External verification and improvement of the Neo-Bioscore staging system in a Chinese cohort

Introduction

Neoadjuvant chemotherapy (NAC) is an application of systematic pre-operative chemotherapy designed to shrink malignant lesions and eliminate potentially micrometastases. NAC allows oncologists to evaluate treatment response and guide subsequent adjustments to treatment regimens based upon tumor burden change. NAC has become the gold-standard treatment for locally advanced breast cancer.

A high pathologic complete response (pCR) rate post-NAC is associated with improved disease-free survival and overall survival (1). To further quantify the relationship between residual tumor and the prognosis of breast cancer patients who do not achieve pCR (non-pCR) after NAC, Jeruss et al. (2) developed a staging system that combined pretreatment clinical stage (CS), posttreatment pathological stage (PS), estrogen receptor (ER) status, and nuclear grade (NG) (CPS + EG). Patients with different scores were divided into more refined prognostic subgroups than those of the American Joint Committee on Cancer (AJCC) staging system. Mittendorf et al. (3) integrated human epidermal receptor 2 (HER2) status into the previous CPS + EG to develop the Neo-Bioscore staging system, which can be applied in HER2–positive patients treated with trastuzumab. The superior performance of CPS + EG or Neo-Bioscore over PS and CS has been demonstrated in the American and Japanese breast cancer patients receiving NAC (4,5). However, the PS, CPS + EG, and Neo-Bioscore staging systems were proven to be equivalent in predicting Brazilian patients' prognosis (6). Another study based on a Chinese population also did not observe the advantages of CPS + EG and Neo-Bioscore over PS (7). Differences in clinical characteristics of breast cancer patients in different countries may explain these results. The median age of diagnosis of patients in the United States was over 60 years old, while the age of diagnosis of Chinese patients is less than 50 years old. A study evaluating survival differences between Chinese and Caucasian women with breast cancer based on the Surveillance, Epidemiology, and End Results (SEER) database indicated that Chinese breast cancer patients have superior breast cancer-specific survival than Caucasians (8). At present, prognostic models applied to the Chinese population remain limited. It is vital to further validate the performance of Neo-Bioscore in Chinese people to identify a favourable prognostic model for these patients.

Several strategies have been utilized to optimize different prognostic models, such as incorporating other prognostic factors or genetic data, developing interpolation models using continuous, square, and cubic covariates, and optimizing the operation mode. Unfortunately, the high costs of genetic models often make it prohibitive to apply clinically. Xu et al. (9) incorporated an additional factor of poor prognosis, HER2-positive status without trastuzumab treatment, to determine an accurate prognosis. We attempted to integrate other prognostic factors into Neo-Bioscore to achieve more precise prognosis predictions.

Researchers have confirmed that pathological nodal status after NAC is significantly correlated to survival and is an independent prognostic factor for local-regional recurrence (10-13). Another staging system, Residual Cancer Burden (RCB), developed based upon the primary tumor and axillary lymph node status after NAC was also reported to be a significant predictor of remote recurrence-free survival. RCB could further categorize patients into different prognostic subgroups when applied to posttreatment AJCC pathologic stage groups (14). Therefore, we would like to explore whether it has an improved predictive ability to replace PS with posttreatment pathological lymph node (ypN) stage and posttreatment pathological tumor (ypT) stage after NAC in Neo-Bioscore.

The objective of the study was to perform an external verification of the Neo-Bioscore in a Chinese cohort from a single institution and then modify the model by replacing PS with two risk factors, ypT and ypN, to improve the prediction of survival outcomes for patients receiving NAC. We present the following article in accordance with the TRIPOD reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-21-6738/rc).

Methods

Participants

Clinicopathologic and survival data were collected from a total of 602 patients diagnosed with primary breast cancer who received NAC and underwent subsequent surgical treatment from February 2005 to August 2018 in the First Medical Center of the PLA General Hospital. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and approved by the Ethics Committee of the Chinese PLA General Hospital (S2021-582-01). Informed consent was waived due to the retrospective nature of the study.

