuable clinical marker for subtype classification, prognostication, and predicting treatment responses in breast cancer. Specifically, Ki67 has been extensively studied in relation to chemotherapy as a predictive factor. Notably, it has been reported that tumors with a high pre-treatment Ki67 index are associated with a higher pCR rate[23]. Despite this strong association, the application of Ki67 in neoadjuvant settings is variable. Wajid et al. [20] concluded that higher Ki-67 values in patients with breast cancer were associated with good pCR in a small group of patients in Pakistan. Similarly, Fasching et al. [21] reported that Ki67 has independently predicted treatment response and prognosis in breast cancer patients. Luminal breast cancer exhibits a variable Ki67 score; Luminal A is typically associated with low scores, whereas Luminal B generally shows the opposite[22]. Kim et al. [24] published a study investigating the potential benefits of NACT in clinically node-positive luminal subtype breast cancer, particularly regarding the facilitation of breast-conserving surgery. In their analysis, higher Ki67 expression was significantly correlated with pCR (OR, 3.715; 95% CI, 1.356–10.177; p = 0.011), a finding further validated through multivariate logistic regression analysis. Likewise, our study confirmed that higher Ki67 scores were also associated with pCR. When comparing the nomogram we developed to the one based on Ki67 alone, the AUC value of our nomogram demonstrates superior predictive capability for pCR than the nomogram using only Ki67.
Furthermore, the literature has explored estrogen and progesterone receptor status as pCR-related factors across different breast cancer subtypes. In luminal-type breast cancer, there is a notable expression of the ER receptor, with or without the presence of the PR receptor, and a negative HER2 status. The relationship between ER and PR receptor statuses and pCR remains controversial. Our study demonstrated that ER and PR statuses were significantly associated with pCR, as evidenced by multivariate logistic regression analysis (OR, 1.174; 95% CI, 1.060–1.301; p = 0.002 and OR, 1.079; 95% CI, 1.029–1.131; p = 0.002, respectively). According to a study by Lips et al., which involved 117 patients with luminal-type breast cancer undergoing NACT, PR-negative cancers were significantly linked to a higher pCR rate compared to PR-positive cancers (35.3 vs. 11.7%; p < 0.001)[24].
Lastly, the published data regarding clinical tumor stage in luminal breast cancer as a factor associated with complete response remains controversial. Our study demonstrated that higher clinical T (T2 or T3) stages were strongly associated with pCR in the multivariate logistic regression analysis (OR, 2.040; 95% CI, 1.431–2.911, p-value < 0.001). Similarly, Bozdogan et al. [34] identified clinical T1 and T2 as independent risk factors that increased pCR (OR = 4.2, CI = 1.2–14.5; p-value = 0.024). On the other hand, Collins et al. [25] evaluated factors predicting pCR and oncologic outcomes in patients with luminal-type breast cancer receiving NACT.
Our study has several limitations. Firstly, it is a retrospective single-center study; therefore, a prospective comparison is warranted to evaluate whether NACT offers a benefit in this subset of patients. Secondly, although our nomogram has been validated by an external cohort, it has not been validated in a cohort from another center. Furthermore, we have observed differences in both the validation and TGs, which could be attributed to several factors, such as the smaller sample size of VG and the evolving criteria for NACT over time. Notably, one of the strengths of our study is its large scale, encompassing 1,571 patients, which surpasses the scope of previously published studies.
In summary, our retrospective study suggests that the combination of pre-NACT clinicopathological features, including clinical T stage, ER and PR Allred scores, and Ki67 index, are powerful predictors of pCR. Our nomogram has demonstrated good predictive ability for pCR in luminal breast cancer. Further prospective and global studies are warranted to validate this nomogram and identify additional clinicopathological features that may help predict pCR in this patient subset.