Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver cancer, with increasing incidence and mortality worldwide (1, 2). ICC is a malignant tumor originating from the proximal secondary bile ducts within the liver parenchyma, often associated with primary sclerosing cholangitis (PSC), liver fluke infection and viral hepatitis are risk factors of ICC (3–5). Surgical resection remains the primary treatment. However, owing to its insidious onset and high invasiveness, liver resection only provides 5-year overall survival ranging from 20% to 40% (6–8). Therefore, identifying prognostic factors is crucial for improving clinical outcomes.

The TNM staging system established by the American Joint Committee on Cancer (AJCC) is considered the gold standard for ICC prognostication (9), but its limitations, especially in T staging (10, 11), underscore the necessity for novel prognostic markers. Emerging evidence highlights the role of malnutrition, inflammation and coagulation in the occurrence and progression of tumors (12–14). Albumin (ALB) is a widely utilized clinical indicator for assessing nutritional status, and previous study have indicated that decreased preoperative serum albumin levels are associated with poor postoperative outcomes in ICC patients (15). In addition, elevated fibrinogen (FIB) may contribute to hypercoagulability and tumor progression across various malignancies, with high fibrinogen linked to poor prognosis in cholangiocarcinoma patients (16). However, not all patients exhibit both nutritional deficiencies and coagulopathy. In recent years, the albumin-to-fibrinogen ratio (AFR), a novel composite biomarker, has been reported to be associated with the prognosis of various cancers, including gallbladder cancer, pancreatic cancer and hepatocellular carcinoma (17–19). However, there are currently no studies investigating the prognostic significance of AFR in ICC patients undergoing surgery.

In this study, we evaluated the prognostic value of preoperative AFR in ICC patients after curative liver resection and developed nomograms to predict their overall survival (OS) and disease-free survival (DFS).

Although ICC is a relatively rare disease, its incidence and mortality have been increasing worldwide in the past few decades (30, 31). Most patients were diagnosed in advanced stages, resulting a 5-year OS of only approximately 30%, while the recurrence rate is as high as 60%, which seriously affects the quality of life and survival time of patients (32–34). Therefore, identifying preoperative prognostic factors for ICC is highly important for improving patient outcomes.

TNM staging system is valuable for ICC prognosis evaluation. However, owing to the high heterogeneity of ICC, TNM staging cannot fully consider the biological behavior of tumors and other factors, especially in T stage. Guo et al. pointed out that T1 ICC was considered a tumor without vascular invasion, but it usually companied by distant metastasis and thus cannot predict distant metastasis in patients with early-stage tumors (11). Our previous study revealed that the optimal cut-off values of T1 staging for isolated ICC with and without vascular invasion were 8 and 3cm, respectively, which are more helpful in predicting the prognosis of ICC (10). The degree of tumor differentiation can directly reflect the morphological and functional similarity between tumor cells and normal tissue cells, but it is affected by subjective factors, and the boundary is blurred, with an error of approximately 20% (30). CA199 is also a commonly used indicator of the prognosis in ICC patients, but it is not expressed in about 10%-15% of Lewis antigen-negative individuals and is also increased in many benign bile duct diseases such as cholelithiasis and cholangitis (35). Inflammatory markers are closely related to the tumor microenvironment and the prognosis of ICC patients. However, it is susceptible to interference from nontumor factors such as basic diseases and infection (36, 37). As a composite biochemical marker, the AFR is relatively stable in circulation and has been used to evaluate the prognosis of a variety of cancers, including colorectal cancer (38), non-small cell lung cancer (39), and prostate cancer (40).

In ICC patients, ALB levels are often reduced due to tumor-related consumption and impaired liver function, which affects the normal metabolism and function of cells and thereby weakens the body’s ability to surveil tumor cells. Moreover, decreased ALB lowers plasma colloid osmotic pressure, leading to tissue edema and a hypoxic tumor microenvironment. This not only promotes adaptive changes in tumor cells, such as the activation of the hypoxia-inducible factor signaling pathway (41, 42), but also promotes malignant behaviors, including tumor angiogenesis, cell proliferation and invasion, enhancing the cytokine-induced inflammatory cancer-associated fibroblast phenotype (43). In addition, ALB has antioxidant capacity. In the process of tumor development and progression, excessive free radicals are produced, which damaging cell DNA, leading to gene mutations and promoting tumor development (44). Low ALB levels cannot effectively remove free radicals, which exacerbates tumor cells proliferation and progression.

As an important part of the coagulation system, fibrinogen is involved in thrombosis and inflammatory responses (45, 46). In cancer patients, elevated fibrinogen is often associated with a hypercoagulable state and increased thrombotic risk, while abnormal fibrinogen expression and dysfunction are related to tumor cell proliferation, invasion and metastasis. The hypercoagulable state caused by high fibrinogen levels promotes the formation of fibrin clots around tumor cells, which not only provide a physical barrier protecting tumor cells from immune surveillance but also serves as a scaffold for tumor cell adhesion and metastasis (47). Fibrinogen also plays an important role in the inflammatory response. In the tumor microenvironment, it acts as an inflammatory mediator and participates in the recruitment and activation of inflammatory cells. These inflammatory cells release interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α) in tumor tissues, which activate multiple signaling pathways to promote tumor cell proliferation, survival, invasion and metastasis (48, 49).

