We conducted an institutional review board-approved retrospective study on advanced unresectable and recurrent iCCA patients treated at the First Affiliated Hospital of Wenzhou University between December 1, 2019, and April 30, 2023. This study followed Good Clinical Practice guidelines and the Declaration of Helsinki ethical principles.

The inclusion criteria were as follows (1): patients older than 18 years with histopathologically confirmed unresectable or metastatic iCCA, (2) patients who had at least one measurable tumor lesion at baseline per the Response Evaluation Criteria 1.1 in Solid Tumors (RECIST criteria 1.1), (3) Patients who were treated with either systemic chemotherapy or chemotherapy plus PD-1/L1, (4) patients who had Child-Pugh class A or B liver function status (score ≤7), (5) Patients who presented with an Eastern Cooperative Oncology Group performance status (ECOG PS) value of 0–1(Using a scale from 0 to 5, where lower scores indicate higher levels of functioning).

Exclusion criteria were set for individuals with severely impaired liver function, specifically those of Child-Pugh class C. Patients who did not receive more than one treatment regimen, either a combination therapy or chemotherapy-alone. The patients who had additional coexisting malignant tumors, patients who did not have any measurable tumors on both computerized tomography (CT) and magnetic resonance imaging (MRI), and patients with incomplete data.

All pathological diagnosis data were obtained from the electronic system of the hospital’s pathology department, whilst all clinical data and treatment histories were independently extracted from both the hospitals’ inpatient and outpatient electronic systems. Current picture archiving computer software (PACS) system was used to obtain electronic linear, as well as circumferential tracings of tumors on either CT or MRI scans, images were regularly evaluated every two to three months during treatment. Tumor measurement data were independently reviewed and assessed by two radiologists per the RECIST criteria 1.1. Patient demographic and basic characteristics including age, gender, BMI, HBV, HCV, CA19–9 levels, AFP levels, Child-Pugh, ECOG score, tumor number, tumor differentiation, and recurrence were obtained and recorded in ( Table 1 ). Data was last updated on April 30, 2023. Written informed consent was waived since this study was conducted retrospectively, however, the patients’ identities were anonymized to ensure their confidentiality. Patients who did not have any documented deaths were censored from the analysis at the cut-off date, April 30, 2023.

The primary outcome evaluated was overall survival (OS), which is defined as the duration from the commencement of treatment to the occurrence of death due to any cause compared between the chemotherapy combined with PD-1/L1 therapy group versus the chemotherapy-alone group. Progression-free survival (PFS), was defined as the duration from the initiation of treatment to the occurrence of disease progression or death from any cause between the combined therapy group and the chemotherapy group. Additionally, secondary endpoints included the objective response rate (ORR), measuring the percentage of patients who had a confirmed complete or partial response according to RECIST version 1.1, and the disease control rate (DCR), measuring the proportion of patients who achieved a complete or partial response or stable disease according to RECIST version 1.1.

Safety was additionally assessed as a secondary outcome, with all adverse events reported by patients or clinicians recorded by the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. The study was conducted per the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) statement checklist.

The baseline characteristics and efficacy data of the two treatment groups were compared using either the Chi-square test or Fisher’s exact test. Survival analyses to generate PFS and OS curves were conducted via the Kaplan-Meier method. The Log-rank test was used to determine the P value and compare curves from the two treatment groups. Hazards ratios were estimated using the Cox proportional hazards model. The statistical significance was set at P<0.05. The statistical analyses were conducted using SPSS 20.0 software (IBM, SPSS, Chicago, IL, USA) and GraphPad Prism, Version 8.2.0 (GraphPad, Inc.).

Between December 2019 and April 2023, 135 patients had been diagnosed with advanced or recurrent iCCA and had received either chemotherapy alone or combined with PD-1/L1. Of the 135 patients, 54 patients were excluded based on the eligibility criteria leaving 81 patients for analysis. Two cohort groups were formed with the following sample size: 51 in the combined therapy (chemotherapy plus PD-1/L1) group and 30 in the chemotherapy-alone group ( Figure 1 ). The combined therapy group consisted of patients treated with chemotherapy plus; durvalumab (26 patients); pembrolizumab (17 patients); and camrelizumab (8 patients) ( Supplementary Table S1 ). Patients in the chemotherapy group were treated with; GemCis (13 patients); gemcitabine and oxaliplatin (GEMOX) (8 patients); gemcitabine and S-1 (GS) (5 patients); FOLFIRINOX (4 patients) ( Supplementary Table S2 ).

Patient demographics and baseline characteristics were recorded ( Table 1 ). Patients consisted of 50.6% (n = 41) males and 49.4% (n = 40) females with a median age of 63 years (range 29–89 years). Most patients had a normal BMI and an ECOG score of 1. 16 patients had HBV while none had HCV. Furthermore, 29 patients had poorly differentiated tumors while 22 had moderately differentiated tumors. Liver and lymph nodes were the most frequent metastatic sites across all patients. There was an insignificant difference between the number of patients that had recurrent iCCA post-resection in the combined therapy group (12 patients) and the chemotherapy-alone group (18 patients).