Only patients who underwent core needle biopsy and were diagnosed with invasive breast cancer were included. All patients were treated with an anthracycline- and/or taxane-based NAC regimen. The number of chemotherapy cycles ranged from 2 to 8. After NAC patients received surgical treatment, including breast-conserving surgery, mastectomy, and axillary assessment. Postoperative adjuvant therapy regimens and radiotherapy were based on the patient's preoperative treatment and postoperative pathological results. During the screening, 166 patients were excluded: 32 had metastatic breast cancer, 12 with bilateral breast cancer, 13 with inflammatory breast cancer, and 4 with other malignant tumors; 105 had incomplete medical records. The final analysis included 436 eligible patients.

Clinicopathologic data

Data collected included basic demographic data (age, gender), menopausal stage, CS, PS, ypT, ypN, ER status, HER2 status, progesterone receptor (PR) status, pretreatment Ki67 level, posttreatment Ki67 level, and tumor grade. CS was based on physical examination, mammography, ultrasonography, and magnetic resonance imaging of the breast and regional nodal basins before NAC. CS and PS were determined according to the 8^th edition AJCC guidelines. ER, PR, HER2 status, pretreatment Ki67 and histological grading were obtained from the core needle biopsy at diagnosis. We utilized the Nottingham histological grading (rather than nuclear grade) due to its widespread use in clinical practice, which was consistent with the Mayo Clinic validation cohort (15). Posttreatment Ki67 was determined from surgical pathology reports.

ER/PR positive status was defined as 1% or greater of the tumor cells stained for the protein by immunohistochemical. HER2 status was defined as positive when 3 + on immunohistochemistry or fluorescence in situ hybridization demonstrated gene amplification. Nottingham histological grade was determined by evaluating tubule formation, nuclear pleomorphism, and mitotic count under the microscope, with a score of 1 to 3 assigned for each parameter. In this system, a combined score of 3–5 points indicated grade 1 (G1), 6–7 points for grade 2 (G2), and 8–9 points for grade 3 (G3). Ki67 level was evaluated by immunohistochemical analysis, and the cutoff value was defined as 14%. pCR was defined as no invasive lesions in the primary breast and axillary.

Outcomes

Disease-specific survival (DSS) (16) was calculated as the time from diagnosis to death caused by breast cancer; overall survival (OS) was defined as the time from diagnosis to death from any cause; disease-free survival (DFS) was defined as the time from diagnosis to the first disease progression (locoregional, distant relapse, ipsilateral or contralateral invasive breast cancer, ductal carcinoma in situ, or an invasive second primary cancer) or death from any cause. Data for patients who had no event at the cutoff date for the final analysis were censored at the time of the last disease assessment for DFS and at the last follow-up date for OS and DSS.

All patients were followed every six months by telephone or clinic. The follow-up was completed until July 29, 2021. The follow-up personnel blinded the patients' identity information and clinicopathologic information.

Statistical analysis

Based on the Neo-Bioscore staging system published by Mittendorf et al. (3), the point assignment of each component, including pretreatment clinical stage, posttreatment pathologic stage, ER status, tumor grade 3, and the HER2 status, is shown in Table 1. Scores of each patient were calculated by summarizing the scores of each component, and patients with the same score were classified into identical subgroups. Continuous variables were converted to categorical variables with a cutoff value of 40 for age and 14% for Ki67. Cox proportional hazards model was used to determine risk factors for DSS. PS was replaced with the ypT and ypN stage in the Neo-Bioscore to create the MNeo-Bioscore (modified Neo-Bioscore) staging system. Here ypT and ypN were each assigned 2 points, and the score distribution of the MNeo-Bioscore is depicted in Table 1.

DSS, OS, and DFS rates were calculated for each subgroup of the four models (Neo-Bioscore, MNeo-Bioscore, PS, and CS) utilizing the Kaplan-Meier method and compared by log-rank test, and Kaplan-Meier curves were plotted. The concordance indexes (C-index) of the four models were calculated; the receiver operating characteristic (ROC) curves were plotted for the four models, and the area under the curve (AUC) values were calculated to assess discrimination. The calibration curves for 3- and 5-year DSS of Neo-Bioscore and MNeo-Bioscore were drawn to evaluate the calibration. For clinical utility, decision curve analysis (DCA) with the threshold probability and net benefit delineated as the horizontal and vertical coordinates to assess the net benefits of the model was performed (17). For the internal validation of the MNeo-Bioscore, C-index and AUC of MNeo-Bioscore were calculated and calibration curves were plotted using the bootstrap resampling method. To compare the four staging systems, the C-index was compared by the “compareC” package. We further drew a time-dependent receiver operating characteristic curve (tdROC) based on the AUC at each time point for the categorized models.

