Abstract
Background & Aims
Microsatellite instability–high/mismatch repair–deficient (MSI-H/dMMR) cholangiocarcinoma (CCA) represents a rare molecular subtype with potential sensitivity to immune checkpoint inhibitors (ICIs). However, real-world evidence regarding the efficacy and safety of ICIs in MSI-H/dMMR CCA remains limited.
Methods
We conducted a single-center, retrospective real-world study of patients with advanced CCA treated with ICI-based regimens. MSI-H/dMMR status was determined by next-generation sequencing and/or immunohistochemistry. A contemporaneous microsatellite-stable (MSS) cohort was included as a control group. Propensity score matching (PSM) was performed to balance baseline characteristics. The primary endpoint was overall survival (OS), and secondary endpoints included progression-free survival (PFS), objective response rate, disease control rate, and safety.
Results
Among 331 patients, 23 (6.9%) were identified as MSI-H/dMMR. After 1:3 PSM, 23 MSI-H/dMMR patients were matched with 69 MSS/preserved MMR (pMMR) patients, achieving well-balanced baseline characteristics. MSI-H/dMMR patients demonstrated significantly prolonged survival compared with MSS/pMMR patients. Median PFS was 64.1 months vs. 7.2 months, and median OS was 70.5 months vs. 14.0 months, respectively (both p <0.001). MSI-H/dMMR status remained the strongest independent prognostic factor for both OS and PFS in multivariable Cox analyses. Objective response rate and disease control rate were significantly higher in MSI-H patients than in matched MSS/pMMR patients. Subgroup analyses showed consistent survival benefits of MSI-H/dMMR across tumor subtype, PD-L1 expression, and CA19-9 strata. Genomic profiling revealed frequent alterations in ARID1A, PBRM1, and KMT2D among MSI-H/dMMR tumors. ICI-based therapy was generally well tolerated; grade 3–4 adverse events occurred predominantly in chemotherapy-containing regimens.
Conclusions
In this real-world PSM study, MSI-H/dMMR CCA was associated with markedly improved survival and durable clinical benefit from ICI-based therapy, with a manageable safety profile. These findings support MSI-H/dMMR as a key actionable biomarker and underscore the importance of routine MSI testing in CCA.
Impact and implications
This study addresses a critical knowledge gap regarding the clinical benefit of immune checkpoint inhibitors in microsatellite instability–high/mismatch repair–deficient (MSI-H/dMMR) cholangiocarcinoma, a rare but biologically distinct subtype characterized by heightened immunogenicity. Our findings demonstrate that MSI-H/dMMR status is a strong and consistent predictor of durable survival across diverse clinical contexts, providing valuable evidence for clinicians, researchers, and stakeholders involved in the management of biliary tract cancers. These results support routine MSI-H/dMMR testing and personalized immunotherapy strategies in clinical practice, while also offering a framework for future translational research and policy decisions aimed at improving access to molecular profiling and precision oncology approaches.
Keywords: Microsatellite instability, Cholangiocarcinoma, Immune checkpoint inhibitors, Real-world evidence, Propensity score matching, Overall survival
Graphical abstract
Highlights
-
•
Large real-world cohort shows exceptional survival benefit in MSI-H/dMMR cholangiocarcinoma.
-
•
MSI-H/dMMR patients derive consistent benefit across multiple treatment lines and regimens.
-
•
Real-world survival in MSI-H/dMMR exceeds outcomes reported in clinical trials.
-
•
MSI-H/dMMR remains a strong prognostic biomarker across clinical subgroups.
Introduction
The DNA mismatch repair (MMR) system plays a fundamental role in maintaining genomic stability by correcting base–base mismatches and insertion–deletion loops generated during DNA replication.1 Deficiency of the MMR pathway results in the accumulation of replication errors within repetitive microsatellite regions, leading to a distinct molecular phenotype termed microsatellite instability–high (MSI-H).[2], [3], [4] Tumors with deficient mismatch repair (dMMR)/MSI-H are typically characterized by a markedly elevated tumor mutational burden and the generation of abundant frameshift-derived neoantigens.3,5,6 These neoantigens enhance tumor immunogenicity, promote immune cell infiltration, and render MSI-H/dMMR tumors highly sensitive to immune checkpoint inhibitors (ICIs).7
Accumulating evidence has demonstrated that MSI-H/dMMR status serves as a robust, tumor-agnostic predictive biomarker for response to programmed cell death protein 1 (PD-1)–based immunotherapy across multiple solid tumor types.4,[8], [9], [10] In the phase II KEYNOTE-158 trial, pembrolizumab demonstrated an objective response rate (ORR) of 34.3%, with a median progression-free survival (PFS) of 4.1 months and a median overall survival (OS) of 23.5 months in patients with previously treated MSI-H/dMMR non-colorectal cancers.11 On the basis of this compelling evidence, PD-1 inhibitors have received regulatory approval for the treatment of advanced MSI-H/dMMR cancers regardless of their tissue of origin, establishing MSI-H/dMMR as one of the most clinically relevant molecular determinants for immunotherapy selection.
Cholangiocarcinoma (CCA) is an aggressive malignancy arising from the biliary epithelium and is anatomically classified into intrahepatic cholangiocarcinoma (iCCA) and extrahepatic cholangiocarcinoma (eCCA).12 CCA is characterized by insidious onset, rapid progression, and extremely poor prognosis, with the majority of patients being diagnosed at an advanced stage and therefore not eligible for curative resection.13 Although gemcitabine plus cisplatin remains the historical backbone of first-line systemic therapy, long-term survival outcomes remain dismal. MSI-H/dMMR constitutes a rare molecular subtype in CCA, with a reported prevalence of approximately 1%–5%.3 Recently, two phase III trials, TOPAZ-1 and KEYNOTE-966, demonstrated that the addition of programmed death-ligand 1 (PD-L1)/PD-1 inhibitors to platinum-based chemotherapy improved survival in unselected patients with advanced CCA; however, MSI-H/dMMR tumors comprised less than 1% of the study population, and no dedicated subgroup analyses were conducted for this rare molecular subtype.14,15 Emerging evidence suggests that MSI-H/dMMR CCA is highly responsive to ICIs. In the KEYNOTE-158 trial, pembrolizumab achieved an ORR of approximately 40% in a limited number of patients with MSI-H/dMMR biliary tract cancer (BTC).11 Nevertheless, the clinical benefit and safety profile of ICIs in patients with MSI-H/dMMR CCA remain insufficiently characterized in real-world practice, particularly in Asian cohorts.
Given the rarity of MSI-H/dMMR CCA and the paucity of real-world evidence, further investigation is urgently warranted to clarify the therapeutic value of ICIs in this distinct molecular subset. Therefore, we conducted a single-center, retrospective real-world study enrolling patients with pathologically confirmed advanced CCA, including both eCCA and iCCA, who were molecularly identified as MSI-H/dMMR. A contemporaneous microsatellite-stable (MSS) cohort was further incorporated as a control group, and propensity score matching (PSM) was performed to enable balanced comparisons of survival and treatment outcomes between MSI-H/dMMR and MSS/preserved MMR (pMMR) populations.
