Abstract
Background/Aim: To retrospectively analyze the relationship between local control (LC) and the minimum dose covering 95% of the planning target volume (PTV D95%) in patients with hepatocellular carcinoma (HCC) treated with stereotactic body radiotherapy (SBRT).
Patients and Methods: Between May 2020 and March 2024, 72 HCC tumors in 59 patients were evaluated at Yokohama City University Medical Center. The prescribed dose ranged from 30-50 Gy in five fractions in accordance with the RTOG 1112 trial. However, PTV D95% was used for evaluation because of dose reductions near organs at risk.
Results: The median follow-up period was 20 months. During this period, 13 patients died, and nine local recurrences were observed. The median PTV D95% was 38.9 Gy (range=10.6-48.5 Gy). The 1- and 2-year LC rates were 93.7% and 82.3%, respectively. Both univariate and multivariate analyses identified PTV D95% as a significant prognostic factor for LC [univariate: hazard ratio (HR)=0.882, 95% confidence interval (CI)=0.797-0.977, p=0.016; multivariate: HR=0.891, 95%CI=0.891-0.989, p=0.031]. Notably, no local recurrences were observed in patients with PTV D95% ≥40 Gy. Receiver operating characteristic analysis determined that 38 Gy was the optimal PTV D95% cut-off to prevent local recurrence, based on Youden’s index.
Conclusion: PTV D95% was a significant prognostic factor for LC in patients with HCC treated with five-fraction SBRT. These findings suggest that PTV D95% ≥40 Gy should be considered as a cut-off dose to prevent recurrence.
Introduction
Primary liver cancer is the sixth most commonly diagnosed cancer, and it was the third leading cause of cancer-related death worldwide in 2020, with approximately 906,000 new cases and 830,000 deaths (1). Hepatocellular carcinoma (HCC) accounts for 75%-85% of primary liver cancers, and the main risk factor is cirrhosis caused by chronic viral hepatitis, alcohol, and other causes (1, 2).
According to current guidelines, surgical resection, liver transplantation, or radiofrequency ablation (RFA) is recommended for early-stage HCC (3, 4). However, fewer than 20% of patients are eligible for surgical resection or liver transplantation because of factors such as poor performance status or a shortage of liver donors (5). The efficacy of RFA is considered to be comparable to that of surgical resection for small HCC (6). However, the efficacy of RFA is lower for tumors located beneath the diaphragm, adjacent to major blood vessels, or near the gallbladder because these locations are associated with higher procedural risks and lower response rates compared with other sites (7).
Stereotactic body radiotherapy (SBRT) is a less invasive and alternative treatment to surgery. Additionally, SBRT is less affected by tumor location compared with RFA (8). Numerous prospective trials and large-scale retrospective studies have demonstrated the ability of SBRT to achieve high rates of local control (LC) and overall survival (OS) for patients with HCC (9-19). However, the doses and treatment planning strategies vary greatly, and no standardized protocol regimens have been established (9, 14, 20-22). Notably, even within individual studies, multiple fractionation regimens were often used; therefore, comparisons using biologically effective doses are common (14-16, 21, 23). Additionally, few previous studies have used a uniform five-fraction regimen for all cases (24). In this study, we exclusively evaluated patients treated with a five-fraction SBRT protocol. The aim of this study was to evaluate the impact of radiation dose on OS and LC in patients with HCC.
Patients and Methods
Patient selection. Data were collected from the medical records of 63 patients (76 HCC tumors) who underwent SBRT between May 2020 and March 2024 at Yokohama City University Medical Center. This retrospective study was approved by the Yokohama City University Clinical Research Review Committee (approval number: F241200008).
Patients were managed in multidisciplinary settings with all possible treatment options available. The inclusion criteria were: 1) HCC, pathologically proven or clinically diagnosed on the basis of typical enhancement patterns on either dynamic computed tomography (CT), dynamic magnetic resonance imaging (MRI), or perflubutane-enhanced ultrasonography; 2) difficulty or refusal to undergo surgery or percutaneous ablative therapies; and 3) Child-Pugh score ≤7.
