Biologically effective dose (BED) of stereotactic body radiation therapy (SBRT) was an important factor of therapeutic efficacy in patients with hepatocellular carcinoma (≤5 cm)

Hepatocellular carcinoma (HCC) is the sixth most common malignancy and the third most common cause of cancer mortality worldwide, and only approximately 20–30% of patients with HCC are eligible for surgical treatment, including liver resection and liver transplantation [1, 2]. Accumulating data have shown that stereotactic body radiation therapy (SBRT) is a safe and effective treatment for HCC, especially in patients with inoperable or recurrent HCC [3‐5]. Furthermore, Su et al. [6] compared the efficacy of stereotactic ablative radiation therapy (SABR) versus liver resection for treating small HCC (< 5 cm) patients with Child-Pugh class A (CP-A) cirrhosis and concluded that SABR has local effects that are similar to those of liver resection. Wahl et al. [7] reported that SBRT and radiofrequency ablation (RFA) were equally effective for treating small HCCs. We conducted this retrospective study to evaluate the efficacy of SBRT and identify prognostic factors related to the efficacy of SBRT in patients with HCC.

In other cancers, such as lung and cervical cancers, with the delivery of increasing biologically effective doses (BEDs) to lesions, the OS of patients increased [8‐10]. However, due to the scarcity of data on HCC, the relationship between the BEDs and the efficacy of SBRT in HCC patients was included in our study.

Radiation-induced liver injury (RILD) is often a fatal complication of SBRT and should be avoided in patients with HCC; furthermore, the CP classification is an important predictive factor of RILD [11‐13]. Therefore, we investigated the incidence of RILD in the included patients and searched for predictive factors among their clinical data, including biochemical parameters and the peripheral platelet (PLT) and white blood cell (WBC) counts.

The median follow-up period was 42 months (range, 6–77 months). By June 2018, 45 patients had experienced relapse or metastasis; 36 patients experienced liver metastasis, 2 experienced lymph node metastasis, 1 experienced bone metastasis, 1 experienced brain metastasis, 1 experienced lung metastasis and 4 experienced multiple organ metastases. The treatment was ultimately selected by the patient (Fig. 1).

In this study, we retrospectively reviewed a large group of CP-A HCC patients treated with SBRT over a three-year span as an extended analysis of our previous observations [25]. Compared to prior studies, the OS, PFS, LC and DMFS rates were all satisfactory in this study. Su et al. [4] reported 3- and 5-year OS rates of 73.5 and 64.3% and 3- and 5-year PFS rates of 58.3 and 36.4%, respectively, with doses of 42-46Gy administered in 3–5 fractions and single-fraction doses of 28–30 Gy. Among 77 HCC patients treated with SBRT, Andres et al. [3] found 1- and 2-year OS rates of 81.8 and 56.6%, respectively, and 1- and 2-year LC rates of 99%. The total dose of 45 Gy in 3 fractions was prescribed to the 80% isodose line. Their OS rates were lower than ours, which have been caused by the inclusion of CP-B patients and/or patients with previous treatments. This finding was confirmed by their conclusion that the CP-B classification was associated with a poor prognosis. Moreover, previous treatments may also be an affect factor. Ueno et al. [26] reviewed 296 patients with single nodular HCC ≤ 5 cm with Child-Pugh A between 2001 and 2011 who underwent surgical resection (SR, n = 136) and radiofrequency thermal ablation (RFA, n = 160), and they found that 5-year OS rates of SR and RFA among all patients were 70.1 and 69.8%, respectively (P = 0.14). The 5-year OS was similar to the 5-year cumulative OS in our result.