Patients and methods
Patients
We conducted a single-center, retrospective real-world study of patients with pathologically confirmed CCA and MSI-H/dMMR who received ICI–based therapy at our institution between 2017 and 2023. There were no restrictions on disease stage or line of treatment. A contemporaneous control cohort of patients with MSS/pMMR CCA was derived from the same institutional database during the same study period. Patients with unknown MSI-H/dMMR status, mixed histology, severe missing data, or loss to follow-up were excluded (Fig. 1). PSM was subsequently applied to generate balanced MSI-H/dMMR and MSS/pMMR cohorts for comparative survival analyses. The study obtained written informed consent from all participants and was approved by the Institutional Ethics Review Committee at Peking Union Medical College Hospital (PUMCH) (JS-1391). This study followed the guidelines of the Declaration of Helsinki and Good Clinical Practice.
Fig. 1.
Study flowchart.
CCA, cholangiocarcinoma; MSS, microsatellite stable; dMMR, mismatch repair deficient; MSI-H, microsatellite instability–high; pMMR, preserved mismatch repair; PSM, propensity score matching.
Outcomes
The primary endpoint of this study was OS, which was defined as the interval from initiation of ICI therapy to death from any cause. PFS was defined as the time from the first administration of ICIs to the occurrence of radiographic disease progression or death, whichever occurred first. ORR was defined as the proportion of patients achieving complete or partial response, and disease control rate (DCR) as the proportion achieving complete or partial response, or stable disease, according to RECIST version 1.1.
MSI-H/dMMR status
MSI-H/dMMR status was determined via next-generation sequencing (NGS) and/or immunohistochemistry (IHC), reflecting diagnostic integration in clinical practice.
For NGS-based assessment, genomic DNA was extracted from FFPE tumor tissues and matched normal controls, and analyzed using a comprehensive 824-gene targeted sequencing panel (OrigiMed, Shanghai, China). MSI status was determined by a validated bioinformatic algorithm that evaluates the distribution of somatic mutations and microsatellite loci across the sequenced regions; tumors were classified as MSI-H based on the manufacturer’s predefined threshold.
Concurrently, IHC was performed on FFPE specimens to evaluate the expression of four MMR proteins: MLH1, MSH2, MSH6, and PMS2. Tumors were defined as dMMR if there was a complete loss of nuclear expression in at least one of these proteins, whereas those with preserved expression in all four proteins were classified as pMMR.
In this study, patients identified as either MSI-H via NGS or dMMR via IHC were collectively defined as the MSI-H/dMMR cohort. Notably, in a subset of patients who underwent both diagnostic modalities (n = 9), a 100% concordance was observed between the NGS and IHC findings, reinforcing the diagnostic reliability of our inclusion criteria.
PD-L1 expression
PD-L1 status was estimated by IHC staining of FFPE tumor sections using anti–PD-L1 antibodies. Tumors were considered PD-L1 positive when the combined positive score was ≥1.16 In a subset of patients who underwent targeted NGS of tumor tissue, PD-L1 status was additionally derived from sequencing-based assays and harmonized as a binary variable (positive vs. negative) for all analyses.
TMB assessment
Tumor mutational burden (TMB) was determined using the aforementioned 824-gene targeted sequencing panel (OrigiMed, Shanghai, China). TMB was defined as the number of somatic nonsynonymous mutations (including single-nucleotide variants and small insertions/deletions) per megabase (Mb) of the sequenced genomic regions. To ensure the accuracy of TMB estimation, synonymous mutations, known driver mutations (based on the COSMIC database), and suspected germline variants (by filtering against matched blood samples and public databases such as 1000 Genomes and ExAC) were excluded from the final count. High TMB (TMB-H) was defined using a pre-specified threshold of ≥10 mutations per megabase (mut/Mb), consistent with established clinical criteria for pan-cancer genomic profiling.
Statistical analysis
Continuous variables are presented as median (IQR), and categorical variables as frequencies and percentages. Group comparisons were performed using the Mann–Whitney U test, chi-square test, or Fisher’s exact test, as appropriate. To reduce baseline confounding between patients with MSI-H/dMMR and MSS/pMMR CCA, PSM was performed using a logistic regression model including sex, tumor type, ECOG performance status, PD-L1 expression, line of therapy, TNM stage, and Child–Pugh class. Patients with MSI-H/dMMR CCA were matched to those with MSS/pMMR CCA at a 1:3 ratio using optimal nearest-neighbor matching without replacement. Covariate balance before and after matching was assessed using standardized mean differences and visualized using a Love plot. Standardized mean differences account for differences in both scale units and absolute group sizes, enabling direct effect size comparisons irrespective of these methodological variations.17 Median follow-up time was estimated using the reverse Kaplan-Meier method to account for the disparity in survival between groups. Differences in follow-up distribution were assessed via the log-rank test. OS and PFS were estimated using the Kaplan–Meier method and compared using the log-rank test. Median survival times are reported with 95% CIs.
All statistical analyses were performed using R software (version 4.5.2). Survival curves were generated using R and GraphPad Prism (version 9.51; GraphPad Software, San Diego, CA, USA). All tests were two-sided, and a p value <0.05 was considered statistically significant.
Results
Baseline characteristics before/after PSM
As illustrated in the study flowchart (Fig. 1), a total of 331 eligible patients were included in the original cohort. Of these, 23 (6.9%) were classified as MSI-H/dMMR. Compared with MSS/pMMR patients, MSI-H/dMMR patients were younger, more frequently PD-L1–positive, more likely to receive first-line therapy, and more commonly presented with eCCA histology and non–stage IV disease. Notable imbalances were observed between the two groups across several clinicopathologic variables, including tumor subtype, PD-L1 expression, line of therapy, and TNM stage.
To account for these confounding factors, we performed a 1:3 PSM analysis. After 1:3 PSM, 23 patients with MSI-H/dMMR CCA were successfully matched with 69 patients with MSS/pMMR CCA. Baseline characteristics were well balanced between the two matched groups. All covariates achieved acceptable balance with standardized mean differences below 0.30, indicating that an adequate balance of baseline characteristics was achieved after matching (Table 1). Covariate balance before and after PSM is further illustrated by the Love plot in Fig. S1.
Table 1.
