Mutational spectrum for guiding the decision of adjuvant treatment in patients with resected biliary tract carcinoma

Background Systemic chemotherapy or chemoradiation therapy has proven to be effective in treating advanced biliary tract carcinoma (BTC). However, its efficacy in the adjuvant setting remains controversial. Therefore, this study aimed to determine the prognostic significance of genomic biomarkers in resected BTC and their potential role in stratifying patients for adjuvant treatment. Methods We retrospectively reviewed 113 BTC patients who underwent curative‐intent surgery and had available tumor sequencing data. Disease‐free survival (DFS) was the primary outcome examined and univariate analysis was used to identify gene mutations with prognostic value. Favorable and unfavoratble gene subsets were distinguished from the selected genes through grouping, respectively. Multivariate Cox regression was used to identify independent prognostic factors of DFS. Results Our results indicated that mutations in ACVR1B, AR, CTNNB1, ERBB3, and LRP2 were favorable mutations, while mutations in ARID1A, CDKN2A, FGFR2, NF1, NF2, PBRM1, PIK3CA, and TGFBR1 were unfavorable mutations. In addition to age, sex, and node positive, favorable genes (HR = 0.15, 95% CI = 0.04–0.48, p = 0.001) and unfavorable genes (HR = 2.86, 95% CI = 1.51–5.29, p = 0.001) were identified as independent prognostic factors for DFS. Out of the 113 patients, only 35 received adjuvant treatment whereas the majority (78) did not. For patients with both favorable and unfavorable mutations undetected, adjuvant treatment showed negative effect on DFS (median DFS: S441 vs. 956 days, p = 0.010), but there was no significant difference in DFS among those in other mutational subgroups. Conclusions Genomic testing might be useful in guiding the decisions regarding adjuvant treatment in BTC.


| INTRODUCTION
Biliary tract carcinoma (BTC), which mainly includes gallbladder cancer and cholangiocarcinoma, is characterized by late-stage diagnosis, a poor prognosis, and low survival rate. [1][2][3] The pathogenesis of BTC is complex and not yet fully understood. In 2018, the incidence and mortality rates of BTC were 4.12 and 3.25 per 100 thousand people in China, respectively. 4 Despite the low incidence rate of BTC, the disease remains a significant burden in China due to the country's large population.
Currently, surgery is the only curative approach for treating BTC. However, only a quarter of BTC patients are eligible for surgery, and the postoperative recurrence rates are high. 5 Several clinical trials have investigated the use of adjuvant chemotherapy to improve the prognosis in BTC patients, but most have been unsuccessful. The initial phase III randomized trial aimed to evaluate adjuvant chemotherapy with fluorouracil and mitomycin C or mycin C in located patients (139 with CCA and 140 with gallbladder cancer included). This trial revealed a significantly improved 5-year disease-free survival (DFS) for patients with gallbladder cancer undergoing chemotherapy (20.3% vs. 11.6%, p = 0.021), which supporting the notion that adjuvant chemotherapy may offer survival benefits for patients with gallbladder cancer who are undergoing resection. 6 The BILCAP study, 7,8 which compared adjuvant capecitabine treatment with observation only in BTC patients, was another successful trial in postoperative adjuvant chemotherapy. Patients who received adjuvant capecitabine had both a longer recurrence-free survival (RFS) (25.9 vs. 17.6 months, hazard ratio (HR) = 0.71) and overall survival (OS) (52.7 vs. 36.1 months, HR = 0.75 in the per-protocol population) compared to those in the observation arm. Based on this study, clinical guidelines from organizations such as ASCO and CSCO recommend capecitabine as the standard adjuvant chemotherapy regimen for BTC. 9,10 Other major trials, such as the PRODIGE 12-ACCORD 18 study, using a gemcitabine-oxaliplatin combination regimen, 11 or the ASCOT study using an S-1 regimen, 12 failed to show significant improvement in prognosis or are still ongoing. In the past decade, immune checkpoint inhibitor (ICI) therapies have been one of the major breakthroughs in the clinical treatment of cancer. Clinical studies have shown that second or late-line ICIs monotherapy can prolong the medium OS of advanced BTC patients, with a range from 5.2 to 14.2 months. 13 However, the field of postoperative adjuvant immunotherapy for BTC has been scarcely explored. Given the clinical and pathological heterogeneity of BTC and its inadequate response to systemic therapies, biomarkers may be necessary to identify appropriate patients for adjuvant treatment.
The genomic landscape of BTC has been extensively studied. 14,15 However, only a few studies have explored the prognostic value of genomic biomarkers in BTC. Conci et al 16 reported that TP53 and ARID1A mutations were associated with a higher risk of recurrence after surgery. Churi et al 17 suggested that TP53, KRAS, and BAP1 mutations predicted dismal prognosis in advanced BTC patients, unlike mutations in FGF pathway that predicted a prolonged prognosis. In intrahepatic cholangiocarcinoma, TP53, KRAS, and CDKN2A mutations were found to cause poor prognosis, but IDH mutated patients had a favorable prognosis. 18,19 Regardless of treatment, Jolissaint et al 20 stated that node-positive intrahepatic cholangiocarcinoma patients had worse OS when TP53, KRAS or CDKN2A/B mutations were detected. Intriguingly, a study recruiting advanced BTC patients receiving platinum-based chemotherapy found that mutations of the DNA damage repair (DDR) pathway, which occurred in 63.5% of patients, predicted a better prognosis progression free survival (PFS) (6.9 vs. 5.7 months, p = 0.013; OS: 21.0 vs. 13.3 months, p = 0.009). 21 Although these studies disclosed the prognostic significance of genomic biomarkers in BTC, no study had interrogated its clinical value in guiding adjuvant treatment decision. Therefore, in this study, we aimed to examine the possibility of utilizing genomic biomarkers in stratifying surgically-resected BTC patients.

