Gallbladder cancers (GBCs) typically manifest late with less than 20% being resectable at diagnosis. Surgical resection remains the primary treatment for loco-regional GBCs [1-3]. However, local and/or distant recurrence rates are high, particularly in patients with positive lymph nodes, positive resection margins, and poor differentiation. Loco-regional recurrence (LRR) after radical surgery occurs at sites such as the hepatic hilum, bilio-enteric anastomosis, liver resection margins, and retro-portal lymph nodes [4]. Adjuvant chemotherapy (CT) improves survival compared to observation alone, yet fails to lower the relapse rate, which exceeds 70%, suggesting that chemoradiation could reduce LRR [5]. Evidence from level 1 studies on adjuvant radiotherapy (RT) in GBC is scarce. Nonetheless, randomized studies addressing biliary tract cancer (BTC) also incorporate GBC. A meta-analysis by Ma et al. [6] (involving 3,191 patients from observational studies) noted a significant survival benefit with adjuvant CT, especially in patients with disease stage 2 or higher or following R1 resection. Another meta-analysis (comprising 27 studies) also endorsed adjuvant CT in GBC [7]. The role of RT postoperatively is endorsed for patients with node-positive and R1 resections, based on results from the SWOG phase 2 trial, GECCOR-GB, and observational studies [8,9]. The GECCOR-GB phase 2 randomized trial reported preliminary outcomes comparing adjuvant chemoradiotherapy (CTRT) and CT following curative resection, showing comparable disease-free survival (DFS) between CT and CTRT, with mature data pending [9].

In this region with a high incidence of GBC, we have employed chemoradiation for two decades, and the use of CT was introduced following the BILCAP trial. This study aims to evaluate the real-world efficacy of different adjuvant modalities (CT vs CTRT) in GBC patients after curative surgery (radical cholecystectomy).

We audited the prospectively maintained records of radically resected GBC patients considered for adjuvant therapy during the study period (January 2007 to December 2017). The study received approval from the Institute's ethics committee (2019-25-IP-EXP-7). Initially (2007–2012), the majority of patients underwent concurrent chemoradiation as adjuvant therapy. Recently (2013–2017), the adjuvant protocol for biliary malignancies changed to CT alone for T2, T3 tumors with node-negative disease. Node-positive patients and those with high-risk features (HRFs) (R1: positive margins, LVI: lympho-vascular invasion, PNI: perineural invasion) were offered CTRT. Due to the absence of randomized evidence regarding the efficacy of various adjuvant modalities in different subsets, there was heterogeneity in the practice of adjuvant strategies, resulting in significant overlap in the final treatment. Some physicians preferred CTRT in most cases referred for adjuvant therapy. Patients at low risk or those who declined CT were kept under observation. To evaluate efficacy, treatment outcomes were divided into three groups: observation group, CT group, and CTRT group. CTRT was administered using the 3D-conformal RT technique. The target dose of RT (50.4 Gy in 28 fractions) was delivered to the postoperative tumor bed (marked by metallic clips) and lymphatics (periportal, coeliac, superior mesenteric, and retroperitoneal lymph nodes [RPLN] up to L2 level). In the absence of delineation guidelines for RT in GBC, some physicians did not include RPLN. Concurrent weekly 5-fluorouracil (before 2014) or daily capecitabine 1,250 mg/m² (Monday–Friday) was administered along with RT. Initially, CT consisted of cisplatin-gemcitabine, which later changed to capecitabine alone following the publication of BILCAP results. Patients in the observation arm were monitored until disease progression.

All surgeries were performed in the surgical gastroenterology unit by skilled surgeons aiming for complete microscopic clearance. Less than 1mm clearance on histopathology was classified as R1 resection. Staging utilized the AJCC (American Joint Committee on Cancer 8th edition). Overall survival (OS) was calculated from the time of diagnosis to death or last follow-up. DFS was defined as the period from diagnosis to the detection of disease recurrence—clinical or radiological.

The SPSS software (version 20.0) was utilized for data analysis. Continuous variables were presented as median and interquartile range (IQR). Categorical variables were reported as frequencies with percentages. The Kaplan-Meier method and log-rank test were applied for survival analysis. Demographic data and tumor characteristics were analyzed through univariate and multivariate analyses to determine the impact of various treatment modalities on survival. A p value of < 0.05 was deemed statistically significant. The effects of treatment groups were assessed in established prognostic categories (age groups, sex, stage, nodal status, and risk groups). The low risk group included stage 2, the intermediate risk group included stage 3a, and the high-risk group included stages 3b and 4. Given that the baseline characteristics of patients in the CT and CTRT groups were not identical (Table 1), drawing conclusions from comparisons between these two treatment groups was unjustified. To address this limitation, propensity score matching for age groups, sex, medical risk factors, and risk groups was conducted between the two treatment groups (CT and CTRT) with a margin of error of 0.05 on the estimated overall propensity score. From the initial 37 and 83 subjects in the CT and CTRT groups, respectively, 32 patients from each group were matched perfectly. Survival outcomes were also reported following propensity score analysis (PSA). The comparison of OS and DFS between CT and CTRT groups was performed using the Kaplan-Meier method with a log-rank test.