Baseline clinical data between the MSS/pMMR group and the MSI-H/dMMR group.
| Variables | Before matching |
After matching |
||||||
|---|---|---|---|---|---|---|---|---|
| MSS/pMMR n = 308 | MSI-H/dMMR n = 23 | p value | SMD△ | MSS/pMMR n = 69 | MSI-H/dMMR n = 23 | p value | SMD△ | |
| Age (mean (SD)) | 58.88 (11.48) | 52.48 (13.28) | 0.034 | -0.482 | 54.62 (11.48) | 52.48 (13.28) | 0.493 | -0.162 |
| Sex (%) | ||||||||
| Female | 115 (37.3) | 9 (39.1) | 0.864 | 0.037 | 20 (29.0) | 9 (39.1) | 0.715 | -0.089 |
| Male | 193 (62.7) | 14 (60.9) | -0.037 | 49 (71.0) | 14 (60.9) | 0.089 | ||
| Tumor type (%) | ||||||||
| eCCA | 52 (16.9) | 10 (43.5) | 0.004 | 0.536 | 26 (37.7) | 10 (43.5) | 0.454 | 0.175 |
| pCCA | 41 (13.3) | 8 (34.8) | 23 (33.3) | 8 (34.8) | ||||
| dCCA | 11 (3.6) | 2 (8.7) | 3 (4.3) | 2 (8.7) | ||||
| iCCA | 256 (83.1) | 13 (56.5) | -0.536 | 43 (62.3) | 13 (56.5) | -0.175 | ||
| ECOG PS (%) | ||||||||
| 0 | 157 (51.0) | 13 (56.5) | 0.614 | 0.112 | 42 (60.9) | 13 (56.5) | >0.999 | -0.029 |
| 1 | 125 (40.6) | 10 (43.5) | 0.058 | 27 (39.1) | 10 (43.5) | 0.029 | ||
| 2 | 24 (7.8) | 0 (0.0) | -0.301 | 0 (0.0) | 0 (0.0) | 0.000 | ||
| 3 | 2 (0.6) | 0 (0.0) | -0.084 | 0 (0.0) | 0 (0.0) | 0.000 | ||
| PD-L1 expression (%) | ||||||||
| Positive | 43 (14.0) | 12 (52.2) | <0.001 | 0.765 | 29 (42.0) | 12 (52.2) | 0.331 | 0.232 |
| Negative | 265 (86.0) | 11 (47.8) | -0.765 | 40 (58.0) | 11 (47.8) | -0.232 | ||
| Line of therapy (%) | ||||||||
| 1 | 192 (62.3) | 20 (87.0) | 0.018 | 0.731 | 55 (79.7) | 20 (87.0) | >0.999 | 0.043 |
| ≥2 | 116 (37.7) | 3 (13.0) | -0.731 | 14 (20.3) | 3 (13.0) | -0.043 | ||
| TNM stage (%) | ||||||||
| II | 17 (5.5) | 2 (8.7) | 0.006 | 0.113 | 12 (17.4) | 2 (8.7) | 0.940 | -0.103 |
| III | 97 (31.5) | 14 (60.9) | 0.602 | 38 (55.1) | 14 (60.9) | 0.148 | ||
| IV | 194 (63.0) | 7 (30.4) | -0.707 | 19 (27.5) | 7 (30.4) | -0.094 | ||
| Child-Pugh (%) | ||||||||
| A | 251 (81.5) | 19 (82.6) | >0.999 | 0.029 | 61 (88.4) | 19 (82.6) | >0.999 | -0.038 |
| B | 56 (18.2) | 4 (17.4) | -0.021 | 8 (11.6) | 4 (17.4) | 0.038 | ||
| C | 1 (0.3) | 0 (0.0) | -0.059 | 0 (0.0) | 0 (0.0) | 0.000 | ||
dCCA, distal cholangiocarcinoma; dMMR, mismatch repair deficient; ECOG PS, Eastern Cooperative Oncology Group performance status; eCCA, extrahepatic cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; MSI-H, microsatellite instability–high; MSS, microsatellite stable; pCCA, perihilar cholangiocarcinoma; PD-L1, programmed death-ligand 1; pMMR, preserved mismatch repair; SMD, standardized mean difference.
P values were calculated using the Mann-Whitney U test for continuous variables and the Chi-square test or Fisher's exact test for categorical variables.
Treatment regimens
Treatment regimens of the MSI-H/dMMR CCA cohort are summarized in Table S1 and Fig. S2. ICI-based combination therapy was predominantly adopted in this real-world cohort. Specifically, 11 patients (47.8%) received ICIs combined with targeted therapy, 8 patients (34.8%) received ICIs combined with chemotherapy, and 4 patients (17.4%) were treated with a triple-combination regimen consisting of ICIs, targeted agents, and chemotherapy. Detailed treatment regimens for both MSI-H and matched MSS cohorts are categorized in Table S2. The utilization of chemo-free regimens (combining ICI with anti-angiogenic tyrosine kinase inhibitors [TKIs]) was comparable between the MSI-H (47.8%) and MSS (49.3%) groups, with the majority administered as first-line therapy.
Efficacy and survival
The significantly longer follow-up in the MSI-H/dMMR group primarily reflects the markedly prolonged survival and treatment duration within this subgroup, as early mortality in the MSS/pMMR group naturally limited the observation window for those patients.
At the time of data cut-off, the median follow-up time duration estimated by the reverse Kaplan-Meier method was 63.5 months (95% CI 35.2-NA) in the MSI-H/dMMR group and 26.9 (95% CI 19.7-51.5) months in the MSS/pMMR group (p = 0.016). Given the 7-year enrollment span (2017–2023) required for this rare population, we verified the enrollment symmetry. The distribution of enrollment years was statistically comparable between the two cohorts (median 2019 vs. 2019, p = 0.137, Fig. S3A), confirming that both groups were managed within the same therapeutic era. The longer follow-up in the MSI-H/dMMR group primarily reflects the survival-driven observation prolongation inherent to this hypermutable phenotype (Fig. S3B). After PSM, from the initiation of immunotherapy, patients with MSI-H/dMMR tumors experienced markedly superior survival compared to those with MSS/pMMR tumors. The median PFS reached 64.1 months in the MSI-H/dMMR group, compared to 7.2 months in the MSS/pMMR group (Fig. 2A, p <0.001). The median OS in the MSI-H/dMMR group was 70.5 months, whereas the median OS in the MSS/pMMR group was significantly shorter at 14.0 months (Fig. 2B, p <0.001).
Fig. 2.
Survival outcomes according to MSI/MMR status.
(A) Kaplan–Meier curves for PFS comparing MSI-H/dMMR and MSS/pMMR cohorts. MSI-H showed significantly prolonged PFS (HR 0.29; 95% CI 0.17–0.51; p <0.001). (B) Kaplan–Meier curves for OS. MSI-H/dMMR similarly demonstrated a marked survival advantage (HR 0.32; 95% CI 0.18–0.59; p <0.001). p values were determined using the log-rank test. HR, hazard ratio; MSS, microsatellite stable; dMMR, mismatch repair deficient; MSI-H, microsatellite instability–high; OS, overall survival; PFS, progression-free survival; pMMR, preserved mismatch repair.
In the unmatched cohort, the ORR was significantly higher in the MSI-H/dMMR group than in the MSS/pMMR group (65% vs. 26%, p <0.001). After PSM, the difference in ORR remained statistically significant, with an ORR of 65% in MSI-H/dMMR patients compared to 38% in matched MSS/pMMR patients (p = 0.029). Similarly, the DCR was markedly higher in the MSI-H/dMMR group than in the MSS/pMMR group before matching (100% vs. 70%, p = 0.002), and this superiority persisted after matching (100% vs. 72%, p = 0.003) (Table S3).
