Postoperative recurrent patterns of gallbladder cancer: possible implications for adjuvant therapy

Gallbladder cancer (GBC) is an uncommon malignancy with high recurrent rate and poor prognosis. This study investigates the recurrent patterns of postoperative GBC, with the aim to guide the adjuvant treatments, including the radiotherapy. Retrospectively analyzed the 109 GBC patients who underwent surgery in our institution from January 2013 to 2018. Clinical follow-up revealed 54 recurrent cases, of which 40 had detailed locations of recurrence. The sites of recurrence were recorded and divided into the tumor bed, corresponding lymphatic drainage area, intrahepatic recurrence, and the other distant metastasis. The median follow-up time is 34 months (IQR: 11–64). The median disease-free survival (DFS) and overall survival (OS) were 48.8 months and 53.7 months, respectively. Through univariate analysis, risk factors for DFS and OS include tumor markers (CA199 and CEA), hepatic invasion, perineural invasion, lymphovascular invasion, TNM staging and tumor differentiation. Through multivariate analysis, risk factors for DFS include hepatic invasion and TNM staging, and for OS is TNM staging only. Of the 40 cases with specific recurrent sites, 29 patients (29/40, 72.5%) had recurrence in the potential target volume of postoperative radiotherapy (PORT), which include tumor bed and corresponding lymphatic drainage area. The common recurrent lymph node groups included abdominal para-aortic lymph node (No.16, 15/29), hepatoduodenal ligament lymph node (No.12, 8/29), retro-pancreatic head lymph node (No.13, 7/29) and celiac axis lymph node (No.9, 4/29). Twenty cases with recurrences inside the potential PORT target volume were accompanied by distant metastasis. Another 11 cases had distant metastasis alone, so totally 31 cases developed distant metastasis (31/40, 77.5%), including 18 cases with hepatic metastasis. The recurrence and metastasis rates are high in GBC and adjuvant therapy is needed. Up to 75% of the recurrent cases occurred in the potential target volume of postoperative radiotherapy, suggesting that postoperative radiotherapy has the possible value of improving local-regional control. The potential target volume of radiotherapy should include the tumor bed, No.8, No.9, No.11, No.12, No.13, No.14, No. 16a2, No. 16b1 lymph node groups.


Introduction
Gallbladder cancer (GBC) is an uncommon type of tumor [1], with the highest incidence in biliary tract tumors [2]. The occurrence of GBC is significantly related to chronic gallstone, and the malignancy incidence is positively correlated with the size of gallstone [3]. Other risk factors include gallbladder polyps (> 1 cm), chronic cholecystitis, porcelain gallbladder, anomalous pancreaticobiliary ductal junction and chronic typhoid infection [4][5][6][7]. Patients with Early stage of GBC has no obvious typical clinical symptoms [8], and some cases are incidentally discovered during or after surgery [9]. At the time of diagnosis, most patients are with advanced diseases, only 10-30% of patients are surgically resectable [10]. However surgical resection is still the only way of radical cure for gallbladder cancer [11]. As the postoperative recurrence rate and metastasis rate are high, adjuvant therapeutics should should be emphasized that adjuvant therapy can improve survival in GBC patients [12].
The best adjuvant treatment for GBC remains controversial [13]. Only a few retrospective studies and randomized phase III trials on postoperative adjuvant therapy of GBC can be found through PubMed. Postoperative radiotherapy for GBC was proposed by Bosset et al. for the first time14. Todoroki et al. suggested that surgical resection combined with intraoperative radiation therapy with or without external radiation therapy (with 2-and 3-year survival rate of 20.2% and 10.1%) has a better survival rate than surgical treatment alone for stage IV GBC patients (1-and 2-year survival rate of 11.1% and 0%) (P < 0.05) [15]. A retrospective analysis of 3187 GBC patients was conducted by Mojica et al., using the Surveillance, Epidemiology and End Results (SEER) database, patients who received adjuvant radiotherapy had a better median survival (14 months vs. 8 months, P < 0.001). Among the numbers of predictors, the only favorable factors for survival were local lymph node metastasis (P = 0.0001) and liver invasion (P = 0.011) [16]. Similarly, Wang et al. obtained a result that adjuvant radiotherapy was one of the prognostic factors [17]. Yang et al. pointed out that the 5-year survival rate and average survival time of GBC patients in the stage III and IV in adjuvant radiotherapy groups were significantly better than those of the control group (P < 0.05), but patients in stage II have no significant difference between two groups [18]. In several postoperative chemoradiotherapy studies of GBC, Jeong et al. suggested that adjuvant radiotherapy might have benefit in local control of GBC [19] and most of the studies revealed that adjuvant chemoradiotherapy can achieve a good long-term survival rate [20,21], especially it may be useful for the tumor patients undergoing R0 resection without lymph node dissection. However, some studies have shown that adjuvant therapy has no significant effect on improving disease-free survival (DFS) [22]. Due to the lack of phase III randomized controlled trials (RCT) data, the role of postoperative adjuvant radiotherapy remains unclear [23,24]. According to the National Comprehensive Cancer Network (NCCN) Guidelines, postoperative chemoradiotherapy or chemotherapy is feasible for GBC, especially in patients with lymph node-positive disease [25]. Therefore, the recurrent patterns after GBC surgery may provide valuable information for adjuvant therapeutic options.
In this study, failure patterns of 109 postoperative GBC patients were retrospectively analyzed, with the aim to explore the sites with high risk of local recurrence and metastasis. Such information would be of great value to adjuvant therapy, including the design of radiotherapy target volume and the application of systematic therapy.

