Published online Dec 21, 2026. doi: 10.3748/wjg.121762
Revised: May 10, 2026
Accepted: June 2, 2026
Published online: December 21, 2026
Processing time: 209 Days and 16.6 Hours
Malignant biliary obstruction (MBO) caused by biliary tract cancer (BTC) is considered to has a poor prognosis, and there is substantial survival heterogeneity after percutaneous biliary decompression (PBD). Although it’s widely agreed that a standardized prognostic evaluation system is needed, a comprehensive and universally accepted prediction model remains lacking.
To develop and validate a practical nomogram for individualized survival pre
A total of 284 patients were retrospectively enrolled with BTC-induced MBO who underwent PBD between June 2015 and July 2021. These patients were randomly assigned to a training cohort (n = 204) and a validation cohort (n = 80). The Cox proportional hazards regression analysis was used to find independent prognostic factors, and these factors were included to build the modified Clinical Outcomes of MBO-Biliary tract cancer Survival (mCOMBO-BaS) nomogram. Model per
Seven prognostic indicators were identified: Tumor subtype, subsequent surgery, total bilirubin, total cholesterol, hemoglobin, sodium, and carbohydrate antigen 19-9 level. The mCOMBO-BaS model showed good discriminatory ability, with C-indices over 0.65 for all BTC subtypes. In the training group, the AUCs for predicting 3-month, 6-month, and 1-year survival were 0.809, 0.798, and 0.733 respectively. In the validation group, they were 0.886, 0.804, and 0.793 respectively. Calibration plots indicated that the predicted survival probabilities were close to the actual observed ones.
The mCOMBO-BaS nomogram can accurately predict the survival of individual BTC patients with MBO un
Core Tip: This study developed and validated the modified Clinical Outcomes of Malignant biliary obstruction-Biliary tract cancer Survival nomogram, a practical tool for predicting the survival of individual patients with biliary tract cancer (BTC)-related malignant biliary obstruction undergoing percutaneous biliary decompression. By integrating seven prognostic variables (tumor subtype, subsequent surgery, total bilirubin, total cholesterol, hemoglobin, sodium, and carbohydrate antigen 19-9 level), the model showed robust discriminatory ability across all BTC subtypes for predicting 3-month, 6-month, and 1-year survival in both training and validation cohorts. This nomogram addresses the lack of personalized risk stratification in this patient population, helping clinicians optimize treatment decisions and improve overall management.
- Citation: Zhang YZ, Tang YZ, Pan ST, He YX, Dai HC, Zhou WZ, Zhou HF. Survival prediction for patients with malignant biliary obstruction caused by biliary tract cancer undergoing percutaneous biliary decompression. World J Gastroenterol 2026; 32(47): 121762
- URL: https://www.wjgnet.com/1007-9327/full/v32/i47/121762.htm
- DOI: https://dx.doi.org/10.3748/wjg.121762
Malignant biliary obstruction (MBO) caused by biliary tract cancer (BTC), hereafter referred to as MBO-BTC, including intrahepatic cholangiocarcinoma (ICC), perihilar cholangiocarcinoma (PCC), distal cholangiocarcinoma (DCC), and gallbladder carcinoma (GBC), accounts for approximately 3% of gastrointestinal malignancies worldwide, with an increasing incidence. Due to challenges in early diagnosis, 60%-70% of patients present with advanced disease, and the median survival with conventional chemotherapy is < 12 months[1]. Obstructive jaundice secondary to MBO markedly impairs quality of life, with symptoms such as pruritus and cholangitis; it exacerbates hepatic and systemic metabolic disturbances, thereby reducing treatment tolerance and short-term survival. Only approximately 20% of patients are eligible for curative resection at diagnosis. The 1-year postoperative recurrence or metastasis rate ranges from 60% to 80%, with an overall 5-year survival rate of 5%-19%, and only 2.5%-2.8% in cases with distant metastasis[2].
Percutaneous biliary decompression (PBD) is one of the standard treatments for unresectable MBO-BTC. By establishing effective biliary drainage, PBD alleviates jaundice, corrects cholestasis, controls infection, and improves clinical status and quality of life. It provides both a temporary measure to extend life and a transition to scheduled surgical care. However, clinical observations indicate that patients with MBO-BTC undergoing PBD have different survival outcomes. The substantial heterogeneity in prognosis presents major challenges for clinicians in formulating customized treatment plans, assessing treatment effectiveness, and delivering accurate prognostic information to patients and their families[3-6].
Although the necessity of establishing a standardized prognosis predictive system is widely recognized, a comprehensive and universally accepted prediction model remains lacking. Most existing models are applicable to MBO secondary to various malignancies, with few specifically developed for BTC. The clinical application is frequently limited by insufficient validation[7,8]. Moreover, although systemic therapy following PBD has been shown to improve overall survival (OS) in patients with unresectable BTC, risk stratification criteria to guide individual strategies still needs exploration[9].
