Stereotactic Body Radiation Therapy for the Management of Hepatocellular Carcinoma: Efficacy and Safety

Simple Summary This study aimed to describe treatment efficacy and safety in patients with hepatocellular carcinoma (HCC) undergoing stereotactic body radiation therapy (SBRT). In one of the largest retrospective studies to date, we analyzed the data of 318 patients. The median follow-up period was 70.2 months. The local control rate at 24 and 60 months was 94% (91–97%) and 94% (91–97%), respectively. Relapse-free survival at 12, 24, and 60 months was 62% (55–67%), 29% (23–36%), and 13% (8–19%), respectively. OS at 12, 24, and 60 months was 72% (95%CI 67–77%), 44% (38–50%), and 11% (7–15%), respectively. The outcome is highly related to the natural evolution of the underlying cirrhosis. Child-Pugh score B-C, high BCLC score, portal thrombosis, GTV volume, and higher PTV volume reported on total hepatic volume ratio were significantly associated with OS. SBRT is efficient for the management of HCC with a favorable toxicity profile. Abstract This study aimed to describe patient characteristics, treatment efficacy, and safety in patients with hepatocellular carcinoma (HCC) undergoing stereotactic body radiation therapy (SBRT). We retrospectively analyzed data of 318 patients with 375 HCC treated between June 2007 and December 2018. Efficacy (overall survival [OS], relapse-free survival, and local control) and acute and late toxicities were described. The median follow-up period was 70.2 months. Most patients were treated with 45 Gy in three fractions. The median (range) PTV volume was 90.7 (2.6–1067.6) cc. The local control rate at 24 and 60 months was 94% (91–97%) and 94% (91–97%), respectively. Relapse-free survival at 12, 24, and 60 months was 62% (55–67%), 29% (23–36%), and 13% (8–19%), respectively. OS at 12, 24, and 60 months was 72% (95%CI 67–77%), 44% (38–50%), and 11% (7–15%), respectively. Approximately 51% and 38% experienced acute and late toxicity, respectively. Child-Pugh score B-C, high BCLC score, portal thrombosis, high GTV volume, and higher PTV volume reported on total hepatic volume ratio were significantly associated with OS. SBRT is efficient for the management of HCC with a favorable toxicity profile. The outcome is highly related to the natural evolution of the underlying cirrhosis.


Introduction
The incidence of hepatocellular carcinoma (HCC) has been increasing over the past few years. It is now the third most common cause of cancer-related deaths worldwide [1].
HCC usually occurs in the context of cirrhosis, which is mainly caused by a viral infection or chronic alcoholism. However, the emergence of non-alcoholic fatty liver disease (NAFLD) due to lifestyle changes leads us to believe that the number of cases will continue to rise.
With surgery (resection or liver transplantation) or percutaneous destruction (ethanol injection or radiofrequency), HCC should be curable [2,3]. However, due to comorbidities or characteristics of the disease itself (i.e., its anatomical relations), many patients cannot benefit from these treatments. In these cases, other treatments have proven beneficial. Chemoembolization has improved survival rates [4], and is recommended as the first-line treatment for advanced HCC in patients with preserved hepatic function, good Eastern Cooperative Oncology Group (ECOG) performance score, no metastases, and no abnormality of portal flow [5,6]. Since 2007, targeted therapies have emerged as possible palliative treatments for HCC. The kinase inhibitors sorafenib and lenvatinib have been shown to improve survival versus placebo [7,8], and appear to be a good treatment option for patients ineligible for specific treatments or after failure of the treatment [9]. More recently, the results of the Check Mate 040 and Keynote 224 studies have permitted nivolumab and pembrolizumab, respectively, to obtain FDA approval for the treatment of advanced HCC [10,11]. After the publication of the results of the IMbrave150 trial demonstrating an overall survival benefit compared to sorafenib, the combination of atezolizumab with bevacizumab has become a standard in France in advanced HCC [12].
Stereotactic body radiation therapy (SBRT) has also appeared to be a possible ablative treatment for HCC. To date, this therapy has been evaluated in retrospective and phase 2 studies [13][14][15][16][17]. These have shown promising results, with good local control, survival rates, and a favorable toxicity profile. Despite these advantages, SBRT remains mainly proposed for HCC that is inaccessible to curative treatment and after the failure of other therapies.
SBRT is a technique that has been developed in our center in 2007, with a large regional recruitment since then. This gave us the opportunity to gather a large cohort of patients homogeneously treated, with a unique epidemiology regarding the cause of cirrhosis, allowing us to provide evaluation data on SBRT in this disease.
The aim of this retrospective study was to describe the population characteristics and to estimate the safety and efficacy in terms of overall survival (OS), relapse-free survival (RFS), and local control in a large cohort of patients treated with SBRT for HCC, since the establishment of the technique in our center. We also aimed to identify prognostic factors associated with OS and toxicity after SBRT. Finally, we examined the pathological response of patients treated with SBRT as a bridge to liver transplantation.

