Currently, the most common local treatment approach utilized for unresectable cholangiocarcinoma is fractionated EBRT +/− chemotherapy. A phase II trial of 128 patients with unresectable intrahepatic malignancies treated with concurrent hepatic artery floxuridine and high-dose radiation (median dose 60.75 Gy in 1.5-Gy fractions BID) included 46 patients with cholangiocarcinoma [4]. The median survival for all patients was 15.8 months, with an actuarial 3-year survival of 17%. In that study, 13 of 36 patients (36%) with cholangiocarcinoma who were available for evaluation of treatment response experienced in-field progression, while 15 others progressed in other intrahepatic sites. Another single institution series reported on 52 patients with unresectable extrahepatic cholangiocarcinoma treated with concurrent chemoradiation over a time period of 60 years [5]. Radiation doses and techniques varied widely among the cohort of patients; however, the first site of disease progression was local in 72% of cases, and the median survival was 10 months. Although cholangiocarcinoma is not considered to be curable without surgery, these high rates of in-field progression with standard EBRT +/− chemotherapy imply a more aggressive local treatment approach may be beneficial in properly selected patients.

Newer options showing promise as local therapy for unresectable or recurrent cholangiocarcinoma include TACE, RFA, and PDT [6-8]. A series of 49 patients with unresectable intrahepatic cholangiocarcinoma treated with TACE reported a median survival of 10 months from the time of treatment [7]. The authors found that patients with hypovascular tumors had inferior outcomes compared to those with hypervascular tumors. Another interesting finding was that median survival for tumors <8 cm was 37.2 months and 10.4 months for larger tumors, though this difference was not statistically significant. RFA has shown promise as an effective local therapy, particularly for smaller tumors. For example, a study of 13 patients with 17 primary intrahepatic cholangiocarcinomas treated with RFA reported a crude local control rate of 88% at a median follow-up of 19.5 months. Two local failures occurred, both in tumors >5 cm in diameter. The median overall survival after RFA was 38.5 months. PDT, which involves the interaction of light with photosensitive agents to produce an energy transfer and a local chemical effect, has been tested in a number of small prospective studies with mixed results. A recently-published article summarizes the current medical literature for PDT for unresectable cholangiocarcinoma [8]. Additionally, a small prospective randomized controlled trial of 39 patients comparing PDT after biliary stenting to stenting alone showed a statistically-significant survival benefit favoring PDT [18]. One of the major drawbacks to each of these approaches is that they typically involve an invasive procedure, a problem not encountered with SBRT.

Other groups have reported results using SBRT for cholangiocarcinoma, some of which are summarized in Table ​Table44[12-16]. The largest current series includes 27 patients with unresectable cholangiocarcinoma treated with SBRT [14]. The dose used on all patients in this study was 45 Gy in 3 fractions, prescribed to isocenter. While the authors do not comment specifically on LC, they note median progression-free and overall survivals of 6.7 and 10.6 months, respectively. Interestingly, they found increased rates of duodenal/pyloric ulceration (n = 6, 22%) and duodenal stenosis (n = 3, 11%) compared to other contemporary series, but they were not able to establish a dose-volume relationship for bowel injury. This increased reported rate of bowel injury may be due to the fact that all but one tumor was hilar in location; therefore the high-dose volume was likely located in closer proximity to small bowel. Additionally, the dose and fractionation schema utilized by the Danish group was more aggressive than that of other series. Yet the lack of dose-volume relationship for toxicity implies other underlying and undiscovered factors may also contribute to bowel injury. Other smaller series report one-year LC rates of 65-77% using various dose and fractionation schemes (Table ​(Table4).4). Median survival times vary greatly among these series, likely reflecting patient selection bias. In spite of the smaller patient numbers in the current series, our results compare favorably with those from other institutions, as no patient in our report has experienced a local failure with a median follow-up of more than a year.

The excellent LC in this series of patients must be balanced against potential toxicity, particularly in light of the Grade 5 liver failure that occurred in one patient. Our institutional liver constraint for 5-fraction liver SBRT, which mirrors constraints provided by the American Association of Physicists in Medicine (AAPM) Task Group 101 (TG101), is to keep ≥700 cm³ of normal liver <21 Gy [19,20]. Constraints from the Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC) are slightly more stringent, with recommended mean liver doses of <13 Gy for three-fraction SBRT and <18 Gy for six-fraction SBRT, and ≥700 cm³ of normal liver to receive ≤15 Gy [21]. In this patient with a liver volume of 1980.4 cm³, 1231.0 cm³ of normal liver received <21 Gy and 1051.4 cm3 received <15 Gy, easily meeting both our institutional constraint and the QUANTEC recommendation. The mean liver dose was 19.3 Gy in five fractions, which is slightly higher than the QUANTEC recommendation of 18 Gy for six-fraction SBRT. We have not routinely employed a mean liver dose constraint at our institution when treating with SBRT as the mean dose does not take the size of a patient’s liver into account. Another factor that may have contributed to the patient’s liver failure was her extensive history of previous systemic therapy, which we did not consider during the treatment planning process. This consisted of neoadjuvant doxorubicin, cyclophosphamide, and 5-fluorouracil (5-FU) followed by concurrent 5-FU and postoperative chest wall and nodal irradiation at an outside institution for stage III breast cancer 16 years prior to SBRT. A year later, the patient developed recurrent disease and underwent autologous bone marrow transplantation. This was followed by two years of tamoxifen therapy. The patient was malignancy-free for 13 years, when a new increase in liver function tests led to the work-up and diagnosis of unresectable, intrahepatic cholangiocarcinoma. She received two doses of chemoembolization with mitomycin-C and cisplatin, and then went on to receive nine 28-day cycles of gemcitabine. She had not received any systemic therapy within the previous 12 months of SBRT, nor had she been previously treated with radiotherapy. SBRT was delivered to the previous site of chemoembolization, meaning the previously embolized volume was not considered as part of the ≥700 cm³ dose constraint. The patient was Child-Pugh Class A at the time of SBRT though bilirubin, serum albumin, and international normalized ratio (INR) were never checked in the post-SBRT period as she never developed symptoms or side effects of liver failure until she presented at her local emergency department with signs of disseminated intravascular coagulation (DIC). We hypothesize that the patient’s previous systemic therapy and transplant may have resulted in subclinical liver injury, and the additional radiation injury from SBRT depleted what remained of the patient’s already diminished hepatic reserve, enough to cause fulminant liver failure. Indeed, altered liver function is a common complication of stem cell transplantation, with complications ranging from increased liver function enzymes to graft-versus-host disease, hepatic veno-occlusive disease, and death due to liver failure [22]. Interestingly, at the time of liver failure the patient had no evidence of active disease, neither cholangiocarcinoma nor breast cancer.