This study aimed to identify easy-obtain clinical prognostic factors for patients with MBO-BTC following PBD and to develop and validate a personalized survival prediction nomogram that integrates multimodal clinical indicators[10]. The proposed nomogram is expected to serve as a quantitative tool to facilitate accurate risk stratification in patients with MBO-BTC after PBD. As other previous models[11-14], this tool could guide individualized palliative treatment decisions and provide a basis for formulating active intervention strategies early for high-risk patients.
This study was designed as a retrospective investigation used the preliminary results from the “Clinical Outcomes of MBO (COMBO)” study approved by the institutional review board at The First Affiliated Hospital with Nanjing Medical University and registered with the Chinese Clinical Trial Registry (ChiCTR2000038614). This study was named “The COMBO-03 study”. Due to its retrospective nature, there was no need to obtain informed consent from patients, and the patient data was anonymous before analysis. Initially, 334 cases met the inclusion criteria, but 11 cases underwent surgical or endoscopic biliary drainage procedures within one month before PBD interventions, 6 cases underwent other anticancer therapy with unclear treatment, 8 cases present of severe conditions before the surgery, 3 cases had Eastern Cooperative Oncology Group (ECOG) score of four, and 22 cases had incomplete records or loss to follow-up after the procedure. A total of 284 consecutive patients admitted between June 2015 and July 2021 were enrolled, including 174 men and 110 women, with a mean age of 65.94 ± 10.3 years. All patients were diagnosed with MBO-BTC and underwent PBD.
The histological diagnosis of BTC was confirmed by histopathological examination of biopsy specimens obtained during endoscopic retrograde cholangiopancreatography, percutaneous transhepatic biliary drainage, or surgical procedures. For patients who had progressed to the advanced stage and thus lost surgical eligibility, the diagnosis was supported by clinical assessment, including jaundice and other symptoms or signs, and supplemented by tumor biomarkers such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9), along with imaging evaluations (computed tomography and magnetic resonance cholangiopancreatography) performed by experienced radiologists. MBO was diagnosed based on cross-sectional imaging demonstrating biliary stricture with upstream ductal dilatation consistent with malignancy.
Among these patients, 55 were diagnosed with ICC, 145 with PCC, 62 with DCC, and 22 with GBC. Patients were randomly allocated to a training cohort (n = 204) or a validation cohort (n = 80).
Inclusion criteria: (1) Age ≥ 18 years; (2) A confirmed histological or clinical diagnosis of BTC; (3) Symptomatic MBO requiring drainage; (4) PBD as the primary decompressive method, including placement of plastic catheters or metal stents with or without 125I irradiation seeds; and (5) Availability of clinical, laboratory, and follow-up data.
Exclusion criteria: (1) Surgical or endoscopic biliary drainage procedures within 1 month prior to PBD interventions; (2) Any other anticancer therapy, except for supportive treatment after interventions; (3) PBD performed for benign biliary diseases; (4) Presence of severe conditions before the surgery, such as shock or multiple organ failure; (5) ECOG score of 4 before interventions; and (6) Incomplete records or loss to follow-up after the procedure.
Collected variables were categorized as follows: (1) Baseline characteristics: Age, gender, ECOG performance status (ECOG-PS), primary tumor location (intrahepatic, perihilar, or distal bile duct; gallbladder), and presence of distant metastases; (2) Preprocedural laboratory data: Total bilirubin (TB), direct bilirubin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, γ-glutamyl transpeptidase, albumin, globulin, total cholesterol (TC), triglycerides, blood urea nitrogen, creatinine, glucose, potassium, sodium (Na), calcium, chloride, white blood cell count, neutrophil percentage, lymphocyte percentage, neutrophil-to-lymphocyte ratio, hemoglobin (Hb), platelet count, CA19-9, CEA, and smoking status; (3) Imaging and procedure-related variables: Bismuth-Corlette classification for MBO; and (4) Postprocedural data: Subsequent anticancer therapies (such as chemotherapy and radiotherapy) and PBD strategies (including plastic catheters or metal stents, with or without 125I irradiation seeds), subsequent surgery (refer to curative-intent resection performed after PBD). The primary endpoint was OS, defined as the time interval from the date of the first PBD procedure to mortality from any cause or the last follow-up.
Data analysis and model construction were performed using SPSS software (version 27.0) and R software (www.r-project.org; version 3.6.3). For descriptive analyses, normally distributed continuous variables were expressed as mean ± SD, non-normally distributed variables as median and interquartile range (IQR), and categorical variables as n (%). For intergroup comparisons, the independent-samples t-test was used for normally distributed continuous variables, and the Mann-Whitney U test for non-normally distributed continuous variables. The χ2 test was applied to categorical variables, and Fisher’s exact test was used when expected frequencies were less than five.
The Kaplan-Meier method was used to estimate survival, and the log-rank test compared differences among subgroups. Univariate Cox proportional hazards regression analysis was performed to evaluate associations between each variable and OS, and to identify potential prognostic factors. Variables with P < 0.05 in the univariate analysis were entered into a multivariate Cox regression model, and stepwise regression was used to identify independent prognostic factors affecting OS in patients with MBO-BTC undergoing PBD.
