As a reliable, objective and easily accessible indicator, survival of cancer patients, particularly cancer specific survival (CSS), has been widely used in evaluating and analyzing the long-term prognosis of breast cancer patients[18]. Based on the analysis of a large cohort of SEER database from 2000–2020, patients in the Neo-AT group were found to have worse BCSS than those who underwent AT, utilizing an integrated range of factors incorporated into the PSM method and a nomogram prediction model. To our knowledge, this is the first large cohort study to investigate the survival benefit of Neo-AT therapy among women who underwent BRS and to construct a nomogram to predict the survival of the related population.
Clinicopathological features, such as age, race, tumor grade level, marital status, molecular subtype and treatment methods (chemotherapy/ radiation therapy) were widely recognized as objective and reliable prognostic indicators that can guide clinical therapy in patients with breast carcinoma[22–28]. In the current study, patients in the Neo-AT group tended to be younger, more likely to be unmarried, and comprised a higher proportion of the Black race. Additionally, they exhibited worse tumor grade, more advanced T and N stage, and a higher prevalence of HER2 + or TNBC subtypes compared to those in the AT group. These characteristics of population were consistent with the previous researches [16, 29]. Young women patients often presented with aggressive molecular subtypes of breast cancer with unfavourable prognosis [30, 31]. Furthermore, unmarried patients tended to have fewer financial resources and less psychosocial support compared to married individuals [32]. Therefore, our findings confirmed that patients in Neo-AT group exhibit multiple features related to poor outcomes.
To investigate the true efficacy arose from the administration of Neo-AT, our exploration primarily focused on factors within the PSM cohort. As expected, the confounding variables in both the Neo-AT and AT groups were eliminated after stringent matching and controlling for interference. Kaplan-Meier curve analysis indicated that subjects in the Neo-AT group still had inferior BCSS and OS when compared with those in the AT group. The result demonstrated that the administration of Neo-AT was an independent risk factor for patients undergoing BRS. The underlying reason for this observation may be that the tumors in the Neo-AT group were lager than those in the AT group [29], even though the differences in T/N stage were balanced through PSM analysis.
To mitigate the estimation bias and further investigate the efficiency of treatment methods for patients with BRS, multivariate cox analysis was performed. Tumor grade Ш - Ⅳ in the degree of cell differentiation, advanced T/N stage, more aggressive molecular subtype, and the administration of Neo-AT were identified as independent adverse factors affecting BCSS for patients with BRS. These findings underscored the importance of carefully considering the choice of treatment method for patients with varying pathological characteristics.
According to current researches, BRS followed by Neo-AT has been deemed safe and effective compared to mastectomy [9–11]. Consequently, an increasing number of breast cancer patients are opting for BRS over mastectomy[33]. However, the precise prognosis value of BRS after Neo-AT for patients remained uncertain. The aforementioned results didn’t sufficiently underscore the emphasis we placed on this topic. To provide more accurate evidence on prognosis estimation and therapy evaluation in these patients, a nomogram was constructed based on the predictive factors selected by multivariate Cox analysis in the PSM cohort.
The model developed in the current study exhibited a strong predictive performance for both 3- and 5-year BCSS. Calibration and validation were conducted to verify the model in both the training and validation cohorts, revealing good consistency between the predicted outcomes and the actual results. For instance, for a Black woman who received AT after BRS with a poorly differentiated/undifferentiated TNBC at T2N2 stage, the BCSS probabilities at 3- and 5-years were as high as 77% and 65%, respectively. The OS could be easily calculated in the same manner as depicted in Supplement Fig. 4, and the validation of OS model presented a well effect (Supplement Fig. 5, 6). Therefore, the nomogram developed in the present study appears to be valuable for assessing short- and long-term prognosis, aiding clinicians in developing therapeutic strategies, and enhancing patient compliance among those who undergo BRS.
Reflecting on the limitations of this study and the inherent disadvantages of using the SEER database [34], certain conclusions must be interpreted with caution. Firstly, being a retrospective cohort study conducted using data from the national cancer institute's SEER program database, the reliability of the results could be affected by the uncertain integrity and authenticity of the records [18]. Confounding factors and selection bias may have influenced the findings. Secondly, the selection of primary data may have inevitable effects on the results. The deletion of missing data could alter the real hazard ratio, while the strict matching process may have excluded important variables such as age at diagnosis and marital status from the model, potentially impacting the efficacy of the model. Thirdly, only observable confounders could be balanced and controlled by the administration of PSM, whereas missing data may result in the loss of important confounders, leading to deviations from the real situation. Furthermore, the absence of information on potential confounders such as family history, insurance coverage, patient anxiety, detailed regimens of endocrine therapy, frailty, or comorbid conditions known to be related to receipt of specific treatments, may impact clinical decisions and breast cancer prognosis [35–37]. However, such factors were not available in the SEER database. Additionally, there was a lack of information about local recurrence and adverse effects of normal tissue after radiotherapy, which are essential for guiding the optimization of treatment options [17]. Prospective cohort studies, which may address these deficiencies, needed to be conducted, though it would require considerable time.