A deep learning-based prognostic prediction model for distal cholangiocarcinoma incorporating the metabolism-inflammation marker monocyte-to-high-density lipoprotein cholesterol ratio

Introduction

Cholangiocarcinoma (CCA) represents to be the most prevalent malignancy of the biliary tract, with a global incidence ranging from 0.3 to 6.0 per 100,000 individuals (1). Distal CCA (dCCA) refers to extrahepatic CCA (eCCA) arising beyond the hepatic hilum, specifically from the mid-to-distal portion of the common bile duct, which accounts for 30–40% of extrahepatic biliary tumors and 10–14% of periampullary malignancies (2). Despite advancements in treatment strategies, CCA remains largely refractory to chemotherapy, and early radical surgical resection continues to be the cornerstone of curative treatment (3). However, the prognosis of dCCA remains dismal, with a median survival of only 24 months following diagnosis (4). While research on CCA has predominantly focused on intrahepatic CCA (iCCA), clinical investigations specifically targeting dCCA remain scarce. As a result, effective prognostic biomarkers capable of predicting tumor recurrence and mortality remain lacking, impeding the development of risk-adapted therapeutic strategies to improve patient outcomes.

Accumulating evidence highlights the critical role of inflammation-metabolism interactions in tumor immunity and the pathogenesis of CCA. Chronic inflammation of the biliary epithelium and bile stasis are well-established risk factors for disease initiation and progression (5). Various cytokines, growth factors, tyrosine kinases, and bile acids modulate the cell cycle dynamics of biliary epithelial cells, while inflammatory cytokines activate inducible nitric oxide synthase (iNOS), leading to excessive nitric oxide (NO) production, oxidative DNA damage, and inhibition of DNA repair mechanisms, thereby fostering tumorigenesis and malignant progression (4). Among immune cell populations, monocytes have emerged as pivotal regulators of cancer progression. Peripheral blood monocyte levels not only reflect systemic inflammatory status but also provide insights into the inflammatory landscape within the tumor microenvironment (TME) (6). Within the TME, monocytes constitute a major infiltrating leukocyte subset and are recruited via the CCL2/CCR2 axis. Regulatory T cells (Tregs) further modulate monocyte differentiation into tumor-associated macrophages (TAMs) through the secretion of immunoregulatory cytokines (7).

TAMs exert context-dependent effects on tumor biology. Recent single-cell RNA sequencing analyses by Yao et al. revealed that monocyte-derived TNFSF13B interacts with TFRC receptors on tumor cells, triggering ferroptosis in CCA cells (8). Conversely, findings by Chen et al. demonstrated that CCA cells engage in paracrine SHH signaling to modulate TAM polarization and transforming growth factor beta 1 (TGF-β1) secretion, thereby fostering tumor proliferation (9). Beyond immune cell regulation, metabolic factors such as high-density lipoprotein (HDL) have been implicated in oncogenesis across multiple malignancies, including pancreatic, thyroid, breast, gastric, and colorectal cancers (10,11). HDL exerts anti-inflammatory and antioxidant effects, modulates angiogenesis, regulates intracellular signaling cascades, and enhances antitumor immune responses, collectively contributing to tumor suppression.

Monocyte-to-HDL ratio (MHR) has recently emerged as a novel metabolism-inflammation-associated immune biomarker, demonstrating significant prognostic implications across various malignancies (12-16). However, its role in dCCA remains unexplored, and its prognostic significance in this malignancy remains undefined. Given the lack of established biomarkers for dCCA prognosis, this study aims to elucidate the prognostic relevance of MHR in dCCA patients and to construct a novel predictive model incorporating inflammation-metabolism-related immune biomarkers. Through a refined understanding of MHR’s impact on dCCA progression and patient survival, these findings may provide a framework for enhanced risk stratification and precision oncology strategies in this aggressive malignancy. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-968/rc).

Methods

Ethics approval and consent to participate:

The protocol adhered to the Declaration of Helsinki and its subsequent amendments, and received approval from the Ethics Committee of The First Affiliated Hospital of Bengbu Medical University (No. 2024-D-302). Owing to its retrospective design, the requirement for written informed consent was waived.

Patients and clinicopathological factors

Patient selection

We performed a retrospective review of dCCA cases undergoing surgery at The First Affiliated Hospital of Bengbu Medical University (Bengbu, China) from January 2011 through July 2023. Eligibility was determined according to predefined inclusion and exclusion criteria. All authors accept responsibility for the accuracy and integrity of the work and will address any related concerns.