Based on these factors, a survival prediction nomogram was constructed using the rms package in R software. The concordance index (C-index) assessed model discrimination, and calibration curves evaluated agreement between predicted and observed outcomes. The receiver operating characteristic (ROC) curves and the area under the curve (AUC) were generated using the pROC package to assess predictive performance for short- and long-term survival. A two-sided P < 0.05 was considered statistically significant.
A total of 284 patients with MBO-BTC who underwent PBD were included. Participants were randomly divided into a training cohort (n = 204) and a validation cohort (n = 80) using a 7:3 ratio via simple random sampling by SPSS (version 27.0) to achieve balanced distribution and reduce selection bias. The training cohort included 119 men (58.3%) and 85 women (41.7%), with a mean age of 66.0 years, while the validation cohort comprised 55 men (68.8%) and 25 women (31.3%) with a mean age of 65.8 years. In the training cohort, 153 patients (75.0%) reached the endpoint, with a median OS of 201 days (IQR: 88-483). In the validation cohort, 66 patients (82.5%) reached the endpoint, with a median OS of 155 days (IQR: 80-463). Baseline characteristics were comparable between cohorts, with no significant differences observed (all P > 0.05; Table 1).
| Characteristics | Total (n = 284) | Training (n = 204) | Validation (n = 80) | P value |
| Gender | 0.105 | |||
| Female | 110 (38.7) | 85 (41.7) | 25 (31.3) | |
| Male | 174 (61.3) | 119 (58.3) | 55 (68.8) | |
| Age | 65.9 (10.3) | 66.0 (10.4) | 65.8 (10.2) | 0.917 |
| Bismuth type | 0.158 | |||
| I | 44 (15.5) | 31 (15.2) | 13 (16.3) | |
| II | 116 (40.8) | 76 (37.3) | 40 (50.0) | |
| III | 116 (40.8) | 90 (44.1) | 26 (32.5) | |
| IV | 8 (2.8) | 7 (3.4) | 1 (1.3) | |
| Interventions | 0.995 | |||
| Plastic catheter | 168 (59.2) | 121 (59.3) | 47 (58.8) | |
| Metal stent | 91 (32.0) | 65 (31.9) | 26 (32.5) | |
| Irradiation stent | 25 (8.8) | 18 (8.8) | 7 (8.8) | |
| Subtype of tumor | 0.509 | |||
| ICC | 55 (19.4) | 37 (18.1) | 18 (22.5) | |
| PCC | 145 (51.1) | 102 (50.0) | 43 (53.8) | |
| DCC | 62 (21.8) | 49 (24.0) | 13 (16.3) | |
| GBC | 22 (7.7) | 16 (7.8) | 6 (7.5) | |
| M stage | 0.676 | |||
| 0 | 235 (82.7) | 170 (83.3) | 65 (81.3) | |
| 1 | 49 (17.3) | 34 (16.7) | 15 (18.8) | |
| History of surgery | 0.405 | |||
| No | 214 (75.4) | 151 (74.0) | 63 (78.8) | |
| Yes | 70 (24.6) | 53 (26.0) | 17 (21.3) | |
| Subsequent surgery | 0.817 | |||
| No | 261 (91.9) | 187 (91.7) | 74 (92.5) | |
| Yes | 23 (8.1) | 17 (8.3) | 6 (7.5) | |
| Subsequent chemotherapy | 0.594 | |||
| No | 251 (88.4) | 179 (87.7) | 72 (90.0) | |
| Yes | 33 (11.6) | 25 (12.3) | 8 (10.0) | |
| Subsequent radiotherapy | 0.293 | |||
| No | 242 (85.2) | 171 (83.8) | 71 (88.8) | |
| Yes | 42 (14.8) | 33 (16.2) | 9 (11.3) | |
| Subsequent targeted therapy | 0.527 | |||
| No | 278 (97.9) | 199 (97.5) | 79 (98.8) | |
| Yes | 6 (2.1) | 5 (2.5) | 1 (1.3) | |
| Gallstones | 0.603 | |||
| No | 236 (83.1) | 171 (83.8) | 65 (81.3) | |
| Yes | 48 (16.9) | 33 (16.2) | 15 (18.8) | |
| Hydrothorax | 0.470 | |||
| No | 257 (90.5) | 183 (89.7) | 74 (92.5) | |
| Yes | 27 (9.5) | 21 (10.3) | 6 (7.5) | |
| Ascites | 0.293 | |||