A two-sided αof 0.05 was used for all tests. Statistical analyses were conducted with R version 3.3.1 (The R Foundation of Statistical Computing, Vienna, Austria).

Results

Patient characteristics

Of the 436 patients included, 66.3% had clinical stage II disease (IIA, 31.0%; IIB, 35.3%), and 25.5% had clinical stage III disease (IIIA, 13.3%; IIIB, 5.5%; IIIC, 6.7%). pCR rate was 12.6% in our cohort, with the remaining patients having pathologic stage I (17.0%), IIA (28.9%), IIB (11.7%), IIIA (18.1%), IIIB (0%), or IIIC (11.7%) disease. The pathological T stages of the patients were T0 (17.7%), T1 (39.7%), T2 (33.3%), T3 (9.4%), and T4 stage (0%); the pathological N stages were N0 (43.3%), N1 (28.9%), N2 (16.1%), and N3 (11.7%). Additionally, 76.8% of patients had Ki67≥14% pretreatment and 50.2% posttreatment. Clinical characteristics are displayed in Table 2.

The median follow-up time was 67 months (2-192 months). During the entire follow-up period, 59 patients (13.5%) died, (56 died from breast cancer), and 101 patients (23.1%) recurred or metastasized.

External validation of the Neo-Bioscore

The Kaplan-Meier curves of CS, PS, and Neo-Bioscore are shown in Figure 1A-1C, respectively, depicting that the three staging systems could significantly distinguish the DSS among different groups (P<0.0001). The ROC curves for 3-year and 5-year DSS are in Figure 2. The C-index for the Neo-Bioscore was 0.78 [95% confidence interval (CI): 0.72–0.83], and the AUC for 3-year and 5-year DSS was 0.76 (95% CI: 0.68–0.85) and 0.81 (95% CI: 0.76–0.87), respectively. The C-index of PS and CS was 0.75 (95% CI: 0.69–0.82) and 0.68 (95% CI: 0.62–0.74), and the C-index of the Neo-Bioscore was significantly superior to CS (P<0.001) though not significantly different from PS (P=0.399). The calibration curves of the Neo-Bioscore at 3- and 5-year DSS are shown in Figure 3A,3B. The Neo-Bioscore fit well based on our patient cohort.

Figure 1 Kaplan-Meier survival curves for disease-specific survival (16) in patients with breast cancer receiving NAC. (A) DSS determined based on the CS. (B) DSS determined based on the PS. (C) DSS determined based on the Neo-Bioscore. pCR, pathologic complete response; NAC, neoadjuvant chemotherapy; DSS, disease-specific survival; CS, pretreatment clinical stage; PS, posttreatment pathologic stage.

Figure 2 Receiver operating characteristic curves in predicting the 3-year disease-specific survival and 5-year disease-specific survival for the Neo-Bioscore, MNeo-Bioscore CS, PS. (A) predicting the 3-year disease-specific survival for four staging systems. (B) Predicting the 5-year disease-specific survival for four staging systems. AUC, area under the receiver operating characteristic curve; CS, pretreatment clinical stage; PS, posttreatment pathologic stage.

Figure 3 Calibration curve for all patients based on the 3-year disease-specific survival and 5-year disease-specific survival in the Neo-Bioscore. (A) 3-year disease-specific survival in the Neo-Bioscore. (B) 5-year disease-specific survival in the Neo-Bioscore. DSS, disease-specific survival.

Development and internal validation of the MNeo-Bioscore

Univariate analysis disclosed that age, CS, ypT stage, ypN stage, tumor grade 3, and posttreatment Ki67 were significantly correlated with DSS (Table 2). In further multivariate analysis, age, ypN, and tumor grade 3 were independent prognostic factors related to DSS (Table 3).