Radiation treatment technique. Patients were immobilized using a BlueBAG™ (Elekta, Stockholm, Sweden), with abdominal compression and/or exhale breath holding. Planning CT images were obtained with dynamic CT. Treatment plans were created using Pinnacle3 version 14.0 (Philips Healthcare, Amsterdam, The Netherlands).
The gross tumor volume (GTV) was delineated by referring to the dynamic CT images. When a patient was intolerant to CT contrast media or a tumor was better visualized on MRI, a dynamic MR image was fused to the planning CT image. The clinical target volume was equated to the GTV, and an internal target volume was added to account for respiratory motion observed under X-ray fluoroscopy. For the planning target volume (PTV), a 5-mm margin was applied around the internal target volume as a set-up margin. All treatments were performed using 6-MV flattening filter-free photon beams and volumetric modulated arc therapy with auto-planning (25). Five fractions were delivered on alternate days, with dose constraints in accordance with the RTOG 1112 trial (26) (Table I). Prescription doses ranged from 30 to 50 Gy in five fractions, also in accordance with the RTOG 1112 trial. However, the minimum dose covering 95% of the PTV (PTV D95%) was used for evaluation because of dose reductions near organs at risk.
Dose volume constraints for organs at risk.
Statistical analysis. Clinical and imaging evaluations were performed approximately every 3 months after SBRT. Dynamic CT or MRI of the liver was performed every 3 months unless contraindicated. Local recurrence was defined as an increase in the size of the target lesion or tumor recurrence within the PTV (Response Evaluation Criteria in Solid Tumors 1.1) (27). OS was defined as the time from the start of the initial SBRT to the date of death or censoring. Time to local recurrence was defined as the time from the start of SBRT to the date of local recurrence or censoring. OS and time to local recurrence were estimated using the Kaplan-Meier Method. Intergroup comparisons of OS and LC by dose levels were performed using the log-rank test. Univariate analysis using Cox proportional hazards models was performed for all potential risk factors for survival and recurrence. Multivariate analysis adjusted for tumor size, a known factor affecting LC, was then performed. Receiver operating characteristic (ROC) curves were used to determine the optimal PTV D95% cut-off value to predict local recurrence, using Youden’s index. Data were analyzed using SPSS version 29.0 (IBM Corp., Armonk, NY, USA), and p<0.05 was considered statistically significant. Toxicity was evaluated using the Common Terminology Criteria for Adverse Events version 5.0 (28). Acute toxicities were defined as adverse events occurring within 3 months after SBRT, and late toxicities were defined as adverse events occurring thereafter. Liver function deterioration was defined as Grade ≥3 elevation in transaminases or an increase of ≥2 points in the Child-Pugh score within 3 months after SBRT.
Results
Between May 2020 and December 2024, data for 59 patients with 72 HCC tumors were retrospectively analyzed, after excluding one patient who discontinued treatment because of cerebral infarction and two patients who were lost to follow-up before imaging evaluation (Table II). The median follow-up time for local recurrence of censored tumors and the OS of censored patients was 20 months (range=3-52 months) and 20 months (range=6-52 months), respectively. The median age was 78.5 years (range=55-93 years), and the median tumor size was 1.9 cm (range=0.5-6.1 cm). Among the 59 included patients, six received SBRT to the liver at our institution more than once (twice, four patients; thrice, two patients).
Tumor characteristics (n=72).
During the follow-up period, 13 patients died, and nine local recurrences were observed. Eight patients died of HCC; the remaining five died of other causes: myocardial infarction, aortic aneurysm rupture, lung cancer, esophageal cancer, and an accident (one each). Four patients received SBRT for two tumors simultaneously. Twelve tumors (18%) were local recurrences or residual tumors following treatments, such as RFA, transcatheter arterial chemoembolization, or surgery. The median PTV D95% was 38.9 Gy (range=10.6-48.5). The 1- and 2-year OS rates were 96.5% and 76.5%, respectively, while the 1- and 2-year LC rates were 93.7% and 82.3%, respectively (Figure 1).
Kaplan-Meier estimates of overall survival (OS) and local control (LC). (A) OS from the time of the first stereotactic body radiotherapy (SBRT); (B) LC from the time of SBRT.