| No | 242 (85.2) | 171 (83.8) | 71 (88.8) | |
| Yes | 42 (14.8) | 33 (16.2) | 9 (11.3) | |
| Hydrops | 0.285 | |||
| No | 234 (82.4) | 165 (80.9) | 69 (86.3) | |
| Yes | 50 (17.6) | 39 (19.1) | 11 (13.8) | |
| ALT (IU/L) | 116 (102) | 117 (105) | 120 (93) | 0.690 |
| AST (IU/L) | 122 (108) | 123 (103) | 119 (122) | 0.806 |
| γ-GT (IU/L) | 513 (367) | 505 (355) | 533 (397) | 0.568 |
| AKP (IU/L) | 626 (413) | 603 (394) | 687 (455) | 0.125 |
| ALB (g/L) | 32.1 (5.11) | 32.3 (5.12) | 31.6 (5.06) | 0.271 |
| GLB (g/L) | 26.4 (IQR: 23.5, 31.8) | 25.7 (IQR: 23.4, 31.9) | 27.6 (IQR: 24.7, 30.7) | 0.122 |
| TB (μmol/L) | 238 (151) | 239 (155) | 234 (139) | 0.798 |
| DB (μmol/L) | 169 (106) | 169 (109) | 170 (101) | 0.977 |
| TB/DB | 1.41 (IQR: 1.37, 1.46) | 1.41 (IQR: 1.37, 1.46) | 1.41 (IQR: 1.36, 1.44) | 0.331 |
| TC (mmol/L) | 5.93 (3.93) | 6.09 (4.22) | 5.51 (3.05) | 0.265 |
| TG (mmol/L) | 2.29 (IQR: 1.41, 3.12) | 2.24 (IQR: 1.36, 3.03) | 2.37 (IQR: 1.57, 3.33) | 0.137 |
| BUN (mmol/L) | 5.35 (4.65) | 5.10 (4.16) | 5.99 (5.68) | 0.147 |
| Cr (μmol/L) | 62.6 (37.6) | 59.0 (19.0) | 71.7 (63.1) | 0.010 |
| K (mmol/L) | 3.72 (0.50) | 3.68 (0.48) | 3.82 (0.53) | 0.024 |
| Na (mmol/L) | 136 (8.68) | 136 (3.66) | 134 (15.25) | 0.112 |
| Cl (mmol/L) | 101 (4.96) | 101 (4.65) | 101 (5.69) | 0.564 |
| Ca (mmol/L) | 2.37 (2.92) | 2.44 (3.44) | 2.19 (0.14) | 0.519 |
| Glu (mmol/L) | 5.87 (1.98) | 5.90 (1.81) | 5.81 (2.38) | 0.740 |
| WBC (× 109/L) | 8.15 (3.80) | 8.12 (3.70) | 8.22 (4.07) | 0.836 |
| N percentage | 73.3 (12.3) | 73.2 (11.6) | 73.5 (14.0) | 0.863 |
| L percentage | 16.3 (9.13) | 16.5 (8.71) | 16.0 (10.2) | 0.703 |
| NLR | 7.49 (12.7) | 7.84 (14.7) | 6.59 (5.00) | 0.458 |
| Hb (g/L) | 110 (18.2) | 110 (18.6) | 109 (17.0) | 0.652 |
| PLT (× 109/L) | 227 (98.0) | 230 (98.3) | 219 (97.1) | 0.397 |
| CEA (ng/mL) | 0.872 | |||
| < 5 | 149 (55.8) | 106 (55.5) | 43 (56.6) | |
| ≥ 5 | 118 (44.2) | 85 (44.5) | 33 (43.4) | |
| CA19-9 (U/mL) | 0.977 | |||
| < 1000 | 152 (56.7) | 109 (56.8) | 43 (56.6) | |
| ≥ 1000 | 116 (43.3) | 83 (43.2) | 33 (43.4) | |
| HBP | 0.310 | |||
| No | 180 (63.4) | 133 (65.2) | 47 (58.8) | |
| Yes | 104 (36.6) | 71 (34.8) | 33 (41.3) | |
| Diabetes | 0.420 | |||
| No | 242 (85.2) | 176 (86.3) | 66 (82.5) | |
| Yes | 42 (14.8) | 28 (13.7) | 14 (17.5) | |
| Smoke | 0.574 | |||
| No | 233 (82.0) | 169 (82.8) | 64 (80.0) | |
| Yes | 51 (18.0) | 35 (17.2) | 16 (20.0) | |
| Alcohol | 0.487 | |||
| No | 241 (84.9) | 175 (85.8) | 66 (82.5) | |
| Yes | 43 (15.1) | 29 (14.2) | 14 (17.5) | |
| ECOG PS score | 0.600 | |||
| 0 | 25 (8.8) | 17 (8.3) | 8 (10.0) | |
| 1 | 214 (75.4) | 157 (77.0) | 57 (71.3) | |
| 2 or 3 | 45 (15.8) | 30 (14.7) | 15 (18.8) |
Three categories of variables potentially associated with prognosis were evaluated: General condition (age, ECOG-PS, comorbidities), tumor-related factors (tumor stage, tumor markers, treatment modalities), and laboratory parameters (hematologic and biochemical indicators). Univariate analysis identified candidate prognostic variables. Multivariate Cox regression analysis subsequently identified seven independent predictors included in the final model: Tumor subtype, subsequent surgery, TB, TC, Hb, Na, and CA19-9 level (P < 0.05; Table 2).