Combining the biological effects of aggressive tumor lesions after NAC and the risk ratio of patients in each substage, the single PS variable in Neo-Bioscore was replaced with the ypT stage and ypN stage, each of which was assigned 2 points to create the modified prognostic model (MNeo-Bioscore), and the distribution of scores are presented in Table 1. The total score for each patient was the sum of the scores of each component. The DSS curves based on ypT and ypN are shown in Figure 4. The MNeo-Bioscore extended the scoring system to a total of 10 distinct 5-year DSS, OS, and DFS subgroups ranging from 0 to 9 points. Note that there were no patients in the study cohort with a score of 9. The DSS curves of distinct groups are shown in Figure 5, with significant differences between the groups (P<0.0001). The DSS of patients in the 7/8 group was significantly worse than that of the other groups.

Figure 4 Kaplan-Meier survival curves for disease-specific survival (16) in patients with breast cancer receiving NAC. (A) DSS determined based on the ypT stage. (B) DSS determined based on the ypN stage. NAC, neoadjuvant chemotherapy; ypT stage, posttreatment pathological tumor stage; ypN stage, posttreatment pathological lymph node stage; DSS, disease-specific survival.

Figure 5 Kaplan-Meier survival curves for disease-specific survival (16) in patients with breast cancer receiving NAC based on the MNeo-Bioscore. NAC, neoadjuvant chemotherapy.

The MNeo-Bioscore was internally verified through the bootstrap resampling method with a C-index of 0.82 (95% CI: 0.78–0.87); the ROC curves for 3- and 5-year DSS were shown in Figure 2, and the AUC was 0.85 (95% CI: 0.78–0.92) and 0.85 (95% CI: 0.81–0.90), respectively. The calibration curves of MNeo-Bioscore for 3- and 5-year DSS are shown in Figure 6A,6B, with the MNeo-Bioscore showing a good fit.

Figure 6 Calibration curve for the bootstrap resampling patients based on the 3-year disease-specific survival and 5-year disease-specific survival in the MNeo-Bioscore. (A) 3-year disease-specific survival in the MNeo-Bioscore. (B) 5-year disease-specific survival in the MNeo-Bioscore. DSS, disease-specific survival.

Comparison of CS, PS, Neo-Bioscore, and MNeo-Bioscore

The 5-year DSS, OS, and DFS of the study population were 88.0% (95% CI: 84.6–91.5%), 87.9% (95% CI: 84.5–91.4%), and 76.8% (95% CI: 72.5–81.4%). The 5-year DSS, OS, and DFS of the different subgroups based upon CS, PS, Neo-Bioscore, and MNeo-Bioscore are depicted in Table 4. Due to the small number of patients, those with scores 5 and 6 in Neo-Bioscore were combined into the score 5/6 group; scores 7 and 8 in the MNeo-Bioscore were combined into the score 7/8 group, and score 0 and 1 in the MNeo-Bioscore were combined into the score 0/1 group. No patients had 9 points in the MNeo-Bioscore and none with 7 in the Neo-Bioscore. Compared with the other three staging systems, the MNeo-Bioscore staging system could identify a subgroup of very high-risk patients with a lower 5-year DSS (47.8% 95% CI: 27.0–84.4%).

The C-index of the MNeo-Bioscore based on 5-year DSS was 0.82 (95% CI: 0.78–0.87), and the differences between MNeo-Bioscore and the other three staging systems were significant (P<0.05). The ROC curves of PS and CS for 3- and 5-year are displayed in Figure 2. The AUC value of MNeo-Bioscore was significantly improved compared to Neo-Bioscore and PS in predicting 3-year DSS (P<0.05), while in predicting 5-year DSS, MNeo-Bioscore was superior to PS (P<0.05) but was not statistically different to Neo-Bioscore (P=0.056). The tdROC based on the four staging systems is shown in Figure 7. The MNeo-Bioscore performed the best, followed by the Neo-Bioscore, PS, and CS models in order.

Figure 7 Time-dependent receiver operating characteristic curve based on clinical stage, pathological stage, Neo-Bioscore, and MNeo-Bioscore staging systems. AUC, area under the curve; PS, posttreatment pathological stage; CS, pretreatment clinical stage.