Inclusion and exclusion criteria

Inclusion criteria: (I) surgical resection for a biliary space-occupying lesion between January 2011 and July 2023; (II) preoperative assessment confirmed the absence of contraindications to surgery; (III) complete tumor resection achieved intraoperatively; (IV) postoperative pathology confirmed CCA of the adenocarcinoma subtype, with the tumor located distal to the cystic duct confluence; (V) availability of complete clinical and follow-up data.

Exclusion criteria: (I) patients who declined participation or withdrew from the study; (II) patients with preexisting hyperlipidemia or long-term use of lipid-lowering agents before surgery; (III) patients with concurrent malignancies; (IV) patients who died during the perioperative period.

Patients grouping and definition

For all included patients, preoperative routine blood tests and biochemical analyses conducted within one week before surgery were retrieved. Monocyte count and serum HDL cholesterol (HDL-c) levels were extracted, and the MHR was calculated using the following formula:

MHR=Monocyte count(109/L)/HDL-c(mmol/L)[1]

We generated receiver operating characteristic (ROC) curves to evaluate the predictive value of preoperative MHR for 1-year survival and estimated the AUC and optimal threshold, and the corresponding area under the curve (AUC) and optimal cut-off value were determined. Patients were then categorized into groups based on the optimal MHR cut-off value for intergroup comparisons.

Variable analysis & follow-up strategy

Baseline preoperative characteristics, intraoperative parameters, and postoperative recovery metrics were extracted from the institutional electronic medical record system. Perioperative differences among patient subgroups were analyzed. Patients were followed up at one and 3 months postoperatively, every 3 months for the first two years, and every 6 months thereafter. Follow-up was conducted via telephone interviews and outpatient visits. Primary endpoint: overall survival (OS), defined as time from surgery to death. Secondary endpoint: tumor recurrence, including both locoregional and distant recurrence. Follow-up assessments included hematological tests [complete blood count, biochemical panels, carbohydrate antigen 19-9 (CA19-9)] and computed tomography (CT) evaluations to assess disease status, recurrence, and ongoing treatment regimens.

Statistical analysis

Continuous data are presented as mean ± standard deviation (SD) or median [interquartile range (IQR)] as appropriate and compared using t-tests or Mann-Whitney U tests; categorical variables were assessed with χ² or Fisher’s exact tests. Survival was analyzed by Kaplan-Meier with log-rank comparisons. Risk factors were evaluated using univariable and multivariable Cox models. Variable importance was ranked with the “randomForestSRC” package and least absolute shrinkage and selection operator regression (“LASSO regression”). Concordance index (c-index) was used to compare the new model’ prognostic ability. All statistical analyses were conducted using SPSS 26.0 software, and a P value <0.05 was considered statistically significant.

Construction and validation of the prognostic nomogram

Predictors retained in multivariable models and ranked highly by recurrence-free survival (RSF)/LASSO were combined to build a nomogram. Performance was assessed by ROC (constructed using the “pROC” package in R)/AUC and bootstrap-corrected c-index (“rms” package in R), calibration was inspected with calibration plots, and clinical utility was examined using decision-curve analysis (DCA) (“ggDCA” package in R).

Results

General characteristics & perioperative outcomes

A total of 188 patients were included in this study, comprising 124 males and 64 females, with a median age of 66.0 (60.0, 73.0) years. Regarding the initial clinical presentation, 159 patients (84.6%) presented with jaundice, 18 (9.6%) with abdominal pain, 5 (2.7%) with nonspecific gastrointestinal symptoms, and 6 (3.2%) were asymptomatic, with tumors incidentally detected during routine health examinations. Among patients with jaundice as the predominant symptom, 96 (51.1%) underwent preoperative biliary decompression, including 19 cases managed via endoscopic retrograde cholangiopancreatography (ERCP) and 77 cases via percutaneous transhepatic biliary drainage (PTBD).

All patients underwent curative-intent resection, with 176 cases undergoing standard pancreatoduodenectomy, 8 cases undergoing hepatopancreatoduodenectomy, and 4 cases undergoing total pancreatoduodenectomy. The median operative duration was 10.0 (8.0, 11.8) hours, with a median intraoperative blood loss of 500.0 (400.0, 775.0) mL. Intraoperative blood transfusion was required in 74 patients (39.4%). Postoperative complications occurred in 54 patients (28.7%), with the most common being pancreatic fistula (n=36, 19.1%), intra-abdominal infection (n=29, 15.4%), delayed gastric emptying (n=16, 8.5%), intra-abdominal hemorrhage (n=12, 6.4%), gastrointestinal bleeding (n=4, 2.1%), and biliary fistula (n=4, 2.1%). The median length of postoperative hospitalization was 22.0 (17.0, 33.0) days.