As of April 30, 2023, 62 (76.5%) deaths had occurred, 38 (74.5%) patients in the combined therapy group and 24 (80%) patients in the chemotherapy-alone group. The median OS was 11 months (95% CI 9.57–12.43) for the patients in the combined therapy group and 8 months (95% CI 6.34–9.64) for those in the chemotherapy-alone group. The hazard ratio (HR) for OS was 0.53 (95% CI 0.30–0.94, P = 0.008) for the combined therapy group versus the chemotherapy-alone group. The median PFS was significantly longer in the combined therapy group at 7 months (95% CI 5.88–8.12) compared to the chemotherapy-alone group at 4 months, (95% CI 3.027–4.97). The HR for PFS of the combined therapy cohort versus the chemotherapy-alone cohort was 0.48 (95% CI 0.27–0.87, P = 0.002) ( Figure 2 ).

More patients attained radiologically confirmed CR and PR in the combined therapy group as compared to the chemotherapy-alone group, 5.9% (n = 3/51) versus 0% and 23.5% (n = 12/51) versus 13.3% (n = 4/30) respectively. Patients in the combined therapy group exhibited a significantly higher ORR (CR + PR), 29.4% (n = 15/51), compared to those in the chemotherapy-alone group, 13.3% (n = 4/30). Similarly, the combined therapy group had a higher DCR (CR+PR+SD) when compared to the chemotherapy-alone group, 78.4% (n = 40/51) and 73.3% (n = 22/30) respectively. 3 of the 12 patients who achieved PR in the combined therapy group underwent curative tumor resection following a preoperative evaluation ( Table 2 ). Overall, patients in the combined therapy group showed a better tumor percentage change relative to baseline size compared to the patients in the chemotherapy-alone group ( Figure 3 ).

Treatment-related adverse effects (TRAEs) occurred in 98.0% (50 of 51) of the patients from the combined therapy group while 90.0% (27 of 30) of the patients in the chemotherapy-alone group experienced TRAEs. Nausea and leukopenia were the predominant grade 1–2 TRAEs in the combined therapy group, 37.3% and 33.3% respectively, while anemia (20%) and leukopenia (16.7%) were predominant in the chemotherapy-alone group. Malaise (7.8%), vomiting (3.9%), and diarrhea (2%) were only recorded in the combined therapy group whereas elevated AST (10%) was only recorded in the chemotherapy-alone group ( Table 3 ). The chemotherapy-alone group had fewer grade 3–4 TRAEs, 7 of 30 (23.3%) patients, compared to the combined therapy group, 34 of 51 (66.7%) patients. The most frequent grade 3–4 TRAEs were leukopenia (17.6% in the combined therapy group and 3.3% in the chemotherapy-alone), thrombocytopenia (13.7% in the combined therapy group and 10.0% in the chemotherapy-alone), and elevated bilirubin levels (9.8% in the combined therapy group and 6.7% in the chemotherapy-alone). The elevated bilirubin levels, AST levels, and ALT levels were likely due to disease progression of lesions occupying the bile ducts. In both groups, no patient discontinued therapy due to any TRAEs and no cases of severe adverse events or treatment-related deaths occurred.

In the combined therapy group, comparing the various anti-PD-1/L-1 drugs used, revealed no significant differences in either OS or PFS (P values = 0.8543 and = 0.6221, respectively) ( Figures 4A, B ). Similarly, no significant difference was observed in both OS and PFS in the various chemotherapy regimens used in the chemotherapy-alone group (P values = 0.0553 and = 0.3488, respectively) ( Figures 4C, D ). However, in the chemotherapy-alone group, patients treated with the GS regimen exhibited a higher OS (12 months) compared to the other regimens (GEMOX 9.5 months, GemCis 8 months, and FOLFORINOX 8.5 months) ( Figure 4C ).

In the combined therapy group, 3 patients achieved CR. Of these 3, 2 patients were treated with chemotherapy plus durvalumab while 1 was treated with chemotherapy plus camrelizumab ( Figures 5A, C ). Additionally, of patients treated with chemotherapy plus durvalumab 6 patients achieved PR, 11 had SD, and 7 had PD. In the chemotherapy plus pembrolizumab subgroup, 4 patients had PD, 9 had SD, and 4 achieved PR ( Figure 5B ). Of the patients treated with chemotherapy plus camrelizumab 2 achieved PR and 4 had SD. Interestingly all patients (n = 7) treated with chemotherapy plus camrelizumab had no tumor progression ( Figure 5A ).

Due to the lack of a prospective design, our study is subject to several limitations. The retrospective approach allows for the potential selection and recall bias. Furthermore, this study analyzed data from a single center and had a relatively limited sample size, as such, larger multicenter prospective investigations are necessary to consolidate our findings. This study did not assess a specific treatment regimen and a standardized dosage control, thus, potentially reducing the internal validity. Future investigations focusing on a single treatment regimen may be required. Since our study focused of advanced, unresectable and recurrent iCCA patients, our findings might not be generalized to other types of CCA. Finally, our study lacks pre-treatment testing for PD-L1 expression, tumor mutation burden (TMB), and microsatellite instability (MSI). These biomarkers may have a substantial impact on therapeutic outcomes. Although these limitations existed that questioned the validity of our findings, our study presents a real-world analysis that might be useful for future prospective studies.