Introduction

The systemic therapy strategy for patients with unresectable hepatocellular carcinoma (HCC) begins with sorafenib (SORA)1 as the first-line therapy. For a decade, SORA remained the only approved molecular-targeted agent (MTA) for unresectable HCC, with patients showing a poor prognosis2. However, there has been a significant change in systemic therapy for patients with HCC in the last few years. There have been rapid developments in systemic therapy, with various MTAs including regorafenib (REGO)3, lenvatinib (LEN)4, ramucirumab (RAM)5, and cabozantinib (CAB)6. Moreover, atezolizumab plus bevacizumab (Atezo/Beva)7 as a combination therapy with an immune checkpoint inhibitor and an anti-angiogenic agent was approved in 2020, with further expansion of treatment strategies for patients with unresectable HCC.

Systemic therapy is the main strategy for patients with intermediate-stage HCC who are refractory or unsuitable for transarterial chemoembolization as well as for patients with advanced-stage HCC8,9. Sequential therapy involving switching across MTAs is currently the primary evidence-based treatment strategy10. Continuous approved MTA treatment significantly prolongs overall survival (OS) in patients with unresectable HCC11,12. Additionally, studies have reported the efficacy and safety of second-line treatment with MTAs after disease progression under treatment with Atezo/Beva13. Therefore, sequential systemic therapy is considered to prolong the prognosis for patients with unresectable HCC.

In the era of sequential therapy using various systemic therapies, we should first familiarize ourselves with the adverse events (AEs) associated with each systemic therapy14,15; further, careful management of AEs is necessary to avoid treatment discontinuation in patients receiving systemic therapy. Sequential therapy has been shown to prolong OS in patients with unresectable HCC16,17, with preserved hepatic function being an important factor for successful sequential therapy16. However, the relationship among preserved hepatic function, sequential therapy, and treatment discontinuation due to AEs remains unclear. Further, there is a need to clarify whether these factors influence sequential systemic therapy as well as to determine the eligibility criteria for sequential therapy.

This study aimed to investigate whether sequential systemic therapy improved OS in patients with unresectable HCC. Additionally, we aimed to evaluate the clinical characteristics of patients eligible for sequential systemic therapy.

Results

Patient characteristics

Table 1 summarizes the baseline characteristics of the included patients. The median age was 73 years; further, 289 (79.2%) patients were men. The etiology of liver disease was hepatitis C virus in 177 (48.5%) patients, and 18.6% of hepatitis C patients (33/177) obtained sustained virologic response. The m-ALBI grades 1, 2a, and 2b were observed in 149 (40.8%), 120 (32.9%), and 96 (26.3%) patients, respectively. The median tumor size was 31 mm; additionally, 174 (47.7%) patients were BCLC stage C. Extrahepatic spread and macrovascular invasion were observed in 129 (35.3%) and 50 (13.7%) patients, respectively. Regarding the first-line therapy, 194 (53.2%), 127 (34.8%), and 44 (12.0%) patients were treated using LEN, SORA, and Atezo/Beva, respectively. The median observation time was 15.7, 16.9, and 6.1 months in LEN, SORA, and Atezo/Beva, respectively. Moreover, 156, 35, 6, 4, and 1 patient transitioned to second-, third-, fourth-, fifth-, and sixth-line therapies, respectively (Table 1).

Table 1 Patient characteristics.
Full size table

Evaluation of the therapeutic response to each first-line drugs

Therapeutic responses to each first-line drug are shown in Supplementary Fig. 1. The objective response rate (ORR) was observed in 36% of patients (16/44), 43% of patients (84/194), and 8% of patients (11/127) in the Atezo/Beva group, in the LEN group, and in the SORA group, respectively (Supplementary Fig. 1). On the other hand, the disease control rate was observed in 84% of patients (37/44), 79% of patients (154/194), and 44% of patients (56/127), respectively. Although there was no significant difference in the response between the Atezo/Beva group and the LEN group, there was a significant difference in the response between the SORA group and the other treatment groups (p = 0.001).

Overall survival with each first-line drugs

The median survival time (MST) was not reached, 19.3 months, and 17.3 months among patients who received Atezo/Beva, LEN, and SORA, respectively (Supplementary Fig. 2). There was no difference in OS among the LEN, SORA, and Atezo/Beva groups as first-line therapy.

The reasons for treatment discontinuation in the first-line and second-line therapy

The reasons for discontinuation observed during systemic therapy are shown in Table 2. Progressive disease was seen in 165 patients (54.0%), fatigue in 32 patients (10.4%), appetite loss in 22 patients (7.2%), and hepatic decompensation in 16 patients (5.1%) in the first-line therapy. On the other hand, progressive disease was seen in 68 patients (51.8%), fatigue in 10 patients (7.6%), appetite loss in 10 patients (7.6%) and hepatic decompensation in 12 patients (9.2%) in the second-line therapy (Table 2).

Table 2 The reasons for treatment discontinuation in the first-line and second-line therapy.
Full size table

Univariate and multivariate analyses of factors for discontinuation of AEs

Multivariate analysis revealed that advanced age and modified-ALBI grade 2b were identified as independent factors associated with treatment discontinuation due to AEs (Supplementary Table 1).

The difference in the reason for treatment discontinuation between the < 75 years and ≥ 75 years groups in the first-line and second-line therapy

The difference in the reason for treatment discontinuation between the < 75 years and ≥ 75 years groups are shown in Supplementary Tables 2 and 3.