Histopathological analysis confirmed R0 resection in 176 patients (93.6%), while the remaining cases underwent R1 resection. All tumors were histologically classified as dCCA of the adenocarcinoma subtype, with 24 cases (12.7%) exhibiting well-differentiated histology, 98 cases (52.1%) moderately differentiated, and 66 cases (35.1%) poorly differentiated. Additionally, 28 patients (14.9%) had portal venous system invasion, and 86 patients (45.7%) had lymph node metastases.

Disease-free survival (DFS) & overall prognosis

As of January 2024, the median DFS for the cohort was 22 months, with 1-, 2-, and 3-year DFS rates of 63.2%, 48.3%, and 38.0%, respectively. The median OS was 26 months, with 1-, 2-, and 3-year OS rates of 76.5%, 51.8%, and 36.4%, respectively.

Patient stratification based on MHR

To evaluate the prognostic value of preoperative MHR in predicting 1-year postoperative survival, an ROC curve was constructed (Figure 1), yielding an AUC of 0.6512 [95% confidence interval (CI): 0.5538–0.7485]. The optimal MHR cut-off value for prognostic stratification was determined to be 0.74, with a sensitivity of 67.4% and specificity of 63.4% in predicting 1-year survival outcomes. Based on this cut-off threshold, patients were categorized into a low MHR group (MHR ≤0.74, n=106) and a high MHR group (MHR >0.74, n=82) for further comparative analysis.

Figure 1 ROC curve of preoperative MHR for predicting 1-year postoperative survival in patients with distal cholangiocarcinoma. AUC, area under the curve; MHR, monocyte-to-high-density lipoprotein cholesterol ratio; ROC, receiver operating characteristic.

Comparison of prognostic outcomes in different groups

The comparison of baseline perioperative characteristics between the low and high MHR groups is presented in Table 1. Patients in the low MHR group exhibited significantly lower preoperative neutrophil count, monocyte count, and total serum bilirubin levels, whereas their preoperative serum albumin and HDL-c levels were significantly higher compared to those in the high MHR group (P<0.05). No significant differences were observed between the two groups regarding tumor-related characteristics. Patients in the low MHR group had a median DFS of 42 months, whereas those in the high MHR group had a median DFS of 18 months. The 1-, 2-, and 3-year DFS rates were 71.4%, 53.5%, and 45.7% in the low MHR group and 52.7%, 35.9%, and 27.1% in the high MHR group (P=0.002, Figure 2A). In terms of OS, the median OS in the low MHR group was 36 months, significantly longer than the 17 months observed in the high MHR group. The 1-, 2-, and 3-year OS rates were 86.4%, 64.1%, and 46.5% in the low MHR group, compared to 63.9%, 35.9%, and 23.2% in the high MHR group (P<0.001, Figure 2B).

Figure 2 Comparison of long-term disease-free survival (A) and overall survival (B) between the low MHR group and the high MHR group. MHR, monocyte-to-high-density lipoprotein cholesterol ratio.

High MHR as an independent risk factor for tumor recurrence in dCCA

To determine the independent predictors of postoperative tumor recurrence, univariate & multivariate analysis was conducted using tumor recurrence as the dependent variable and preoperative, intraoperative, pathological, and postoperative variables as independent factors. multivariate analysis identified the following as independent risk factors for tumor recurrence (Table 2): preoperative high MHR [hazard ratio (HR) =1.218, 95% CI: 1.053–1.408], elevated preoperative carcinoembryonic antigen (CEA) (HR =1.761, 95% CI: 1.152–2.692), increased intraoperative blood loss (HR =1.054, 95% CI: 1.010–1.100), poor tumor differentiation (HR =1.765, 95% CI: 1.155–2.695), greater tumor invasion depth (HR =2.339, 95% CI: 1.162–4.707), lymph node metastasis (HR =2.269, 95% CI: 1.509–3.414) and portal system invasion (HR =2.428, 95% CI: 1.407–4.189). Patients with high preoperative MHR and CEA levels, increased intraoperative blood loss, poor tumor differentiation, tumor invasion depth >12 mm, lymph node metastasis, and portal vein invasion had a significantly higher risk of postoperative tumor recurrence.