Patients
All patients who started SBRT on hepatic lesions in our center between June 2007 and December 2018 were retrospectively included in the current study, leading to a total of 318 patients. The patient selection was performed using the CONSORE tool.
The diagnosis of HCC was made using liver biopsy or dynamic imaging, according to Barcelona diagnosis criteria [18]. In this case, HCC diagnosis was established by the detection of contrast enhancement in the arterial phase (wash-in), followed by hypoenhancement in the portal phase (wash-out) on magnetic resonance imaging and/or computed tomography. Indication of SBRT was based on the decision by the institutional multidisciplinary consultation committee of hepatic tumors, which considered among other factors, patient comorbidities, ECOG score, anatomical relations of the HCC, and multiplicity and site of lesions. SBRT was discussed for patients with up to four synchronous hepatic lesions without distant lesions. Failure of other local treatments, such as radiofrequency (RFA) or trans-arterial chemoembolization (TACE) was also an indication for SBRT. Patients could have previously undergone surgical resection, RFA, TACE, radiotherapy, chemotherapy, and targeted therapy. Any previous treatment was not recorded if the target lesion was different from the one treated with SBRT. Child-Pugh, Model For End-Stage Liver Disease (MELD) and Barcelona Clinic Liver Cancer (BCLC) scores were calculated based on clinical, biological and imagery data collected during the first consultation with the radiotherapist. All clinical data were collected from computerized patient records.
The study complies with the "reference methodology" adopted by the French Data Protection Authority (CNIL), all patients provided informed consent for data collection and analysis during the follow-up survey and the study was approved by the relevant Institutional Ethics Committee.

Treatment
For each patient, hepatic SBRT therapy with Cyberknife ® was preceded by placement of fiduciaries to track the lesions. Our process has been previously described in Jarraya et al. [19]. Planning CT is realized during a brief apnea. Additionally, if possible, patients underwent MRI, which was registered with CT, when fiducials were visible on MR, to help with delineation of the target volumes and organs at risk, performed by a radiation oncologist. T2 and dynamic portal venous phases were used to delineate tumors. Clinical Target Volume (CTV) was defined as Gross Target Volume (GTV) + 5 mm margins and Planning Target Volume (PTV) as CTV + 3 mm margins. For multiple lesions, the nearby targets were merged into the same volume.
A total dose of 21-54 Gy in 3-6 fractions was prescribed on the 80% isodose line. Dose to organs at risk (duodenum, stomach, and liver) adhered to the dose constraints in effect at our center. If necessary, target volumes coverage, the total dose, and the number of fractions was modified to fulfill these constraints (from TG 101 AAPM [20]) and adapted for six fractions. In rare cases, the CTV could be suppressed to meet the dose constraints. Dosimetry data was collected from Accuray Precision ® software according to International Committee for Radiological Units (ICRU) 91 criteria [21].
Intra-fraction motion management was dealt with the Synchrony TM system, which correlates the movement of the fiduciaries with external diodes placed on the patient's chest. This enables the system to follow translations of the target during free breathing The use of an abdominal compression belt could be combined to limit respiration-induced movements to facilitate the creation of a correlation model.

Follow-Up
Follow-up was planned every three months in the first three years, and every six months afterwards. A clinical examination was performed by either a radiation oncologist or gastroenterologist. Each consultation was preceded by imaging (MRI or CT scan). Responses were assessed according to the modified RECIST (mRECIST) criteria for hepatocellular carcinoma [22]. Blood work was also performed, collecting data on toxicity and alpha fetoprotein levels. Post-treatment Child-Pugh scores and alpha fetoprotein levels were recorded at three months post-treatment, which corresponded to the first medical follow-up consultation.
We also reported whether the patient had undergone liver transplantation after SBRT; for these patients, the pathological response was evaluated by analyzing the explanted liver. Complete response was defined as complete necrosis of the lesion without any remaining malignant cells.

Endpoint Definition
OS was estimated from the start of treatment until death from any cause. Patients alive at the date of the last news event were censored at this date. RFS was estimated from the start of treatment until the date of the first relapse (local, hepatic or distant) or death from any cause. Patients alive at the date without relapse of the last news were censored at this date. Local control was estimated from the start of the treatment until the first local relapse. Other types of relapse and death were considered competitive events.
Toxicities were graded using the Common Terminology Criteria for Adverse Events v5.0. Acute toxicity was defined as adverse events occurring in the first three months following treatment, based on the first follow-up visit; while late toxicity was defined as events occurring after at least three months of follow-up.