| Study design
A total of 113 patients with BTC who underwent curativeintent surgery between April 2016 and July 2021 at Hunan Provincial People's Hospital were reviewed. The inclusion criteria were as follows: (1) aged 18 years or older; (2) pathologically diagnosed with BTC, which included cholangiocarcinoma and gallbladder cancer; (3) clinical stage I-III and underwent curative-intent surgery; and (4) tumor sequencing data were available for analysis. The exclusion criteria were as follows: (1) mixed hepatobiliary cancer, ampullary cancer, neuroendocrine cancer, or sarcomatoid K E Y W O R D S adjuvant therapy, bile duct cancer, cholangiocarcinoma, gallbladder cancer, next-generation sequencing cancer; (2) R2 resection margin status; and (3) missing clinical and/or pathological information required for analysis. According to the NCCN guidelines, 22 the postoperative adjuvant therapy requirements included the following: (1) for patients who have undergone R0 resection for gallbladder or extrahepatic bile duct cancer and have negative regional nodes or carcinoma in situ at the margin, the options of systemic therapy (preferred), clinical trial (preferred), observation alone, or fluoropyrimidine-based chemoradiation may be considered; and (2) for patients with intrahepatic CCA who have undergone R0 resection, the options of systemic therapy (preferred), clinical trial (preferred), or observation may be considered. The Ethics Committee of the Hunan Provincial People's Hospital approved this study (Approval No.: Research 2022-53), and informed consent was obtained from all enrolled patients.

| Data collection
Clinical data, including information on the patients' age, sex, cancer type, resection margin status, pathological T and N staging, tumor differentiation, microvascular and perineural invasion status, and adjuvant treatment regimens, were collected through medical record. The most commonly used adjuvant treatment regimens were S-1 based, platinum-based, ICI-based, and capecitabine (see Table S1 for details). On average, patients underwent adjuvant treatment was 180 days (with a range of 30-625 days). Follow-up was conducted in accordance with CSCO guideline. 10

| Sequencing data
Tumor tissue specimens obtained from surgery were processed into formalin-fixed paraffin-embedded (FFPE) samples and were collected. All the FFPE samples were reviewed by an independent pathologist to ensure the percentage of tumor cells was sufficient for sequencing. The targeted sequencing panel, Yuansu®450, was applied to FFPE samples and the matched peripheral blood samples were used to detect germline mutations. The experiment was conducted in Shanghai OrigiMed Co., Ltd., Shanghai, China, a laboratory certified by the College of American Pathologists (CAP) and the American Clinical Laboratory Improvement Amendments (CLIA). The procedures for DNA extraction, library preparation, sequencing, and bioinformatics analysis were described in detail in a previous study. 23 In this study, germline or somatic mutations in ATM, ATR, BAP1, and BRCA2 were classified as DDR pathway mutations, and somatic mutations in FGF3 and FGFR2 were classified as fibroblast growth factor (FGF) pathway mutations.

| Statistical analysis
The statistical analysis was performed with R software version 4.1.2. Continuous variables were presented as mean ± SD or median (range) and compared by using t test or Mann-Whitney U test as appropriate. Categorical variables were presented as frequency (percentage) and analyzed using Chi-square test or Fisher's exact test as appropriate. DFS was defined as the time from surgery to recurrence, death of any cause, or last consultation day. Survival analysis was performed utilizing Kaplan-Meier curves and compared using the log-rank test. Multivariate Cox regression analysis was conducted to identify independent prognostic factors. A two-sided p < 0.05 was considered statistically significant.

