Surgical management of recurrent intrahepatic cholangiocarcinoma: predictors, adjuvant chemotherapy, and surgical therapy for recurrence: A multi‐institutional study by the Kyushu Study Group of Liver Surgery

Abstract Objectives of the present study were to identify predictors of the recurrence of intrahepatic cholangiocarcinoma (ICC), and to evaluate the survival benefit of adjuvant chemotherapy and surgical treatment for ICC recurrence. A multi‐institutional retrospective study was carried out in 356 patients with ICC who underwent curative surgery at one of 14 institutions belonging to the Kyushu Study Group of Liver Surgery. A total of 214 patients (60%) had recurrence. Predictors of ICC recurrence were as follows: positive for pathological intrahepatic metastasis (im), positive for lymph node metastasis (n), positive for pathological lymphatic infiltration (ly), pathological bile duct invasion (b), and tumor size ≥4.4 cm. Adjuvant chemotherapy was given to 120 patients (34%) and, in the patients with im or tumor size ≥4.4 cm, adjuvant chemotherapy showed a survival benefit. Only 37 patients (17%) underwent surgical treatment for ICC recurrence. The surgical treatment resulted in a good 5‐year survival rate (44%), which is similar to the rate obtained by the first operation for primary ICC. Prognosis of patients with primary im after the second operation was significantly worse (5‐year survival 18%) compared to patients without primary im. Primary im+ should be considered a contraindication for surgical treatment for ICC recurrence.


| INTRODUCTION
Intrahepatic cholangiocarcinoma (ICC) is a rare disease; however, it has a relatively highly prevalence in Asia and in the USA, at over 1/100 000 population ratio. 1 In the USA, the incidence of ICC has increased by 165% over the past 30 years, and surgical resection remains the only curative treatment option. 2 Surgical results for ICC remain unsatisfactory because of the high rate of recurrence, reported to be 53-71%. [3][4][5] Several studies showed that cure after resection of ICC is an elusive goal, and there is an increasing awareness of the need for predictors of recurrence such as tumor size ≥5.0 cm, 3,6 lymph node metastasis, 3,7 macrovascular invasion, 3 satellite liver nodules, 7 and pathological perineural invasion (pn). 7 To improve patient survival after ICC resection, optimal treatment strategies must be identified for both ICC recurrence and prevention of recurrence such as adjuvant chemotherapy.
The major recurrent focus of ICC has been reported to be the liver, and the rate of ICC recurrence in the liver is approximately 60%. 3,5 Considerable interest has been paid to various treatment options against ICC recurrence such as surgical treatment, [6][7][8] chemotherapy, 9 radiation therapy, 7 radiofrequency ablation (RFA), 10,11 and transarterial chemotherapy 12,13 with various degrees of success. In recent reports, aggressive surgical treatment for ICC recurrence led to good patient survival after recurrence, with a 3year overall survival (OS) rate of 25%, 14 40%, 6 and 100%. 7 However, the efficacy of this strategy remains unclear because of the small numbers of patients who underwent surgical treatment for ICC recurrence in those studies (ie from four to 10 patients). 6,7,14 In the present study, we attempted to identify predictors of ICC recurrence after curative surgeries, and we evaluated the survival benefit of adjuvant chemotherapy and surgical treatment for ICC recurrence in a multi-institutional retrospective study conducted by

| Surgical techniques and follow-up methods
Details of our surgical techniques and patient follow-up methods have been reported previously. [15][16][17] Major hepatectomies with bile duct resection were carried out when bile duct invasion of ICC was suspected to affect the first branch of the hepatic duct. Partial hepatectomies were carried out in cases of peripheral ICC without bile duct invasion. When surgeons believed it would be better to confirm the surgical margins, the resected stump was used for frozen pathology. 15 The right and left lobes of the liver have different routes of lymphatic drainage, and thus the style of lymph node dissection was different according to tumor location on the right or left lobe. 16 We generally did not carry out regional lymph node dissection in patients with peripheral ICC without macroscopic swelling lymph nodes.
The application of adjuvant chemotherapy was determined by each physician in charge based on the patient's age, activities of daily life, and the reported presence or absence of poor prognostic factors such as lymph node metastasis (n), pathological lymphatic infiltration (ly), pathological intrahepatic metastasis (im), and pathologically poor ICC differentiation. 15 Starting in 2006, gemcitabinebased chemotherapy was generally applied and, from 2008, S-1 has been the alternative option. 18,19 Before 2006, oral uracil-tegafur (UFT) and venous 5-fluorouracil (5-FU) were the major options in ICC adjuvant settings.
Indications for surgical treatment for ICC recurrence were determined by each physician or each institutional cancer board. Essentially, surgical treatment was applied only when the patients was considered able to achieve macroscopically curative status by surgical treatment for ICC recurrence, irrespective of the recurrent focuses or the number of recurrences.