The DCA curves based on CS, PS, Neo-Bioscore, and MNeo-Bioscore are shown in Figure 8A,8B. MNeo-Bioscore showed the best clinical usefulness, followed by the Neo-Bioscore and PS, while the CS demonstrated the worst performance.

Figure 8 Decision curve of the PS, CS, Neo-Bioscore, and MNeo-Bioscore staging systems at 3 and 5 years. (A) Neo-Bioscore, and the MNeo-Bioscore staging systems at 3 years. (B) Neo-Bioscore, and MNeo-Bioscore staging system at 5 years. PS, posttreatment pathology stage; CS, pretreatment clinical stage.

Discussion

It is critical for clinicians to formulate precise individualized treatment plans for patients with breast cancer who fail to achieve pCR after NAC. Developing an effective prognostic model and dividing patients into appropriate prognostic subgroups according to the risk of disease progression is imperative for guiding individualized follow-up treatment. In this study, based upon a cohort composed of Chinese patients treated in one institution, the Neo-Bioscore exhibited superior discrimination and accuracy in the classification of DSS prediction. However, no noticeable improvement was observed over PS. This finding was consistent with the results of Xu et al.’s (9) study based on another Chinese population. Since trastuzumab for the treatment of HER2-positive tumors has been proven to prolong the survival of patients, the Neo-Bioscore defined HER2 negative as an adverse prognostic factor and assigned a score of 1 point (3). HER2-positive patients in China frequently are not treated with trastuzumab due to its high cost. Therefore, the Neo-Bioscore has limited accuracy in predicting the prognosis of HER2-positive patients in that population. Xu et al. (18) incorporated HER2-positive status without trastuzumab treatment into Neo-Bioscore, regarding it as more of a risk factor than a HER2-negative status and assigning it a score of 2 points. Interestingly, the improved Neo-Bioscore showed a better 5-year DSS stratification ability than PS (AUC 0.79 vs. 0.65; P=0.03) (9).

Considering the independent prognostic role of axillary lymph node burden on patient survival after NAC, we incorporated ypT and ypN as independent factors into the Neo-Bioscore, replacing the pathological stage, to refine the Neo-Bioscore scoring system. Compared with the Neo-Bioscore, the MNeo-Bioscore model included the proportion of residual tumor lesions by superimposing the independent prognostic effects of residual tumor and axillary lymph node burden after NAC. As a result, the MNeo-Bioscore identified a subgroup of extremely high-risk patients with a 5-year disease-specific survival of 47.8 (95% CI: 27.0–84.4). Kantor et al. (19) found that RCB, which consists of residual tumor burden and axillary lymph node burden as 2 separate independent factors which had better discrimination of estimated 7-year DFS and OS after NAC compared with Neo-Bioscore. Moreover, Laas et al. (20) further demonstrated that RCB also had superior discriminatory performance to the Neo-Bioscore in breast cancer subtypes based on ER, PR, and HER2 status. These results are consistent with our findings and suggest that the direct additive effect of residual tumor burden and axillary lymph node burden alone is superior to the AJCC-TNM anatomic staging.

The advantages of MNeo-Bioscore were reflected in the prognostic stratification of stage III tumors. For patients with PS stage 0, I, IIA, and IIB, the scores remained unchanged or only adjusted by 1 point before and after model revision in the pathological staging module, and the scores remained the same within subgroups, except for patients with stage IIA whose scores were adjusted from 1 point to 1–2 points. Meanwhile, the score of IIIA patients was changed from 1 to 1–3 points, the score for patients with stage IIIB disease was adjusted from 1 point to 2–3 points, and the stage IIIC patient score was modified from 2 to 2–4 points. Thus, there was a broader range of patient scores in the MNeo-Bioscore. The staging system identified extremely high-risk patients with stage T_3-4N₃ disease with a score of 4 points which are more likely to benefit from postoperative adjuvant chemotherapy. Some disadvantages to the MNeo-Bioscore were also evident. Since tumor size and nodal status were treated entirely independently, the synergistic effect of tumor aggressiveness was ignored. As a result, a subset of patients with stage IIIA scored 1 point, lower than stage IIB or even equal to stage I. This phenomenon does not correspond with the current perceptions regarding tumor staging. Therefore, the improvement of MNeo-Bioscore was mainly reflected in identifying extremely high-risk patients, with relatively attenuated risk stratification for intermediate risk patients. Subsequent work should likely include developing a clustered MNeo-Bioscore, in which patients will be classified into low, intermediate, and high risk based on the total score obtained by the MNeo-Bioscore, after which the predictive efficacy of the modified model should be verified.

