Managing hepatocellular carcinoma across the stages: efficacy and outcomes of stereotactic body radiotherapy

Liver cancer, ranking as the 6th most prevalent malignancy globally and standing as the third leading cause of cancer-related death, is a challenging malignancy and is surpassed only by lung and colorectal cancer [1]. Hepatocellular carcinoma (HCC) is the most common primary liver cancer, accounting for > 80% of cases [2]. Traditionally, surgical resection and orthotopic liver transplantation have been considered the cornerstone of curative interventions for HCC, delivering survival rates exceeding 5 years in cases of early stage HCC [3]. However, their applicability is restricted by stringent criteria. A common predicament is the unsuitability of certain patients for surgical intervention, primarily due to poor hepatic reserve, concurrent comorbidities, or presenting in an advanced stage [4].

In response to these limitations, locoregional therapies such as radiofrequency ablation (RFA), microwave ablation (MWA), or transarterial arterial chemoembolization (TACE) have been employed as minimally invasive alternatives in early and intermediate stages [3].

Meanwhile, the recent advancement in radiation therapy technology has enabled the precise delivery of lethal high radiation doses to localized tumors in one or a few fractions [5]. Consequently, SBRT has gained widespread acceptance as a treatment modality for various primary tumors, including those in the lungs, prostate, liver, and pancreas, or as an ablative tool for various metastatic sites in oligometastatic disease [6‐10]. An expanding body of evidence supports the safe and effective application of SBRT in diverse stages in HCC, including early, intermediate, and advanced stages or in the setting of planned orthoptic liver transplantation as a bridging therapy, yielding impressive high control rates of 70–100% [8, 11‐16].

Nevertheless, ongoing debates regarding the use of SBRT in HCC warrant further exploration. These include determining the optimal radiation dose that can achieve effective local control while minimizing additional radiation exposure to the cirrhotic liver tissue. In addition, there is a need to identify specific patient subsets that may derive substantial benefit from the high local control rates offered by SBRT and those subsets at higher risk for hepatic and extrahepatic progression who might need or benefit more from combined therapeutic approaches.

In the current retrospective study, we aimed to report on the single institution experience of stereotactic body radiotherapy in HCC, emphasizing local control outcomes as well as the overall and progression-free survivals. Additionally, we aim to determine the optimal radiation dose for HCC that achieves a high control rate while maintaining an acceptably low toxicity profile for SBRT. Finally, we also aimed to investigate the prognostic factors that influence survival to enable us to define suitable candidates for combined therapies in the future.

Following approval from the local ethics committee (RWTH Aachen University, Faculty of Medicine, EK 23–264), we conducted a retrospective analysis of patients diagnosed with HCC who had undergone SBRT as part of their disease management between January 2013 and June 2023. Each patient’s case was discussed in a multidisciplinary tumor board comprising experts from various disciplines, including hepatology, medical oncology, hepatobiliary and transplant surgery, interventional and diagnostic radiology, and radiation oncology. Nonsurgical local treatments, including SBRT, RFA, MWA, TACE, and selective internal radiation therapy (SIRT), were offered to patients deemed ineligible for surgical resection, still suitable for local therapy. The choice of SBRT over other treatment options was based on specific factors such as tumor location, including those near the liver dome or blood vessels, lesion size exceeding 3 cm, contraindications for anesthesia, and the need for adjuvant or salvage therapy following TACE.

Only patients who received liver-directed SBRT, as defined by the German Society for Radiation Oncology [5], were included in the analysis. Exclusion criteria were (1) radiation was delivered in symptom palliation, (2) SBRT to lesion outside the liver, (3) none of the survival parameters could be retrieved, and (4) active second malignancy at the time of SBRT application.

A cohort of 54 patients diagnosed with HCC who underwent liver-directed SBRT as part of their disease management was initially identified. However, 2 patients were subsequently excluded from the analysis: one patient was undergoing active treatment for non-small cell lung cancer at the time of SBRT delivery, and no retrievable survival data were accessible for the second patient. Therefore, our final analysis encompassed a total of 52 patients. Among them, comprehensive survival and imaging data from 49 patients, including a total of 62 treated lesions, were available for this study. For the remaining 3 patients, only overall survival data could be analyzed.

Table 1 provides an overview of patient and disease characteristics. The median follow-up period for the entire cohort was 16.7 months. Sixteen patients had early-stage HCC, and 36 had intermediate and advanced stages (Barcelona clinic liver cancer “BCLC” stage B and C), with 25 patients with BCLC B and C having received prior therapy before SBRT. Sixty-two lesions were evaluable for the analysis with a median volume of 8.15 (0.5–539) cm³. The median of the prescribed physical dose was 40 Gy in a median 5 fractions (EQD2_α/β10 60 Gy).

Notably, 4 patients received SBRT as bridging for a liver transplant and were subsequently transplanted at 1.5, 2.4, 6.2, and 10 months after SBRT. Three patients had a pathological complete response (pCR) at the time of the transplant (2.4, 6.2, and 10 months after SBRT), and 1 patient had a pathological partial response (viable cells in 20% of the lesion 1.5 months after SBRT), the survival of the 4 patients was 9, 10.5, 12.3 and 17.7 months at the time of the analysis (supplementary figure 1).

Previously, the treatment results of HCC with radiation were unsatisfactory due to the limited accuracy in the application of the radiation dose to the target volume, which was compensated with the irradiation of a large volume of healthy liver parenchyma, resulting in high toxicity and a lower response rate [21]. Lately, stereotactic radiotherapy (SRT), initially developed for treating cranial tumors, emerged as one of the curative methods for early HCC, delivering ablative radiation doses to tumors while sparing the adjacent healthy liver tissues and yielding a high local control rate of around 98–90% [22‐24]. Comparative studies with thermal ablation showed similar efficacy for small lesions and better results for tumors larger than 3 cm in diameter or the subphrenic area of the liver [15, 25, 26]. Nonetheless, several retrospective and prospective studies showed the superiority of SBRT regarding the local control for HCC compared to TACE [27‐30].