| Statistical analysis
Continuous variables are expressed as means AE standard deviation (SD) and were compared using Student's t-test. Categorical variables were compared using the v 2 -test. Multivariate logistic regression models were used to determine independent predictors of ICC recurrence. Any death that occurred in the hospital after surgery was recorded as a mortality. Complications were evaluated with the Clavien's classification, 20     Using all significant variables in the univariate analysis, we used multivariate logistic regression models to determine independent predictors of ICC recurrence ( Table 2). Receiver operating character-

| Survival benefit of adjuvant chemotherapy
Adjuvant chemotherapy was given to 120 patients (34%), including gemcitabine-based (67 patients; 56%), S-1 (19 patients; 16%), and other chemotherapies (34 patients; 28%). Figure 1A demonstrates that adjuvant chemotherapy did not provide a survival benefit in the series of all patients (P=.5898). To analyze the survival benefit of adjuvant chemotherapy especially in patients with independent predictors of ICC recurrence, we carried out subgroup analyses concerning patient survival after surgery, and the results are summarized in Table 3. In patients with im (P=.0110) and in patients with tumor size ≥4.4 cm (P=.0224), adjuvant chemotherapy had a significant survival benefit.

| Clinical results of surgical treatment for ICC recurrence
Surgical treatment for ICC recurrence was carried out in only 37 patients (17%) with 43 focuses. Various focuses were targeted by the surgical treatment: liver (25 patients, 58%), lymph node (eight patients, 19%), lung (seven patients, 16%), bone (one patient, 2%), adrenal gland (one patient, 2%), and brain (one patient, 2%). There were six patients with two resected focuses: four with liver and lung, and two patients with liver and lymph node focuses. Among 25 patients who underwent hepatectomy, two patients also had ablation therapy such as RFA or microwave coagulation therapy (MCT). Seventeen patients (46%) underwent adjuvant chemotherapy prior to a second surgery. However, this rate was not significantly different from that of the patients who underwent non-surgical treatment against ICC recurrences (58%, P=.1738). Median period between the first and second surgeries was 1.85 (0.56-9.09) years. Non-surgical treatment (n=123) for ICC recurrence in our series consisted of chemotherapy (n=93), chemoradiotherapy (n=10), radiotherapy (n=9), ablation therapy (n=6), and transcatheter arterial chemoembolization (n=5). As for patient survival, Figure 2B demonstrates that there was no significant difference between the patients with a first operation  Prognostic factors for overall survival in the second operations were analyzed, and the results are summarized in Table 5. Only one patient with primary n (1.5%), and two patients with primary ly (4.2%) had a second operation. Factors associated with recurrent ICC such as multiple recurrences, extrahepatic metastasis, and disease-free interval<1 year cannot predict patient survival after the second operation; however, the factor "im (+)" associated with primary ICCs is the only poor prognostic factor in the second operation. The 3-year survival rate of patients without primary im (-) was 81%, whereas that of patients with primary im (+) was 18%.
The survival curves after the second operation related to solitary or multiple recurrence (P=.8256) and to the presence or absence of primary im are provided in Figure 3A and B, respectively.

| DISCUSSION
There have been several reports, including our own, concerning the poor prognostic factors or predictors of recurrence after curative operation for ICC. 3,6,7,15 In this report, we identified five indepen-  survival benefit of adjuvant chemotherapy using propensity scorematched modeling. 9 In their series, the propensity score- In the present series, ICC recurrence occurred in 214 patients (60%), and the major recurrent focuses were the liver (37%), lymph nodes (29%), lung (14%), and peritoneum (12%). Recent reports also mentioned that the major ICC recurrent focus was the liver, and the rate of intrahepatic recurrence was relatively high (60%) compared to our results. 3,5 In our 81 patients with lymph node recurrence, regional lymph node recurrence occurred in 13 patients only (16%). Therefore, we propose that aggressive and routine lymph node dissection in patients with ICC cannot always contribute to prevention of lymph node recurrence after operation. 15,16 Although our series is the larger patient population compared to previous reports, 6 and there was no significant difference between the survival curve of patients with first and second operations (Fig 2B). Short-term surgical results such as mortality (4% vs 0%; P<.0001) and morbidity (36% vs 5%; P<.0001) in patients with a second operation were significantly better than those in patients with the first operation (  of the many physicians. However, the number of patients with recurrent ICC who undergo a second operation is quite small and, in this multi-institutional study, we therefore made it a priority to collect as large a number of such patients as we could. As a result, 37 patients with a second operation were included. This number is by far largest reported of such patients. Of course, prospective studies with the same therapeutic policies for ICC recurrence are necessary to confirm our results. In conclusion, the rate of ICC recurrence after curative operation in our series was high at 60%. Predictors of ICC recurrence were as follows: im (+), n (+), ly (+), b (+), and tumor size ≥4.4 cm. Adjuvant chemotherapy for patients with im or tumor size ≥4.4 cm would have a survival benefit. Only a few of our patients (17%) underwent surgical treatment for ICC recurrence; however, this treatment led to good 5-year survival (44%). Primary im (+) should be considered a contraindication for the surgical treatment of ICC recurrence.

DISCLOSURE
Conflict of Interest: Authors declare no conflicts of interest for this article.