High MHR as an independent risk factor for long-term survival in dCCA

To evaluate factors influencing long-term survival in dCCA, univariate & multivariate analysis was also performed using postoperative survival as the dependent variable, and the following factors were identified as potential prognostic determinants (Table 3): preoperative high MHR (HR =1.230, 95% CI: 1.060–1.427), elevated preoperative CA19-9 (HR =1.010, 95% CI: 1.002–1.018), poor tumor differentiation (HR =1.576, 95% CI: 1.045–2.377), greater tumor invasion depth (HR =2.011, 95% CI: 1.065–3.798), presence of lymph node metastasis (HR =1.906, 95% CI: 1.289–2.817). Patients with high preoperative MHR and CA19-9 levels, poor tumor differentiation, tumor invasion depth >12 mm and lymph node metastasis exhibited significantly worse long-term survival outcomes. In order to pinpoint the most impactful prognostic factors, we applied a random survival forest algorithm to assess the importance of each variable. This analysis revealed that MHR, CA19-9, lymph node metastases, portal system invasion, and tumor differentiation were among the top-ranking predictors (Figure 3). Additionally, we incorporated a LASSO regression to further strengthen the methodological rigor of our deep learning model. Ranking the variables by their LASSO coefficients corroborated that MHR, lymph node metastasis, portal system invasion, CA19-9 and tumor differentiation consistently emerged among the top five predictors (Figure 4). This finding aligns well with the results of the multivariate analysis and the final set of risk factors included in our nomogram.

Figure 3 Random survival forest plot for patients with distal cholangiocarcinoma. ALB, albumin; ALT, alanine aminotransferase; AST, aspartate transaminase; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; HDL-c, high-density lipoprotein cholesterol; MHR, monocyte-to-high-density lipoprotein ratio; TB, total bilirubin.

Figure 4 LASSO-Cox regression for variable selection. (A) LASSO coefficient profiles of the 23 variables. The x-axis represents the log-transformed lambda (λ), and the y-axis shows the coefficients. (B) Ten-fold cross-validation for tuning parameter selection in the LASSO-Cox model. The x-axis represents log(λ), and the y-axis shows the partial likelihood deviance. LASSO, least absolute shrinkage and selection operator.

Subgroup analysis of MHR as an independent risk factor on the long-term survival in dCCA

In the above experiments, we know that lymph node metastasis and portal vein system invasion are important prognostic indicators of dCCA patients. In order to exclude the influence of the above two factors on the long-term survival of dCCA patients, we conducted a subgroup analysis. Consistent with the previous conclusion, preoperative MHR is an important prognostic indicator of dCCA patients without lymph node metastasis (Figure 5): Patients with a high preoperative MHR had a poorer OS (χ²=5.806, P=0.02). Among dCCA patients without portal vein system invasion, those with a high preoperative MHR had significantly poorer OS (χ²=18.62, P=0.001) and DFS (χ²=8.002, P=0.005). Subgroup analysis suggested that MHR is an independent and potentially clinically valuable important prognostic factor.

Figure 5 Subgroup analysis of the impact of MHR on the long-term survival of patients with dCCA. dCCA, distal cholangiocarcinoma; MHR, monocyte-to-high-density lipoprotein cholesterol ratio.

Development and validation of the prognostic model

A nomogram for OS was developed using data from the entire patient cohort. Figure 6 illustrates the final nomogram, which integrates preoperative MHR, CA19-9, lymph node metastases, portal system invasion, and tumor differentiation, accompanied by its corresponding calibration plots and ROC curves for evaluation. These independent prognostic factors were integrated to predict individualized OS probabilities, providing a clinically relevant tool for risk stratification and decision-making. We employed benefit curves to evaluate the predictive performance of the model. As shown in Figure 7, these benefit curves provide insight into how the model performs across various threshold levels.

Figure 6 Prognostic prediction model nomogram and calibration & AUC curve based on preoperative MHR, CA19-9, LNM, PSI and tumor differentiation. The asterisks (*) indicate the statistical significance levels: *, P<0.05; **, P<0.01; ***, P<0.001. AUC, area under the curve; CA19-9, carbohydrate antigen 19-9; CI, confidence interval; LNM, lymph node metastases; MHR, monocyte-to-high-density lipoprotein cholesterol ratio; OS, overall survival; PSI, portal system invasion.

Figure 7 Benefit curves at different time points. CA19-9, carbohydrate antigen 19-9; LNM, lymph node metastases; MHR, monocyte-to-high-density lipoprotein cholesterol ratio; PSI, portal system invasion.

Discussion

This study is the first to investigate the prognostic significance of the MHR, a novel metabolism-inflammation-associated immune biomarker, in dCCA. Our findings establish that preoperative MHR serves as a novel prognostic factor for both DFS and OS in dCCA patients. Elevated MHR correlates with an increased risk of tumor recurrence and poorer long-term survival outcomes, highlighting its potential clinical relevance.