| Characteristics | Univariate analysis | Multivariate analysis | ||
| HR (95%CI) | P value | HR (95%CI) | P value | |
| Gender | 0.255 | |||
| Female | ||||
| Male | 1.209 (0.872-1.677) | |||
| Age | 1.020 (1.003-1.036) | 0.019 | ||
| Bismuth type | 0.078 | |||
| I | ||||
| II | 0.388 (0.162-0.926) | |||
| III | 0.466 (0.212-1.025) | |||
| IV | 0.606 (0.277-1.325) | |||
| Interventions | 0.522 | |||
| Plastic catheter | ||||
| Metal stent | 0.754 (0.447-1.274) | |||
| Irradiation seeds | 0.853 (0.491-1.481) | |||
| Subtype of tumor | 0.002 | 0.002 | ||
| ICC | ||||
| PCC | 0.781 (0.418-1.457) | 0.535 (0.344-0.833) | ||
| DCC | 0.445 (0.250-0.790) | 0.479 (0.285-0.804) | ||
| GBC | 0.403 (0.213-0.762) | 1.266 (0.648-2.477) | ||
| M stage | 0.022 | |||
| 0 | ||||
| 1 | 1.596 (1.069-2.383) | |||
| History of surgery | 0.267 | |||
| No | ||||
| Yes | 1.231 (0.853-1.776) | |||
| Subsequent surgery | < 0.001 | < 0.001 | ||
| No | ||||
| Yes | 0.230 (0.101-0.526) | 0.226 (0.096-0.530) | ||
| Subsequent chemotherapy | 0.002 | |||
| No | ||||
| Yes | 0.376 (0.203-0.695) | |||
| Subsequent radiotherapy | 0.152 | |||
| No | ||||
| Yes | 0.714 (0.450-1.132) | |||
| Subsequent targeted therapy | 0.041 | |||
| No | ||||
| Yes | 0.234 (0.058-0.944) | |||
| Gallstones | 0.683 | |||
| No | ||||
| Yes | 0.915 (0.596-1.404) | |||
| Hydrothorax | 0.367 | |||
| No | ||||
| Yes | 1.254 (0.767-2.053) | |||
| Ascites | 0.008 | |||
| No | ||||
| Yes | 1.714 (1.149-2.558) | |||
| Hydrops | 0.043 | |||
| No | ||||
| Yes | 1.483 (1.012-2.173) | |||
| ALT (IU/L) | 1.000 (0.998-1.001) | 0.713 | ||
| AST (IU/L) | 1.001 (1.000-1.003) | 0.037 | ||
| γ-GT (IU/L) | 1.000 (0.999-1.000) | 0.298 | ||
| AKP (IU/L) | 1.000 (1.000-1.001) | 0.208 | ||
| ALB (g/L) | 0.922 (0.893-0.952) | < 0.001 | ||
| GLB (g/L) | 1.009 (0.986-1.033) | 0.433 | ||
| TB (μmol/L) | 1.002 (1.001-1.003) | < 0.001 | 1.002 (1.001-1.004) | < 0.001 |
| DB (μmol/L) | 1.003 (1.002-1.005) | < 0.001 | ||
| TB/DB | 0.655 (0.280-1.531) | 0.329 | ||
| TC (mmol/L) | 0.944 (0.894-0.997) | 0.039 | 0.913 (0.856-0.973) | 0.005 |
| TG (mmol/L) | 0.965 (0.855-1.090) | 0.569 | ||
| BUN (mmol/L) | 1.054 (1.023-1.086) | < 0.001 | ||
| Cr (μmol/L) | 1.005 (1.001-1.008) | 0.005 | ||
| K (mmol/L) | 0.621 (0.429-0.898) | 0.011 | ||
| Na (mmol/L) | 0.897 (0.859-0.936) | < 0.001 | 0.918 (0.876-0.963) | < 0.001 |
| Cl (mmol/L) | 0.939 (0.903-0.977) | 0.002 | ||
| Ca (mmol/L) | 0.153 (0.050-0.462) | < 0.001 | ||
| Glu (mmol/L) | 0.943 (0.856-1.039) | 0.235 | ||
| WBC (× 109/L) | 1.099 (1.049-1.152) | < 0.001 | ||
| N percentage | 1.017 (1.001-1.033) | 0.032 | ||
| L percentage | 0.968 (0.949-0.988) | 0.002 | ||
| NLR | 1.008 (1.000-1.015) | 0.055 | ||
| Hb (g/L) | 0.988 (0.981-0.995) | 0.001 | 0.984 (0.974-0.994) | 0.001 |
| PLT (× 109/L) | 0.999 (0.998-1.001) | 0.344 | ||
| CEA (ng/mL) | < 0.001 | |||
| < 5 | ||||
| ≥ 5 | 1.758 (1.269-2.436) | |||
| CA19-9 (U/mL) | < 0.001 | < 0.001 | ||
| < 1000 | ||||
| ≥ 1000 | 2.552 (1.835-3.548) | 1.898 (1.296-2.781) | ||
| HBP | 0.017 | |||
| No | ||||
| Yes | 1.480 (1.074-2.038) | |||
| Diabetes | 0.826 | |||
| No | ||||
| Yes | 1.052 (0.669-1.655) | |||
| Smoke | 0.331 | |||
| No | ||||
| Yes | 1.208 (0.826-1.768) | |||
| ECOG PS score | 0.002 | |||
| 0 | ||||
| 1 | 0.369 (0.192-0.712) | |||
| 2 or 3 | 0.524 (0.349-0.788) | |||
A nomogram predicting OS at 3-, 6-, and 12-months in patients with MBO-BTC was constructed based on seven variables: Tumor subtype, subsequent surgery, TB, TC, Hb, Na, and CA19-9 (Figure 1). The scoring formula was as follows: Total score = (ICC = 20, PCC = 2, DCC = 0, GBC = 27) + (subsequent surgery: Yes = 0, No = 42) + TB × 0.07 + (100-TC × 2.5) + (336-Na × 2.3) + (73-Hb × 0.43) + (CA19-9 level: Over 1000 = 18, less than 1000 = 0).