Patients
The study was approved by the Institutional Review Board of the Second Affiliated Hospital, Zhejiang University School of Medicine (SAHZU). The medical records from January 2013 to 2018 were retrospectively analyzed, including 109 patients who were followed up for at least 2 months after resection. All the patients achieved histopathology diagnosis after surgery. Patients with previous or contemporaneous tumor history were excluded. The clinical variables collected in the retrospective analysis included gender, age, T stage, N stage, clinical TNM staging, tumor differentiation, recurrence time, tumor biomarkers, history of gallstones, bile duct stones, hepatitis, diabetes, hypertension, schistosomiasis and postoperative adjuvant treatment. Tumor clinical staging was performed according to AJCC staging system, 8th Edition. Follow-up period ended on September 28, 2020.

Recurrent patterns
According to the 8th edition of the AJCC staging system, inter-aortocaval lymph node and para-aortic lymph node metastasis are regarded as distant metastasis. Therefore, in this study, the term "local-regional recurrence" was not used. We defined "the potential target volume of postoperative radiotherapy (PORT)" which include the tumor bed (surgical margin) and high-risk lymphatic drainage area. Long term follow-up was performed by specialist physicians. The initial recurrent sites were classified to the potential target volume of PORT recurrence, intrahepatic recurrence, and other distant recurrence. Tumor recurrences were detected by imagiological examination (most cases will be monitored by continuous postoperative evaluation). For recurrent local lymph nodes, it can have the features of greater or equal to 10 mm in short diameter, obvious necrosis, obvious enhancement, and eccentric calcification. If possible, PET/CT can assist in diagnosis, and the presence of distant metastasis can be determined at the same time. There are only 2 cases were diagnosed as tumor recurrence by biopsy, though we encouraged to get definite pathological diagnosis. The initial recurrent pattern, DFS and OS will be analyzed. DFS was from the day of operation to the time of first tumor recurrence, OS was from the day of operation to the death or the last follow-up time.

Statistical analysis
The correlation of patients' characteristics with DFS and OS were analyzed by Kaplan Meier analysis. Significance was evaluated with the log-rank test. Cox proportional hazards models was applied for multivariate survival analysis. Statistical significance was defined as the P value < 0.05. IBM SPSS Statistics 23.0 and Graphpad Prism 5.0 was used for statistical analyses.

Patient characteristics
The characteristic statistics of 109 patients are shown in Table 1. There were 37 males (33.9%) and 72 females (66.1%). The mean age was 64.5 years, and the median age was 65 years (IQR: 56-73). Preoperative CA-199

Follow-up and survival
The      Table 3.

