Introduction

Hepatocellular carcinoma (HCC) has become one of the leading causes of death in cirrhotic patients [1]. Tumour stage, liver function and patient general condition are the main prognostic factors in HCC. Among the available classifications, the Barcelona Clinic Liver Cancer (BCLC) classification is endorsed by both US and European associations for the study of liver diseases and oncology [2]. Surgery is the standard of care for these patients, but only 10–30% of patients are eligible for this treatment. When surgery is not possible, therapeutic options include percutaneous alcohol injection, radiofrequency ablation (RFA) [3] and transarterial chemoembolization (TACE) [4]. When these options are not feasible, recommendations include best supportive care, sorafenib [5] or radiotherapy [6], performed alone or in combination [7].

Three local modalities are actually recognized as curative treatment options for patients with early-stage hepatocellular carcinoma: liver transplant, surgical resection, and percutanuous radiofrequency ablation. Palliative treatments include sorafenib, radiotherapy and TACE [8]. Conformal radiotherapy and stereotactic body radiotherapy (SBRT) are not included as curative options in the European Association for the Study of the Liver practical guidelines [9] and are consequently rarely discussed during tumor boards.

Stereotactic body radiotherapy (SBRT) delivers high radiation doses with precision and has been used to treat patients without other therapeutic options in clinical situations as diverse as inoperable recurrent head and neck squamous-cell carcinoma [10] or pelvic recurrences in a previously irradiated area [11]. Liver SBRT, pioneered almost 20 years ago by Blomgren et al. [12], can be used to treat secondary [13] or primary tumors and is an option for local treatment of HCC patients ineligible for surgery, TACE, chemotherapy, or RFA. This treatment technique has shown enticing rates of local control and low toxicity [14]. Ultimately, SBRT could be an alternative to existing local treatments. CyberKnife^® (Accuray Incorporated, Sunnyvale, California, U.S.A.) is a robotic image-guided system that delivers hypofractionated SBRT with intrafraction motion detection and correction. Other possible techniques include Novalis Tx^® (Brainlab AG, Feldkirchen, Germany) and TrueBeam^® (Varian, Palo Alto, California, U.S.A.).

We already published results in patients treated for primary and secondary liver lesions [15]. This study reports the results of SBRT in an updated and expanded population of patients treated for hepatocellular carcinoma only. The primary objective was to determine the prognostic factors of tumor local control and overall survival for these patients, in order to better distinguish patients who are most likely to benefit from this treatment option from those that are not.

Materials and Methods

Seventy-five HCC patients (12 women and 63 men) were treated with SBRT from July 2007 to November 2011. The patients included in the study had HCC, an Eastern Cooperative Oncology Group (ECOG) performance score of less or equal to 2, pre-treatment Child-Turcotte-Pugh (CTP) (Table 1) scores A5–B8, and ineligibility for surgical resection, TACE, RFA, or sorafenib. Main reasons for ineligibility were inoperable patients and size or location of the tumor incompatible with RFA or TACE procedures. Diagnosis was established based on biopsy or according to the Barcelona criteria [16]. Patients could have no more than three lesions in order to qualify. Any previous treatments could include TACE, chemotherapy, surgery, or RFA. Patients did not have any distant metastases. Median follow-up for surviving patients was 10.0 months (range, 3–49 months).

Cases were discussed in a multi-disciplinary HCC board consisting of a hepatologist, a hepatic surgeon, a radiation oncologist, a medical oncologist, and a radiologist. This retrospective, single-institution, cross-sectional study was approved by our Institutional Committee on Human Research.

Results

Tumor local control rate

Ninety-six targets were treated and individually evaluated for 75 patients. The actuarial 1- and 2-year local control rate was 89.8%. Seven targets (7.3%) in six patients (8%) locally progressed after treatment. Eighteen hepatic recurrences (24%) at a distance from the target were observed. Six patients had distant metastasis after treatment. Kaplan-Meier curve for tumor local control is presented in Figure 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Kaplan Meier curves for local control and overall survival.

https://doi.org/10.1371/journal.pone.0077472.g001

We performed a univariate analysis to identify prognostic factors affecting local control. Patients with a Cancer of the Liver Italian Program (CLIP) score higher than 2 were more likely to relapse (HR=10.127; CI95=2.182, 47.12; p=0.0031). A lower risk of relapse was observed when a higher radiation dose per fraction was administered (HR=0.6473; CI95=0.492, 0.851; p=0.0018) and a higher total dose was delivered (HR=0.8634; CI95=0.789, 0.945; p<0.0001). In multivariate analysis, a higher total dose remained significantly associated with a lower risk of relapse (HR=0.866; CI95=0.753, 0.996; p=0.0441). AFP was also associated with higher risk of local failure (HR=1.0011; CI95=1.001, 1.002; p<0.0001). Results are presented in Table 4.