The role of Ki67 as a predictor of response to neoadjuvant hormonal therapy is well-established (21,22). However, its value in predicting response to NAC remains controversial (23). Several studies have shown that high Ki67 expression before NAC is significantly associated with the pCR rate after NAC (24,25), while others have found that the Ki67 index after NAC, but not before NAC, to be the only prognostic marker of long-term outcomes (26,27). Furthermore, Matsubara et al. (28) and Cabrera-Galeana et al. (29) found that the reduction rate of Ki67 after NAC was significantly associated with recurrence-free survival (RFS), and the change of Ki67 before and after NAC was also a significant predictor of patient outcome. Sheri et al. (30) combined Ki67 with RCB after NAC as the residual proliferative cancer burden (RPCB), which provided significantly more prognostic information than used alone. This effect might be associated with resistant and highly proliferative disease in patients with high levels of Ki67 in resected tumors that cannot be eradicated by chemotherapy. In this study, 14% was used as the cutoff value of Ki67 before/after NAC, which is also the cutoff value to distinguish between luminal A and luminal B breast cancer subtypes. The results suggested that posttreatment, instead of pretreatment, the Ki67 level was significantly associated with DSS, and neither were independent prognostic factors. Two main reasons were considered to explain this result. Firstly, the cutoff value of the Ki67 level in this study was small. Tanei et al. (26) found that using 50% as the cutoff value of Ki67 before NAC had the highest combined sensitivity for predicting prognosis. Secondly, pretreatment Ki67 and posttreatment KI67 were analyzed as two independent factors in this study. However, considering that Ki67 in different patients responds differently to NAC, the change of Ki67 before and after NAC may provide more significant prognostic implications. Future work will require the introduction of Ki67 into the MNeo-Bioscore and exploring the optimal source and cutoff value of Ki67, which could provide additional prognostic information.

Mittendorf et al. (3) confirmed that the CPS + EG staging systems fit nearly identically when ER status was defined at 2 different cutoff values: 1% and 10%. In our study, 1% was selected as the cutoff value. Although ER and HER2 status were not independent prognostic factors for 5-year DSS in this study, ER-negative status was still an essential factor for a poor prognosis considering the molecular subtypes of breast cancer. Additionally, our study applied the Nottingham (also known as Elston-Ellis) histological grade to define tumor grade rather than the nuclear grade. The Nottingham histological grade modified the Scarff-Bloom-Richardson grading system, which has lower inter-observer and intra-observer variability than the nuclear grading system (31), and which evaluates three parameters: tubule formation ratio, nuclear pleomorphism, and mitotic activity counting. This grading system has become the most widely used histological grading system globally for breast cancer and is accepted by a significant number of authoritative bodies, including the World Health Organization (WHO) (32), Union for International Cancer Control (UICC), and the AJCC (33). Abdelsattar et al. (15) validated the stratified prognostic ability of the CPS + EG staging system using the Nottingham grade instead of nuclear grading in an external cohort. In our study, Nottingham histological grade exhibited an independent association with 5-year DSS in multivariate analysis, and the Neo-Bioscore scoring system still showed an excellent ability to stratify prognosis.

Despite promising findings, some limitations to our study should be addressed. First, the data examined were limited, and retrospective in nature and thus, our study was prone to selection and recall bias. Therefore, these results should be confirmed by prospective studies. Second, the number of cases in the higher and lower score groups was small, and adjacent score groups had to be combined for survival analysis. Future studies should incorporate a more significant number of cases for internal validation and employ external validation of the MNeo-Bioscore in combination with patient data from divergent medical centers.

Conclusions

The Neo-Bioscore staging system could precisely stratify the prognosis of patients with breast cancer after NAC in a single-institution Chinese population. Incorporating the ypN and ypT stages in the Neo-Bioscore staging system may optimize the prognostic stratification of patients receiving NAC to improve patient outcomes.