Inflammation has been recognized as a critical driver of cholangiocarcinogenesis, with chronic biliary inflammation and bile stasis playing key roles in tumor initiation and progression. Inflammatory cells contribute to tumor proliferation by secreting interleukin-6 (IL-6), which activates the mitogen-activated protein kinase (MAPK) cascade, leading to the transcription of oncogenic genes. Additionally, activation of epidermal growth factor receptor (EGFR) has been implicated in CCA pathogenesis, and the overexpression of ERBB2, a member of the EGFR family, has been shown to induce biliary epithelial tumorigenesis in murine models.

Monocyte levels in peripheral blood reflect both systemic inflammatory status and the inflammatory microenvironment within tumors. Circulating monocytes infiltrate tumor tissues and differentiate into TAMs, which exist in two distinct polarization states: classically activated M1 macrophages and alternatively activated M2 macrophages. M1 macrophages exhibit pro-inflammatory and antitumor properties, responding to lipopolysaccharides (LPS), interferon-gamma (IFN-γ), and tumor necrosis factor-alpha (TNF-α). In contrast, M2 macrophages, primarily stimulated by IL-4 and IL-13, exhibit pro-tumorigenic activity. In the TME, TAMs predominantly acquire an M2-like phenotype. This polarization is driven by factors secreted from tumor and stromal cells, such as monocyte chemoattractant protein-1 (MCP-1/CCL2), colony-stimulating factor-1 (CSF-1), and vascular endothelial growth factor-A (VEGF-A) (17). Dwyer et al. showed that CCA exhibits an upregulation of the TWEAK (TNF-like weak inducer of apoptosis)/Fn14 signaling axis, which promotes monocyte recruitment and TAM differentiation through NF-κB-mediated MCP-1 expression, ultimately driving tumor progression (18).

Dyslipidemia has also been implicated in tumorigenesis, particularly in cancer metastasis. Low levels of HDL-c have been associated with an increased risk of distant metastases in multiple malignancies (19). Tumor cells actively uptake cholesterol, leading to a systemic reduction in HDL levels, which in turn facilitates tumor progression (20). In addition, HDL exerts immunomodulatory effects by depleting cholesterol in TAMs, which diminishes the tumor-promoting activity of these cells. Concurrently, HDL elevation enhances neutrophil-mediated anti-inflammatory responses, promotes the activation of CD8⁺ and CD4⁺ T cells, and modulates the function of antigen-presenting cells and the complement system (19). ApoA-1, the primary protein component of HDL, has been shown to influence tumor and immune cell proliferation within the TME (21). Ma et al. demonstrated that ApoA-1 downregulates the MAPK signaling pathway, induces apoptosis, and suppresses hepatocellular carcinoma proliferation (22).

In summary, MHR represents a novel biomarker that integrates both monocyte-mediated pro-tumorigenic effects and HDL-associated tumor-suppressive mechanisms, offering superior clinical utility compared to monocyte or HDL levels alone. MHR has been widely reported as a prognostic indicator in various malignancies, strongly correlating with tumor initiation, progression, and prognosis. Our study demonstrates that dCCA patients with high preoperative MHR exhibit significantly shorter RSF and OS, which indicate that MHR may be a valuable prognostic biomarker for dCCA. Specifically, when MHR exceeds 0.74, patients face an increased 1-year mortality risk, indicating that early intervention with anti-inflammatory and lipid-lowering therapies may help mitigate recurrence and mortality risk, ultimately improving survival outcomes. Furthermore, tailored postoperative treatment strategies and follow-up protocols for high-MHR patients may enhance OS rates.

CA19-9 represents to be a classic gastrointestinal tumor biomarker and is widely used as a blood-based marker for both the clinical diagnosis of CCA and the assessment of its prognosis (23,24). A study by Jiang et al. (25) demonstrated the high sensitivity and specificity of CA19-9 in diagnosing CCA (74% and 82%, respectively). Furthermore, Tella et al. (26) reviewed the United States National Cancer Database, which included data on 2,100 patients with eCCA. Among these patients, approximately 1,500 (over 70% of the cohort) had elevated CA19-9 levels. Notably, the median survival for patients with elevated CA19-9 was significantly shorter than that for patients with normal CA19-9 levels (8.5 vs. 16.0 months, P<0.01). Moreover, an elevated CA19-9 level was confirmed to be an independent prognostic factor for long-term survival (HR =1.72, 95% CI: 1.46–2.02).

