Therapeutic outcomes and prognostic factors in unresectable gallbladder cancer treated with gemcitabine plus cisplatin

Gallbladder cancer (GBC) is likely to be diagnosed at progressive stages and shows a very poor prognosis. Combination therapy with gemcitabine and cisplatin (GEMCIS) has been widely used as first-line palliative chemotherapy for advanced GBC. This study was designed to investigate the efficacy of GEMCIS and identify prognostic factors in patients with unresectable GBC. Patients with GBC who were treated with GEMCIS from January 2008 to June 2017 in a single tertiary hospital were included. All cases of GBC were diagnosed by pathologic findings and extent of the tumour was assessed by imaging tests. Combination chemotherapy consisted of cisplatin 25 mg/m2 and gemcitabine 1000 mg/m2 intravenously on days 1 and 8 every 3 weeks. To determine factors affecting prognosis, Kaplan–Meier survival analysis, log-rank test and the Cox proportional hazard regression linear model were used. All variables with P < 0.1 in univariable analysis were included in the multivariable model. A total of 173 patients received a median of 5.3 ± 4.4 cycles of chemotherapy over 3.8 ± 3.9 months. Most of the patients (94.8%) were stage IVB at the time of diagnosis and the most common site of metastasis was the liver (42.8%). Disease control rate was 59.5%: 2 (1.2%) patients with complete response, 26 (15.0%) patients with partial response and 75 (43.4%) patients with stable disease. Overall survival (OS) and progression-free survival were 8.1 (95% confidence interval [CI], 7.1–10.2) and 5.6 (95% CI 4.5–6.8) months, respectively. Multivariable regression model indicated that metastasis to liver (hazard ratio [HR] = 1.63, 95% CI 1.11–2.40; P = 0.013), neutrophil-to-lymphocyte ratio (NLR) ≥3 (HR 1.65, 95% CI 1.09–2.49; P = 0.017), CEA ≥ 5 ng/mL (HR 1.50, 95% CI 1.02–2.19; P = 0.038), and CA19–9 ≥ 500 U/mL (HR 1.59, 95% CI 1.01–2.50; P = 0.043) were significantly associated with OS. GEMCIS demonstrated a high disease control rate in patients with unresectable GBC. Factors independently related to OS were metastasis to liver, NLR ≥ 3, CEA ≥ 5 ng/mL and CA19–9 ≥ 500 U/mL.


Background
Gallbladder cancer (GBC) accounts for 80 to 90% of carcinomas in the biliary system. Cancer-related mortality among patients with GBC is increasing and expected to continue to increase until 2030 worldwide [1]. Since there are no specific symptoms in the early stage, GBC is often not diagnosed until advanced stages. Surgery is the only curative treatment method; however, fewer than 10% of patients can undergo surgery, and 50% are found to have metastasis to lymph nodes at the time of diagnosis [2]. GBC has an abysmal prognosis and a median survival of 6 months if untreated [3].
Gemcitabine plus cisplatin (GEMCIS) is widely used as first-line chemotherapy for unresectable GBC based on a recent clinical trial showing favourable outcomes of the combination chemotherapy in patients with biliary tract cancer (BTC) [4]. Median overall survival (OS) and progression-free survival (PFS) were 11.7 and 8.0 months, respectively, in patients with BTC treated with GEMCIS [4]. However, in a more recent phase II clinical study that included only patients with unresectable GBC, median OS and PFS were 6.2 and 3.1 months, respectively [5]. Another study revealed that gallbladder cancer responds poorly to chemotherapy compared to other subtypes of biliary tract cancer [6]. Even though prognosis and response to chemotherapy in patients with GBC are different from those in patients with other subtypes of BTC, there is a lack of studies evaluating the efficacy of GEMCIS only in patients with GBC.
Several studies have identified various prognostic factors in patients with GBC, including jaundice, metastasis to lymph nodes, and metastasis to liver [7][8][9][10]. Tumour markers, particularly carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9), are also well-known prognostic factors in GBC [11][12][13]. In addition, it has been shown that several variables of systemic inflammation response such as neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) have prognostic value in BTC [14][15][16]. However, previous studies are based on retrospective data from patients with BTC or GBC who were treated surgically, and prognostic factors in patients with unresectable GBC treated with GEMCIS are largely unknown.
Although GEMCIS is widely used in patients with advanced GBC based on the clinical trial in 2010, there may be differences in treatment efficacy between GBC and other subtypes of BTC. Moreover, previous studies regarding treatment efficacy and prognostic factors have focused mainly on patients with BTC or GBC who were treated primarily by surgery. Therefore, this study aimed to evaluate the efficacy of GEMCIS and identify prognostic factors in only patients with unresectable GBC.