Subgroup analyses further supported the robustness of the findings. Across all prespecified subgroups – including tumor type, PD-L1 expression, and carbohydrate antigen 19-9 (CA19-9) level – MSI-H/dMMR patients consistently demonstrated numerically superior OS and PFS compared with MSS/pMMR patients, without evidence of contradictory trends. The prognostic effect of MSI-H/dMMR appeared particularly notable in patients with low CA19-9 levels, whereas the survival advantage remained directionally consistent even in those with elevated CA19-9 (Figs. S4 and S5). To rigorously rule out selection bias and guarantee-time bias, we performed landmark analyses at multiple time points. The survival benefit of the MSI-H/dMMR group remained significant at the 3-month (hazard ratio [HR] 0.244, p <0.001), 6-month (HR 0.302, p = 0.003), and 12-month (HR 0.374, p = 0.032) landmarks (Table S4). Similarly, a 3-month landmark analysis for PFS confirmed a consistent advantage (HR 0.243, p <0.001), indicating that the superior outcomes were not driven by early attrition in the MSS/pMMR cohort. To further isolate the direct therapeutic efficacy from the confounding effects of prior treatment lines, we restricted the analysis to the homogeneous subgroup of patients treated with first-line immunotherapy (n = 79). In this setting, the MSI-H/dMMR group maintained a profound survival advantage (OS HR 0.146; PFS HR 0.147; p <0.001) compared to the MSS/pMMR group. Notably, therapeutic efficacy in this first-line cohort was characterized by a 65.0% ORR and a 100% DCR in the MSI-H/dMMR group (Fig. S6).
Multivariable Cox regression analysis demonstrated that MSI-H/dMMR status was the strongest independent prognostic factor for both OS and PFS. Compared with MSS/pMMR tumors, MSI-H/dMMR was associated with a markedly reduced risk of death (OS: adjusted HR 0.27, 95% CI 0.12–0.60; p = 0.001) and disease progression (PFS: adjusted HR 0.28, 95% CI 0.13–0.57; p <0.001). Serum CA19-9 <37 U/ml also independently predicted improved outcomes (OS: adjusted HR 0.47, 95% CI 0.25–0.86; p = 0.015; PFS: adjusted HR 0.44, 95% CI 0.25–0.78; p = 0.005) (Tables S5 and S6; Figs. S7 and S8). Although PD-L1 positivity was associated with numerically favorable OS and PFS, the associations did not reach statistical significance in multivariable Cox analyses (OS: adjusted HR 0.70, p = 0.221; PFS: adjusted HR 0.61, p = 0.063).
Genomic landscape and TMB analysis
Genomic profiling was available for 14 patients with MSI-H/dMMR CCA, and the mutational landscape of the top 30 most frequently altered genes is summarized in Fig. 3. Consistent with the hypermutated phenotype of MSI-H/dMMR tumors, a broad spectrum of genetic alterations was observed, predominantly consisting of frameshift, nonsense, missense, and splice-site mutations.
Fig. 3.
Comprehensive genomic landscape and clinicopathological correlations in MSI-H cholangiocarcinoma.
The oncoplot visualizes the somatic mutation profile of the top 30 most frequently altered genes across 14 patients with MSI-H CCA and available NGS data. Columns represent individual patients, stratified primarily by treatment strategy (Chemo-free vs. Chemo-containing) and secondarily by overall survival duration. Rows represent genes ranked by mutation frequency. Different colors in the central matrix indicate specific types of genomic alterations. Integrated annotation tracks at the bottom display matched clinical and molecular characteristics, including TMB, treatment group, survival outcomes, and tumor staging. Notably, LRP1B mutations were observed to be enriched in the Chemo-free cohort, correlating with the "hyper-mutated" phenotype. The level of significance for LRP1B enrichment was p = 0.029, calculated using Fisher's exact test. CCA, cholangiocarcinoma; CNV, copy number variations; eCCA, extrahepatic cholangiocarcinoma; ECOG PS, Eastern Cooperative Oncology Group performance status; iCCA, intrahepatic cholangiocarcinoma; MSI-H, microsatellite instability–high; OS, overall survival; PD-L1, programmed death-ligand 1; PFS, progression-free survival; TMB, tumor mutational burden.
The most frequently altered genes included ARID1A, PBRM1, and KMT2D, as well as MSI-related genes such as MLH1, RNF43, TGFBR2, and ACVR2A. Mutations in TP53, NOTCH3, PIK3CA, AXIN1, and APC were also commonly observed. Concurrent alterations involving multiple functional gene categories were detected in several patients. Integrated annotation of clinical and survival information demonstrated marked inter-patient heterogeneity in OS and PFS within the MSI-H/dMMR subgroup. Tumor subtype, TNM stage, PD-L1 expression, ECOG performance status, and CA19-9 level were distributed across different mutational backgrounds.
Notably, we observed a significant correlation between genomic profile and treatment regimen: Mutations in LRP1B were significantly enriched in the favorable outcome group treated with chemo-free regimens compared to the chemo-containing group (85.7% vs. 14.3%, p = 0.029). This suggests that the robust efficacy of ICI plus TKI in the absence of chemotherapy may be biologically driven by the hyper-immunogenic phenotype associated with LRP1B loss.
To further quantify the genomic instability associated with mismatch repair deficiency, we compared the TMB between the MSI-H/dMMR cohort (n = 16) and the matched MSS/pMMR control group (n = 45). TMB assessment revealed a striking biological divergence between the two subtypes (Fig. S9A). The median TMB in the MSI-H/dMMR group was 15.8-fold higher than that in the MSS/pMMR group (30.8 [IQR 22.4-35.2] vs. 1.95 [IQR 0.96-3.8] mut/Mb; p <0.001). The distribution of TMB values in the MSI-H/dMMR subset was consistently elevated with a minimum value of 10.1 mut/Mb, whereas the MSS/pMMR group generally exhibited a low mutational load. Notably, when applying the clinically established threshold of 10 mut/Mb or more, 100% (16/16) of MSI-H/dMMR patients were classified as TMB-High. In contrast, only 11.1% (5/45) of patients in the MSS/pMMR group met this criterion (p <0.001, Fig. S9B). This perfect concordance within the dMMR cohort underscores the robust mechanistic link between mismatch repair deficiency and a hypermutable phenotype in CCA.
Safety and adverse events
Treatment-related adverse events (AEs) are summarized in Table 2. Among the 23 treated patients, 22 (95.7%) experienced AEs of any grade, and grade 3–4 events occurred in 9 patients (39.1%). Severe toxicities occurred predominantly in the chemotherapy-containing cohort (7/12, 58.3%), compared with only 2 patients (18.2%) in the non-chemotherapy cohort. In the chemotherapy cohort, the predominant grade 3–4 AEs were hematologic (thrombocytopenia and anemia, each 25.0%; leukopenia, 16.7%) and hepatic enzyme elevation (ALT/AST, 16.7%). In contrast, chemo-free regimens exhibited a much more favorable safety profile, associated mainly with grade 1–2 toxicities, with only rare occurrences of grade 3–4 events. No treatment-related deaths occurred. These findings indicate substantially higher toxicity with chemotherapy-containing combinations, while immunotherapy monotherapy maintained a favorable safety profile.
