The Significance of Transarterial Chemo(Embolization) Combined With Tyrosine Kinase Inhibitors and Immune Checkpoint Inhibitors for Unresectable Hepatocellular Carcinoma in the Era of Systemic Therapy: A Systematic Review

Background and Aims Regardless of great progress in early detection of hepatocellular carcinoma (HCC), unresectable HCC (uHCC) still accounts for the majority of newly diagnosed HCC with poor prognosis. With the promising results of a double combination of transarterial chemo(embolization) and tyrosine kinase inhibitors (TKIs), and TKIs and immune checkpoint inhibitors (ICIs), a more aggressive strategy, a triple combination of transarterial chemo(embolization), TKIs, and ICIs has been tried in the recent years. Hence, we aimed to conduct a systematic review to verify the safety and efficacy of the triple therapy for uHCC. Methods PubMed, MedLine, Embase, the Cochrane Library, and Web of Knowledge were used to screen the eligible studies evaluating the clinical efficacy and safety of triple therapy for patients with uHCC up to April 25th 2022, as well as Chinese databases. The endpoints were the complete response (CR), objective response rate (ORR), disease control rate (DCR), conversion rate, progression-free survival (PFS) rate, overall survival (OS) rate, and the incidence of adverse events (AEs). Results A total of 15 studies were eligible with 741 patients receiving transarterial chemoembolization (TACE) or hepatic arterial infusion chemotherapy (HAIC) combined with TKIs and ICIs. The pooled rate and 95% confidence interval (CI) for CR, ORR, and DCR were 0.124 (0.069–0.190), 0.606 (0.528–0.682), and 0.885 (0.835–0.927). The pooled rates for PFS at 0.5 years and 1 year were 0.781 (0.688–0.862) and 0.387 (0.293–0.486), respectively. The pooled rates for OS at 1, 2, and 3 years were 0.690 (0.585–0.786), 0.212 (0.117–0.324), and 0.056 (0.028–0.091), respectively. In addition, the pooled rate and 95%CI for the conversion surgery was 0.359 (0.153–0.595). The subgroup analysis of control studies showed that triple therapy was superior to TACE+TKIs, TKIs+ICIs, and TKIs in CR, ORR, and DCR, conversion rate; PFS; and OS. No fatal AEs were reported, and the top three most common AEs were elevated ALT, elevated AST, and hypertension, as well as severe AEs (grading ≥3). Conclusion With the current data, we concluded that the triple therapy of TACE/HAIC, TKIs, and ICIs would provide a clinical benefit for uHCC both in short- and long-term outcomes without increasing severe AEs, but the conclusion needs further validation. Systematic Review Registration http://www.crd.york.ac.uk/PROSPERO/, Review registry: CRD42022321970.


INTRODUCTION
Primary liver cancer is the sixth most common cancer worldwide with approximately 906,000 newly diagnosed patients per year and more than 90% patients having hepatocellular carcinoma (HCC) (1). The prognosis of HCC patients remains far from satisfactory with the median overall survival (OS) of 25-30 months (2,3). Radical surgery is still the most cost-effective curative treatment for HCC patients, but the majority have lost the chance of surgery, so called "unresectable" HCC (uHCC), mainly due to the absence of symptoms in the early stage of HCC (4)(5)(6).
There are no guidelines or consensus on the management of uHCC patients up to now (4)(5)(6) because this population is too heterogeneous. For uHCC patients at intermediate stage according to the Barcelona Clinic Liver Cancer stage (BCLC) (6), transarterial chemoembolization (TACE), as a classical modality of transarterial chemo(embolization), is strongly recommended with the overall objective response rate (ORR) beyond 50% (7,8). In the recent years, another modality of transarterial chemo(embolization), hepatic arterial infusion chemotherapy (HAIC), has been identified as non-inferior to TACE in the management of HCC, and particularly, the advantage of HAIC over TACE has been verified among those with macrovascular invasion (9,10). Nonetheless, the prognosis of patients receiving TACE/HAIC remains poor with the median progression-free survival (PFS) of 2.8-9.6 months and most of the patients will get resistant after repeated TACE (9,11). For advanced uHCC patients including those with extrahepatic metastasis, systemic therapy is the preferred strategy (12,13). With the advent of the novel tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs), sorafenib is not the only option for advanced HCC (14,15). In addition, the doublecombination modality of systemic therapy, such as atezolizumab and bevacizumab (15), lenvatinib and pembrolizumab (16), durvalumab and tremelimumab (17), apatinib and camrelizumab (18), and novolumab and ipimumab (19), have exhibited promising results with manageable toxicity. However, the objective response rate (ORR) of systemic therapy is still poor, and the time to response might be too long.
The combination of locoregional and systemic treatments is another option for uHCC (3). In theory, locoregional treatment, such as TACE or HAIC, is efficient to achieve satisfactory local control (LC); however, it has not always translated into a longterm survival benefit. On the other hand, systemic therapy is the key to improve the long-term prognosis, but unsatisfactory LC will impair the long-term survival advantage. Preclinical studies have identified the synergistic effect of TACE/HAIC and systemic therapy (20,21), which was also confirmed in practice of the combination of TACE and sorafenib (22), TACE and lenvatinib (23), HAIC and sorafenib (24), and HAIC plus lenvatinib and toripalimab (25), but it is not the end.
With the publication of the IMbrave 150 trial (15), HCC has entered the era of molecular and immune therapy. It is surprising that approximately 40% patients were found to receive previous TACE before randomization in the IMbrave 150 trial, which shed light on a more aggressive modality, the triple therapy of TACE/HAIC, TKIs, and ICIs. Furthermore, the current strategy is far from enough to satisfy the increasing demands of "conversion therapy" for uHCC. In the past 2 years, the triple therapy of TACE/HAIC, TKIs, and ICIs has been tried with encouraging results (20,25,26), but most of the studies were retrospective with a small sample size. Therefore, in this study, we comprehensively reviewed all the literature on the triple therapy of TACE/HAIC, TKIs, and ICIs and aimed to provide substantial clues for the subsequent studies.