The current analysis provides some insights into managing HCC using SBRT. The assessment of local tumor control is paramount when evaluating the efficacy of any local treatment modality for HCC. In this study, the local tumor control achieved with SBRT was convincingly high, with FFLP rates at 1, 2 and 3 years of 93.4%, 93.4%, and 80.1%, respectively, with local progression in only 4 cases, underscoring the potential of SBRT in achieving durable local control. In the univariate analysis, the tumor volume, or intrahepatic tumors versus PVTT did not differ for the local control. Notably, PD as EQD2_α/β10 ≥ 61 Gy showed a trend for better local control (p-value 0.078), and D98 as EQD2_α/β10 > 61 was associated with a statistically significant improvement of local FFLP (p-value 0.034).

Previously, various studies tried to find a dose–response relationship in HCC; some could not establish one [31], while others claimed an advantage for higher doses over lower ones for the local control [32, 33]. In our analysis, the logistic regression model relating PD to FFLP showed a weak dose–response relationship, with β1 of 0.043 and a nonsignificant p-value “0.35”. This suggests a limited direct correlation between PD in the reported dose range (median 40 Gy in 5 fractions: EQD2_α/β10 : 60 Gy, interquartile range 56–70 Gy) and local tumor control, a finding that aligns with Ohri et al. [31]. The analysis involving D98% presented a more pronounced dose–response curve, with a β1 of 0.06 and a marginally significant p-value “0.085” with TCP values for 95 and 98%, projected at 56.9 and 73.1 Gy, respectively. This enhanced relationship suggests that D98% may serve as a more reliable predictor for local tumor control, potentially providing a more precise dosimetric criterium for the evaluation of the radiation therapy plan in HCC. The discrepancy in the predictive value of PD and D98% could be attributed to the PD parameter not accounting for the spatial distribution of the dose within the PTV, unlike D98%, which is more reflective of the actual dose coverage in the target volume. These considerations are pivotal and may underscore the implication of an individualized dose prescription in radiation therapy for HCC. Based on the current analysis, an EQD2_α/β10 ≥ 61 Gy (equal to biological effective dose “BED_α/β10” 73.3 Gy) proved to be sufficient to reach a high probability of tumor control, but a higher EQD2_α/β10 > 73.1 Gy (BED_α/β10 85 Gy) may not translate into a meaningful advantage for FFLP.

These results should be considered in the light of the assessment of the treatment-related toxicity, which is a critical aspect of any therapeutic intervention. In this study, cRILD was not observed, suggesting the safety of SBRT to spare the uninvolved liver parenchyma. However, changes in CPS were noted; namely, 4 patients (7.7%) experienced meaningful CPS progression ≥ 2 points (ncRILD), and 1 of these patients (1.9%) died of refractory variceal bleeding after CPS progression. Dawson et al. suggested that the dose mean to the liver is the most useful parameter in predicting cRILD in liver SBRT, with cRILD not observed when the dose mean to the liver is below 31 Gy [34]. Nonetheless, ncRILD is mostly underestimated in SBRT of HCC, with some studies reporting its incidence around 20% and suggesting a dose mean to liver below 15 Gy to reduce its incidence [35, 36]. We further generated a logistic regression model to predict ncRILD using the dose mean to the liver. The logistic regression model showed a 10% probability of ncRILD at 12.8 and 12.6 Gy, as EQD2_α/β3 and EQD2_α/β8, respectively. It is important to note that the dose mean to the liver for patients with CPS B and C was further kept below 8 Gy, with few exceptions at the initial phase of our SBRT experience.

Regarding survival outcomes, the median OS for patients with “BCLC A” was not reached at the time of the analysis, while the “BCLC B” and “C” were 29.3 and 7.1 months, respectively, despite the high local control rate for the treated lesions. We investigated the possible tumor characteristics associated with worse PFS and OS. Predictably, advanced CP score (B and C) or performance status were associated with reduced OS. Also, the presence of PVTT or a larger tumor ≥ 37 cm³ strongly predicted shorter PFS and OS. These results may be useful in a number of aspects, such as the proper selection of patients with good performance status who profit from local therapy. Further, the local therapy tool, such as SBRT, alone may not be sufficient for larger tumors or PVTT to achieve intra- and extrahepatic tumor control [37], and a possible combination for these stages (BCLC B or C) with systemic therapy could be advantageous. Indeed, NRG 11112 recently showed that combining SBRT with sorafenib for stage BCLC B and C resulted in a durable response with higher OS and PFS than sorafenib alone [38]. Novel combinations with immunotherapy are currently being investigated with promising results [39].

Also, the utilization of repeated SBRT after hepatic progression was feasible in 5 patients without increased toxicities and aligns with similar previous reports [40].

This study demonstrates the potential of stereotactic body radiotherapy (SBRT) as an effective treatment modality for hepatocellular carcinoma (HCC), achieving promising local control rates with manageable toxicity across different stages of HCC. SBRT with EQD2_α/β10 61–73.1 Gy may be sufficient to achieve durable local control and avoid unnecessary radiation dose exposure to the liver. The dose mean to the liver is the main parameter to predict the Child–Pugh score (CPS) progression after SBRT and should be held below 12.8–12.6 Gy (EQD2_α/β3–8) and may be lower for CPS B and C. Finally, tumors ≥ 37 cm³ and portal vein tumor thrombus (PVTT) may profit more from combining the SBRT with systemic therapy to delay the intra- and extrahepatic progression. Current prospective studies are investigating such combinations.