Table 2.
Adverse events in the MSI-H/dMMR CCA cohort.
| Total (n = 23, %) |
Chemotherapy-containing (n = 12, %) |
Non-chemotherapy (n = 11, %) |
||||
|---|---|---|---|---|---|---|
| All grade | 3-4 grade | All grade | 3-4 grade | All grade | 3-4 grade | |
| Any adverse event | 22 (95.7) | 9 (39.1) | 12 (100.0) | 7 (58.3) | 10 (90.9) | 2 (18.2) |
| Hematologic | ||||||
| Neutropenia | 5 (21.7) | 1 (4.3) | 5 (41.7) | 1 (8.3) | 0 (0) | 0 (0) |
| Thrombocytopenia | 6 (26.1) | 3 (13.0) | 6 (50.0) | 3 (25.0) | 0 (0) | 0 (0) |
| Anemia | 8 (34.8) | 3 (13.0) | 7 (58.3) | 3 (25.0) | 1 (9.1) | 0 (0) |
| Leukopenia | 3 (13.0) | 2 (8.7) | 3 (25.0) | 2 (16.7) | 0 (0) | 0 (0) |
| Hepatobiliary | ||||||
| ALT/AST increased | 7 (30.4) | 3 (13.0) | 5 (41.7) | 2 (16.7) | 2 (18.2) | 1 (9.1) |
| Hyperbilirubinemia | 2 (8.7) | 1 (4.3) | 1 (8.3) | 1 (8.3) | 1 (9.1) | 0 (0) |
| Hypoalbuminemia | 2 (8.7) | 1 (4.3) | 2 (16.7) | 1 (8.3) | 0 (0) | 0 (0) |
| Gastrointestinal | ||||||
| Nausea/vomiting | 2 (8.7) | 0 (0) | 2 (16.7) | 0 (0) | 0 (0) | 0 (0) |
| Diarrhea | 2 (8.7) | 0 (0) | 2 (16.7) | 0 (0) | 0 (0) | 0 (0) |
| Anorexia | 2 (8.7) | 0 (0) | 2 (16.7) | 0 (0) | 0 (0) | 0 (0) |
| Endocrine/metabolic | ||||||
| Hypothyroidism | 5 (21.7) | 0 (0) | 3 (25.0) | 0 (0) | 2 (18.2) | 0 (0) |
| Hyperuricemia | 1 (4.3) | 0 (0) | 1 (8.3) | 0 (0) | 0 (0) | 0 (0) |
| Dermatologic | ||||||
| Rash | 5 (21.7) | 3 (13.0) | 4 (33.3) | 2 (16.7) | 1 (9.1) | 1 (9.1) |
| Constitutional | ||||||
| Fatigue | 7 (30.4) | 1 (4.3) | 6 (50.0) | 1 (8.3) | 1 (9.1) | 0 (0) |
| Insomnia | 2 (8.7) | 0 (0) | 2 (16.7) | 0 (0) | 0 (0) | 0 (0) |
| Cardiovascular | ||||||
| Hypertension | 3 (13.0) | 1 (4.3) | 2 (16.7) | 1 (8.3) | 1 (9.1) | 0 (0) |
| Neurologic | ||||||
| Headache | 1 (4.3) | 0 (0) | 1 (8.3) | 0 (0) | 0 (0) | 0 (0) |
| Hoarseness | 1 (4.3) | 0 (0) | 1 (8.3) | 0 (0) | 0 (0) | 0 (0) |
| Other | ||||||
| Fever | 1 (4.3) | 0 (0) | 1 (8.3) | 0 (0) | 0 (0) | 0 (0) |
AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CCA, cholangiocarcinoma; dMMR, deficient mismatch repair; MSI-H, microsatellite instability–high.
Discussion
In this real-world study with a PSM control cohort, we comprehensively evaluated the efficacy, survival outcomes, and safety of ICI-based therapy in patients with MSI-H/dMMR CCA. By incorporating a PSM MSS/pMMR control cohort, our results demonstrated that MSI-H/dMMR CCA was associated with a markedly higher ORR and DCR, as well as significantly prolonged PFS and OS compared with MSS/pMMR counterparts, while maintaining a manageable safety profile. These findings provide robust real-world evidence that MSI-H/dMMR represents a strong predictive biomarker for immunotherapy responsiveness in CCA and reinforce the unique therapeutic relevance of MSI-H/dMMR status in CCA and support routine MSI/MMR testing in this disease. The observed incidence of MSI-H/dMMR of 6.9% in our cohort was higher than the 1%–2% reported in several historical series.18,19 However, historically, MSI detection in CCA has been confounded by the limitations of traditional 5-marker PCR assays, which were optimized for colorectal cancer and often fail to capture the unique mutational patterns of non-CRC tumors.20 Furthermore, the characteristic desmoplastic stroma and low tumor cellularity of CCA can attenuate molecular signals, leading to false-negative results in conventional IHC or small-panel assays.21 In contrast, our use of a high-depth (900 × ) 824-gene NGS panel allows for the assessment of hundreds of microsatellite loci, significantly enhancing sensitivity in detecting subtle instability signals – a reliability confirmed by the 100% concordance observed in our dual-platform tested subset (n = 9). This high detection rate is further contextualized by the institutional role of PUMCH as a national tertiary referral center. Patients referred to our center often present with refractory disease and are more likely to undergo comprehensive genomic profiling, resulting in a natural enrichment of rare molecular phenotypes. Thus, our reported prevalence reflects the real-world molecular landscape of a screened, ICI-treated CCA population in a high-volume precision medicine setting, aligning with contemporary diagnostic recommendations for large-panel NGS in advanced malignancies.
We have thoroughly evaluated the potential contribution of MSI-H/dMMR status as an independent prognostic factor to the observed survival benefit. As noted by Göppert et al., MSI-H CCA is frequently characterized by a distinct immune microenvironment, including dense infiltration of CD8+ tumor-infiltrating lymphocytes, which may confer a more favorable clinical outcome independent of therapeutic intervention.19 This prognostic effect is particularly pronounced in early-stage, resectable cases, where an intact immunosurveillance system can effectively restrain micrometastatic foci.
However, it is imperative to rigorously distinguish the prognostic value in early-stage disease from the predictive value in the advanced setting. Recent evidence indicates that in advanced or metastatic CCA, the natural survival advantage conferred by dMMR/MSI-H status is significantly attenuated due to the establishment of immune escape mechanisms, such as PD-L1 upregulation and T-cell exhaustion; in the absence of immunotherapy, outcomes may even be indistinguishable from those of MSS patients.22
In the current study, the median OS of the MSI-H/dMMR cohort reached 70.5 months, representing a 5- to 6-fold increase compared to the concurrent matched MSS/pMMR cohort (14.0 months) and historical chemotherapy controls.23,24 Furthermore, the high ORR of 65% and DCR of 100% confirm that this survival benefit is driven by treatment-induced tumor regression rather than slow growth resulting from biological indolence. Therefore, while the immunogenicity of MSI-H/dMMR status provides a biological substrate for prognosis, only intervention with ICIs can unleash this potential in the advanced stage, translating a modest biological advantage into a profound clinical survival benefit.