Furthermore, our data indicate a significant correlation between CA19-9 levels and lymph node metastasis, which may explain the poor long-term survival observed in patients with elevated CA19-9 levels. In patients eligible for surgery, lymph node metastasis is well recognized as one of the most critical prognostic factors influencing long-term survival (27,28). Kiriyama et al. (29) conducted a retrospective study on 370 patients with dCCA who underwent pancreaticoduodenectomy (PD) across 24 Japanese hospitals. Their findings revealed that patients without lymph node metastasis had a significantly higher 3-year survival rate compared to those with lymph node metastasis (66.3% vs. 36.1%, P<0.001). Similarly, Byrling et al. (30) identified lymph node metastasis as the only independent risk factor for long-term survival in dCCA, with an associated risk approximately 2.88 times higher (P=0.016). Our study also demonstrated that patients without lymph node metastasis had a significantly better long-term prognosis compared to those with lymph node metastasis (P=0.001). Therefore, accurate preoperative assessment of lymph node metastasis is crucial for predicting long-term survival outcomes. Meanwhile, given the anatomical proximity of the distal common bile duct to the lymphatic and portal venous systems, patients with dCCA and portal vein invasion are at a higher risk of lymph node metastasis. Miura et al. (31) retrospectively analyzed 129 patients with dCCA who underwent PD and found that the rate of lymph node metastasis was significantly higher in patients with portal vein invasion compared to those without vascular involvement (87% vs. 37%, P=0.005). This evidence underscores the critical interplay between CA19-9 elevation, lymph node metastasis, and portal vein invasion, highlighting the importance of comprehensive preoperative evaluation to guide risk stratification and treatment decisions in dCCA patients.

Tumor differentiation has been widely recognized as an independent prognostic factor in various malignancies, including dCCA. Poorly differentiated tumors exhibit more aggressive biological behavior, including enhanced proliferative capacity, increased invasive potential, and a higher likelihood of metastasis, ultimately leading to worse clinical outcomes. Histopathologically, dCCA is classified into well-, moderately-, and poorly-differentiated tumors based on glandular architecture, nuclear atypia, and mitotic activity. The degree of differentiation significantly correlates with patient survival. A retrospective study demonstrated that dCCA patients with well-differentiated tumors had a significantly higher 5-year OS rate compared to those with poorly differentiated tumors (P=0.027) (32). Poor differentiation is associated with higher rates of early recurrence and resistance to chemotherapy, contributing to the unfavorable prognosis. Another systematic review and meta-analysis also showed that a poor/moderate histological grade was associated with poor survival of patients with CCA (33). Several molecular pathways have been implicated in the association between poor differentiation and worse prognosis in dCCA. Epithelial-mesenchymal transition (EMT) is a well-established mechanism contributing to the loss of epithelial characteristics and acquisition of mesenchymal properties, leading to enhanced invasiveness and metastatic potential. Poorly differentiated dCCA tumors exhibit upregulation of N-cadherin, Snail, and ZEB1, alongside a downregulation of E-cadherin, reflecting an active EMT program (34). Additionally, TP53 mutations and overexpression of MDM2 have been associated with poor differentiation, genomic instability, and resistance to apoptosis, further exacerbating disease aggressiveness (35), which is along with our study, suggesting that the differentiation status of dCCA is closely associated with prognosis.

However, this study is not without limitations. First, as a single-center retrospective cohort study with a relatively small sample size (n=188), further large-scale, multicenter investigations and high-level prospective studies are warranted to validate our findings. Second, loss to follow-up in some patients may introduce selection bias, potentially influencing the results. Nonetheless, CCA, characterized by poor prognosis, has emerged as a major global health concern, with many patients facing early recurrence and high mortality rates. dCCA, as a distinct subtype, exhibits differences in treatment response and prognosis compared to iCCA. While previous research has largely focused on iCCA, dCCA remains understudied, and no reliable biomarkers exist to predict recurrence and mortality risks in these patients. Our study suggests that preoperative MHR is a simple, cost-effective, and readily accessible biomarker that may aid in optimizing treatment decisions. Early therapeutic intervention targeting monocyte-driven inflammation and HDL metabolism may provide a novel avenue for improving prognosis and quality of life in dCCA patients. However, further research is required to elucidate the molecular mechanisms underlying MHR’s role in tumor progression and prognosis.

