This study was a meta-analysis to further evaluate the treatment efficacy and safety of RHR versus RFA in early-stage RHCC. The present study demonstrated that RHR was more effective than RFA for extending the OS and PFS of RHCC patients, especially for patients who had a single tumor > 3 cm and ≤ 5 cm. However, both RHR and RFA were suitable for single RHCC ≤ 3 cm. Moreover, the results suggested that the incidence of complications was lower for patients treated by RFA. To our knowledge, this study included the largest study population and presents the latest meta-analysis including new studies published within the last five years. In addition, a randomized clinical trial was included in the meta-analysis, which contributed to a high evidence level. Therefore, the results of this study can provide important data with which guidelines for the management of RHCC after initial treatment could be established.
Several previous meta-analyses have been performed to evaluate the outcomes of RHR versus RFA for RHCC patients, of which Gavriilidis et al concluded that the RHR and RFA groups had similar OS and PFS rates[9], while only five retrospective studies were included in the study[15, 16, 26, 27, 30]. On the other hand, another previous meta-analysis showed that RHR was associated with comparable OS rates and higher PFS rates[8, 10, 19]. The differences between the findings of previous meta-analyses and those of our study might be explained by the following reasons. First, in previous studies, the number of included studies was small, and non-RCTs were included in the evaluation. Second, though several studies reported generally comparable outcomes between RHR and RFA, a tendency toward longer OS and PFS was observed in the RHR group compared with the RFA group[22, 27, 29]. Third, two recent high-quality studies revealed that RHR was still the most effective treatment, followed by RFA for RHCC[17, 18], and the results may play an important role in the meta-analysis. However, these two studies had never been included in previous meta-analyses.
According to the outcomes, tumor recurrence may be one of the most important factors affecting OS in patients with RHCC. There are many factors associated with tumor recurrence, and the completeness and safety margin of treatment are key elements. HCC has a tendency to invade portal branches and thus cause tumor dissemination along the liver segment[31]. Segment-based anatomic partial hepatectomy can remove both the primary tumor and microvascular invasion, together with at least 1 cm of the rim of normal hepatic parenchyma[32]. However, in the RFA procedure, it is hard to create a sufficient safety margin precisely in the 3-dimensional liver with the guidance of 2-dimensional ultrasonography[33]. There is also a lack of objective evaluations of the safety margin and ablation effect. In addition, some risk factors for recurrence are associated with RFA but not with resection. For example, difficult locations, such as a tumor located on the liver surface or near the main hepatic vessels or hilum, are a worsening indicator in ablation[34]. Moreover, the complete ablation rate is affected not only by tumor location but also by the experience of the operator. Therefore, it is not surprising that RFA has been frequently reported to have higher recurrence rates than resection for the treatment of HCC[34].
With regard to treatment-related complications, RHR was associated with a greater incidence of major complications than RFA, which should be attributed to the minimally invasive characteristic of RFA. Compared with RHR, RFA can be performed percutaneously, thus greatly minimizing the surgical impact. In addition, RFA preserves as much liver parenchyma as possible and causes minor damage to the remnant liver[35]. Hence, RFA can serve as an alternative choice of treatment for early-stage RHCC with the advantage of less invasiveness.
In the subgroup analysis of patients in China, the results concerning OS and PFS were similar to the outcomes of the meta-analysis without regional restriction. This finding was also confirmed by the study of Chen et al. [10]. According to the latest data, approximately 46.71% of new cases of HCC are diagnosed in China, and over 85% of patients with HCC are linked with hepatitis B virus infection[36]. Therefore, the results of the subgroup analysis add weight to the current clinical decision in the Chinese population.
Another subgroup analysis was performed in RHCC ≤ 3 cm, of which RFA achieved equivalent OS and PFS rates compared with RHR. After initial resection, RHCC is usually smaller than 3 cm under intensive screening[20]. Previous studies have demonstrated that a smaller tumor size is closely related to an increased chance of complete ablation[37]. This may be because RFA can achieve a greater safety margin than RHR for RHCC ≤ 3 cm. As expected, subgroup analyses demonstrated better OS after RHR than after RFA among patients with an RHCC diameter greater than 3 cm. Unfortunately, these results should be further explored because of the limited number of included research studies in the subgroup analysis.
There is no doubt that for patients with reserved liver function, RHR is preferred over RFA if the tumor is resectable. However, the reported rate of RHR for RHCC in clinical practice was less than 30%[38]. As an effective alternative for surgery, RFA has some advantages when compared with RHR in treating RHCC. First, as a minimally invasive treatment modality, RFA can greatly decrease the incidence of major complications. Second, repeatability is a major advantage of RFA[26]. For patients with limited liver remnants, RFA may serve as an ideal treatment choice. Therefore, for those who are unsuitable for RHR or have a tumor size smaller than 3 cm, RFA may be a replacement therapy for resection because of its safety and feasibility.
Moderate heterogeneity was found in the meta-analysis of major complications. Sensitivity analysis was conducted by eliminating each study in turn. Finally, we found that the heterogeneity of the meta-analysis mainly came from the study of Liang et al. [15]. In the study of Liang et al., major complications were defined as complications with Clavien–Dindo classification grade II or higher, while in the other included studies, grade III or higher was applied[16–18, 20–22, 26, 27, 29]. Consequently, the definition in Liang’s study overestimated the incidence of major complications, which led to heterogeneity.
This meta-analysis has several limitations. First, only a small number of studies examined the treatment options for RHCC. A total of 10 studies were included in this meta-analysis, and only 6 reported PFS. Second, indirect data acquisition obtained from survival curves may have an effect on our outcomes. Third, only 2 or 3 studies in the subgroup analysis covered the tumor size of RHCC, and more evidence is needed in future studies. Furthermore, many studies have demonstrated that the number of lesions and Child-Pugh class are important prognostic factors[39, 40]. However, the data on tumor number and Child-Pugh class were not sufficient for meta-analysis in subgroups.