Initial disease recurrence
Among the 54 cases, the imaging of 40 cases were able to identify the sites of initial recurrence. The other 14 cases had follow-up imaging in local hospital and came to our institution with hardcopy diagnostic report only, so the exact locations of tumor recurrences were unavailable. Only 2 recurrent cases were diagnosed  Of the 29 patients whose initial recurrence inside the potential PORT volume, 9 (9/40, 22.5%) developed recurrence only inside the potential PORT volume, including 2 tumor bed recurrences, 7 lymph node metastases, and none had recurrence in both sites simultaneously (Table 4). Twenty had synchronous distant metastasis, including 9 in remnant liver, 16 in other distant sites, and of these, 5 had synchronous recurrences at remnant liver and other distant sites.
Among the 31 patients with distant metastasis, 11 patients did not develop recurrence in the potential PORT volume, including 6 hepatic metastasis, 1 abdominal wall metastasis, 1 multiple pelvic implantation   metastasis and 3 multiple metastasis in remnent liver and abdominal wall. The other 20 patients had distant metastasis and recurrence in the potential PORT volume at the same time.

Discussion
The study showed the factors affecting the survival status of patients with gallbladder cancer after resection and the recurrence mode after operation. According to the research results, 49.5% of the patients had recurrence after surgery, which is consistent with the recurrence rate of some previous studies [26,27].
In this study, postoperative recurrence rate varied with different stages. Postoperative recurrence mainly occurs in the middle and advanced stages of gallbladder cancer. For early gallbladder cancer (TNM stage 0&I), the postoperative recurrence rate is 0%, and the five-year survival rate is 100%. This result was the same as that of lee et al. [28]. Of the 40 cases with detail follow-up records of recurrence, there was no significant difference between relapsing in radiotherapy potential volume only (9 cases, 22.5%) and distant metastasis only (11 cases, 27.5%). According to the different recurrent patterns of previous trials, this result is consistent with the fact [22,26,29,30]. Although there is no plentiful enough large-scale phase III clinical trial, many previous trials and this study suggested that specific adjuvant therapy should be taken for GBC with high recurrence rate.
Postoperative radiotherapy is recommended to combine with chemotherapy in all feasible patients with locally advanced gallbladder cancer. According to the analysis of the initiate recurrent pattern of 29 cases, the common recurrent lymph nodes were para aortic lymph nodes (No.16), hepatoduodenal ligament lymph nodes (No.12), retro-pancreatic head lymph node (No.13), celiac axis lymph node (No.9), right lymph node of cardia (No.1), lesser curvature lymph node (No.3), splenic artery lymph node (No.11) and superior mesenteric artery and vein lymph node (No.14). According to Uesaka and Ito's anatomical description of gallbladder lymphatic drainage, there were four routes of gallbladder lymphatic drainage: gallbladder pancreaticoduodenal route (mainly), gallbladder hepatoduodenal ligament route, gallbladder mesenteric route and porta hepatis route [31,32]. Finally, the first three pathways converged at the level of the left renal vein in the para aortic lymph nodes. The porta hepatis pathway may be related to liver metastasis. The left renal vein is the dividing line between No.16a2 and No.16b1, which is consistent with the recurrent pattern of para aortic lymph nodes in 15 cases. Lymph nodes recurred at the level of left renal vein in every case. In addition, a small number of No.16a1 and No.16b2 recurrences were considered as secondary lymph node metastasis. J Socha et al. performed systematic review and metaanalysis of lymph node metastases in biliary tumors of different T stages [33]. It is suggested that the adjuvant radiotherapy target volume of T3-4 GBC should include No.8, No.9, No.12, No.13, No.14, No.16 lymph node groups. In our retrospective study, only 1 case of recurrent metastatic lymph node was located at No.8. For cross-regional lymph nodes, the center of lymph node was used as the localization marker for statistical analysis during imaging review. The No.9, No.12 and No.13 were all distributed around No.8, and each landmark tissues and organs were close to each other. Therefore, combined with the literature, No.8 should be included in the postoperative radiotherapy target volume. Combined with other research reports and structural relationships, No.1 and No.3 have a small probability of metastasis, and the greater toxic side effects of PORT can be predicted. So the inclusion of the above two groups of lymph nodes in the PORT target volume is not considered in this study. No.14 metastases of locally advanced gallbladder cancer have also been observed in other studies, and the same number of No.11 metastases occurred in our study. Anatomically, we considered that the PORT target volume should include No.11 (corresponds to the lymphatic region of the proximal splenic artery) and No.14, which may improve local control of the disease. At the same time, attention should be paid to patients' radiotherapy tolerance, and appropriate adjustments can be made for different patients and different stages. Therefore, after comprehensive consideration, the target volume of adjuvant radiotherapy should include the tumor bed, No. 8,No.9,No.11,No.12,No.13,No.14,No.16a2 and No.16b1 lymph node groups.