Statistical Analysis
The median follow-up was estimated using the reverse Kaplan-Meier method. OS and PFS curves were estimated using the Kaplan-Meier method. To describe the local control rate, we estimated the cumulative incidence of local recurrence using Kalbfleisch and Prentice method. The association between the risk of death and prognostic factors was analyzed using a Cox model after checking the proportional hazard assumption. The association between prognostic factors and acute toxicity of grade ≥2 and late toxicity of grade ≥2 was analyzed using logistic regression models. The software used for the statistical analysis was Stata version 15.0 (StataCorp. 2017. Stata statistical software: Release 15 StataCorp LLC, College Station, TX, USA).
Child Pugh score has been calculated after treatment in 97 patients. Child Pugh score was stable or improved in 53% of the patients. The score increased by 1, 2, and 3 points in 23%, 9%, and 9% of the patients, respectively. Toxicities are described in Table 3. Table 3. Description of observed acute toxicities with grades in columns (N = 318).

Response after Liver Transplant
Eleven patients (3%) benefited from liver transplantation. The median time between SBRT and transplantation was 13 months (range, 1. 8-24.5). Based on the analysis of the explanted liver, six patients had a complete response to the targeted HCC and five had a partial response. The median time to transplantation for these five latter patients with a partial response was 5 months (range 1.9-13.1), whereas the median time to transplantation for patients with a complete response was 15.2 months (range 1.8-24.5). Pathological analysis revealed disease outside the target volume in eight patients.

