This retrospective study included patients with histologically or cytologically confirmed advanced BTC who experienced disease progression following GP-based chemotherapy, including neoadjuvant, adjuvant, or palliative settings. Medical records were reviewed from Seoul St. Mary’s Hospital and St. Vincent’s Hospital. The study was conducted in accordance with Korean regulations and the Declaration of Helsinki, and was approved by the Institutional Review Board (IRB) of The Catholic University of Korea, Seoul St. Mary’s Hospital (approval ID: KC25RASI0352). The ethics committee waived the requirement for written informed consent due to the retrospective nature of the study.

Patients received the Nal-IRI/FL regimen, which consisted of intravenous liposomal irinotecan 70 mg/m² administered over 90 minutes, followed by leucovorin 400 mg/m² intravenously over 30 minutes, and fluorouracil 2400 mg/m² as a continuous intravenous infusion over 46 hours, repeated every 2 weeks. Treatment was continued until radiological or clinical disease progression, as assessed by the treating physician, or the occurrence of unacceptable toxicity. Dose and schedule modifications were permitted at the physician’s discretion to manage adverse events.

Relative dose intensity (RDI) was defined as the ratio of the delivered dose intensity to the planned dose intensity over a specified period, expressed as a percentage. RDI was evaluated in patients who received at least two cycles of chemotherapy.

Tumor response was evaluated based on the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Imaging assessments, including computed tomography or magnetic resonance imaging of the chest, abdomen, and pelvis, were performed every 6 to 8 weeks after treatment initiation. Additional imaging studies were conducted when clinically indicated. Serum carbohydrate antigen 19-9 (CA19-9) levels were assessed concurrently with radiologic evaluations. Adverse events were assessed at each clinical encounter and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. At each visit, patients were assessed with physical examination, Eastern Cooperative Oncology Group (ECOG) performance status, symptom review, and laboratory tests. Patients were considered evaluable for radiologic response if they had received at least one cycle of treatment and had undergone at least one post-baseline imaging assessment. Fluorouracil-refractory disease was defined as progression occurring during fluorouracil-based chemotherapy or within 3 months of its discontinuation. Fluorouracil-resistant disease was defined as progression occurring between 3 and 6 months after completing such therapy.

In a subset of patients, molecular profiling was performed using formalin-fixed, paraffin-embedded tumor tissue obtained either at the time of diagnosis or upon disease recurrence. Next-generation sequencing (NGS) was conducted using the Oncomine™ Comprehensive Assay Plus (Thermo Fisher Scientific, Waltham, MA, USA), which analyzes 517 cancer-related genes, including oncogenes, tumor suppressor genes, gene fusions, microsatellite instability status, and tumor mutational burden. The therapeutic outcomes were compared according to the presence or absence of specific molecular alterations, including RAS and TP53 mutations.

Descriptive statistics are presented as proportions for categorical variables and medians with interquartile ranges (IQR) for continuous variables. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Continuous variables such as age were not statistically compared, as only summary-level data were available from previously published clinical trials. Fisher’s exact test was used for pairwise comparisons of objective response rate (ORR) and disease control rate (DCR) between the current study and prior prospective trials. OS was defined as the time from initiation of chemotherapy to death from any cause. PFS was defined as the time from the initiation of chemotherapy to either radiographic disease progression or death, whichever occurred first. Kaplan–Meier methodology was used to generate non-parametric estimates of median OS and PFS. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using unstratified Cox proportional hazards regression to assess the effects of treatment and other potential prognostic variables on survival outcomes. Variables considered for univariable analysis were selected based on clinical relevance and potential association with survival, while those included in multivariable analysis were limited to variables that showed statistical significance in univariable analysis or were recognized as established prognostic factors. All statistical tests were two-sided, and p-values less than 0.05 were considered statistically significant. Statistical analyses were performed using SPSS for Windows, version 24.0 (IBM Corp., Armonk, NY, USA), and GraphPad Prism, version 10.3 (GraphPad Software, San Diego, CA, USA).

Between January 2022 and December 2024, a total of 93 patients with advanced BTCs who received Nal-IRI/FL as second- or third-line therapy were retrospectively identified and included in the analysis. Of these, treatment was discontinued in 91 patients, while 2 were still receiving Nal-IRI/FL at the time of data cutoff ( Figure 1 ). Baseline characteristics of this cohort were compared with those reported in the NIFTY and NALIRICC trials ( Table 1 ). The median age was 65 years (IQR, 57–72), comparable across all three cohorts. Although gender distribution showed a statistically significant difference (p = 0.042), the overall proportions of male and female patients were broadly similar. A significantly higher proportion of patients with ECOG performance status 2 was observed in our cohort (p < 0.001). The distribution of primary tumor sites also differed significantly (p = 0.004), with intrahepatic cholangiocarcinoma being predominant in NALIRICC, whereas our cohort and NIFTY exhibited a more balanced distribution across intrahepatic, extrahepatic, and gallbladder cancers. Prior curative surgery was more commonly reported in our study (54.8%) and NALIRICC (61.2%) compared to NIFTY (29.5%, p = 0.001). Regarding treatment history, approximately one-quarter of patients in our cohort received Nal-IRI/FL as third-line therapy (25.8%, p < 0.001).