Acknowledgments

Funding: None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://atm.amegroups.com/article/view/10.21037/atm-21-6738/rc

Data Sharing Statement: Available at https://atm.amegroups.com/article/view/10.21037/atm-21-6738/dss

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://atm.amegroups.com/article/view/10.21037/atm-21-6738/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and approved by the Ethics Committee of the Chinese PLA General Hospital (S2021-582-01). Informed consent was waived due to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cortazar P, Zhang L, Untch M, et al. Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014;384:164-72. [Crossref] [PubMed]
2. Jeruss JS, Mittendorf EA, Tucker SL, et al. Combined use of clinical and pathologic staging variables to define outcomes for breast cancer patients treated with neoadjuvant therapy. J Clin Oncol 2008;26:246-52. [Crossref] [PubMed]
3. Mittendorf EA, Vila J, Tucker SL, et al. The Neo-Bioscore Update for Staging Breast Cancer Treated With Neoadjuvant Chemotherapy: Incorporation of Prognostic Biologic Factors Into Staging After Treatment. JAMA Oncol 2016;2:929-36. [Crossref] [PubMed]
4. Mittendorf EA, Jeruss JS, Tucker SL, et al. Validation of a novel staging system for disease-specific survival in patients with breast cancer treated with neoadjuvant chemotherapy. J Clin Oncol 2011;29:1956-62. [Crossref] [PubMed]
5. Hasegawa Y, Matsubara N, Kogawa T, et al. Neo-Bioscore in Guiding Post-surgical Therapy in Patients With Triple-negative Breast Cancer Who Received Neoadjuvant Chemotherapy. In Vivo 2021;35:1041-9. [Crossref] [PubMed]
6. Soares ICS, Bello MA, Bergmann A, et al. Comparison of the performance of four staging systems in determining the prognosis of breast cancer among women undergoing neoadjuvant chemotherapy. Breast Cancer Res Treat 2021;187:547-55. [Crossref] [PubMed]
7. Xu L, Duan X, Zhou B, et al. Validation of the CPS+EG and Neo-Bioscore staging systems after preoperative systemic therapy for breast cancer in a single center in China. Breast 2018;40:29-37. [Crossref] [PubMed]
8. Lim DW, Giannakeas V, Narod SA. Survival Differences in Chinese Versus White Women With Breast Cancer in the United States: A SEER-Based Analysis. JCO Glob Oncol 2020;6:1582-92. [Crossref] [PubMed]
9. Xu L, Liu Y, Fan Z, et al. Assessment of CPS + EG, Neo-Bioscore and Modified Neo-Bioscore in Breast Cancer Patients Treated With Preoperative Systemic Therapy: A Multicenter Cohort Study. Front Oncol 2021;11:606477. [Crossref] [PubMed]
10. Bear HD, Anderson S, Smith RE, et al. Sequential preoperative or postoperative docetaxel added to preoperative doxorubicin plus cyclophosphamide for operable breast cancer:National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol 2006;24:2019-27. [Crossref] [PubMed]
11. Choi M, Park YH, Ahn JS, et al. Assessment of pathologic response and long-term outcome in locally advanced breast cancers after neoadjuvant chemotherapy: comparison of pathologic classification systems. Breast Cancer Res Treat 2016;160:475-89. [Crossref] [PubMed]
12. Werutsky G, Untch M, Hanusch C, et al. Locoregional recurrence risk after neoadjuvant chemotherapy: A pooled analysis of nine prospective neoadjuvant breast cancer trials. Eur J Cancer 2020;130:92-101. [Crossref] [PubMed]
13. Wang X, He Y, Li J, et al. Impact of response of primary tumor and positive axillary lymph node to neoadjuvant chemotherapy on breast cancer survival. Zhonghua Yi Xue Za Zhi 2015;95:256-9. [PubMed]
14. Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007;25:4414-22. [Crossref] [PubMed]