The clinical efficacy observed in our real-world cohort of patients with MSI-H/dMMR CCA is largely consistent with previously reported data from pivotal immunotherapy trials.25,26 Previous findings highlight a critical knowledge gap regarding the true magnitude of benefit conferred by ICIs specifically in the MSI-H/dMMR CCA population, particularly in real-world settings where patient heterogeneity, subgroup analyses, treatment patterns, and long-term outcomes are insufficiently characterized.8 Unlike the BTC subgroup analysis in the KEYNOTE-158 study, which pooled heterogeneous BTC subtypes and lacked a molecularly matched control group, our study incorporated a contemporaneous MSS/pMMR CCA cohort with PSM, enabling a more rigorous evaluation of the independent prognostic and predictive value of MSI-H/dMMR status.
Compared with the AGEO multicenter cohort conducted in France, our study represents a real-world, single-center Asian cohort exclusively focusing on CCA, thereby providing ethnicity-specific evidence for the efficacy and safety of ICIs in MSI-H/dMMR CCA. This design highlights the unique clinical relevance and translational significance of MSI-H/dMMR as a therapeutic biomarker specifically in Chinese and Asian patients with CCA.26
MSI-H/dMMR consistently conferred survival benefit across all subgroups. These findings align with the established tumor biology of MSI-H/dMMR malignancies, which are characterized by hypermutated phenotypes resulting from dMMR pathways.27 The accumulation of numerous somatic mutations generates a high TMB, leading to the production of immunogenic neoantigens that facilitate enhanced T-cell recognition and antitumor immune responses.28 This immunogenic environment is further amplified by the upregulation of multiple immune checkpoint molecules and increased tumor-infiltrating lymphocytes, creating a milieu particularly susceptible to ICIs.29
At the molecular level, the distinction between MSI-H/dMMR and MSS/pMMR CCA extends beyond the core dMMR pathway deficiency to encompass specific mutational patterns that may influence both tumor behavior and treatment response. The frequent occurrence of mutations in epigenetic regulators such as ARID1A, PBRM1, and KMT2D in MSI-H/dMMR tumors suggests a synergistic relationship between defective DNA repair and chromatin remodeling disruptions.30 ARID1A mutations, particularly common in CCA, contribute to genomic instability and may alter transcriptional programs that affect cell cycle regulation and DNA damage response.31 Similarly, PBRM1 loss impacts the SWI/SNF chromatin remodeling complex, potentially enhancing immunogenicity through modified antigen presentation mechanisms.32 KMT2D alterations, involved in histone methylation and transcriptional regulation, have been associated with distinct immune microenvironment profiles in various malignancies.33 The significant enrichment of LRP1B mutations in the chemo-free group suggests that these patients possessed an ultra-immunogenic phenotype. Mechanistically, LRP1B deficiency has been widely reported as a surrogate biomarker for elevated TMB and enhanced tumor immunogenicity across multiple solid tumors.34,35 This implies that for MSI-H patients harboring LRP1B mutations, cytotoxic chemotherapy may provide limited additional benefit, supporting a chemo-free de-escalation strategy. Future prospective trials stratified by LRP1B status are warranted. We acknowledge that given the limited sample size (n = 14) of our NGS sub-cohort, these observations regarding concomitant molecular alterations remain hypothesis-generating rather than conclusive. These co-occurring mutations likely create a permissive landscape for immune surveillance and attack, potentially underpinning the superior outcomes with immunotherapy despite the aggressive nature of BTC.
The heterogeneity of treatment regimens in this real-world cohort reflects the evolving therapeutic paradigm for CCA over the study period. The current standard of care – gemcitabine/cisplatin plus immunotherapy – was established in late 2022;15 consequently, the majority of patients in our cohort were treated prior to the global adoption of this guideline. In the absence of a unified standard for MSI-H CCA at the time, ICI monotherapy was rarely selected in our practice. This decision was driven by the historically modest ORR observed with single-agent PD-1 blockade in unselected BTC populations, which raised concerns regarding primary resistance and rapid progressive disease.11 Instead, the combination of ICI with anti-angiogenic TKIs was prioritized as a potent chemo-sparing strategy. Mechanistically, anti-angiogenic agents can reprogram the immunosuppressive tumor microenvironment and normalize tumor vasculature to enhance T-cell infiltration, thereby synergizing with ICIs.36
The safety analysis revealed a striking dichotomy in toxicity profiles. Severe AEs were driven predominantly by chemotherapy, with a 58.3% incidence of grade 3–4 toxicities in thechemotherapy-containing group vs. 18.2% in the chemo-free cohort. This observation confirms that cytotoxic agents substantially amplify treatment burden, whereas immunotherapy-based combinations maintain a good tolerability profile. Crucially, this superior safety did not come at the cost of efficacy. Given the exceptional survival outcomes observed in the chemo-free group, our findings provide a compelling rationale for a paradigm shift: for MSI-H/dMMR CCA, cytotoxic chemotherapy may be safely omitted in the first-line setting, consistent with the established standard in other MSI-H malignancies.23 This approach maximizes the therapeutic index, leveraging the tumor's intrinsic immunogenicity to achieve durable disease control while sparing patients from severe hematologic and hepatic toxicities.
We acknowledge that retrospective studies may be subject to selection bias, where patients with indolent tumor biology survive longer to undergo testing. However, our comprehensive sensitivity analyses strongly refute this as the primary driver of our results. First, the survival benefit persisted across 3-, 6-, and 12-month landmark analyses, confirming that the advantage is sustained even after conditioning on long-term survival. Second, in the first-line setting, the MSI-H/dMMR group achieved a 65% ORR and a nearly nine-fold extension in median PFS compared to the MSS/pMMR group. Such rapid tumor regression and durable disease control are characteristic of active therapeutic efficacy rather than mere biological indolence. Collectively, these data confirm that MSI-H/dMMR status is a robust predictive biomarker for ICIs, independent of recruitment era or selection artifacts.
Given the rarity of MSI-H CCA, we benchmarked our outcomes against pivotal global cohorts to contextualize efficacy (Table S7). Our cohort achieved a median OS of 70.5 months and a 100% DCR, numerically surpassing benchmarks from both ICI monotherapy trials and the recent western AGEO study.26 This survival advantage likely stems from the high utilization of ICI plus anti-angiogenic TKIs in our first-line setting, suggesting that this synergistic combination may offer superior efficacy compared to both monotherapy and standard-of-care chemo-immunotherapy, thereby validating a tailored, chemotherapy-free strategy.[14], [15]
Nevertheless, several limitations of our study should be acknowledged. First, this was a single-center retrospective analysis with a relatively small sample size of MSI-H/dMMR patients, reflecting the rarity of this molecular subtype in CCA. Second, although genomic profiling was available for a subset of patients, not all MSI-H/dMMR cases underwent comprehensive NGS, which limited deeper integrative analyses of molecular determinants of response and resistance. Third, treatment regimens were heterogeneous, reflecting real-world clinical practice rather than protocol-driven therapy, which may have introduced additional confounding. Finally, despite the use of rigorous PSM, residual unmeasured confounders cannot be completely excluded.