Conclusions

By combining a readout of systemic inflammation with HDL-related lipid status, the MHR emerges as an informative prognostic index in dCCA. In parallel, CA19-9 concentration, nodal involvement, portal system invasion, and histologic differentiation are each independently associated with survival. Integrating these variables within our deep-learning-based prognostic model enables earlier risk triage and more targeted postoperative management, with potential to improve clinical outcomes.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-968/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-968/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-968/prf

Funding: This study was supported by the Bengbu Medical University Natural Science Youth Project (No. 2024BYZD089 to H.B.Z.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-968/coif). L.W. reports funding from the Bengbu Medical University Natural Science Youth Project (No. 2024BYZD089). The other 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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and approved by the Ethics Committee of The First Affiliated Hospital of Bengbu Medical University (No. 2024-D-302). Due to the retrospective nature of the study, participant informed consent was waived.

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. Banales JM, Marin JJG, Lamarca A, et al. Cholangiocarcinoma 2020: the next horizon in mechanisms and management. Nat Rev Gastroenterol Hepatol 2020;17:557-88. [Crossref] [PubMed]
2. Brindley PJ, Bachini M, Ilyas SI, et al. Cholangiocarcinoma. Nat Rev Dis Primers 2021;7:65. [Crossref] [PubMed]
3. Muttillo EM, Ciardi A, Troiano R, et al. Pancreatic ductal adenocarcinoma and distal cholangiocarcinoma: a proposal of preoperative diagnostic score for differential diagnosis. World J Surg Oncol 2021;19:10. [Crossref] [PubMed]
4. Ilyas SI, Gores GJ. Pathogenesis, diagnosis, and management of cholangiocarcinoma. Gastroenterology 2013;145:1215-29. [Crossref] [PubMed]
5. Elvevi A, Laffusa A, Scaravaglio M, et al. Clinical treatment of cholangiocarcinoma: an updated comprehensive review. Ann Hepatol 2022;27:100737. [Crossref] [PubMed]
6. Fang L, Yan FH, Liu C, et al. Systemic Inflammatory Biomarkers, Especially Fibrinogen to Albumin Ratio, Predict Prognosis in Patients with Pancreatic Cancer. Cancer Res Treat 2021;53:131-9. [Crossref] [PubMed]
7. Olingy CE, Dinh HQ, Hedrick CC. Monocyte heterogeneity and functions in cancer. J Leukoc Biol 2019;106:309-22. [Crossref] [PubMed]
8. Yao W, Liu X, He Y, et al. ScRNA-seq and bulk RNA-seq reveal the characteristics of ferroptosis and establish a risk signature in cholangiocarcinoma. Mol Ther Oncolytics 2022;27:48-60. [Crossref] [PubMed]
9. Chen Z, Li H, Li Z, et al. SHH/GLI2-TGF-β1 feedback loop between cancer cells and tumor-associated macrophages maintains epithelial-mesenchymal transition and endoplasmic reticulum homeostasis in cholangiocarcinoma. Pharmacol Res 2023;187:106564. [Crossref] [PubMed]
10. Ben-Aicha S, Badimon L, Vilahur G. Advances in HDL: Much More than Lipid Transporters. Int J Mol Sci 2020;21:732. [Crossref] [PubMed]
11. Qin L, Sun K, Shi L, et al. High-Fat Mouse Model to Explore the Relationship between Abnormal Lipid Metabolism and Enolase in Pancreatic Cancer. Mediators Inflamm 2023;2023:4965223. [Crossref] [PubMed]
12. Xu H, Pang Y, Li X, et al. Monocyte to high-density lipoprotein cholesterol ratio as an independent risk factor for papillary thyroid carcinoma. J Clin Lab Anal 2021;35:e24014. [Crossref] [PubMed]
13. Wu H, Zhang J, Zhou B, et al. Preoperative monocyte to high-density lipoprotein ratio as a predictor of survival outcome of gastric cancer patients after radical resection. Biomark Med 2023;17:123-31. [Crossref] [PubMed]
14. Liu Q, Wang H, Chen Q, et al. Nomogram incorporating preoperative pan-immune-inflammation value and monocyte to high-density lipoprotein ratio for survival prediction in patients with colorectal cancer: a retrospective study. BMC Cancer 2024;24:740. [Crossref] [PubMed]