15. Abdelsattar JM, Al-Hilli Z, Hoskin TL, et al. Validation of the CPS + EG Staging System for Disease-Specific Survival in Breast Cancer Patients Treated with Neoadjuvant Chemotherapy. Ann Surg Oncol 2016;23:3206-11. [Crossref] [PubMed]
16. Sigurdsson H, Baldetorp B, Borg A, et al. Indicators of prognosis in node-negative breast cancer. N Engl J Med 1990;322:1045-53. [Crossref] [PubMed]
17. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making 2006;26:565-74. [Crossref] [PubMed]
18. Xu L, Zhang Z, Liu Q, et al. Validation of CPS+EG, Neo-Bioscore, and modified Neo-Bioscore staging systems after preoperative systemic therapy of breast cancer: Protocol of a retrospective multicenter cohort study in China. Thorac Cancer 2018;9:1565-72. [Crossref] [PubMed]
19. Kantor O, Laws A, Pastorello RG, et al. Comparison of Breast Cancer Staging Systems After Neoadjuvant Chemotherapy. Ann Surg Oncol 2021;28:7347-55. [Crossref] [PubMed]
20. Laas E, Labrosse J, Hamy AS, et al. Determination of breast cancer prognosis after neoadjuvant chemotherapy: comparison of Residual Cancer Burden (RCB) and Neo-Bioscore. Br J Cancer 2021;124:1421-7. [Crossref] [PubMed]
21. Dowsett M, Smith IE, Ebbs SR, et al. Short-term changes in Ki-67 during neoadjuvant treatment of primary breast cancer with anastrozole or tamoxifen alone or combined correlate with recurrence-free survival. Clin Cancer Res 2005;11:951s-8s. [Crossref] [PubMed]
22. Ellis MJ, Tao Y, Luo J, et al. Outcome prediction for estrogen receptor-positive breast cancer based on postneoadjuvant endocrine therapy tumor characteristics. J Natl Cancer Inst 2008;100:1380-8. [Crossref] [PubMed]
23. Nielsen TO, Leung SCY, Rimm DL, et al. Assessment of Ki67 in Breast Cancer: Updated Recommendations From the International Ki67 in Breast Cancer Working Group. J Natl Cancer Inst 2021;113:808-19. [Crossref] [PubMed]
24. Alba E, Lluch A, Ribelles N, et al. High Proliferation Predicts Pathological Complete Response to Neoadjuvant Chemotherapy in Early Breast Cancer. Oncologist 2016;21:150-5. [Crossref] [PubMed]
25. Prihantono P, Hatta M, Binekada C, et al. Ki-67 Expression by Immunohistochemistry and Quantitative Real-Time Polymerase Chain Reaction as Predictor of Clinical Response to Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer. J Oncol 2017;2017:6209849. [Crossref] [PubMed]
26. Tanei T, Shimomura A, Shimazu K, et al. Prognostic significance of Ki67 index after neoadjuvant chemotherapy in breast cancer. Eur J Surg Oncol 2011;37:155-61. [Crossref] [PubMed]
27. Jones RL, Salter J, A'Hern R, et al. The prognostic significance of Ki67 before and after neoadjuvant chemotherapy in breast cancer. Breast Cancer Res Treat 2009;116:53-68. [Crossref] [PubMed]
28. Matsubara N, Mukai H, Masumoto M, et al. Survival outcome and reduction rate of Ki-67 between pre- and post-neoadjuvant chemotherapy in breast cancer patients with non-pCR. Breast Cancer Res Treat 2014;147:95-102. [Crossref] [PubMed]
29. Cabrera-Galeana P, Muñoz-Montaño W, Lara-Medina F, et al. Ki67 Changes Identify Worse Outcomes in Residual Breast Cancer Tumors After Neoadjuvant Chemotherapy. Oncologist 2018;23:670-8. [Crossref] [PubMed]
30. Sheri A, Smith IE, Johnston SR, et al. Residual proliferative cancer burden to predict long-term outcome following neoadjuvant chemotherapy. Ann Oncol 2015;26:75-80. [Crossref] [PubMed]
31. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991;19:403-10. [Crossref] [PubMed]
32. Böcker W. WHO classification of breast tumors and tumors of the female genital organs: pathology and genetics. Verh Dtsch Ges Pathol 2002;86:116-9. [PubMed]
33. Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin 2017;67:93-9.