| Clinical characteristics
In the present study, a total of 113 BTC patients were analyzed (see Figure S1), of whom 35 received adjuvant treatment and 78 did not. Patients' demographic and clinical information were listed in Table 1. The baseline characteristics, including age, sex, resection margin status, T staging, N staging, tumor differentiation, microvascular invasion, and perineural invasion were comparable for patients who received adjuvant treatment and those who did not (p > 0.05), except for cancer type (p < 0.01) and survival condition (p = 0.013), with patients who received adjuvant therapy having a better prognosis.
The median follow-up duration was 788 days (with a range from 6 to 1869 days), moreover, there was no statistical difference in median follow-up time between patients received adjuvant therapy and those who did not (338 vs. 200 days, p = 0.186). At the time of analysis, 55 patients (48.7%) had experienced a recurrence. In the univariate analysis, the use of adjuvant treatment was not found to be associated with DFS (p = 0.320). However, N staging (p = 0.044), tumor differentiation (p = 0.010), and microvascular invasion (p = 0.005) were found to be correlated with DFS (as seen in Figure S2). Other clinical factors, such as age, sex, cancer type, resection margin status, T staging, and perineural invasion, were not found to be associated with DFS (p > 0.05).
In our cohort, we conducted univariate analysis on the top 50 frequently mutated genes (Table S2). As different mutations have diverse effects on DFS, we categorized gene mutations with prognostic value into two subsets: "Favorable genes," which had a positive effect on DFS (with survival curves of the mutated group above that of the wild-type group), including mutations in ACVR1B, AR, CTNNB1, ERBB3, and LRP2 ( Figure S4A-E); and "Unfavorable genes," which had a negative effect on DFS (with survival curves of the mutated group below that of the wild-type group), including mutations in ARID1A, CDKN2A, FGFR2, NF1, NF2, PBRM1, PIK3CA, and TGFBR1 ( Figure S5A-H). "Favorable genes" demonstrated a significantly better DFS (median DFS: not reached vs. 347 days, p < 0.001; Figure 2A), while "Unfavorable genes" exhibited a significantly worse DFS (median DFS: 219 vs. 848 days, p = 0.003; Figure 2B). Furthermore, patients with high TMB demonstrated a trend for better DFS, although this difference was not statistically significant (median DFS: not reached vs. 528 days, p = 0.145; Figure 2C).

| Multivariate prognostic analysis
We performed a multivariate Cox regression analysis to comprehensively compare the prognostic value of genomic biomarkers and conventional clinical factors. As shown in Table 2

| The prognostic significance of mutational subgroup
To better stratify BTC patients based on mutations, we simultaneously considered both favorable and unfavorable genes. We classified the patients into four mutational subgroups: The Good-Prognosis group, with only favorable gene mutations detected; the Poor-Prognosis group, with only unfavorable gene mutations detected; the Unclassifiable group, with both favorable and unfavorable gene mutations detected; and the Wild-Type group, without any favorable or unfavorable gene mutations. Figure 3A displayed the survival curves of these mutational subgroups. The Good-Prognosis and Unclassifiable groups showed the best prognosis, with a median DFS not reached, and followed by the Wild-Type group (median DFS 585 days), while the Poor-Prognosis group had the worst prognosis, with a median DFS of 182 days. Given the small sample size and the comparable prognosis of the Good-Prognosis and Unclassifiable groups, we opted to merge them into a single group for subsequent analysis ( Figure 3B).

| The efficacy of adjuvant treatment in each mutational subgroup
To explore the feasibility of utilizing genomic biomarkers to guide the decision-making process for adjuvant treatment in BTC patients, we conducted a comparison of DFS between patients who received adjuvant treatment and those who did not within each mutational subgroup. Our findings revealed that within the Good-Prognosis/Unclassifiable group, there was no significant difference in DFS (median DFS not reached, p = 0.152; Figure 3C), which suggested that the benefit of adjuvant treatment was somewhat limited. Similarly, in the Poor-Prognosis group, we observed uncompilable results (median DFS 338 vs. 148 days, p = 0.444; Figure 3D). Conversely, patients in the Wild-Type group who did not receive adjuvant treatment experienced a significantly longer DFS (median DFS 441 vs. 956 days, p = 0.010; Figure 3E), implying that adjuvant treatment could potentially be detrimental to this group of patients.

| The prevalence of other treatment-guiding biomarkers in each mutational subgroup
The TMB value could represent a possible biomarkers for predict the efficacy of ICIs. 24 Among the good-prognosis/ unclassifiable group, wild-type group, and poor-prognosis group, there were 4 (17%), 1 (2%), and 3 (6%) patients with high TMB, respectively. Interestingly, among the 5 patients who received adjuvant ICI, the one with the highest TMB value (6.2 mutations/Mb) did not experience recurrence after a 792 days follow-up. Conversely, all the other 4 patients with relatively lower TMB values (range 1.6-5.1 mutations/Mb) experienced disease recurrence at 141 to 585 days post-surgery (see Table S1).

