pCCA is a relatively rare malignancy that lacks clinical evidence on its prognosis evaluation. Surgery is a potentially curative treatment for patients with pCCA, with surgery removing the tumor and obtaining malignant tissue to guide other therapies. However, surgery can affect many organs and systems of body serious complications can occur, or patients may not benefit from the operation in consideration of long-term survival. Thus, the decision of surgery should be made carefully after personalized evaluation. To facilitate the decision-making, we constructed a risk score model including the preoperative blood characteristics to distinguish surgery patients who could be resectable following the actual guidelines(19, 20).
Cancer-related inflammation is a critical component of tumor progression: various growth factors, cytokines, and chemokines are released by inflammatory cells (21). Several studies have revealed that relevant biomarkers can predict the prognosis of patients with cancer, such as C-reactive protein, interleukin-6, and tumor necrosis factor-α receptor 2 in pancreatic cancer (22), SII index in bladder cancer (23), colorectal cancer (24), and prostate cancer (25). Nutrition-related indexes were also widely reported as cancer prognostic predictors. A meta-analysis with 6,320,365 participants demonstrated the association of OS with obesity (defined as a BMI > 30) in patients with cancers (26). Other indexes, like PNI, ALBI, and PINI, were successively constructed and validated in multiple human cancers (14, 27–29). In pCCA, some researchers have verified the predictability of some inflammation-based indexes (PLR, NLR) and nutritional indexes (controlling nutritional status, PNI, C-reactive protein/albumin ratio, and BMI) (17, 30, 31). However, no prognostic prediction models combining inflammatory and nutritional statuses have been reported in pCCA. Therefore, NAB—described in this study—is the first risk score including inflammation- and nutrition-related factors.
The NAB score consists of two biomarkers: neutrophils and ALBI. NAB can also be recognized as the combination of three preoperative peripheral blood contents—neutrophils, TB, and ALB—which could reflect the inflammation and the nutrition status of an individual. Kaplan–Meier analysis showed that patients with a low NAB score had a longer OS and RFS in both datasets. Cox regression similarly demonstrated that NAB score was an independent risk factor for OS and RFS. Subsequently, the AUCs calculated via time-ROC analysis revealed that NAB had a considerable predictive capacity for pCCA. Additionally, the biomarkers factored in the NAB score were all peripheral blood contents and easily obtained in clinical settings before surgery. Overall, NAB provides a new prognostic prediction model based on preoperative examination and reflects both the inflammation and the nutrition status of a patient. This model is different from CA199 (mostly used for diagnosis and surveillance) and other postoperative factors, such as R0 resection, or tumor characteristics (differentiation; lymphovascular, perineural, and/or microvascular invasion; and lymph node involvement) (32). NAB could be recognized as a measurement that indicates whether a patient is suitable for surgery or not.
A nomogram incorporating NAB scores and other independent prognostic factors (CA199, NAB, R0 resection, ASA, perineural invasion, differentiation, LNM, and CCI index) was developed to predict more accurately the outcome of individual patients with pCCA. The nomogram demonstrated potential value in clinical practice in view of C-index, calibration curves, and DCA curves. In addition, compared with the single NAB, the NAB-nomogram exhibited improved net benefits for different OS and RFS prediction. The favorable results were replicated well in the validation cohort. Therefore, our nomogram might be a useful method for evaluating the prognosis of patients with pCCA after resection.
Although NAB and the NAB-based nomogram performed well in this study, there are some limitations in the study. First, the sample size is limited, and the study is a single-center retrospective study. Thus, multicenter studies and large-scale clinical validation are needed to evaluate the external utility of our NAB score and nomogram. Second, our study lacks the exploration of NAB-related and tumor-specific indicators such as characteristic gene sets and their expression profiles.