This study aimed to assess the association between AGR and long-term oncologic survival in patients who underwent R0 resection for HCCA. AGR is calculated by dividing the concentration of albumin by the concentration of globulin and has been used as an effective tool for evaluating liver function. In clinical practice, the normal AGR of patients is > 1.0. Consequently, this value was considered a threshold to divide all patients into the IAGR and NAGR groups in our study. A retrospective study of 123 patients with cholangiocarcinoma from China demonstrated the prognostic value of the preoperative AGR for OS. They found that it is superior to measuring globulin or albumin alone for predicting mortality in cholangiocarcinoma patients. However, compared with our research, they lacked research on RFS, had only a small sample size, and did not analyze the various types of cholangiocarcinoma in subgroups. We used a multicenter database and only selected patients with HCCA to compensate for the abovementioned deficiencies. Notably, for the first time, preoperative IAGR was confirmed to be an independent risk factor for predicting long-term oncologic survival in patients with HCCA. Although the incidence of HCCA is less than 1% of systemic tumors, the long-term prognosis is poor. Therefore, we enrolled as many patients as possible with adequately long periods of follow-up. In addition, we carried out an analysis of multiple prognostic risk factors, including clinical, pathological, biochemical, and operative variables.
As a carrier of sex hormones, globulin is commonly used to reflect the inflammatory status of patients, along with most proinflammatory proteins (including complement components, high-sensitivity C-reactive protein, immunoglobulins, etc.). Chronic inflammation may produce chronic oxidative stress and generate oxygen-free radicals. Oxygen-free radical signaling can activate HIF-1, which in turn upregulates gene expression for glycolysis, angiogenesis, and other cellular metabolisms. This facilitates the transformation of ordinary cells into cancer cells and tumor growth. Based on the above reasons, many scholars believe that chronic inflammation plays a crucial role in various aspects of tumor progression and development, including proliferation, metastasis, and survival, in patients with malignant tumors[20, 21]. In previous studies, serum globulin was suggested to be an independent predictor of long-term outcomes in some diseases, such as lung cancer, gastric cancer, and esophageal squamous carcinoma[22–25]. A Japanese research team found that elevated serum non-albumin proteins and immunoglobulins were strongly related to common variants of tumor necrosis factor receptor superfamily member 13B (TNFRSF13B). TNFRSF13B is closely associated with tumor progression. This finding suggested a potential relationship between the development of malignant tumors and serum globulin.
Albumin may affect the survival of patients with malignant tumors through several mechanisms. Serum albumin is commonly used as a biomarker of liver function and nutritional status. Moreover, in the clinical setting, serum albumin forms an important component in many scores that are designed to reflect liver function, such as Child–Pugh grading and ALBI grading. A low serum albumin level may be related to liver dysfunction. Poor liver function and malnutrition are common reasons for the worse long-term prognosis of patients with hepatobiliary tumors. Furthermore, chronic inflammation may lead to a low level of serum albumin. Mantovani et al. indicated that tumor cells and immune cells stimulate the release of various growth factors and cytokines (including interleukin-6, tumor necrosis factor, etc.) during the process of tumor-associated systemic inflammation. These growth factors and cytokines inhibit serum albumin production and stimulate tumor growth and progression through the stimulation of the migration and subversion of the host immune response[29–31]. In addition, Laursen et al. found several anticancer mechanisms of serum albumin, such as antioxidant function. Previous studies reported that the development and progression of tumor cell lines in vitro, such as human breast cancer cell lines, were regulated by high serum albumin concentrations.
Multivariable Cox regression analyses revealed that preoperative IAGR was the only indicator of liver function that was significantly associated with worse OS and RFS in patients with HCCA who underwent R0 resection. This result should remind hepatic surgeons of the importance of preoperative AGR in predicting survival and recurrence following HCCA R0 resection. A retrospective study of lung adenocarcinoma indicated that serum albumin and globulin concentrations were severely affected by blood volume, which may reduce their predictive value. The hydration status of patients commonly influences the serum levels of albumin and globulin. However, AGR is not easily affected by the hydration status of the patient. Azab et al. found that serum albumin and serum globulin levels are increased or decreased proportionately in conditions such as dehydration and fluid retention. A multicenter study from China observed that the preoperative IAGR was associated with worse survival and recurrence conditions in patients with HCC after surgery and that the performance of AGR in predicting prognosis after liver resection was the best among the key indexes reflecting liver function. In addition, it is worth noting that in this study, AGR was found to surpass other indicators of liver function, such as albumin and ALT, in predicting OS and RFS according to the result of time-dependent AUROC analysis. Considering that AGR is a valuable biomarker reflecting liver function and that AGR is not influenced by hydration status, we concluded that it can serve as a more advantageous biomarker for HCCA prognosis than other liver function indicators alone.
In this study, laboratory parameters, including hemoglobin, albumin, globulin, INR, and CA 19 − 9, showed significant differences between the IAGR and NAGR groups. However, it was not appropriate to use propensity score matching to balance the baseline characteristics between the two groups to examine the association between the preoperative AGR and prognosis, as it may lead to more selection biases between the IAGR and NAGR groups. Consequently, in our study, we used classical statistical approaches (univariable and multivariable Cox regression analyses) to determine the relationship between the preoperative IAGR and poor prognostic outcomes in patients following R0 resection for HCCA while adjusting for the other risk factors.
In addition to the preoperative IAGR, a few other independent risk factors that lead to worse OS and RFS were identified in our study. These independent risk factors included CA19-9 > 150 U/L, maximum tumor size > 3 cm, macrovascular invasion, microvascular invasion, poor tumor differentiation, and lymph node metastasis. All these independent risk factors have been reported in a previous study[37–42].
There are several limitations to this study. First, this study was a retrospective study, which may lead to biases, including selection, confounding, and information biases. Nonetheless, this study used a large sample from a multicenter database to improve the accuracy of the results. Second, this study did not measure specific inflammatory and nutritional indicators, such as neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, sarcopenia, and C-reactive protein. Inflammatory and nutritional indicators may reflect liver function. This study compared the predictive ability of different liver function indicators, including AGR, albumin, and ALT, in predicting long-term oncologic survival. Third, this study used 1.0 as the cutoff value in normal clinical practice to divide all patients into the INGR and NAGR groups. Identifying the best cutoff value for AGR to determine prognosis will be our focus of future research.