The median follow-up duration was 4.1 months (95% CI, 3.2–4.8). Among the 93 patients included in this study, 85 (91.4%) experienced disease progression and 83 (89.2%) had died at the time of analysis. The median PFS and OS in the overall cohort were 2.1 months (95% CI, 1.6–2.6) and 4.2 months (95% CI, 3.2–5.2), respectively ( Figures 2A, B ). A In the subgroup of 76 patients (81.7%) evaluable for radiologic response, median PFS was 2.5 months (95% CI, 2.2–2.8) and median OS was 5.0 months (95% CI, 2.6–7.4) ( Figures 2C, D ). Table 2 summarizes the effectiveness outcomes in comparison with the NIFTY and NALIRICC trials. The median PFS in our cohort was shorter than that observed in NIFTY (4.2 months) and comparable to NALIRICC (2.6 months). The 6-month PFS rate was 13.7%, compared to 31.8% in NIFTY and 23.0% in NALIRICC. Similarly, the median OS in our cohort was 4.2 months, whereas it was 8.6 months in NIFTY and 6.9 months in NALIRICC. The 6-month OS rate was 38.3%, compared to 60.7% in NIFTY and 50.0% in NALIRICC. The ORR was 7.5%, which was numerically lower than that reported in NIFTY (14.8%) and NALIRICC (14.3%), though not statistically significant (p = 0.156 and p = 0.240, respectively). The DCR was 40.8%, which was significantly lower than that of NIFTY (64.8%, p = 0.018) but not significantly different from NALIRICC (51.0%, p = 0.290).

Table 3 summarizes the results of univariable and multivariable analyses evaluating potential prognostic factors for survival outcomes. In the multivariable model for PFS, having three or more metastatic lesions was the only variable significantly associated with poorer PFS (HR = 1.86; 95% CI, 1.16–2.99; p = 0.010). Other clinical factors, including number of prior treatment lines and fluorouracil resistance, were not significantly associated with PFS. Similarly, in the multivariable analysis for OS, the presence of three or more metastatic lesions was independently associated with inferior OS (HR = 2.39; 95% CI, 1.49–3.83; p = 0.001), and poor performance status was also significantly associated with worse outcomes (HR = 1.81; 95% CI, 1.00–3.27; p = 0.049). Moreover, a history of prior curative surgery showed a trend toward improved OS, although it did not reach statistical significance (HR = 0.65; 95% CI, 0.40–1.05; p = 0.080).

Tissue-based NGS was performed in 71 of the 93 patients (76.3%), and molecular profiling results are summarized in Supplementary Table S1 . Among these, RAS mutations were identified in 21 patients (29.5%) and TP53 mutations in 36 (50.7%). Median PFS was 2.8 months (95% CI, 1.3–4.3) in the RAS-mutant group and 1.8 months (95% CI, 1.4–2.3) in the wild-type group (p = 0.475, Figure 3A ). Median OS was 4.3 months (95% CI, 3.1–5.5) in the RAS-mutant group and 4.0 months (95% CI, 2.4–5.5) in the wild-type group (p = 0.865, Figure 3B ). For TP53 status, median PFS was 2.4 months (95% CI, 1.4–3.3) in the mutant group and 1.8 months (95% CI, 1.3–2.3) in the wild-type group (p = 0.686, Figure 3C ); median OS was 4.3 months (95% CI, 2.5–6.1) and 4.2 months (95% CI, 2.8–5.6), respectively (p = 0.378, Figure 3D ). No statistically significant differences in survival outcomes were observed based on RAS or TP53 mutation status.

Adverse events are detailed in Table 4 . Among the 93 patients in the present cohort, the most frequently reported adverse events of any grade were anemia (73.1%), fatigue (51.6%), and neutropenia (47.3%). Neutropenia of grade 3 or higher occurred in 36 patients (38.7%), and febrile neutropenia was observed in 7 patients (7.5%). Notably, the incidence of grade 3 or higher anemia (20.4%) was higher compared to previous clinical trials. Gastrointestinal toxicities such as diarrhea (7.5%) and nausea (19.3%) were relatively infrequent and mostly low-grade, whereas any grade of fatigue was more common (51.6%) compared to prior studies. Biliary complications, including adverse events such as cholangitis and hyperbilirubinemia, were observed in 31 patients (33.3%), with 27 patients (29.0%) experiencing events of grade 3 or higher, likely reflecting the underlying anatomical predisposition to biliary obstruction. In this context, sepsis of grade 3 or higher occurred in 14 patients (15.1%), including 4 cases of grade 5 events.

Supplementary Table S2 presents the treatment exposure and RDI among the 71 evaluable patients who received at least two cycles of chemotherapy. The median treatment duration was 2.1 months (IQR, 1.4–2.3) and was comparable between patients treated in the second-line and third-line or later settings. The median number of treatment cycles was 4 overall, with a slightly lower median in the third-line group (3 cycles) compared to the second-line group (4 cycles). The overall median RDI was 0.69 (IQR, 0.60–0.76), with similar values across subgroups. Only 17 patients (23.9%) achieved an RDI ≥80%, with a higher proportion observed in the second-line group (26.4%) compared to the third-line group (16.7%).