The unit of TB is μmol/L. The range of TC is 0-40 mmol/L, the range of Hb is 20-170 g/L, and the range of Na is 120-146 mmol/L. This model is a modified version of the COMBO-BTC survival model and is referred to as the modified Clinical Outcomes of MBO-Biliary tract cancer Survival (mCOMBO-BaS) model.
Predictive performance was evaluated internally and externally using calibration curves, ROC curves, and the C-index. Calibration curves for 3-, 6-, and 12-month OS in both cohorts closely aligned with the 45° reference line (Figure 2). The AUC values from ROC analyses demonstrated the model’s discriminative ability for OS prediction at 3 months, 6 months, and 12 months (Figure 3). The model exhibited good discriminatory power in both the training (AUCs of 0.809, 0.798, and 0.733, respectively) and validation cohorts (AUCs of 0.886, 0.804, and 0.793, respectively). It also demonstrated robust predictive capability across four tumor subgroups (Figure 4), with C-indices exceeding 0.65 for all subtypes (ICC 0.785, PCC 0.776, DCC 0.655, GBC 0.755).
In the training cohort, tumors classified into four subtypes-DCC, PCC, ICC, and GBC-were ranked in descending order of OS, with DCC showing the longest OS (Figure 5A, training cohort: P = 0.001). Patients who did not undergo subsequent surgery had shorter OS than those who did (Figure 5B, training cohort: P < 0.001). Patients with CA19-9 ≥ 1000 U/mL exhibited shorter OS than those with CA19-9 < 1000 U/mL (Figure 5C, training cohort: P < 0.001). Furthermore, patients with M1 stage had substantially shorter OS than those with M0 stage (Figure 5D, training cohort: P = 0.021). In the validation cohort, although there didn’t appear significant differences of OS among tumor subtypes (Figure 5E, validation cohort: P = 0.130), the trend of survival performance was shown better in DCC subtype than other subtypes. Similar to the training cohort, significant shorter OSs were showed in the groups of patients who did not undergo subsequent surgery (Figure 5F, validation cohort: P = 0.007), patients with CA19-9 ≥ 1000 U/mL (Figure 5G, validation cohort: P = 0.025), and patients with M1 stage (Figure 5H, validation cohort: P < 0.001).
MBO-BTC, a serious clinical disorder, is characterized by a high incidence rate and poor prognosis[15]. PBD serves as an essential palliative intervention to relieve obstructive jaundice and improve quality of life in patients who are temporarily unsuitable for curative resection[16]. It is known that heterogeneity among BTC subtypes and variability in survival outcomes necessitate a unified and robust model to enable more efficient guidance for clinical decision-making and optimize therapeutic strategies.
Based on the mCOMBO-BaS model[10], subsequent curative-intent resection performed after PBD, named as sub
The mCOMBO-BaS model exhibited robust predictive performance, with AUC values ranging from 0.733 to 0.886 across training and validation cohorts. Importantly, the model maintained discriminative ability across all four BTC subtypes (ICC, PCC, DCC and GBC), with C-indices exceeding 0.65. This subtype-independent predictive capability holds significant clinical value, given the distinct biological behaviors and treatment responses associated with different anatomical tumor locations.
Seven predictive factors incorporated into the mCOMBO-BaS model cover multiple dimensions of patient status and disease characteristics. Four biochemical indicators, including TB, TC, Hb, and Na, reflect nutritional and metabolic status. Generally, TB is a core indicator for evaluating the degree of liver injury and cholestasis, with elevated levels predicting mortality in cirrhosis and acute-on-chronic liver failure[17]. TC reflects hepatocellular synthetic capacity. Decreased TC levels are associated with severe liver function impairment and higher tumor aggressiveness[18]. Hb reflects anemia status, an independent prognostic marker in malignancies associated with reduced survival[19]. Na is an indicator representing liver function reserve. Hyponatremia predicts severe complications and increased mortality in decompensated cirrhosis[20].
Especially, among these four BTC subtypes, DCC shows the most favorable survival while GBC consistently demonstrates the poorest prognosis[21]. This discrepancy highlights essential distinctions in anatomical location, tumor biology, and therapeutic response among BTC subtypes. Our study confirmed that GBC is associated with considerably worse outcomes than other BTC subtypes. Compared with chemotherapy or supportive care alone, subsequent surgery following initial biliary drainage substantially prolonged OS[22]. Elevated preoperative CA19-9 levels (≥ 1000 U/mL) are an independent predictor of poor prognosis in BTC, correlating with the degree of aggravated obstructive jaundice and increased tumor burden. This biomarker efficiently stratifies patients into distinct risk categories, with higher levels indicating poorer outcomes[10].