Factor	Hazard Ratio (95% CI)	p value
Univariate analysis
AFP	1.0011 (1.001,1.002)	<.0001
Child-Pugh score	0.6664 (0.061,7.303)	0.7397
MELD score	1.0345 (0.907,1.180)	0.6135
OKUDA score	1.6909 (0.178,16.018)	0.6471
CLIP score = 1	0.8346 (0.082,8.472)	0.8785
CLIP score = 2	10.1277 (2.182,47.012)	0.0031
Previous treatment	1.2377 (0.263,5.835)	0.7875
Previous chemo-embolization	1.3802 (0.237,8.049)	0.7202
Previous chemotherapy	2.6016 (0.398,16.999)	0.3181
Number of treated lesions	0.7643 (0.103,5.652)	0.7923
Sum of GTV volumes	1.0064 ( 1.004,1.009)	<0.0001
Total dose	0.8634 (0.789,0.945)	<0.0001
Dose / fraction	0.6473 (0.492,0.851)	0.0018
Multivariate analysis
AFP	1.001 (1.000,1.002)	0.0063
Total dose	0.866 (0.753,0.996)	0.0441

Table 4. Prognostic factors for local control (n=96 lesions).

CSV

Download CSV

Discussion

Stereotactic body radiotherapy (SBRT) is not included as a curative option in the European Association for the Study of the Liver practical guidelines [9]. Of note, concerns have been raised about the possibility of radiation-induced liver disease for patients treated with external-beam radiotherapy (EBRT), which encumbered the inclusion of EBRT in the therapeutic strategies for cirrhotic patients. This is despite almost 20 years have now passed since Blomgren et al. have pioneered SBRT for liver malignancies [12].

Since then, most studies published about SBRT for liver tumors have contained both liver metastases and hepatocellular carcinoma. Articles about SBRT for HCC alone are scarce. Most of these studies are retrospective, like ours, or with a very limited number of patients. We present here the largest study in terms of both patients (75) and the number of lesions (96) treated.

In Blomgren’s articles, 42 tumors in 31 patients (eight with HCC) were treated. Dose per fraction ranged from 7.7 to 30 Gy (mean, 14.2 Gy). Treatments consisted of one to four fractions delivering a total mean dose of 8 to 66 Gy to the PTV, respectively. The local control rate in the study was 80%. Since then, SBRT for HCC has been explored as a local salvage treatment after incomplete transarterial TRACE [19], as a bridge to transplantation [20] or for recurrence treatment [21].

Four prospective studies have been published regarding patients with HCC treated with SBRT. In the study by Mendez-Romero et al [22], 25 patients with liver tumors, including eight patients with HCC, were treated. Two fractionation regimens were prescribed depending on the tumor diameter, i.e. , patients with tumors smaller than 4 cm received 37.5 Gy in three fractions, while patients with tumors larger than 4 cm had 25 Gy in five fractions. Patients who presented with local failure were all in this latter group. Due to the occurrence of these failures, the dose was increased to 30 Gy in three fractions for patients with tumors larger than 4 cm. One-year local control was 75% and 1-year survival rate 48%. One patient with CTP B cirrhosis experienced grade 5 RILD. Two years later, Tse et al published their own prospective study, a phase 1 trial with 41 patients among whom 31 had HCC [23]. Radiation dose was adapted according to predicted liver toxicity, based on a normal tissue complication model [24]. Median dose was 36 Gy (range, 24–54 Gy) in six fractions. Median tumor volume was 173 cm³ (range, 9–1913 cm³). No patients were diagnosed with RILD, although 26% had grade 3 liver enzymes elevation. One-year local control rate was 65% and median survival was 11.7 months. In 2010, Cardenes published a phase I trial of SBRT for HCC with 17 patients (25 lesions). No dose-limiting toxicity was reported following an escalated 48 Gy in three fractions introduced in one to two fractions per week to patients with Child-Turcotte-Pugh (CTP) A cirrhosis. In line with previous studies, patients with CTP B cirrhosis experienced higher toxicity: four out of the five CTP B patients had grade 3 or higher liver toxicity, with an overall survival less than 6 months.