Despite these limitations, our study represents one of the largest single-center real-world analyses focusing exclusively on MSI-H/dMMR CCA and provides clinically meaningful survival, response, and safety data in this rare subgroup. From a clinical perspective, our findings emphasize that MSI-H/dMMR status should be regarded as a key actionable biomarker in CCA and that ICIs should be strongly considered once MSI-H/dMMR is identified. Future multicenter studies with larger sample sizes and integrated multi-omics profiling are warranted to better refine patient selection, elucidate mechanisms of primary and acquired resistance, and optimize immunotherapy-based treatment strategies in MSI-H/dMMR CCA.
Financial support
This work was supported by the Beijing Natural Science Foundation (L248022), National High Level Hospital Clinical Research Funding [2025-PUMCH-C-015, 2025-PUMCH-D-001], and the Postdoctoral Fellowship Program of CPSF (GZC20240152).
Abbreviations
AE, adverse event; BTC, biliary tract cancer; CA19-9, carbohydrate antigen 19-9; CCA, cholangiocarcinoma; DCR, disease control rate; dMMR, deficient mismatch repair; eCCA, extrahepatic cholangiocarcinoma; HR, hazard ratio; ICIs, immune checkpoint inhibitors; iCCA, intrahepatic cholangiocarcinoma; IHC, immunohistochemistry; MMR, mismatch repair; MSI-H, microsatellite instability–high; MSS, microsatellite-stable; NGS, next-generation sequencing; ORR, objective response rate; OS, overall survival; PD-1, programmed cell death protein 1; PFS, progression-free survival; pMMR, preserved MMR; PSM, propensity score matching.
Authors' contributions
M.P. and X.Y. designed the study, analyzed the data, and drafted the manuscript. C.L. and J.Z. were responsible for data collection and statistical analysis. N.Z., J.L., Z.H., S.L., and B.S. assisted in clinical data acquisition and patient follow-up. X.Y. and H.Z. supervised the study, obtained funding, and critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript.
Data availability
All data supporting the results of the study can be found in the article. Further inquiries can be directed to the corresponding author.
Ethics statement
The study protocol was approved by the institutional review boards and ethics committees of PUMCH (No. JS-1391) and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Please refer to the accompanying ICMJE disclosure forms for further details.
Acknowledgements
We thank the patients who participated in this study and all staff at the hospital for their contributions to this study.
Footnotes
Author names in bold designate shared co-first authorship
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhepr.2026.101816.
Contributor Information
Xiaobo Yang, Email: yangxiaobo67@pumch.cn.
Haitao Zhao, Email: zhaoht@pumch.cn.
Supplementary data
The following are the Supplementary data to this article:
References
- 1.Baretti M., Le D.T. DNA mismatch repair in cancer. Pharmacol Ther. 2018;189:45–62. doi: 10.1016/j.pharmthera.2018.04.004. [DOI] [PubMed] [Google Scholar]
- 2.Ambrosini M., Manca P., Nasca V., et al. Epidemiology, pathogenesis, biology and evolving management of MSI-H/dMMR cancers. Nat Rev Clin Oncol. 2025;22(6):385–407. doi: 10.1038/s41571-025-01015-z. [DOI] [PubMed] [Google Scholar]
- 3.Yang X., Lian B., Zhang N., et al. Genomic characterization and immunotherapy for microsatellite instability-high in cholangiocarcinoma. BMC Med. 2024;22(1):42. doi: 10.1186/s12916-024-03257-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dudley J.C., Lin M.T., Le D.T., Eshleman J.R. Microsatellite instability as a biomarker for PD-1 blockade. Clin Cancer Res Off J Am Assoc Cancer Res. 2016;22(4):813–820. doi: 10.1158/1078-0432.CCR-15-1678. [DOI] [PubMed] [Google Scholar]
- 5.Schwitalle Y., Kloor M., Eiermann S., et al. Immune response against frameshift-induced neopeptides in HNPCC patients and healthy HNPCC mutation carriers. Gastroenterology. 2008;134(4):988–997. doi: 10.1053/j.gastro.2008.01.015. [DOI] [PubMed] [Google Scholar]
- 6.Llosa N.J., Cruise M., Tam A., et al. The vigorous immune microenvironment of microsatellite instable colon cancer is balanced by multiple counter-inhibitory checkpoints. Cancer Discov. 2015;5(1):43–51. doi: 10.1158/2159-8290.CD-14-0863. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Holder A.M., Dedeilia A., Sierra-Davidson K., et al. Defining clinically useful biomarkers of immune checkpoint inhibitors in solid tumours. Nat Rev Cancer. 2024;24(7):498–512. doi: 10.1038/s41568-024-00705-7. [DOI] [PubMed] [Google Scholar]
- 8.Maio M., Ascierto P.A., Manzyuk L., et al. Pembrolizumab in microsatellite instability high or mismatch repair deficient cancers: updated analysis from the phase II KEYNOTE-158 study. Ann Oncol Off J Eur Soc Med Oncol. 2022;33(9):929–938. doi: 10.1016/j.annonc.2022.05.519. [DOI] [PubMed] [Google Scholar]
- 9.Zhang N., Yang X., Piao M., et al. Biomarkers and prognostic factors of PD-1/PD-L1 inhibitor-based therapy in patients with advanced hepatocellular carcinoma. Biomark Res. 2024;12(1):26. doi: 10.1186/s40364-023-00535-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Piao M., Li C., Huang Z., et al. Conversion surgery after immune checkpoint inhibitor-based combination therapy for initially unresectable hepatocellular carcinoma: a retrospective cohort study. Liver Cancer. 2025;14(4):456–473. doi: 10.1159/000543994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Marabelle A., Le D.T., Ascierto P.A., et al. Efficacy of pembrolizumab in patients with noncolorectal high microsatellite instability/mismatch repair-deficient cancer: results from the phase II KEYNOTE-158 study. J Clin Oncol Off J Am Soc Clin Oncol. 2020;38(1):1–10. doi: 10.1200/JCO.19.02105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Komuta M. Intrahepatic cholangiocarcinoma: tumour heterogeneity and its clinical relevance. Clin Mol Hepatol. 2022;28(3):396–407. doi: 10.3350/cmh.2021.0287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Vithayathil M., Khan S.A. Current epidemiology of cholangiocarcinoma in Western countries. J Hepatol. 2022;77(6):1690–1698. doi: 10.1016/j.jhep.2022.07.022. [DOI] [PubMed] [Google Scholar]
- 14.Kelley R.K., Ueno M., Yoo C., et al. Pembrolizumab in combination with gemcitabine and cisplatin compared with gemcitabine and cisplatin alone for patients with advanced biliary tract cancer (KEYNOTE-966): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Lond Engl. 2023;401(10391):1853–1865. doi: 10.1016/S0140-6736(23)00727-4. [DOI] [PubMed] [Google Scholar]