Discussion
This study supports the fact that SBRT provides excellent local control for HCC, with an acceptable tolerance profile. In the current treatment guidelines, SBRT is not universally accepted as a first-line treatment for HCC, and is even absent in the recommendations of the European Association for the Study of the Liver (EASL) [2], the Japanese Society for Hepatology (JSH) [23] and the Asian Pacific Association for the Study of the Liver (APASL) [24]. However, the American Association for the Study of Liver Disease (AASLD) [3] and the European Society for Medical Oncology (ESMO) [25] guidelines mention SBRT as an option for early-stage HCC.
This study evaluated the efficacy and safety of SBRT for HCC in 318 patients with 375 lesions. To our knowledge, this is the largest retrospective study on this subject. Our cohort is unique as the prevalence of alcohol abuse in our population is considerably higher than in other studies, in which viral hepatitis has been identified as the main cause of cirrhosis. This factor increases the risk of tumor progression as well as deterioration of liver function, thus making our population more precarious. In addition, our retrospective study is the only one to extensively report dosimetry results according to the ICRU 91 recommendations. Our results support the fact that SBRT is a safe and efficient treatment option for unresectable HCC.
These outcomes in the present study were comparable to other local therapies. A meta-analysis found similar local control rates for SBRT and RFA [32], while others found better local control in the SBRT group [33,34]. TACE and SBRT have also been shown to be comparable in terms of LC and OS [35], although other studies found superior LC, PFS, and intrahepatic control in the SBRT group, with similar OS [36,37]. For mediumsized HCCs, SBRT appears to be superior to TACE [38]. For small HCCs with Child A cirrhosis, a matched-pair analysis found no difference in OS and PFS between SBRT and liver resection [39]. However, these findings should be validated in large phase III clinical trials.
In this study, we found that Child scores B or C, higher Barcelona Clinic Liver Cancer stage, the existence of portal thrombosis, PTV volume reported on total hepatic volume ratio, and larger GTV volume was significantly associated with a higher risk of death. The findings of previous studies confirm the reliability of these factors in predicting mortality [15,16,28]. These factors reflect more advanced HCC and underlying liver disease and highlight the importance of patient selection before SBRT. However, we must remember that these factors can also contra-indicate other therapies, or be the reason for their failure; in these cases, SBRT can be the only local treatment option. Other prognostic factors for OS found in the literature include radiation dose [40], albumin levels [17], and performance status [41]. Zhang et al. established a nomogram to predict mortality after SBRT for unresectable HCC, and identified race, T stage, N stage, M stage, and chemotherapy as independent risk factors for survival [42].
The uniqueness of our study lies in the very high proportion of patients with cirrhosis caused, at least partially, by alcohol abuse. However, this characteristic did not appear as a statistically significant prognostic factor of global survival or toxicity, whether or not alcohol consumption was pursued.
Given the small number of in-field recurrences in our study, we did not analyze the prognostic factors for local control. The main factors which have been found to be associated with local control were radiation dose and tumor size [28,41].
SBRT is a safe treatment option. In our cohort, no deaths were attributed to treatment toxicity, and we did not have any treatment interruption due to poor tolerance. Two of our patients did not complete the full treatment. One patient suffered from myocardial infarction after two fractions and died 12 days later, while the other patient presented with confusion and a suspicion of hepatic encephalopathy after one fraction and died three weeks later.
Most patients with acute toxicities present with mild symptoms. Seventeen percent of patients presented with grade ≥ 2 acute toxicities, none higher than grade 3, which is similar to what was found in previous studies [14,15,43]. As in the present study, the most commonly reported acute toxicities were fatigue, nausea, and hepatic pain.
We identified high GTV volume and high PTV volume reported on total hepatic volume ratio as prognostic factors of grade > 1 toxicity. We were not able to specifically study risk factors for each type of toxicity given the scarcity of events and the low reliability of retrospective data collection for some events such as fatigue.
Late toxicities were mainly fatigue and ascites. However, both can also be explained due to the natural progression of HCC and underlying cirrhosis, and identifying the contribution of SBRT in the occurrence of these symptoms remains a challenge. Moreover, the assessment of late toxicity is limited by the retrospective nature of the present study.
The main cause of failure after treatment was an out-field hepatic relapse, which has also been reported in previous studies on SBRT. This is explained by the fact that HCC is a multicentric disease that occurs from an underlying pathologic liver, creating a microenvironment favorable for carcinogenesis.
Future perspectives include the association of systemic treatment with local therapies. The emergence of immunotherapies for other indications has led to the question of their interest in the treatment of HCC. The FDA has approved the use of nivolumab and pembrolizumab for HCC previously treated with sorafenib, based on the result of the Check Mate 040 and KEYNOTE-224 studies [10,11]. In a recent phase III, randomized, doubleblind trial, pembrolizumab was compared to placebo and tended to show an improvement in OS and PFS, but the results were not statistically significant [44]. The combination of checkpoint inhibitors with SBRT appears to be an interesting approach, with a rationale based on the antitumor immune responses of these therapies [45][46][47]. A case report of five patients treated with SBRT followed by anti-PD1 antibodies has shown promising results, with 1-year local control and OS of 100% [48]. Further investigations are needed, and trials are currently underway to explore this perspective [49].
Despite advances in local and systemic therapies, the best treatment for HCC remains liver transplantation. With graft shortage, the waiting time can be long, which is why local treatments can be indicated to temporarily prevent tumor progression. SBRT can also be proposed for downstaging to meet the Milan criteria [50]. In our cohort, 11 patients benefited from liver transplantation, and all had a partial or complete response of the target on the pathological analysis of the explanted liver. If we look at the five patients with partial response, their time to transplantation was inferior to the median of the whole cohort. It is possible that the surgical intervention was conducted before the lesion had reached complete necrosis. Out-field HCC was found in eight patients, which is consistent with the above-mentioned fact that HCC is a multicentric disease. Previous studies have found similar results with good local control rates and favorable toxicity profiles [14,51]. In a series of 38 patients treated with SBRT, the histologic response rate (complete plus partial response) was 68% [52]. In a prospective study, patients were enrolled for SBRT under a standardized protocol and compared with a retrospective cohort of patients who underwent TACE or High Intensity Focused Ultrasound (HIFU). One hundred fifty patients were evaluated (SBRT, n = 40; TACE, n = 59; HIFU, n = 51). The tumor control rate at 1 year was significantly higher after SBRT compared with TACE and HIFU (92.3%, 43.5%, and 33.3%, respectively; p = 0.02). With competing risk analysis, the cumulative incidence of dropout at 1 and 3 years was lower after SBRT compared with TACE and HIFU too. The authors concluded that SBRT was safe, reduced the risk of waitlist dropout, and should be used as an alternative to conventional bridging therapies [53]. This strategy has also shown its safety and efficacy in patients with more advanced Child-Pugh B or C cirrhosis [54,55].
The main limitations of this study are those inherent to retrospective studies, mainly incomplete data. Large prospective, phase III trials would be needed to confirm the evidence pointed out in our study and other previous ones.

Conclusions
SBRT is safe and efficient for the treatment of inoperable hepatocellular carcinoma or as a bridge to transplantation. Efficacy is comparable to that of other local therapies, thus, its indication must not only be evoked after the failure of these treatments, but should be considered at any stage in the management of HCC. The pattern of failure is proof of a multicentric disease, and further studies are needed to show the benefits of concomitant systemic therapies.