Univariate Cox proportional hazards analysis identified comparing patients negative for both hepatitis B virus (HBV) and hepatitis C virus (HCV) with those positive for HBV and/or HCV etiologies for HCC as baseline factors, showed that patients in the latter group had a more favorable prognosis for OS [hazard ratio (HR)=0.882, 95% confidence interval (CI)=0.797-0.977; p=0.049] (Table III). For LC, both univariate and multivariate analyses showed that PTV D95% was a significant prognostic factor (univariate: HR=0.882, 95%CI=0.797-0.977, p=0.016; multivariate: HR=0.891, 95%CI=0.802-0.989, p=0.031) (Table IV). The optimal dose for PTV D95% to prevent local recurrence was 38 Gy on the basis of Youden’s index in the ROC analysis (sensitivity=100%, specificity=63.5%, area under the curve=0.73) (Figure 2).
Univariate analysis of the prognostic factors for overall survival (n=59 patients).
Univariate and multivariate analyses of the prognostic factors for local control (n=72 tumors).
Receiver operating characteristic (ROC) curve for the prediction of local recurrence based on the minimum dose covering 95% of the planning target volume. The optimal cut-off dose was 38 Gy based on Youden’s index, yielding a sensitivity of 100%, specificity of 63.5%, and an area under the curve (AUC) of 0.73.
Figure 3 illustrates the number of tumors categorized by PTV D95%; no local recurrences were observed with PTV D95% ≥40 Gy. Patients were stratified into two groups by the PTV D95% value as ≥40 Gy versus <40 Gy. LC was significantly better in the ≥40 Gy group (p=0.009; log-rank test), while OS showed no significant difference between the groups (p=0.844; log-rank test) (Figure 4).
Number of tumors categorized by the minimum dose covering 95% of the planning target volume (PTV D95%). Tumors with local recurrence are shown in orange and those without local recurrence are shown in blue. No local recurrences were observed with PTV D95% ≥40 Gy.
Kaplan-Meier estimates of overall survival (OS) and local control (LC) for the minimum dose covering 95% of the planning target volume (PTV D95%) ≥40 Gy (solid line) versus <40 Gy (dashed line) doses. (A) OS by a PTV D95% dose of ≥40 Gy versus <40 Gy. No significant difference is observed between the groups (p=0.844; log-rank test); (B) LC by a PTV D95% dose of ≥40 Gy versus <40 Gy. PTV D95% ≥40 Gy has better LC compared with PTV D95% <40 Gy (p=0.009; log-rank test).
One (1%) patient developed Grade ≥3 acute toxicity (Grade 4 cholangitis); no Grade ≥ 3 late toxicities were observed. Liver function deterioration within three months of SBRT was noted in two (3%) patients. In one of the two patients, the Child-Pugh score increased from 6 to 8, while the score in the other patient increased from 5 to 9. However, in both cases, the Child-Pugh score subsequently returned to baseline levels.
Discussion
Historically, whole-liver irradiation posed a significant risk of liver toxicity, particularly radiation-induced liver disease (29). However, advancements in radiation delivery techniques have enabled partial liver irradiation, with SBRT providing high-precision, high-dose delivery in a limited number of fractions; thereby, reducing toxicity (29). Despite these advancements, patients with a Child-Pugh score ≥8 remain at considerable risk for liver toxicity, and the American Society for Radiation Oncology guidelines recommend SBRT for HCC patients with a Child-Pugh score ≤7 (30).
Repeated SBRT for HCC is safe when the target lesion differs from previously irradiated sites (31, 32). Furthermore, SBRT has comparable OS and superior LC rates compared with RFA (8, 33, 34). For small lesions, SBRT results in LC rates equivalent to those with surgical resection (35), making SBRT a highly effective treatment option for patients ineligible for RFA or surgery.
Several meta-analyses have identified tumor size as a factor influencing LC (14-16). However, in the present study, tumor size was not significantly associated with LC. Furthermore, even after adjusting for tumor size in the multivariate analysis, only PTV D95% remained a significant predictor of LC in both the univariate and multivariate analyses.