15. Miao T, Lou X, Dong S, et al. Monocyte-to-High-Density Lipoprotein-Cholesterol Ratio Predicts Prognosis of Hepatocellular Carcinoma in Patients with Metabolic-Associated Fatty Liver Disease. J Hepatocell Carcinoma 2024;11:145-57. [Crossref] [PubMed]
16. Ye XW, Wang ZY, Shao YX, et al. Monocyte to high-density lipoprotein ratio based prognostic nomogram for patients following allogeneic vascular replacement pancreaticoduodenectomy. Front Genet 2024;15:1465318. [Crossref] [PubMed]
17. Zhou M, Wang C, Lu S, et al. Tumor-associated macrophages in cholangiocarcinoma: complex interplay and potential therapeutic target. EBioMedicine 2021;67:103375. [Crossref] [PubMed]
18. Dwyer BJ, Jarman EJ, Gogoi-Tiwari J, et al. TWEAK/Fn14 signalling promotes cholangiocarcinoma niche formation and progression. J Hepatol 2021;74:860-72. [Crossref] [PubMed]
19. Zhou L, Li H, Zhang XX, et al. Rapamycin treated tol-dendritic cells derived from BM-MSCs reversed graft rejection in a rat liver transplantation model by inducing CD8(+)CD45RC(-)Treg. Mol Immunol 2021;137:11-9. [Crossref] [PubMed]
20. Ganjali S, Banach M, Pirro M, et al. HDL and cancer - causality still needs to be confirmed? Update 2020. Semin Cancer Biol 2021;73:169-77. [Crossref] [PubMed]
21. Georgila K, Vyrla D, Drakos E. Apolipoprotein A-I (ApoA-I), Immunity, Inflammation and Cancer. Cancers (Basel) 2019;11:1097. [Crossref] [PubMed]
22. Ma XL, Gao XH, Gong ZJ, et al. Apolipoprotein A1: a novel serum biomarker for predicting the prognosis of hepatocellular carcinoma after curative resection. Oncotarget 2016;7:70654-68. [Crossref] [PubMed]
23. Wang J, Lyu SC, Zhou L, et al. Prognostic analysis of pancreatic carcinoma with portal system invasion following curative resection. Gland Surg 2021;10:35-49. [Crossref] [PubMed]
24. Lyu SC, Wang J, Huang M, et al. CA19-9 Level to Serum γ-Glutamyltransferase as a Potential Prognostic Biomarker in Patients with Pancreatic Head Carcinoma. Cancer Manag Res 2021;13:4887-98. [Crossref] [PubMed]
25. Jiang T, Lyu SC, Zhou L, et al. Carbohydrate antigen 19-9 as a novel prognostic biomarker in distal cholangiocarcinoma. World J Gastrointest Surg 2021;13:1025-38. [Crossref] [PubMed]
26. Tella SH, Kommalapati A, Yadav S, et al. Novel staging system using carbohydrate antigen (CA) 19-9 in extra-hepatic cholangiocarcinoma and its implications on overall survival. Eur J Surg Oncol 2020;46:789-95. [Crossref] [PubMed]
27. Chen Q, Cui S, Huang J, et al. Venous thromboembolism in patients undergoing distal cholangiocarcinoma surgery: Prevalence, risk factors, and outcomes. Asian J Surg 2023;46:3648-55. [Crossref] [PubMed]
28. Wang J, Lyu SC, Zhu JQ, et al. Extended lymphadenectomy benefits patients with borderline resectable pancreatic head cancer-a single-center retrospective study. Gland Surg 2021;10:2910-24. [Crossref] [PubMed]
29. Kiriyama M, Ebata T, Aoba T, et al. Prognostic impact of lymph node metastasis in distal cholangiocarcinoma. Br J Surg 2015;102:399-406. [Crossref] [PubMed]
30. Byrling J, Andersson R, Sasor A, et al. Outcome and evaluation of prognostic factors after pancreaticoduodenectomy for distal cholangiocarcinoma. Ann Gastroenterol 2017;30:571-7. [Crossref] [PubMed]
31. Miura F, Sano K, Amano H, et al. Evaluation of portal vein invasion of distal cholangiocarcinoma as borderline resectability. J Hepatobiliary Pancreat Sci 2015;22:294-300. [Crossref] [PubMed]
32. Guilbaud T, Girard E, Lemoine C, et al. Intra-pancreatic distal cholangiocarcinoma and pancreatic ductal adenocarcinoma: a common short and long-term prognosis? Updates Surg 2021;73:439-50. [Crossref] [PubMed]
33. Tang Z, Yang Y, Zhao Z, et al. The clinicopathological factors associated with prognosis of patients with resectable perihilar cholangiocarcinoma: A systematic review and meta-analysis. Medicine (Baltimore) 2018;97:e11999. [Crossref] [PubMed]
34. Montal R, Sia D, Montironi C, et al. Molecular classification and therapeutic targets in extrahepatic cholangiocarcinoma. J Hepatol 2020;73:315-27. [Crossref] [PubMed]
35. Kendall T, Verheij J, Gaudio E, et al. Anatomical, histomorphological and molecular classification of cholangiocarcinoma. Liver Int 2019;39:7-18. [Crossref] [PubMed]