CP score was an influential factor in OS and PFS in our study, which may be related to the extent to which CP reflects cirrhosis, which is associated with occurrence of complication and tumorigenesis. Moreover, we found that the OS, PFS and DMFS rates also increased significantly with BED₁₀ ≥ 100Gy in treating HCC with SBRT, but with no significant difference in the LC rates. The correlation between LC and BEDs in our study was consistent with Ohri’ findings [27], but he didn’t explore the relation between BEDs and survival rates. In addition to initial radiotherapy, the efficacy of subsequent treatment after recurrence was also a factor affecting OS. The relation between PFS and BEDs of our study, we speculate that SBRT could change immune microenvironment, which can be proven by DMFS. This assumption is also supported by some other tumor-related articles. Lee et al. [28] reported that ablative radiotherapy dramatically increased T-cell priming in draining lymphoid tissues, leading to reduction/eradication of the primary tumor or distant metastases. Meanwhile, antitumor immunity significantly contributes to the superior response induced by one dose of 20Gy compared with that induced by 4 doses of 5Gy. Schaue et al. [29] studied the tumor-specific immune response in mice with murine melanoma irradiated with 15Gy administered in different fractionated dose schemes, and the results showed that a single dose of 7.5Gy or higher, but not lower than 5Gy, was immune-stimulatory, as mediated by tumor-reactive T cells. They all showed that under the condition that the total dose remained unchanged, the superior anti-tumor immune response was related with the higher single dose. And it is the same with BEDs, which increase with higher single dose when the total dose remained the same. Therefore, with higher BEDs causing a better anti-tumor immune microenvironment, the recurrence rates of tumor decreased. However, it’s only our hypothesis. In our other studies, we found that in these patients different BED₁₀ values lead to different changes of immune system, such as NK cell functions, and further experiments are in progress (data not shown).

The influential factor of RILD is the pretreatment PLT count, as demonstrated by Velec et al. [11]. Some studies shown that the PLT count indirectly reflects the degree of liver cirrhosis by indicating the degree of portal hypertension and hypersplenism [30, 31]. Nozaki et al. [32] performed in vitro and in vivo studies, and they have proven that thrombopoietin promotes liver regeneration and improves liver cirrhosis by increasing the PLT level, indirectly implying that PLT decrease would worsen liver cirrhosis. We think it is the relation between the PLT count and liver cirrhosis, an influential factor of RILD that makes PLT count an influential factor of RILD [33]. In clinical work, we found Child-Pugh score and PLT were not compatible. Degree of liver cirrhosis could affect liver nutrition supply by affecting liver hemodynamics, which may affect liver function. However, liver cirrhosis degree is not the only influence factor of liver function, which may be affected by many other factors, including effective liver volume, hepatic cell function and compensate ability of liver, etc. Moreover, PLT is not a parameter affecting Child-Pugh score. Our result reminds us that besides Child-Pugh score, PLT is also a non-negligible factor.

Though the limited dose of bile duct was not shown in TG-101, previous studies showed that higher dose radiation of bile duct was related with higher risk of stenosis. Barney et al. [34] reported one case of grade 3 biliary stenosis after SBRT where the patient was treated with a dose of 50Gy in five fractions. Takahisa et al. [35] showed that the true threshold for biliary stenosis would be between 40Gy in 5 fractions and 80Gy in 5 fractions, and a special caution was necessary when treating patients with more than 40Gy in 5 fractions. Before this study, we had 2 cases with obstructive jaundice after 50Gy/5f and 54Gy/6f, and their PET-CT showed bile duct stenosis but no active lesion. Since then, we adopted the dose fraction of 49Gy/7f and 48Gy/8f in SBRT of small HCC when the tumor is near to bile duct to reduce this risk. And until now, no patient has had bile duct stenosis by the two doses and fraction sizes above, so we consider they were safe fractionated regimens. Considering both the relation of BEDs and survival rates in our study, we think the trade-off between the efficacy and risk of SBRT in tumors near the bile duct deserves further study, and the optimal dose fractionated regimens needs large-sized samples to explore.

This study has several limitations. First, this was a retrospective study. Second, the number of patients in the two groups is much different, which may cause bias. A relevant prospective clinical trial is in progress (Clinicaltrails NCT 03295500). We hope to get objective and accurate results.

SBRT is a safe and effective option for CP-A HCC patients, and an increased BED₁₀ and lower CP score are associated with improved OS and PFS. The results indicate that a strategy escalating radiation doses over a limited time frame are worth exploring in a prospective clinical trial. Meanwhile, PLT count should be considered, which is a predictive factor of RILD in SBRT of HCC.