The median age of the patients (n = 142) was 50 years, with an IQR of 42 to 58 years (Table 1). The demographic profiles of patients across the three groups were comparable (Table 1). At a median follow-up of 31 months (IQR 18–78 months), the median OS for all patients was 50 months. The median OS was 93 months for the CT arm, 34 months for the CTRT arm, and 30 months for the observation arm (p = 0.625) (Table 2, Fig. 1). In univariate analysis, the median OS was numerically better in younger patients (< 50 years: 48 vs 42 months, p = 0.291) and females (50 vs 26 months, p = 0.062) (Table 2, Fig. 2), while it was lower in the presence of comorbidities (26 vs 48 months, p = 0.29). T2 patients had the best OS (72 months [T2] vs 40 months [T3] vs 16 months [T4], p = 0.131), and node-negative patients had better OS compared to node-positive patients (72 vs 40 months, p = 0.082). The effects of various adjuvant therapy modalities on OS, based on prognostic factors, are presented in Table 2. The superior efficacy of CT was evident regardless of nodal status and more pronounced in T2 disease compared to T3/T4 disease. The type of CT showed numerically superior OS with capecitabine alone (n = 8) compared to the cisplatin-gemcitabine combination (survival not reached [NR] vs 53 months, favoring capecitabine alone), though the difference was not statistically significant.

Similarly, the best DFS was observed with CT, with numerically superior DFS in patients under 50 years (36 vs 26 months, p = NS [not significant]), females (36 vs 18 months, p = NS), node-negative vs node-positive patients (40 vs 24 months, p = NS), and stage 2 vs stage 3 vs stage 4 (NR vs 33 vs 20 months, p = NS), as well as R0 vs R1 resections (68 vs 12 months, p < 0.001). DFS was, however, worse in those with comorbidities (24 vs 36 months, p = NS) (Table 3). Capecitabine showed a numerically superior DFS of NR compared to 36 months with cisplatin-gemcitabine (p = NS). Patients who underwent multi-visceral resection had the poorest outcomes, with 18 months DFS and 23 months OS.

Multivariate analysis of factors significantly associated with improved OS and DFS identified: resection margin status (hazard ratio [HR] 2.8, p < 0.001, Fig. 2B) and higher disease stage (HR for stages 3 and 4 was 2.5, p = 0.05) (Table 4, 5). The type of treatment was not significant. The slightly better OS (HR 0.56, p = 0.062) in women was marginally significant (Fig. 2A). After propensity score matching, both the proportion of disease-free patients and the median DFS time in the CT groups were higher compared to the CTRT groups, although the difference was not statistically significant (36 months, 95% confidence interval [95% CI]: 29.9–42.0 months vs 20 months, 95% CI: 15.1–24.9 months; p > 0.05). Similarly, the proportion of patients with OS and the median OS time in the CT groups were higher compared to the CTRT groups, although the difference was not statistically significant (median OS 93 months [95% CI: 22–163 months] vs 28 months [95% CI: 13.6–42.4]) (Fig. 3).

LRR was 27% in the observation arm, 13% in the CT arm, and 18% in the CTRT arm, while the respective distant metastasis (DM) rates were 18%, 27%, and 20% (Table 6). Including RPLN in the RT target volume resulted in a numerically lower LRR (17% vs 32%, p = NS), with similar DM rates of 20%. T2 showed almost identical rates of LRR and DM with CT or CTRT while T3 had similar LRR rates with CT/CTRT but a higher DM rate with CT (40%) vs CTRT (27%). Node-negative disease exhibited similar LRR and DM rates with CT or CTRT, while node-positive disease showed lower LRR with CT (6%)/CTRT (14%) but a higher DM rate with CT (35%) vs CTRT (20%).

