In this study we evaluated the role of EBRT +/− concomitant chemotherapy in patients with extrahepatic cholangiocarcinoma in the settings of unresectable locally advanced primary disease and local relapse after resection. Toxicity was mild to moderate and no toxic death was observed, suggesting that EBRT can be safely administered to such patients. The response rate of loco-regional disease to EBRT and the overall disease control rate were high, 52% and 92% respectively. In literature, the response rate to EBRT is rarely described, but seems relatively high. For example, Lu et al. reported a 56% response rate in 18 patients treated with high-dose radiotherapy , and McMasters et al. reported partial (2 patients; 22%) and complete (3 patients; 33%) pathological response in 9 patients treated with neoadjuvant chemoradiation . In a series of 35 patients with unresectable intrahepatic cholangiocarcinoma, Chen et al. reported 37% of objective response and 86% of overall disease control ; Jiang et al. reported a 75% objective response rate after EBRT for concurrent regional lymph node metastases in 24 patients with resected intrahepatic cholangiocarcinoma . Despite the response rate, 66% of our patients experienced progression disease. The first site of disease progression was metastatic, mainly hepatic and peritoneal, in agreement with recent studies [5, 13]: Ghafoori et al. showed that the majority of patients with unresectable extra-hepatic disease treated with EBRT had local control at the time of death with 18 out of 21 disease progressions occurring firstly in metastatic sites and only 3 in loco-regional sites ; similarly, in a German study  including 15 patients treated with primary chemoradiation, 53% of first treatment failures were metastatic and 40% were loco-regional. By contrast, the studies, which have reported a prominence of loco-regional progressions as first treatment failure, mainly concern the adjuvant setting . In the study reported by Ben-David and colleagues, the first site of failure was locoregional but corresponded in fact to the 28 patients treated by adjuvant radiotherapy after R0/R1 surgical resection . In our series, OS was poor (median 12 months) as previously reported in literature.
Regarding clinical outcome, we differentiated two populations, those with a locally advanced disease and those with a local recurrence after curative resection. These two populations had different clinical outcome after treatment with RT or RT-CT alone: the first group of patients treated for a locally advanced primary extrahepatic cholangiocarcinoma, with a median OS of 11 months, and a metastatic progression in 87% of patients who relapsed, and the second group of patients similarly treated for an unresectable local relapse after primary surgery, with a median OS of 21 months and 60% of distant failure. In our study, patients with primary extrahepatic cholangiocarcinoma have a poor prognosis, with a high number of metastatic relapses when the relapse occurs, resulting in a high disease-related mortality. The survival rates described here are in accordance with those published in the literature for unresectable patients treated with RT or RT-CT [5, 9, 10, 13, 15–17] and superior to survival rates observed with historical controls and symptomatic palliation. However, this therapeutic strategy has not been widely adopted for all patients. Considering the frequent metastatic evolution of such tumors, and the potential side effects of RT and RT-CT, many patients are treated with CT alone. Recently, in a phase III randomized study, patients (most had metastases) treated with CT alone (gemcitabine and cisplatin) experienced median OS of 11.7 months . Most of the trials investigating the role of chemotherapy in the management of cholangiocarcinomas include patients with locally advanced tumors as well as those with metastatic disease [19–21]. The survival rates obtained in our study are in the same range as the survival rates published for patients treated with RT-CT alone (9 to 14 months) [4, 5, 9, 10, 13, 15–17, 22, 23]. Such a comparison should be cautiously taken into account, because data for RT-CT come mainly from old studies with small sample sizes, longer accrual periods, and lack of CT scan imaging for staging, as compared with the CT studies. Furthermore, some patients included in the CT studies could have benefited from RT-CT after induction CT and/or second-line therapies.