However, several limitations should be acknowledged. First, the model was established based on a single-center cohort, which may limit its robustness compared with large-scale multicenter models and affect its generalizability. Second, although retrospective validation provides valuable insights, it is inherently less reliable than prospective studies in establishing causal relationships due to the influence of confounding factors and bias. Third, the study population included only patients who underwent PBD as the initial decompression procedure, excluding those with prior endoscopic or surgical drainage such as ERCP or EUS-BD within one month, which might make the research more limited. Moreover, patients underwent subsequent surgery after PBD indicated better survival performance, but the timing of surgery relative to PBD was not analyzed in our study. This omission hinders our ability to identify the optimal time window for surgery after PBD. We should address these gaps by enrolling multicenter cohorts, implementing prospective designs to minimize bias, including patients with diverse initial drainage methods, systematically docu
In summary, the mCOMBO-BaS model represents a meaningful advancement in prognostic assessment for patients with MBO-BTC undergoing PBD, which integrates seven readily accessible predictors including tumor subtypes, subsequent surgeries, TB, TC, Na, Hb and CA19-9 level. This model accurately predicts 3-, 6-, and 12-month survival, with good performance in both training and validation cohorts. This nomogram can easily help to individualize prognoses, stratify risk, and guide treatment decisions across different tumor subtypes.
The authors thank Ms. Yan Shen, Ms Zhong-Ling Pei and Ms. Ling-Ling Wu from Department of Interventional Radiology, The First Affiliated Hospital with Nanjing Medical University, Nanjing, China, for their work in the follow-up of patients.
| 1. | Siegel RL, Kratzer TB, Giaquinto AN, Sung H, Jemal A. Cancer statistics, 2025. CA Cancer J Clin. 2025;75:10-45. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 1828] [Cited by in RCA: 2920] [Article Influence: 2920.0] [Reference Citation Analysis (15)] |
| 2. | Banales JM, Rodrigues PM, Affò S, Andersen JB, Aspichueta P, Boulter L, Bridgewater J, Calvisi DF, Cardenas A, Cardinale V, Carpino G, Coulouarn C, Dopazo C, Edeline J, Fabris L, Folseraas T, Forner A, Goeppert B, Heikenwalder M, Kendall TJ, Khan SA, Klümpen HJ, Koerkamp BG, Lamarca A, Lindsey S, Lleo A, Luedde T, Macias RIR, Morement H, Nault JC, Olaizola P, Perugorria MJ, Raggi C, Rimassa L, Saborowski A, Valle JW, Vithayathil M, Vogel A, Braconi C; International CCA Consensus Consortium. Cholangiocarcinoma 2026: status quo, unmet needs and priorities. Nat Rev Gastroenterol Hepatol. 2026;23:65-96. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 7] [Cited by in RCA: 53] [Article Influence: 53.0] [Reference Citation Analysis (5)] |
| 3. | Yang H, Qin Q, Tang Y, Zhu W. Correlation between functional drainage and survival in malignant biliary obstruction after percutaneous biliary drainage. Heliyon. 2024;10:e24088. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 7] [Reference Citation Analysis (1)] |
| 4. | Chan K, Vigneswaran G, Modi S, Sew Hee C, Maclean D, Stedman B, Bryant T, Maher B. Identifying predictive markers for survival in malignant biliary obstruction following percutaneous transhepatic biliary drainage. Clin Radiol. 2025;82:106772. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 1] [Cited by in RCA: 1] [Article Influence: 1.0] [Reference Citation Analysis (0)] |
| 5. | Rohan T, Cechova B, Matkulcik P, Straka M, Zavadil J, Eid M, Uher M, Dostal M, Andrasina T. Prognostic factors for survival in patients with advanced cholangiocarcinoma treated with percutaneous transhepatic drainage. Sci Rep. 2025;15:2172. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in RCA: 5] [Reference Citation Analysis (0)] |
| 6. | Duzkopru Y, Eren T, Kocanoglu A, Dogan O, Isak OA, Ergun O. Factors affecting survival in patients undergoing percutaneous transhepatic biliary drainage for malignant biliary obstruction. Eur Rev Med Pharmacol Sci. 2024;28:1554-1561. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 1] [Reference Citation Analysis (0)] |
| 7. | Zerem E, Imširović B, Kunosić S, Zerem D, Zerem O. Percutaneous biliary drainage for obstructive jaundice in patients with inoperable, malignant biliary obstruction. Clin Exp Hepatol. 2022;8:70-77. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 22] [Cited by in RCA: 20] [Article Influence: 5.0] [Reference Citation Analysis (0)] |
| 8. | Song W, Zhu ZG, Wu Q, Lv CG, Wang YG, Chen L, Miao DL. A nomogram to predict overall survival for biliary tract cancer. Cancer Manag Res. 2018;10:1535-1541. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 11] [Cited by in RCA: 18] [Article Influence: 2.3] [Reference Citation Analysis (0)] |