Study subjects
Patients with unresectable GBC who were treated with GEMCIS at Seoul National University Hospital between January 2008 and June 2017 were analysed. All patients were diagnosed by pathologic confirmation and medical records of the patients were retrospectively reviewed. Extent of the disease was evaluated by contrast-enhanced computed tomography and 18F-fluoro-2-deoxy-D-glucose positron emission tomography with computed tomography scan. Locations of regional lymph nodes included hepatic hilus along the common bile duct, hepatic artery, portal vein, and cystic duct, whereas involvement of lymph nodes around the celiac and superior mesenteric artery was considered distant metastatic disease [17]. A total of 186 patients were reviewed. Five patients who lacked baseline laboratory findings and eight patients with history of active cancer in another organ within 5 years were excluded (Fig. 1). In total, 173 patients were enrolled and analysed. This study was approved by the Institutional Review Board of the Seoul National University Hospital, Seoul, Korea (1703-004-834).

Treatment and data collection
Each cycle of combination chemotherapy consisted of cisplatin 25 mg/m 2 and gemcitabine 1000 mg/m 2 administered intravenously on days 1 and 8 every 3 weeks. The treatment was repeated until the occurrence of unacceptable toxicity, loss to follow-up, confirmation of disease progression, or death. Patients were followed up regularly and laboratory tests and imaging tests were periodically evaluated to assess therapeutic outcomes.
Demographic and clinical variables included age, gender, Eastern Cooperative Oncology Group (ECOG) performance status, body mass index, Charlson comorbidity index score, cancer stage, location of invasion and metastasis, previous therapeutic history, and total number of chemotherapy cycles. Laboratory variables included glomerular filtration rate, aspartate aminotransferase, alanine aminotransferase, total bilirubin, alkaline phosphatase, CEA, CA19-9, NLR, and PLR. NLR and PLR values were defined as the number of absolute neutrophils and platelets divided by the absolute lymphocyte count from samples of peripheral blood, respectively.
The primary endpoint was OS and the secondary endpoints were PFS and objective best overall tumour response. OS was defined as time from initiation of chemotherapy to the final date of follow-up or death from any cause. PFS was defined as time from initial treatment to the confirmation of disease progression or death. Mortality data were collected by the Ministry of the Interior and Safety. Best overall tumour response was assessed by contrast-enhanced CT scan based on the Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria [18]. Minimum duration for definition of stable disease was 1 month and tumour response was evaluated at intervals of two cycles.

Statistical analysis
Continuous variables with a normal distribution were expressed as the mean and standard deviation, while those with a non-normal distribution were summarised as the median and interquartile range (IQR). Categorical variables were presented with percentage, and the Pearson χ 2 test was applied to determine differences between them. OS and PFS were calculated and compared using the Kaplan-Meier method and the log-rank test.
To determine factors affecting prognosis, hazard ratio (HR) and 95% confidence interval (CI) for OS and PFS were calculated using the Cox's proportional hazard regression model. Best cut-off values of NLR, PLR, and tumour markers were obtained using the Contal and O'Quigley method [19]. Multivariable analysis was performed including variables with P < 0.1 in univariable analysis. Variables with P < 0.05 were considered to indicate statistical significance. All statistical analyses were performed using R ver. 3.3.3 (Institute for Statistics and Mathematics, Vienna, Austria; http://www.R-project.org).

Clinical characteristics of patients
Baseline characteristics are summarised in Table 1. The most common site of adjacent cancer invasion was the liver (27.2%), followed by the extrahepatic bile duct (15.0%) and intestine (11.6%). Invasion of hepatic artery and main portal vein was identified in 15 (8.7%) and 11 (6.4%) patients, respectively. Twenty-seven (15.6%) patients were identified to have invasion to more than two organs. Most of the patients (94.8%) were in stage IVB. Among the total patients, liver metastasis was identified in 74 (42.8%) patients, followed by peritoneum in 46 (26.6%) and lung in 29 (16.8%) patients. There were two patients each with metastasis to abdominal wall and adrenal gland and one patient with metastasis to spleen. Metastasis to distant lymph node was found in 102 (59.0%) patients, pericaval area in 58 (33.6%), paraaortic area in 65 (37.6%), and the area above the diaphragm in 31 (17.9%) patients. Eighty-one (46.8%) patients had metastasis to multiple lymph nodes.