Anyway, in our series as well as in others published series, the prognosis of such patients remains poor, with an important metastatic relapse rate frequently resulting in rapid disease-related mortality, suggesting the necessity to integrate new treatment approaches such as systemic agents [19–21]. For locally advanced pancreatic cancers (LAPC) patients, primary CT helps to identify those who may then potentially benefit from RT-CT, therefore sparing almost 30% of them from this unuseful and heavy treatment; furthermore, the outcome of patients not experiencing disease progression after 3 months of induction CT seems better if subsequent RT-CT is administered as compared with continuation of CT . This therapeutic strategy, currently widely adopted in the management of LAPC and tested in a large randomized trial (LAP 07), could be applied to the management of unresectable main bile duct cholangiocarcinoma, which presents almost the same natural history, with a high and rapid metastatic potential. RT or RT-CT could be reserved to non-progressive patients after induction CT in order to increase the local control. In our study, with the use of a landmark analysis, with the exclusion of early relapse, we did not observe any difference in terms of overall survival, which does not support this hypothesis but the small effective of our study does not allow drawing any definitive conclusion. In this study, RT with or without concomitant CT seems efficient to ensure local control, with 79% of local controls observed at 1 year for such advanced disease, similar to published rates in the literature . The optimal radiation dose in the definitive treatment of biliary malignancies remains unknown, but some authors recommend dose level over 45 to 50 Gy in 5 weeks [25, 26]. Since larger volumes will not tolerate a higher dose administrated with external beam irradiation, internal irradiation with iridium-192 seems to be an attractive approach to boost the area where more tumor burden exists [27–30]. More recently, stereotactic fractionated radiotherapy was shown promising to increase the rate of local control . The role of CT in combination with RT remains undefined. The use of 5FU-based CT in combination with RT is extrapolated from the survival benefit demonstrated with other gastrointestinal malignancies, including pancreatic cancer [32–34]. The number of patients reported to be receiving combined therapy is too small to draw conclusions about the benefit of concomitant CT, with reports showing conflicting results [5, 22, 35, 36]. In our study, RT-CT was well tolerated with an acceptable incidence of acute side effects, not very different from patients treated by exclusive RT. Based on the lack of significant toxicity added with CT in published series, and the proven benefit of this treatment in other gastrointestinal malignancies, the use of concomitant RT-CT can be recommended in biliary tract cancers .
However, our second group of patients, treated by RT for an unresectable local relapse of disease after primary surgery, experienced better outcome, with a median OS of 21 months after RT (+/−CT). The differences in PFS, OS and site of relapses between the 2 groups 1 and 2, were not statistically significant, likely because of an insufficient number of patients, but this difference has still been reported in the literature, in a small series where patients treated for a local relapse after primary surgery had better outcome . Local relapse represents 33% of all relapses of patients treated by primary surgery for extrahepatic cholangiocarcinomas, most of patients developing distant metastasis at the time of recurrence . The median OS described in this study for this specific group of patients is superior to that described in series of patients treated by exclusive surgery [7, 38]. Even though these results should be cautiously taken into account because of the small size of this group, they suggest a potential benefit of such aggressive loco-regional treatment for this specific selected population. Tumors initially resectable and who recurred only locally had certainly a less aggressive behaviour than most of biliary tract cancers, likely explaining in part the better outcome when compared to initially locoregional advanced cancers. These findings are comforted by Yoon and colleagues who reported two cases of patients treated by curative reoperation for recurrent cancer of the extrahepatic bile duct, both alive at 46 and 9 months . The authors concluded that a surgical curative re-resection is possible in selected patients with recurrent bile duct cancer, mostly of the papillary type, and such aggressive treatment should be considered whenever possible in case of recurrence.
Our study presents some limitations. The first one its retrospective nature and the small number of patients included. In fact, this limitation is due to the scarcity of disease: to our knowledge, all studies reported in literature [4, 5, 9, 10, 13, 15–17, 22, 23] are retrospective and small, the largest one including 54 patients . No prospective randomized study is available in literature regarding the role of radiation therapy in unresectable patients. The second limitation is the absence of control group that would include unresectable patients not treated with radiotherapy. In fact, based on several older publications for review see [6, 7] that suggested that EBRT offers effective palliation of symptomatic disease and improves survival, it has been our institutional policy for more than 10 years to routinely recommend radiotherapy to patients with unresectable extrahepatic cholangiocarcinomas, and the aim of our study was to report our experience during the conformal 3D RT period between 1995 and 2008. Similarly, most of published retrospective studies do not include any control group [4, 5, 9, 10, 13, 15–17, 22, 23].