| DISCUSSION
There is a lack of evidences supporting the clinical significance of genomic biomarkers in BTC. [16][17][18][19][20][21] While some studies have examined the efficacy of molecularly guided targeted therapies in the advanced stage setting, 25-28 only few studies have investigated the prognostic value of genetic mutations in the resectable cohort. 16 Currently, according to the various clinical guidelines, 10,22 the consideration of adjuvant therapy for postoperative BTC patients depends mainly on clinical features such as the successful resection degree and the lymph node metastasis status.
As of yet, no study has explored the possibility of guiding adjuvant treatment decisions based on genetic mutations. In agreement with a prior study, 14 TP53 and KRAS mutations were found to be the most frequently mutated genes in BTC. However, our results differed slightly from past studies [16][17][18][19][20][21] in univariate analysis. The negative prognostic values of TP53 and KRAS mutations and the positive prognostic values of DDR pathway, FGF pathway, and IDH1/2 mutations were not statistically significant. These discrepancies could be attributed to the differences in the study population. Previous studies mainly focused on advanced BTCs or a specific subtype of BTC, such as intrahepatic cholangiocarcinoma, whereas our study particularly investigated localized BTC patients who underwent curative-intent surgery. Indeed, the prognostic value of ARID1A and CDKN2A mutations in our study was consistent with previous researches. 16,20 However, when compared to the research that specifically studies surgically resected BTC patients, 16 our results were partially consistent, as ARID1A mutation were demonstrated to have prognostic value, while TP53 mutation did not. We believe that this inconsistency may be due to differences in race. To the best of our knowledge, our study is the first to explore the comprehensive genomic profile of resectable BTCs using a larger panel (450 vs. 52-gene), revealing several novel gene mutations with favorable (i.e., ACVR1B, AR, CTNNB1, ERBB3, and LRP2) and unfavorable (i.e., ARID1A, CDKN2A, FGFR2, NF1, NF2, PBRM1, PIK3CA, and TGFBR1) prognostic values. In multivariate analysis, N staging was found to be an independent prognostic factor for DFS in surgically resected BTC patients. [29][30][31] However, both mutational spectrums-"favorable" and "unfavorable" genes can also significantly predict DFS in localized BTC patients, suggesting that mutations were strong prognostic factors along with conventional clinical factors. Combining the two mutation subsets helps in dividing BTC patients into subgroups with significantly different prognosis.
To examine the clinical value of this mutational subgrouping in guiding the decision of adjuvant treatment in BTC, the DFS of the enroll patients who received or did not receive adjuvant treatment in each mutational subgroup was compared. There was no significant difference in DFS in the Good-Prognosis/Unclassifiable group and the Poor-Prognosis group, indicating that the benefit of adjuvant treatment was limited. However, strikingly, none of the patients receiving adjuvant treatment in the Good-Prognosis/Unclassifiable group had disease recurrence during follow-up, suggesting that adjuvant treatment might be effective in this mutational subgroup. Further study is needed to validate this finding. On the contrary, the DFS was significantly longer for patients who did not receive adjuvant treatment in the Wild-Type group, implying that adjuvant treatment may be detrimental to patients of this mutational group. The lack of available biomarkers for selecting targeted therapy could be a possible reason for the poor outcome of adjuvant treatment in BTCs. There are a few cases supporting the selection of high TMB patients for ICI in our study, but this required further validation in larger sample sizes. The current study had several limitations. First, its retrospective nature may introduce selection biases. Second, the small sample size of certain types of patients (such as those with rare mutations or receiving specific adjuvant regimens) greatly limited the reliability of the corresponding subgroup analysis. Third, due to the relatively small number of cases, we combined the pathological subtypes of BTC, including gallbladder and bile duct cancer in our data analysis. However, there may be differences in the biology among these pathological subtypes of BTC. In future studies, we may consider analyzing the data from different pathological subtypes separately. In all, a prospective and systematic study with a larger sample size and specific subgroups is needed to further validate these findings. The importance of genetic biomarkers is increasingly apparent in various types of cancer, particularly in lung and colon cancer. These biomarkers have found application in precision medicine, including auxiliary treatment planning and molecular profiling. As a result, there has been a surge of research focusing on genetic biomarkers associated with clinical therapeutic efficacy. Additionally, there is an urgent need for more clinical research data in the field of biliary tract tumors to further elucidate the importance of genetic biomarkers in precision therapy.

| CONCLUSIONS
To conclude, this study supports the prognostic significance of mutations in localized BTC patients, and the mutational spectrum might guide the decision for adjuvant treatment. The pre-selection of appropriate patients with genomic biomarkers will be key to the success of upcoming clinical trials studying the efficacy of adjuvant treatment in BTC patients.