- 15.Oh D.Y., He A.R., Bouattour M., et al. Durvalumab or placebo plus gemcitabine and cisplatin in participants with advanced biliary tract cancer (TOPAZ-1): updated overall survival from a randomised phase 3 study. Lancet Gastroenterol Hepatol. 2024;9(8):694–704. doi: 10.1016/S2468-1253(24)00095-5. [DOI] [PubMed] [Google Scholar]
- 16.Kim H., Kim R., Jo H., et al. Expression of PD-L1 as a predictive marker of sensitivity to immune checkpoint inhibitors in patients with advanced biliary tract cancer. Ther Adv Gastroenterol. 2022;15 doi: 10.1177/17562848221117638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schober P., Mascha E.J., Vetter T.R. Statistics from A (agreement) to Z (z score): a guide to interpreting common measures of association, agreement, diagnostic accuracy, effect size, heterogeneity, and reliability in medical research. Anesth Analg. 2021;133(6):1633–1641. doi: 10.1213/ANE.0000000000005773. [DOI] [PubMed] [Google Scholar]
- 18.Ando Y., Kumamoto K., Matsukawa H., et al. Low prevalence of biliary tract cancer with defective mismatch repair genes in a Japanese hospital-based population. Oncol Lett. 2022;23(1):4. doi: 10.3892/ol.2021.13122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Goeppert B., Roessler S., Renner M., et al. Mismatch repair deficiency is a rare but putative therapeutically relevant finding in non-liver fluke associated cholangiocarcinoma. Br J Cancer. 2019;120(1):109–114. doi: 10.1038/s41416-018-0199-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shang Y., Yang L., Hu P., et al. Applying next-generation sequencing to detect microsatellite instability in pan-cancer patients: a retrospective study of 35,563 Chinese cases. Npj Precis Oncol. 2025;9(1):303. doi: 10.1038/s41698-025-01096-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fassan M., Angerilli V., Normanno N., et al. Practical guidelines for molecular testing of cholangiocarcinoma in clinical practice: Italian experts’ position paper. Crit Rev Oncol Hematol. 2024;194 doi: 10.1016/j.critrevonc.2023.104224. [DOI] [PubMed] [Google Scholar]
- 22.Khuntikeo N., Padthaisong S., Loilome W., et al. Mismatch repair deficiency is a prognostic factor predicting good survival of opisthorchis viverrini-associated cholangiocarcinoma at early cancer stage. Cancers. 2023;15(19):4831. doi: 10.3390/cancers15194831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Diaz L.A., Shiu K.K., Kim T.W., et al. Pembrolizumab versus chemotherapy for microsatellite instability-high or mismatch repair-deficient metastatic colorectal cancer (KEYNOTE-177): final analysis of a randomised, open-label, phase 3 study. Lancet Oncol. 2022;23(5):659–670. doi: 10.1016/S1470-2045(22)00197-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bridgewater J., Lopes A., Palmer D., et al. Quality of life, long-term survivors and long-term outcome from the ABC-02 study. Br J Cancer. 2016;114(9):965–971. doi: 10.1038/bjc.2016.64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Winkelmann R., Schneider M., Hartmann S., et al. Microsatellite instability occurs rarely in patients with cholangiocarcinoma: a retrospective study from a German tertiary care hospital. Int J Mol Sci. 2018;19(5):1421. doi: 10.3390/ijms19051421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Campo-Le-Brun I., Grapinet E., Aurillac V., et al. Real-world efficacy of immune checkpoint inhibitors in microsatellite unstable/mismatch repair-deficient biliary tract cancer: an AGEO study. Eur J Cancer. 2025;227 doi: 10.1016/j.ejca.2025.115670. [DOI] [PubMed] [Google Scholar]
- 27.Alouani E., Taieb J., Tougeron D., et al. Efficacy of immunotherapy in gastrointestinal (GI) tumors with mismatch repair deficient (MMRd) unusual phenotype: an AGEO real-world study. J Immunother Cancer. 2025;13(10) doi: 10.1136/jitc-2024-011436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Taieb J., Svrcek M., Cohen R., et al. Deficient mismatch repair/microsatellite unstable colorectal cancer: diagnosis, prognosis and treatment. Eur J Cancer. 2022;175:136–157. doi: 10.1016/j.ejca.2022.07.020. [DOI] [PubMed] [Google Scholar]
- 29.Mlecnik B., Bindea G., Angell H.K., et al. Integrative analyses of colorectal cancer show immunoscore is a stronger predictor of patient survival than microsatellite instability. Immunity. 2016;44(3):698–711. doi: 10.1016/j.immuni.2016.02.025. [DOI] [PubMed] [Google Scholar]
- 30.Astier C., Ngo C., Colmet-Daage L., et al. Molecular profiling of biliary tract cancers reveals distinct genomic landscapes between circulating and tissue tumor DNA. Exp Hematol Oncol. 2024;13(1):2. doi: 10.1186/s40164-023-00470-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhang C., Zhu Y., Ai L., et al. Targeting WEE1 in ARID1A/TP53 concurrent mutant colorectal cancer by exploiting RLoop accumulation and DNA repair deficiencies. Adv Sci Weinh Baden-wurtt Ger. 2025;30 doi: 10.1002/advs.202512074. Published online November. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Li R., He J., Liu C., et al. PBRM1 deficiency enhances PD1 immunotherapeutic sensitivity via chromosomal accessibility in colorectal cancer. Theranostics. 2025;15(8):3316–3331. doi: 10.7150/thno.100793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wang G., Chow R.D., Zhu L., et al. CRISPR-GEMM pooled mutagenic screening identifies KMT2D as a major modulator of immune checkpoint blockade. Cancer Discov. 2020;10(12):1912–1933. doi: 10.1158/2159-8290.CD-19-1448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Armstrong A.J., Dietz H., Balli D., et al. LRP1b loss predicts sensitivity to immunotherapy in patients with NSCLC: an analysis of the phase III CheckMate-026 randomized trial. Clin Cancer Res Off J Am Assoc Cancer Res. 2025;31(24):5198–5210. doi: 10.1158/1078-0432.CCR-25-0952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Eklund E.A., Svensson J., Näslund L.S., et al. Comprehensive genetic variant analysis reveals combination of KRAS and LRP1B as a predictive biomarker of response to immunotherapy in patients with non-small cell lung cancer. J Exp Clin Cancer Res CR. 2025;44(1):75. doi: 10.1186/s13046-025-03342-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Villanueva L., Lwin Z., Chung H.C., et al. Lenvatinib plus pembrolizumab for patients with previously treated biliary tract cancers in the multicohort phase II LEAP-005 study. J Clin Oncol. 2021;39(3_suppl):321. doi: 10.1200/JCO.2021.39.3_suppl.321. 321. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data supporting the results of the study can be found in the article. Further inquiries can be directed to the corresponding author.