Three retrospective studies have been published with HCC-only patients treated with SBRT. Forty-two patients were treated with 30 to 39 Gy in three fractions in the Kwon et al study [25]. Median follow-up was 29 months. At 1 and 3 years, in-field-progression-free survival was 72% and 68%, respectively. Overall 1- and 3-year survival rates were 92.9% and 58.6%, respectively. Patients experienced grades 1–2 toxicity but one patient died with extrahepatic metastasis and radiation-induced hepatic failure. The same year, Seo et al published a study including 38 patients with HCC treated with 33 to 57 Gy in three to four fractions [26]. Local control rate and overall survival were 79% and 68%, respectively. Dose was identified as an independent prognostic factor for survival, i.e., patients treated with 42 Gy in three fractions had an 81% 2-year survival rate, while patients who received a lower dose had a 25% rate. It is, nevertheless, difficult to draw strong conclusions considering the limited number of patients in this study. Six patients experienced a decline in liver function and one patient had a grade 3 musculo-skeletal toxicity. More recently, Andolino et al have reported on 60 patients treated with SBRT for liver-confined HCC [27]. Like in the Mendez-Romero study, dose was adapted to the CTP score of the patients, i.e., the median number of fractions, dose per fraction, and total dose were three, 14 Gy, and 44 Gy, respectively, for those with class A cirrhosis and five, 8 Gy, and 40 Gy, respectively, for those with class B. The 2-year local control and overall survival were 90% and 67%, respectively. Regarding toxicity, 13% of the patients experienced an increase greater than one grade of hematologic and hepatologic dysfunction.

More recently, two studies were reported regarding patients with active HCC unsuitable for standard locoregional therapies treated with SBRT [28]. 102 patients were evaluable : local control at 1 year was 87% (95% CI, 78% to 93%). Like in our study, SBRT dose was associated with local control on univariate analysis. Median overall survival was 17.0 months (95% CI, 10.4 to 21.3 months). This study and our study’s overall survival rates compare favorably with best supportive care and even with sorafenib [29,30] the only other potentially available therapy for these patients.

In these seven studies, SBRT showed high local control rates in the range of 70–90% at 1 year. Our results, regarding local control is similar to what has already been reported in the previous studies, suggesting that SBRT is an effective treatment, even for those previously treated for HCC. We showed that total dose was a prognostic factor for local control in multivariate analysis, further supporting the importance of this factor for the success of the treatment. In the present study, the 2-year survival rate was 50.4%, which is slightly lower than that in the other studies. This could be attributed to the fact that only ineligible patients for surgery, or those with relapsed HCC after a failed prior treatment, were included, indicating that these patients had less favorable prognoses to begin with. This higher mortality cannot be attributed to treatment toxicity, since it was found to be very low. In view of the large proportion of multifocal tumors in our study population and the difficult tumor sites for operation, this slightly lower survival rate is not surprising.

CTP B is recognized as a prognostic factor for overall survival and our analysis confirmed that patients with a higher score had lower overall survival. This could also hint at the fact that patients with advanced cirrhosis are more likely to die from liver failure. The toxicity of SBRT consists mainly of liver dysfunction. Other toxicities (pain, fatigue, nausea, and vomiting) were not dose-limiting, which was also the case in our study. A particular attention should be paid to patients with CTP B score since they are more likely to experience liver failure after SBRT. Our study did not confirm that result, even though we did find that patients with larger healthy liver volumes were less like to decompensate their cirrhosis.

Conclusion

SBRT is an effective and safe treatement. In a population of patients without other curative local treatment options, SBRT shows sustained local control, associated with survival rates higher than historical controls, with a low risk of serious toxicity. Further randomized trials are required to compare his efficacy and safety to RFA and TACE.

Supporting Information

Acknowledgments

The authors acknowledge A. Chatelet and S. Marchant for their valuable contribution during the conduct of this study.