Over the past two decades, advances in imaging and surgical techniques have enhanced the R0 resection rates, leading to improved outcomes. The R0 resection rate in this study was 75%, which is comparable to recent reports (68%–87%). Therefore, the focus has shifted more towards systemic rather than local control. Recent studies, including three randomized trials, have shown a significant improvement in OS with the addition of adjuvant CT [5,10,11]. We adopted a similar strategy of using CT as standard care instead of CTRT in node-negative and T2, T3 tumors without HRF. Our real-world data indicate that OS with CT in stage 2 and node-negative cases is superior compared to CTRT and observation groups. Since the observation arm demonstrated inferior survival across most subsets, similar to findings in randomized controlled trials (RCTs), it should not be offered in any subset. Given the superior outcomes of CT over CTRT in T2 and node-negative disease, we suggest that CTRT may be avoided in this subset. Univariate analysis showed a higher DM rate with CT in T3 and node-positive tumors, suggesting that CTRT may be superior in T3 tumors. Despite the inherent limitations of audit, the numerically higher (statistically non-significant) outcomes with CT support its use in low and intermediate risk groups, except in high-risk groups where CTRT may be more beneficial. The suboptimal results of CTRT alone in our series in high-risk groups can be attributed to a higher disease burden leading to increased DM rates. This may also result from variability in the practice of CTRT, including lack of standardization in target volume delineation and inclusion of RPLN in the target volume. Although including RPLN yields better local control, its effectiveness in enhancing survival remains unproven [12]. Future studies should explore improvements in high-risk groups by incorporating consolidation CTRT after initial CT in a randomized study. The existing literature on PSA of postoperative GBC shows benefits with AT, the superiority of CT over surgery alone, short-term superiority of CTRT over surgery alone, and some studies have found CTRT superior to CT [13-15]. In another PSA of 2,689 patients, no significant difference was found in OS between observation group, CT, and CTRT (p > 0.05) in stage II GBC patients. However, in patients with stage III–IV GBC, the CT group showed a superior OS compared to the observation group (p < 0.001), and the CTRT group exhibited superior OS compared to both the CT and observation groups (p < 0.001). A nomogram was also proposed for predicting survival based on age, T-stage, N-stage, number of nodes removed, grade, and type of adjuvant therapy [16-18].

The phase 2 RCT, which confirmed the efficacy of chemoradiation (SWOG) in stage 2-3 tumors, reported 17% local recurrence and 30% DM rates [8]. Our findings indicate a 13% LRR with CT in mostly T2, node-negative cases and 18% with CTRT in predominantly T3-4 cases. Notably, the most significant factors adversely affecting OS were higher disease stage and margin-positive resection. Elevated DM rates associated with CT and node-positive disease suggest that integrating CTRT with CT could potentially enhance outcomes. Reviews of adjuvant CT and CTRT are detailed in Table 7 [5,10,11,13,15,16,18-21].

Similar to the BILCAP trial, which recorded superior outcomes in women, we noted comparably better results in female and younger patients. This pattern of improved outcomes in women and younger patients has been documented in several other studies (Table 7) [5].

The standardization of adjuvant CT in GBC remains under development. The selection of CT medications varies, with published RCTs (BILCAP, PRODIGE, and ASCOT trials) employing one or more of the following: capecitabine, S-1, gemcitabine, and cisplatin/oxaliplatin [5,10,11]. BILCAP and ASCOT reported positive outcomes, whereas PRODIGE did not–attributable to its exclusion of the HRF population. With 47% node positivity and 38% R1 patients, BILCAP was the largest trial, suggesting that a higher proportion of HRF correlates with increased OS and recurrence free survival benefits. Our data demonstrate numerically superior OS and DFS with capecitabine-based CT in T2 and node-negative disease. The GBC component constituted 20% of the BTC trials, yet our results are exclusively from GBC patients. Updated BILCAP results show a median OS of 49.6 months with capecitabine compared to 36 months with observation, which our real-world data echoes. Despite the comparatively worse demographic characteristics of our cohort (25% stage 2, 52% stage 3, and 22% stage 4 against 26% stage 1, 61% stage 2, and 13% stage 3 in the BILCAP study), our outcomes remain similar [22]. Although our study did not reach statistical significance, adjuvant CT is potentially clinically significant, particularly as CT agents evolve. With ongoing drug development, it is conceivable that adjuvant therapy could offer increased survival benefits. Currently, American Society of Clinical Oncology and European Society of Medical Oncology recommend adjuvant capecitabine as the standard treatment for radically resected BTC [23].

The current analysis has several limitations. It is a retrospective study spanning 11 years, a period during which the institute’s policy on adjuvant treatment changed. Despite robust follow-up, selection bias remains inherent in all retrospective studies. To enhance the reliability of results, PSA was employed. There was variability in the use of modalities of AT (CTRT vs CT) among T2/T3 node-negative patients. Besides margin involvement, other clinical parameters, extent of surgery, and pathological findings were similar across the three groups. Patients with HRF selectively underwent CTRT. Despite the selection bias, patients who received adjuvant therapy experienced better outcomes than those who underwent surgery alone. These results are encouraging and strongly support the use of multi-modality treatment. Future randomized studies should focus on definitively assessing the survival benefits of adjuvant CT vs CTRT or consolidation CTRT in patients with HRFs.

It is important to understand that while GBC is often grouped with BTC, it has distinct characteristic traits, presentation, treatment responses, and recurrence patterns compared to intrahepatic and extrahepatic cholangiocarcinoma. The advantage of adjuvant CT over CTRT was observed in patients with low and intermediate risk groups. PSA indicated a non-statistical benefit of CT across all prognostic subsets. Despite a higher proportion of high-risk feature patients, results from this study are consistent with those from BILCAP and provide encouraging evidence for clinical practice. All radically resected GBC patients should receive adjuvant CT, whereas CTRT should be considered for those with HRFs. In summary, this study provides valuable insights into the role of adjuvant therapy in treating GBC.