MATERIAL AND METHOD
This systematic review was conducted according to the preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline, which was also registered at http://www. crd.york.ac.uk/PROSPERO/ (Review registry 321970) "hepatocellular carcinoma" or "HCC", and "transcatheter arterial chemoembolization" or "transarterial chemoembolization" or "hepatic arterial infusion chemotherapy" or "chemotherapy" or "TACE" or "HAIC", and "tyrosine kinase inhibitors" or "TKIS" and "immune check point inhibitors" or "ICIs" or "programmed cell death protein 1" or "programmed cell death ligand 1" or "PD-1" or "PD-L1" or "B7-H1." The literature searching began in December 2021, and the last searching was April 25th 2022. Considering that TACE or HAIC is preferred in China, the Chinese database of China National Knowledge Infrastructure (CNKI) and Wanfang were also used to identify the eligible studies.

Selection Criteria
The inclusion criteria were as follows: i) patients diagnosed as HCC by image or biopsy, ii) unresectable after Multi-Disciplinary Team (MDT), and iii) received the triple combination modalities of TACE/HAIC, TKIs, and ICIs, regardless of sequence; iv) endpoints must consist at least one of the following items: complete response (CR), ORR, disease control rate (DCR), PFS, OS, conversion rate, and adverse events (AEs).
The exclusion criteria were as follows: i) combined with other treatment such as ablation or radiation, ii) duplicate report derived from the same cohort, iii) protocol, case reports or reviews, and iv) data unavailable.

Data Acquisition
According to the predefined forms, the information of the eligible studies include the surname of the first author, year of publication, design of the study, and study period. In addition, baseline characteristics in each study (sample size, age, sex, hepatitis B virus infection, Child-Pugh grade, AFP level, tumor number, tumor size, macrovascular invasion, extrahepatic metastasis, BCLC stage, mean PFS, mean OS and regimen of TACE/HAIC, TKIs, and ICIs) were extracted directly by two independent researchers (QK and FX). Endpoints including the CR, ORR, DCR, PFS, OS, conversion rate, and adverse events were obtained directly from the main text or supplementary files and were then cross-validated between the researchers. In case of any discrepancy, an MDT discussion including at least one senior doctor was introduced to reach the final decision. Of note, tumor responses were determined by the modified response evaluation criteria in solid tumors (mRECIST) or RECIST in this study, but we would choose the results evaluated by mRECIST if they were also evaluated by RECIST in each included study.

Quality Assessment
Considering that all the included studies were retrospective, the quality of each eligible study was assessed by the modified Newcastle-Ottawa Scale (NOS) (27). Briefly, the risk of bias was graphically presented as a proportion of all included studies. Evaluating elements included the following: i) whether the study reported a definite definition of the objective; ii) whether a clear triple combination of TACE/HAIC, TKIs, and ICIs was offered (including the technique, regimen, and course of TACE/HAIC and dosage and courses of TKIs and ICIs); iii) whether the criteria of response assessment was provided (i.e., RECIST or mRECIST); and iv) whether there was a clear definition of outcomes including THE CR, ORR, and DCR.

