The examination of individual clinical data from the combined therapy and chemotherapy groups included data on the first-line treatment for advanced biliary malignancies in 172 patients (Fig. 1 and Table 1). All patients had ECOG scores ranging from 0 to 1. There were no statistically significant differences in age (p=0.054), gender (p=0.874), height (p=0.376), weight (p=0.454), tumor location (p=0.401), and so forth between the four groups. PFS was significantly lower in group B (11.05 ± 7.00), group C (8.89 ± 8.00), and group D (6.02 ± 6.00) compared to group A (12.42 ± 12.00), with statistical significance (P < 0.0001, Fig. 2).

Flow diagram of study design

(1) Kaplan–Meier curve for progression-free survival according to treatment therapy. (2) Kaplan–Meier curve for overall survival according to treatment therapy mPFS, median progression-free survival;mOS, median overall survival; NR, not reached. A: GEMOX + ICIs(sintilimab or camrelizumab) + lenvatinib B:GEMOX + ICIs(sintilimab or camrelizumab) + bevacizumab C:GEMOX + ICIs(sintilimab or camrelizumab)D:GEMOX

In this study, a minimum of two treatment cycles was administered to 40 patients in group A, followed by an efficacy evaluation. Among these patients, the trial yielded no patients with CR: PR, 45.00% (17/40); CCA, 13.95% (6/40); SD, 45.00% (18/40); CCA, 25.00% (10/40); GCA, 20.00% (8/40); PD, 12.50% (5/40); CCA, 5.00% (2/40); and GCA, 7.50% (3/40). In group A, the DCR was 87.50% (35/40), and the ORR was 45.00% (17/40) (Fig. 3 and Table2).

Maximum percentage reduction of target lesions from baseline in the CI arm (n = 43)

In group B, 46 patients received at least two evaluations, with a mean treatment duration of 16 ± 2 cycles. No CR was observed; 34.78% (16/46) had PR, including 15 (32.61%) with CCA and 1 (2.17%) with GCA; 43.48% (20/46) had SD, including 15 (32.61%) with CCA and 5 (10.87%) with GCA; and 23.26% (10/43) had PD, including 4 (8.70%) with CCA, 2 (4.35%) with GCA, and 4 (8.70%) were GCA. In the combined therapy group, the overall DCR was 78.26% (36/46) and the overall ORR was 34.78% (16/46).

The mean treatment period in group C was 10 ± 2 cycles, and 30 patients were evaluated at least once. The number of patients who achieved CR in the study was 0%; PR was 16.67% (5/30), including 16.67% (5/30) with CCA; SD was 45.00% (18/30), including 33.33% (10/30) with CCA, and 26.67% (8/30) with GCA; PD was 26.67% (8/30), including 37.21% (6/30) with CCA and 20.00% (2/30) with GCA. The ORR in group C was 16.67% (5/30), and the DCR was 76.67% (23/30).

The mean treatment period in group D was 5 ± 2 cycles, and 56 patients were evaluated at least once. The number of patients who achieved CR in the study was 0; PR was 17.86% (10/56), including 10.71% (6/56) with CCA and 7.14 (4/56) with GCA; SD was 41.07% (23/56), including 23.21% (13/56) with CCA and 17.86% (10/56) with GCA; and PD was 58.93% (33/56), including 46.43% (26/56) with CCA and 12.50% (7/56) with GCA. Group D had an ORR of 17.86% (10/56) and a DCR of 58.93% (33/56).

Thirteen out of the 172 patients died during the follow-up period, with A median duration of follow-up for the entire population being 10.2 months (95% CI: 9.7–10.7) at the time of the last follow-up (two patients in group A, two patients in group B, four patients in group C, and seven patients in group D). For the entire population, the median PFS was 9.74 months (95% CI: 8.8–10.6) (Fig. 4). With P < 0.05, the median PFS in groups A, B, C, and D was 12.4 months, 11.05 months, 8.9 months, and 6.0 months, respectively. The median OS in groups A, B, C, and D was 13.63 months, 12.41 months, 11.24 months, and 8.87 months, respectively, and these results were considered statistically significant (P = 0.001 and P < 0.05, Fig. 4).