| 9. | Jansen H, Pape UF, Utku N. A review of systemic therapy in biliary tract carcinoma. J Gastrointest Oncol. 2020;11:770-789. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 12] [Cited by in RCA: 22] [Article Influence: 3.7] [Reference Citation Analysis (0)] |
| 10. | Zhou HF, Wang JL, Yang W, Zhou C, Shen Y, Wu LL, Pei ZL, Zhou WZ, Liu S, Shi HB. Survival prediction for patients with malignant biliary obstruction caused by pancreatic cancer undergoing biliary drainage: the COMBO-PaS model. Surg Endosc. 2023;37:1943-1955. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 4] [Reference Citation Analysis (0)] |
| 11. | Xu X, He M, Wang H, Zhan M, Yang L. Development and validation of a prognostic nomogram for gallbladder cancer patients after surgery. BMC Gastroenterol. 2022;22:200. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 4] [Cited by in RCA: 17] [Article Influence: 4.3] [Reference Citation Analysis (0)] |
| 12. | Bai Y, Liu ZS, Xiong JP, Xu WY, Lin JZ, Long JY, Miao F, Huang HC, Wan XS, Zhao HT. Nomogram to predict overall survival after gallbladder cancer resection in China. World J Gastroenterol. 2018;24:5167-5178. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in CrossRef: 19] [Cited by in RCA: 32] [Article Influence: 4.0] [Reference Citation Analysis (1)] |
| 13. | Tang YY, Zhao YN, Zhang T, Chen ZY, Ma XL. Comprehensive radiomics nomogram for predicting survival of patients with combined hepatocellular carcinoma and cholangiocarcinoma. World J Gastroenterol. 2021;27:7173-7189. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in CrossRef: 5] [Cited by in RCA: 22] [Article Influence: 4.4] [Reference Citation Analysis (7)] |
| 14. | Liu Y, Zhang C, Song M, Han X, Jiao D. Predicting early biliary infection after stenting of malignant biliary obstruction: model development and internal validation. Abdom Radiol (NY). 2023;48:2456-2465. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 3] [Reference Citation Analysis (0)] |
| 15. | Wang Y, Ding M, Zhang Q, Wang J, Yang X, Zhou F, Li L, Yuan Z, Jin H, Qian Q. Activation or suppression of the immune response mediators in biliary tract cancer (BTC) patients: a systematic review and meta-analysis. J Cancer. 2017;8:74-84. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 8] [Cited by in RCA: 13] [Article Influence: 1.4] [Reference Citation Analysis (0)] |
| 16. | Makki M, Bentaleb M, Abdulrahman M, Suhool AA, Al Harthi S, Ribeiro MA Jr. Current interventional options for palliative care for patients with advanced-stage cholangiocarcinoma. World J Clin Oncol. 2024;15:381-390. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in CrossRef: 7] [Cited by in RCA: 8] [Article Influence: 4.0] [Reference Citation Analysis (0)] |
| 17. | Zhang CH, Ni XC, Chen BY, Qiu SJ, Zhu YM, Luo M. Combined preoperative albumin-bilirubin (ALBI) and serum γ-glutamyl transpeptidase (GGT) predicts the outcome of hepatocellular carcinoma patients following hepatic resection. J Cancer. 2019;10:4836-4845. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 9] [Cited by in RCA: 20] [Article Influence: 2.9] [Reference Citation Analysis (0)] |
| 18. | Peng L, Yan Q, Chen Z, Hu Y, Sun Y, Miao Y, Wu Y, Yao Y, Tao L, Chen F, Li H, Xu T. Research progress on the role of cholesterol in hepatocellular carcinoma. Eur J Pharmacol. 2023;938:175410. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 13] [Reference Citation Analysis (0)] |
| 19. | Huang JC, Ma YW, Wang HX, Lang R, Jiang T. Prognostic impact of albumin-bilirubin score in predicting the long-term survival of distal cholangiocarcinoma after radical surgery. World J Surg Oncol. 2025;23:160. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 2] [Reference Citation Analysis (0)] |
| 20. | Rondon-Berrios H, Velez JCQ. Hyponatremia in Cirrhosis. Clin Liver Dis. 2022;26:149-164. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 26] [Cited by in RCA: 25] [Article Influence: 6.3] [Reference Citation Analysis (1)] |
| 21. | Hussain MM, Wang JM, Zhai AQ, Li FY, Hu HJ. Comparison of prognostic factors and their differences in intrahepatic, hilar, and distal cholangiocarcinoma: A systematic review and meta-analysis. World J Gastrointest Oncol. 2025;17:107995. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in RCA: 4] [Reference Citation Analysis (0)] |
| 22. | Jagtap VK, Kumar S, Harris C, Lynser D, Raphael V. Improved Survival With Chemotherapy in Patients With Malignant Biliary Tract Obstruction After Percutaneous Transhepatic Biliary Drainage (PTBD). Cureus. 2024;16:e63218. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 1] [Reference Citation Analysis (0)] |