Statistical Analysis
A meta-analysis of the pooled rate was conducted using Rstudio and R (4.1.2); while a comparison analysis between two groups was conducted using RevMan Version 5.3. The pooled rate for the CR, ORR, and DCR and rates of PFS and OS at different time-points was used as an effect size with 95% confidence interval (CI). The c2 test and I 2 statistics were used to evaluate the heterogeneity among the included studies. If P>0.10 and I 2 <50%, there was no apparent heterogeneity, and the fixed-effect model was used to estimate the effect size; otherwise, the random-effect model was used (28,29). Sensitivity analysis was carried out by removing each of the included studies one by one to determine the reliability of the results. Subgroup analyses were also conducted to decrease the heterogeneity among the included studies. Publication bias was determined using the funnel plot with Egger's and Begg's tests. In this study, a P-value <0.05 was considered significant.
All of the included studies came from China, and all the studies were retrospective, three of which were multi-centered (25,30,34). The baseline characteristics in each study were depicted in Table 1, as well as the results of quality assessment. Of note, the baseline characteristics of the control arm in the comparing cohort studies were also depicted in Table 1.
Considering that there was no consensus on the triple combination of TACE/HAIC, TKIs, and ICIs, the scheme in each study was a little different from each other. Table 2 exhibited the detailed information on the treatment scheme, including the technique of TACE/HAIC, drug regimens, and the sequence of local and systemic therapy.

Adverse Events
The pooled rates for the treatment related AEs were depicted in months, respectively, without severe AEs during the treatment. From then on, more studies have been published with promising results. In this systematic review, 15 studies were identified with 741 patients receiving TACE/HAIC+TKIs+ICIs, and initial analysis showed encouraging results. However, the CR rate ranged from 0.069 to 0.190, the ORR ranged from 0.528 to 0.682, and the DCR ranged from 0.835 to 0.927, as well as the mean PFS (4.0-16.3 months) and mean OS (8.6-24.8 months). The divergences between the included studies might be attributed to the following reasons: 1) the sample size of all eligible studies was small, which meant that II error is hard to avoid; 2) there is substantial heterogeneity among uHCC patients, not to mention primary or recurrent HCC; 3) the regimen of triple therapy was very different from each other, including transarterial therapy modality (conventional TACE, DEB-TACE, or HAIC), TKI agents (sorafenib, Lenvatinib, or apatinib), and ICI agents (pembrolizumab, camrelizumab, tislelizumab, sintilimab, toripalimab, or nivolumab); 4) triple therapy as first-line treatment or not, which greatly influenced on tumor responses, median PFS and OS; 5) different treatment goals, for example, successful conversion and subsequent curative resection would have better prognosis than those with   and Child-Pugh B7, 60 mg/m 2 ), 5-fluorouracil, 1-1.5 g/m 2 HAIC for 4-24 h (Child-Pugh A, 1.5 g/m 2 ; and Child-Pugh B, 1 g/m 2 ) and leucovorin (200 mg, intravenous infusion for 2 h before 5-Fu) was used. HAIC was repeated every 4-6 weeks until the intrahepatic lesions achieved CR, disease progression, or until the toxicity was unacceptable Lenvatinib: 8 mg per day oral Sorafenib: 200 mg oral, twice daily; 400 mg was administered orally twice daily if drug tolerance was acceptable. If patients had received sorafenib or lenvatinib before the study, regorafenib or apatinib was given. Regorafenib: Approximately 80 mg was administered orally once daily; 120 mg was administered orally once daily if drug tolerance was acceptable. Apatinib: 250 mg was administered orally once daily for 28 days as a treatment cycle.
Camrelizumab (200 mg/3 weeks), sintilimab (200 mg/3 weeks), toripalimab (240 mg/3 weeks), and nivolumab (3 mg/kg every 2 weeks). palliative treatment. Hence, the conclusion needs further validation, and Table 7 exhibited ongoing prospective trials evaluating the clinical efficacy of the triple combination modality. Conversion therapy is well concerned nowadays in the field of uHCC (43,44). Evidence suggests that R0 resection is a crucial independent protective factor of long-term survival (45). Shindoh et al. (46) found that advanced HCC patients after conversion therapy receiving R0 resection could have comparable prognosis with initially resectable HCC patients. Previous studies found that the successful conversion rate was 42.4% by lenvatinib and ICIs (47), 14.8% by TACE+sorafenib (48), and 12.8%-14.3% by HAIC +sorafenib (24,49), respectively. Using the triple combination modality of HAIC+TKIs+ICIs, the conversion rate was reported to be as high as 60% by Zhang et al. (36), and in this systematic review, the pooled rate for the conversion surgery was 35.9%. The underlying mechanism of the synergistic effect of the triple combination might be as follows: 1) TACE or HAIC could improve the tumor immune microenvironment, induce    (20) firstly identified that the triple therapy could not only activate cell immunity but also stimulate humoral immunity, and circulating Ig G, Ig l, and Ig k could serve as potential biomarkers of triple therapy. In the future, more attention should be paid on the triple combination modality for uHCC, especially for those with a strong willingness to receive radical resection. It has been yet to be known which is the optimal modality of transarterial chemo(embolization) because there are rare reports comparing TACE, DEB-ATCE, and HAIC in the triple therapy for unresectable HCC. TACE has always been the cornerstone for intermediate-stage HCC (6), which was repeatedly confirmed by a recent systematic review with a median OS of 19.4 months (50). However, repeated TACE may lead to liver function impairment and even TACE resistance, and TACE alone is unsatisfactory for patients in advanced stage, especially portal vein invasion or extrahepatic spread (9, 10). Drug-eluting beads TACE (DEB-TACE) was found to yield better tumor responses and a similar safety profile compared to conventional TACE (5). Ren et al. (51) firstly compared the efficacy of DEB-TACE combined with ICI versus conventional TACE combined with ICI for unresectable HCC. Results showed that DEB-TACE was a safe and well-tolerated treatment and produced better PFS and tumor response in patients with unresectable HCC than conventional TACE. On the other hand, HAIC has been identified to be non-inferior to TACE in local control and even had a weak advantage over TACE in long-term prognosis (9)(10)(11). In this systematic review, a slight advantage of HAIC+TKIs+ICIs over TACE+TKIs+ICIs was observed in CR, ORR, and DCR, but it did not translate into survival benefit in PFS and OS. In addition, an apparent inferiority of HAIC+TKIs+ICIs to TACE+TKIs+ICIs was also found in the conversion rate (33.4% vs. 38.9%). Hence, it remains controversial in the choice of conventional TACE or DEB-TACE or HAIC among the triple combination, and "head-to-head" prospective trials might be the answer in the future.