The ORR and the DCR by the different treatment regimes. (1) ORR by different treatment regimens and (2) DCR by different treatment regimens. Median (range) values across different treatment regimens were calculated from median values reported in individual studies where available. DCR, disease control rate; ORR, objective response rate. A: GEMOX + ICIs (sintilimab or camrelizumab) + lenvatinib B: GEMOX + ICIs(sintilimab or camrelizumab) + bevacizumab C:GEMOX + ICIs(sintilimab or camrelizumab)D:GEMOX

Age, sex, CA19-9 level, the neutrophil-lymphocyte ratio (NLR), and other univariate variables were used as independent variables in a one-way analysis using a Cox regression model. Variables that were significant in the one-way analysis (P < 0.1) were included in the multivariate Cox regression analysis. This revealed that the patients’ metastasis site (HR = 1,145, 95% CI: 0.28–0.73, P = 0.001), whether the NLR ratio was < 2.3 (HR = 0.054, 95% CI: −5.607–2.178, P < 0.001) and whether chemotherapy was administered through hepatic artery embolization (HR = 1.237, 95% CI: 1.26–21.11, P = 0.001) could impact the patients’ PFS (Table 3 and Fig. 5) and have a better prognosis.

A Forest plot of multifactorial risk profiles affecting PFS in group A. B Forest plot of multifactorial risk profiles affecting OS in group A. OR:odds ratio; P:p-value

Analysis of the factors influencing OS, both single and multi-factor.

The Cox regression model was used for univariate analysis, and the variables of univariate components (age, sex, CA199 level, NLR, etc.) were employed as independent variables, with death as the outcome and OS as the time. The multivariate Cox regression analysis included variables that were significant in the univariate analysis (P < 0.1). The findings indicated that the patients’ OS may be impacted by the location of their metastatic foci (95% CI: 0.30–0.78, P = 0.002) and whether the NLR ratio < 2.3 (95% CI: −5.607−2.178, P < 0.001) and whether chemotherapy was administered through hepatic artery embolization (95% CI: 1.13–19.09, P = 0.043) could have an positive impact on the OS (Table 4 and Fig. 5).

Almost all patients had different types and degrees of TRAEs, and the total incidence was 97.9% (126/129), which were 96.0% (38/40), 97.9% (45/46), 96.7% (29/30), and 94.7% (53/56)in the group A, group B, group C, and group D, respectively. All TRAEs were mainly grade 1–2 level, with an incidence of 90.7% (117/129), mainly including 67.5% (87/129) of nausea and vomiting, 66.0% (85/129) of loss of appetite, 62.4% (80/129) of have the runs, etc. The total incidence of grade 3 level TRAEs was 18.6% (24/129), mainly including 6.8% (9/129) of spin, 5.4% (7/129) of fatigue, 5.15% (10/194) of skin and 3.9% (5/129) of hands and feet reactive, and so on (Table 5).

The incidence of grade 1–2 level AEs was 98.0% (39/40), 97.9% (45/46), 96.7% (29/30), and 96.4% (54/56) in the group A, group B, group C, and group D, respectively, while the incidence of grade 3level AEs was 10.0% (4/40), 19.6% (9/46), 40.0% (12/30), and 10.7% (6/56), respectively. The difference was not statistically significant (P = 1.000, P = 3.020, P = 0.703, and P = 0.585, respectively). In the group A, loss of appetite (13/40, 32.5%), nausea and vomiting (11/40, 28.0%), and hypothyroidism (11/40, 25.6%) were the common grade 1–2 level AEs, whereas fatigue (1/40, 4.0%), skin of hands and feet reactive (1/40, 4.0%), and spin (1/40, 4.0%) were common grade 3 level AEs. In the group B, have the runs (15/46, 32.6%) and immunohepatitis (10/46, 21.7%) were common grade 1–2 AEs, while spin (4/46, 7.0%) and skin of hands and feet reactive (3/46, 6.5%) were common grade 3 level AEs. Equally, in the group C, nausea and vomiting (14/30, 46.5%) and loss of appetite (10/30, 23.3%) were common grade 1–2 AEs, while immunohepatitis (6/30, 18.6%) and rashes (2/30, 7.0%) were common grade 3 level AEs. Finally, in the group D, fatigue (20/56, 37.2%) and hypothyroidism (13/56, 23.3%) were common grade 1–2 AEs, while myelosuppression (4/56, 7.0%) and fatigue (2/56, 3.8%) were common grade 3 level AEs (Table 5 and Figure 6).

(1) Frequency of grade 3/4 adverse events in all groups. (2)Frequency of any grade and grade 3/4 adverse events in Group A. A: GEMOX + ICIs(sintilimab or camrelizumab) + lenvatinib B:GEMOX + ICIs(sintilimab or camrelizumab) + bevacizumab C:GEMOX + ICIs(sintilimab or camrelizumab) D:GEMOX