Discussion
The aim of this study was to examine the efficacy of GEMCIS and identify prognostic factors in unresectable GBC. In this study, GEMCIS showed a high disease control rate, and liver metastasis, NLR, CEA, and CA 19-9 were significantly associated with prognosis. Overall, this study showed a median OS of 8.6 months,  in general accordance with the previous phase II study in Japan regarding the efficacy of GEMCIS in 14 patients with unresectable GBC [4]. On the other hand, previous study which investigated efficacy of GEMCIS in patients with BTC reported the median OS of 11.7 months, suggesting that prognosis is worse in GBC than that in other subtypes of BTC [4,5]. Moreover, in this study, there were 76 patients who were not able to undergo chemotherapy, 48 patients loss to follow-up, and 4 patients who died immediately after GBC diagnosis. Since GEMCIS was clinically applied in patients with generally good performance, the actual prognosis of unresectable GBC is likely to be worse than the prognosis reported in this study. The gallbladder has no serosal layer near the liver and its perimuscular connective tissue is in direct contact with the liver. Subsequently, invasion of GBC into liver is very common and liver involvement is known to be associated with poor prognosis [3,6]. In this study, however, liver invasion was not an independent prognostic factor, as opposed to previous studies which mostly included patients with resectable GBC. The discrepancy may stem from the different inclusion criteria. Meanwhile, metastasis to liver, which occurs via portal tracts, is common in advanced GBC and indicates poor prognosis [9,20,21]. These findings coincide well with the results of our study that identified liver metastasis as an independent poor prognostic factor. Currently, there are no standard treatment methods regarding liver metastasis although several chemotherapy regimens can be considered in patients with liver metastasis on the basis of the National Comprehensive Cancer Network Clinical Practice Guidelines [22]. Because of the absence of the specific standard treatment guidelines, further studies are required to evaluate best treatment modalities for liver metastasis.
It is well known that CEA and CA19-9 are helpful not only in diagnosis, but also in predicting prognosis of GBC [11,12]. Wang et al. [13] reported that CA19-9 plays an important role as an independent prognostic factor in GBC. Likewise, Park et al. [23] recently reported that CEA independently predicts prognosis in patients with metastatic BTC. Consistent with previous   [23,24], the results of our study demonstrated that baseline CA 19-9 and CEA level were independent prognostic markers. The association of NLR with prognosis has been widely studied in patients with BTC [15,16,23,25]. Zhang et al. [14] found that patients with NLR ≥ 2.61 had a worse prognosis than those with NLR < 2.61 in a study of 316 patients with GBC treated with surgery. In our study, multivariate analysis revealed NLR ≥ 3 as an independent risk factor for poor OS, which was in general accordance with previous studies. The correlation between NLR and prognosis can be explained by the fact that neutrophils secrete vascular endothelial growth factors and several cytokines to promote tumour development and proliferation whereas lymphocytes play a crucial role in tumour defence by inducing cytotoxic cell death [26][27][28]. Pro-inflammatory and pro-angiogenic cytokines are known to be important causative factors in the development of BTC and cytokine-based therapies have been studied [29]. It is expected that the NLR will play an important role in the prediction of prognosis in future development of therapies targeting cancer-associated inflammation.
A limitation of our study is that it is based on retrospective data from a single tertiary care center. Although the results of our study showed a somewhat longer OS than the previous phase II clinical trial in Japan, there may be several confounding factors stemming from the retrospective study design that affect the results of analysis. Second, since the present study was conducted without a control group, careful interpretation and further validation is needed. Despite these limitations, clinical data from this study are helpful because there are very few studies examining the efficacy of GEMCIS combination chemotherapy and prognostic factors specifically in patients with unresectable GBC.
Third, the cut-off values of the tumour markers, NLR, and PLR were estimated based on a relatively small sample size. Although these cut-off values were found to be statistically significant, a large-scale prospective study to determine the ideal cut-off value is needed. Another limitation is that direct invasion of GBC to surrounding tissues was based on imaging findings because we selected only patients with unresectable disease. Diagnostic aspects of imaging tests in evaluating direct invasion of tumour to surrounding tissues may be less sensitive or less accurate than pathologic examination accompanied by surgical resection. Nevertheless, due to the development of recent imaging techniques, the sensitivity and specificity of the diagnosis by computed tomography are as high as 99 and 76%, respectively, and the interpretation of the results of this study seems to be reasonable [30].

Conclusions
In conclusion, GEMCIS is an effective regimen in patients with unresectable GBC. The prognostic factors identified in this study might help accurate patient risk stratification and decision of a proper treatment plan.