Table V summarizes the results of previous studies that evaluated five-fraction SBRT for HCC. Kibe et al. reported that tumor size and salvage treatment were prognostic factors for LC (36). However, no significant differences were observed for these factors in the present study. Three previous studies reported no significant differences in LC rates between five-fraction doses of 35 Gy and 40 Gy (9, 36, 37); however, the actual PTV D95% doses were not reported. Similarly, Chen et al. reported improved LC with prescribed doses of ≥45 Gy (13); however, the study did not consider a cut-off dose or PTV D95%.
Patient and treatment characteristics in studies of five-fraction stereotactic body radiotherapy in hepatocellular carcinoma.
In this study, the actual measured PTV D95% values were analyzed as continuous variables. We also assessed the SBRT dose cut-off. Moreover, the sample size of 72 tumors in this study exceeded the sample sizes in previous reports. Therefore, this study is currently the largest, to our knowledge, to evaluate optimal dose cut-offs for LC in HCC treated with SBRT delivered exclusively in five fractions.
Chen et al. reported poorer OS in patients with versus without HCV infection (13). In our study, there was no significant difference in OS in patients with versus without HCV infection, only. However, patients with HBV and/or HCV infection had better OS than those without either infection.
Grade ≥3 elevation in transaminases and a ≥2-point increase in the Child-Pugh score are often used as indicators of hepatotoxicity, with an incidence of 10%-30% within 1-3 months after SBRT (25, 38, 39). In this study, these changes were observed in two (3%) patients within three months of SBRT. One patient was receiving anticoagulation therapy, and in the other, the changes were presumed to have resulted from acute cholangitis. The latter patient also developed Grade 4 septic shock, thrombocytopenia, and Grade 3 leukopenia. In both patients, increases in the Child-Pugh score were transient and resolved over time.
In SBRT, dose reductions to PTV D95% are sometimes necessary to minimize GI toxicity. In this study, reduced doses were associated with higher rates of local recurrence. However, no Grade ≥3 GI toxicities were observed in our study, whereas previous reports indicated an incidence range of 5%-10% (11, 19, 40).
In this study, no local recurrences were observed at PTV D95% ≥40 Gy. The fact that this dose did not significantly deviate from the cut-off value (38 Gy) determined by the ROC analysis, and the low rate of toxic effects suggest that PTV D95% 40 Gy should be considered the prescribed dose for SBRT in HCC. Ismael et al. reported that long-term LC was safely achieved by increasing the prescribed PTV D95% dose while reducing doses to the GI tract through the insertion of a spacer (41). This method should be considered when the GI tract is in close proximity to the tumor(s).
This study has several limitations. First, the median patient age was 78.5 years, which is relatively high, and some patients experienced difficulty attending follow-up visits for reasons unrelated to HCC. Second, this was a retrospective, single-institution study with a small sample size, and caution is warranted regarding the generalizability of our findings. Third, the follow-up period was short; therefore, continued monitoring is necessary to evaluate long-term outcomes.
Conclusion
This study demonstrated that PTV D95% was a prognostic factor for LC in both the univariate and multivariate analyses of patients with HCC treated with five-fraction SBRT. Notably, no local recurrences were observed for tumors that received doses of ≥40 Gy, suggesting that a cut-off dose of 40 Gy should be considered. Regarding OS, HBV and/or HCV infection background factors were identified as favorable prognostic factors. As a single-center retrospective study with a limited patient cohort and short follow-up period, these findings should be interpreted with caution when considering generalizability. Long-term outcomes need to be validated in future studies.
Acknowledgements
The Authors thank Jane Charbonneau, DVM, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.
Footnotes
Authors’ Contributions
Conceptualization, Data curation, Formal analysis, and Roles/Writing – original draft: Yuta Nishikawa and Ichiro Ogino. Supervision: Yuta Nishikawa and Ichiro Ogino. All Authors read and approved the final manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest associated with this manuscript.
Funding
None.
Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.
- Received May 28, 2025.
- Revision received June 19, 2025.
- Accepted June 20, 2025.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).