In tumor patients, AFR comprehensively reflects nutritional status, inflammatory response and coagulation function—distinguishing it from single biomarkers like ALB and NLR, where ALB only mirrors nutritional or hepatic synthetic function and NLR solely captures inflammatory status. This integrated nature enables AFR to play a key role in tumor occurrence, development, and prognosis (50, 51). Notably, the biological basis of the AFR’s prognostic value in ICC differs from that in other cancers, primarily due to this ICC’s unique association with cholestasis and liver fibrosis. In ICC, tumor-related biliary obstruction triggers cholestasis, which activates hepatic inflammation and fibrosis. Activated hepatic stellate cells (HSCs) upregulate fibrinogen synthesis through TGF-β/Smad signaling, leading to fibrinogen elevation (52, 53). This mechanism is less prominent in cancers not involving the biliary-liver axis, such as pancreatic cancer where fibrinogen changes are primarily driven by systemic inflammation. Concurrently, liver fibrosis and cholestasis impair albumin synthesis, further reducing AFR. This dual mechanism, involving cholestasis and fibrosis-driven fibrinogen elevation alongside liver dysfunction-induced albumin reduction, endows AFR with ICC-specific prognostic relevance, allowing it to more accurately reflect the tumor’s interplay with hepatic physiology than in other malignancies. Our Spearman correlation analysis confirmed a strong negative association between AFR and Child-Pugh scores (r=−0.243, P<0.001), solidifying AFR’s role in liver function assessment. Unlike the Child-Pugh score, which relies on subjective clinical indicators such as ascites and isolated biomarkers, AFR integrates ALB—a key Child-Pugh score component, with FIB, which serves as a sensitive marker of hepatic synthetic function and fibrosis. This lets AFR capture subtle liver function changes in ICC patients, such as identifying subclinical impairment in low-AFR Child-Pugh A patients who will face higher postoperative mortality and recurrence risks. Thus, AFR is both an independent prognostic factor and a complement to conventional liver function assessments, guiding personalized care. In this study, we determined the optimal AFR cut-off value (8.71) using X-tile analysis and applied PSM to balance confounding factors between the two groups. Our results showed that the AFR is closely associated to the clinicopathological characteristics and prognosis of ICC patients. Patients with low AFR tended to have poorer liver function (high AST, GGT, and TBIL levels, high Child-Pugh scores, and low albumin levels), evaluated tumor marker levels (CEA, CA199, and CA125) and worse tumor conditions (larger tumor volume, poor tumor differentiation, and liver capsule invasion). Survival analysis revealed that OS and DFS were significantly lower in the low AFR group than in the high AFR group (all P<0.001). Additionally, univariate and multivariate Cox regression analyses both identified the AFR as an independent prognostic factor for OS and DFS. Based on the independent prognostic factors for OS and DFS, we constructed nomograms and comprehensively validated their performance. The C index, calibration curve and DCA all revealed that the nomogram outperformed other individual factors in predicting prognosis in almost all ranges and confirmed its clinical utility. ROC curve analysis further confirms that the nomogram provides more accurate prognostic stratification for both OS and DFS compared to other ICC prognostic indicators, including the AJCC 8th edition TNM staging system, TBS, PNI, SII, NLR and PLR. We therefore believe that our nomogram compensates for the limitations of AJCC 8th edition TNM staging system and other widely used prognostic factors, thereby providing a more accurate prediction of prognosis for ICC patients.

Notably, AFR is an inexpensive and easily obtained biomarker, making it highly practical for routine clinical practice. Unlike complex multi-parameter models that often require specialized tests or multiple procedural steps, AFR is derived exclusively from two routine preoperative blood tests, and this reliance on ubiquitous laboratory data eliminates the need for additional invasive procedures or costly analyses, ensures results are typically available within hours, and grants the marker broad applicability even in resource-limited settings where advanced testing is unavailable.

However, the study still has several limitations. First, this is a retrospective study, selection bias is inevitable. Second, AFR was measured only once using a single preoperative blood sample, which cannot fully reveal the dynamic changes throughout the disease course. In future studies, we will incorporate serial AFR measurements during follow-up to further refine its clinical utility as a dynamic monitoring tool. Third, the current study was conducted in Chinese hospitals, with most patients of Chinese ethnicity and a relatively high HBV infection rate, which limits its generalizability of our finding to other populations. In addition, the clinical utility of AFR was not evaluated in the context of contemporary ICC treatments, including patients receiving targeted therapies like FGFR inhibitors or immune checkpoint blockade. Therefore, future international multicenter studies are required to validate the global applicability of AFR and our nomograms, and prospective studies should further systematically evaluate AFR in combination with molecular profiling to determine its role in predicting response to specific therapeutic modalities.

The preoperative AFR is an independent prognostic factor for postoperative OS and DFS in ICC patients, and higher levels of AFR predict better OS and DFS. The nomograms including the AFR provide good prognostic prediction for ICC patients after surgery. On the basis of its availability and low cost, personalized treatment strategies based on the AFR are expected to be developed in the future. In future studies, in addition to prospective studies measuring the AFR at multiple time points, larger scale and multicenter clinical studies can be carried out to further verify the role of the AFR in guiding the treatment decisions of ICC patients.