(Continued)
As a saying goes, one size does not fit for all, and not all unresectable HCCs will benefit from the triple therapy. Ju et al. (39) found that TACE+apatinib+camrelizumab provided a clinical benefit for the subgroups of age <65 years old, men, PS score of 1, Child-Pugh classification of B, liver cirrhosis, hepatitis B infection, and AFP >200 mg/ml (all P<0.05), and similar findings were  Hence, identifying the potential beneficiary of the triple combination modality is an urgent agenda. Safety is a bottleneck of the triple combination modality. The most common AEs are impaired liver function, fever, and abdominal pain related to TACE/HAIC (9-11); hypertension; diarrhea; and hand-foot syndrome to TKIs (52,53) and rash, fatigue, and pruritus to ICIs (14,19), respectively. A combination of TACE/HAIC and TKIs often increases the AEs of hypertension, hand-foot syndrome, and diarrhea, and a combination of TKIs and ICIs increases the risk of fatigue, rash, and hypothyroidism (20,25,31). As for the triple combination modality, safety can never be overemphasized. In this systematic review, the most common AEs were still elevated ALT and/or AST and hypertension, as well as the most severe AEs, but no mortality caused by the triple combination modality was reported. These results indicated that liver function   might be selection criteria for the triple modality, and patients with impaired liver function will be contradicted to this modality. Generally, the triple combination of TACE/HAIC, TKIs, and ICIs for uHCC needs a long way to go. Apart from the triple modality itself including the scheme and sequence, more factors should be of concern: 1) how many additional survival benefit, 2) how much cost-effectiveness, and 3) how about AEs with an intensified regimen. In addition, the accessibility of the medical  care is another decision-making factor. In China, TKIs like sorafenib and lenvatinib and ICIs like camrelizumab and sintilimab have been enrolled into a healthcare insurance, which are much cheaper than nivolumab, pembrolizumab, and atezolizumab. Furthermore, TACE is much more preferred than HAIC, owing to its high compliance. Hence, more factors should be taken into consideration in the future trials. There were several limitations in this systematic review. First, all of the studies were retrospective, in which recalling bias was hard to avoid. Second, considering that all the published studies came from China, the conclusion would not be applicable for the western patients due to the apparent heterogeneity in etiology between the East and the West. Third, data on the TACE/HAIC, TKIs, and ICIs were not available in several included studies; hence, the corresponding subgroup analysis could not be conducted. Fourth, considering that some studies came from the same center, the patient's cohort might be the presence of overlap. Last but not the least, the sequential order of the triple modality was not unified among the included studies, and in the future, an extensive consensus should be reached on this issue.

CONCLUSION
With the current data, we concluded that the triple combination of TACE/HAIC, TKIs, and ICIs would provide a clinical benefit for uHCC both in short-and long-term outcomes without increasing severe AEs. However, more is unknown on the optimal regimen, potential beneficiary, and latent AEs. Future RCTs with a larger sample size and cross-regional centers will aid in better clarifying the role of the triple modality for uHCC.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
QK, FX, and HF: acquisition of data, analysis and interpretation of data. QK and LW: conception and design of the study. QK and LW: drafting the article. YZ and JL: critical revision, final approval. All authors contributed to the article and approved the submitted version.