In this study, we reviewed the results of CGP tests performed in clinical practice for patients with advanced HCC. Our CGP results were consistent with the previous results for the genetic landscape of HCC [16]; furthermore, our analysis of the impact of CGP results obtained under real-world conditions on treatment selection has important clinical significance for certain patients.

CGP tests were initially introduced in clinical practice for patients with various solid tumors at an advanced stage [17], providing a basis for determining alternative treatments or participating in clinical trials after the standard of care based on genetic alterations. In our cohort, one patient harbored the TPM3–NTRK1 fusion gene, which is found at a relatively high frequency in thyroid or central nervous system tumors and is extremely rare in HCC [18]. The patient received a TRK inhibitor as second-line treatment under health insurance coverage based on CGP results and showed a good response. In addition, one of three patients with high TMB selected pembrolizumab. Although pembrolizumab has been shown to be effective in patients with high TMB, there is no evidence for the efficacy of this anti-programmed death receptor-1 antibody, even after progression to treatment using atezolizumab, an anti-programmed cell death ligand-1 antibody. We did not select pembrolizumab for the remaining two patients with tumor progression shortly after 6 weeks of atezolizumab plus bevacizumab therapy considering the relatively low TMB. Although pembrolizumab controlled tumor progression in a patient with 27.1 mutations/Mb, further studies are needed to determine which patients with high TMB are likely to benefit from pembrolizumab. Twenty-nine patients with alterations in several genes, including CTNNB1, MET, PIK3CA, FGFR2, and BRCA1, were potential candidates for clinical trials. We provided information about the clinical trials. However, no patients were enrolled in a clinical trial, although some visited sites of clinical trials. In a recent study in France, treatment selection for 15 patients with HCC and hepato-cholangiocarcinoma was based on the results of genomic testing, and some patients had a good response to treatment after progression under atezolizumab and bevacizumab combination therapy [19]. Health insurance systems and access to medicine differ substantially among countries. Eligibility for clinical trials is also updated occasionally and varies by region. Therefore, the impact of CGP on treatment selection may be constantly changing [20].

Of particular note, our results indicated that CGP tests can provide useful information for selecting second-line treatments after immunotherapy. For advanced HCC, several treatment options have been developed in the past decade, including first-line treatments as well as effective second- and later-line treatments, which are essential for improving prognosis in advanced HCC [21]. Effective second-line treatments are important, because not all patients can receive third- or later-line treatments due to tumor progression, deterioration of liver functional reserve, irreversible adverse events, or worsening of general condition [22]. However, data on the relative efficacies of second- or later-line treatments after immunotherapy are lacking. Furthermore, we do not have a clear picture of the role of CGP tests in clinical decision-making. For at least 12 patients (23%) in this cohort, accessible treatments covered by health insurance were recommended by the molecular tumor board, and those who received the recommended treatment exhibited a clinical benefit. We believed that these results support the clinical significance of CGP tests for the selection of second-line therapy after immunotherapy in patients with HCC, potentially benefiting a large number of patients.

The timing of CGP is a point of contention and depends on the purpose of the test. We performed CGP after the discontinuation of later-line treatment to provide a basis for the identification of alternative treatments based on gene alterations. For several patients, the information was useful for selecting second-line treatments; therefore, we provided information about CGP tests after the discontinuation or at the expected discontinuation of first-line treatment and conducted testing if the patients agreed. The most frequently observed druggable alterations in this cohort were CTNNB1 mutations, for which β-catenin inhibitors are being evaluated in clinical trials (i.e., https://jrct.niph.go.jp/latest-detail/jRCT2080224780). Unfortunately, the study participants were not able to participate in clinical trials; in particular, 10 of 29 patients with CTNNB1 mutations had already received second- or later-line treatment, which was an exclusion criterion. Earlier CGP tests may provide expanded treatment opportunities for patients and facilitate the development of new drugs for HCC. Some studies have identified predictive factors for the response to immunotherapy in HCC. For example, CTNNB1 alterations are associated with a suppressive microenvironment and patients with CTNNB1 alterations have a poorer response to immunotherapy than that of patients with wild-type CTNNB1 [23, 24]. However, the reproducibility of these findings is unclear, and Montironi et al. reported that one of the two types of CTNNB1 mutation positive tumors are considered to be responsive to immunotherapy [25]. Therefore, there is not sufficient evidence to justify CGP testing before first-line treatment.

In biliary tract cancer or other solid cancers, the development of precision medicine for second-line treatment resulted in increased drug availability, and CGP tests are positioned to become standard as companion diagnostic assay [26]. No agent based on gene alterations is available for HCC, and HCC has not yet reached that area. Even if druggable gene alterations can be detected, the recommended agents do not always produce favorable results due to deteriorated liver function and complications of cirrhosis, and therefore, the availability of agents for advanced HCC is currently limited [27]. We evaluated the effectiveness of the treatments recommended by the molecular tumor board using response to treatment, because this study included only patients who underwent CGP and the single-arm nature requires caution in interpreting time-to-event results such as PFS and OS. Moreover, because of the variety of treatment lines of the included patients, we selected anti-tumor effect as efficacy endpoint to directly assess the efficacy of treatment. However, we believed that the results of this study, coupled with successes in studies of other solid cancers and an increasing understanding of the molecular basis of HCC, suggest that treatment strategies based on gene alterations will be useful in the future.

CGP testing for patients without access to tumor tissue is one of the issues regardless of cancer [28]. Various factors including the kinds of sample, quality of sample, tumor cell content rate, intra-tumor and inter-tumor heterogeneity, and treatment can affect the sensitivity of detection of gene alterations [29]. It is important to note that TMB assessed via liquid biopsy can be also influenced by as the quantity of circulating tumor DNA (ctDNA) present in the blood or the potential interference from clonal hematopoiesis. When ctDNA levels are low, particularly in early or intermediate-stage cancers or in patients with minimal residual disease, the TMB estimation may be less reliable. Additionally, clonal hematopoiesis, which involves age-related mutations in hematopoietic stem cells, can introduce mutations unrelated to the tumor, potentially affecting the interpretation of TMB in liquid biopsy samples. In cases where tissue biopsy is available, TMB is typically considered more reliable due to the higher quantity and quality of tumor DNA, allowing for a more comprehensive assessment of the mutational burden. Therefore, as mentioned above, we prefer to collect tissue samples as much as possible or to use previously resected tissues. However, liquid biopsy has the advantages of easy access, non-invasiveness, and avoidance of biopsy complications, and with the rapid development of ctDNA technology [30], liquid biopsy remains a valuable, albeit potentially less precise, alternative for patients where only liquid biopsy is feasible.

The present study had several limitations, including a relatively small sample size restricted to the Japanese population, selection bias, and a short observation period. Therefore, further investigations and eventually large-scale studies are needed to definitely conclude the usefulness of CGP tests and precision medicine based on gene alterations for improving outcomes in patients with advanced HCC.

In conclusion, we reviewed the genomic profiles of patients with advanced HCC and subsequent treatment selection in clinical practice. A subset of patients with advanced HCC had druggable gene alterations, providing useful information not only for proposing alternative treatments but also for selecting more promising targeted treatments, especially second-line treatments after immunotherapy.