The prevalence of discontinuation due to progressive disease was significantly higher in the < 75 years group compared to in the ≥ 75 years group in the first and second-line therapy, respectively. On the other hand, the prevalence of discontinuation due to fatigue was significantly higher in the ≥ 75 years group compared to in the < 75 years group in the first-line therapy (Supplementary Table 2). In addition, the prevalence of discontinuation due to appetite loss was significantly higher in the ≥ 75 years group compared to in the < 75 years group in the second-line therapy (Supplementary Table 3).

We also evaluated the baseline characteristics according to age (75 years) in patients with end first-line therapy. Regarding ALBI score, it was significantly different between the < 75 years group and ≥ 75 years group; however, there was no significant difference in the rate of initial dose reduction between the < 75 years and ≥ 75 years group in SORA treatment and LEN treatment (Supplementary Table 4). The cut-off age (< 75 years and ≥ 75 years) was based on the result of the data-mining analysis to determine the profiles associated with sequential therapy rates.

Conversion rate to later-line sequential therapy and therapeutic agents

Figure 1 shows the conversion rate to later-line sequential MTA therapy. Among the included patients, 58 (15.9%) and 307 (84.1%) patients continued and discontinued first-line therapy, respectively. Among the patients who discontinued first-line therapy, 159 (51.8%) patients underwent second-line systemic therapy. LEN, Atezo/Beva, SORA, REGO, and RAM were used as second-line therapy in 52 (32.7%), 43 (27.0%), 26 (16.4%), 22 (13.8%), and 16 (10.1%) patients, respectively. Among the patients who received second-line therapy, 28 (17.6%) and 131 (82.4%) patients continued and discontinued treatment, respectively. Among the patients who discontinued second-line therapy, 46 (35.1%) patients underwent third-line systemic therapy. LEN, RAM, Atezo/Beva, SORA, REGO, and CAB were used as third-line therapy in 17 (36.9%), 9 (19.6%), 9 (19.6%), 6 (13.0%), 3 (6.5%), and 2 (4.4%) patients, respectively (Fig. 1). The most common sequential therapy until second-line and third-line therapy were SORA-LEN sequential therapy (27%:43/159) and SORA-REGO-LEN sequential therapy (34.8%:15/43), respectively.

Figure 1
figure 1

Conversion rate to later-line sequential therapy and therapeutic agents. The yellow, red, blue, orange, green, and gray blocks indicate LEN, SORA, atezolizumab plus bevacizumab, SORA, REGO, RAM, and CAB, respectively. Abbreviations: LEN, lenvatinib; SORA, sorafenib; REGO, regorafenib; RAM, ramucirumab; CAB, cabozantinib.

Full size image

Relationship between discontinuation due to AEs and transition to the subsequent systemic therapy

Table 3 shows the rates of discontinuation of first- and second-line therapies due to AEs. There were 138 (44.9%) and 59 (45.0%) patients who discontinued first-line and second-line treatment, respectively, due to AEs (Table 3). The conversion rate to second-line and third-line therapies was significantly lower among patients with discontinuation due to AEs than among those without (30.4% vs. 69.2%, p < 0.001; 6.7% vs. 58.3%; p < 0.001, Table 3).

Table 3 Relationship between discontinuation due to AE and transition to next systemic therapy.
Full size table

Overall survival with systemic therapy according to first-, second-, and third-line or later therapies

The median survival time (MST) was 13.7 months, 19.2 months, and 35.6 months among patients who received first-line, second-line, and third-line or later therapies, respectively (Fig. 2). In the baseline characteristics, younger age and better ALBI scores were detected in the group who could reatch to the third-line or later-line therapy, when compared to the first-line or second-line groups (Supplementary Table 5).

Figure 2
figure 2

The overall survival of patients with HCC treated with systemic therapy. The red, green, and blue lines indicate the first-line, second-line, and third-line or later-line therapy groups, respectively. Abbreviations: HCC, hepatocellular carcinoma.

Full size image

ALBI score after first-, second-, and third-line or later therapies

Figure 3 shows the changes in the ALBI score from baseline. The median ALBI scores at the end of first-line, second-line, and third-line treatment were − 2.14, − 1.95, and − 1.92, respectively. The ALBI score was significantly worse among patients who discontinued treatment due to AEs during the sequential systemic therapy than among patients who did not [(− 2.01 vs. − 2.34, p < 0.001: first-line therapy), (− 1.74 vs. − 2.16, p < 0.001: second-line therapy), (− 1.64 vs. − 2.05, p < 0.001: third-line therapy)].

Figure 3
figure 3

The albumin-bilirubin (ALBI) score over time in systemic therapy. The red and blue lines indicate the treatment discontinuation and no treatment discontinuation due to AEs, respectively. Abbreviations: AE, adverse event.

Full size image

Decision tree analysis for transition to sequential therapy

We performed a decision tree analysis to determine the profiles associated with sequential therapy rates. It showed that the m-ALBI grade was the first splitting variable for sequential therapy rates. The sequential therapy rates were 31.4% and 59.6% in patients with m-ALBI grade 2b and m-ALBI grade 1 or 2a, respectively (Fig. 4). Among patients with m-ALBI grade 1 or 2a, the second splitting variable was age. The sequential therapy rate among patients with m-ALBI grade 1 or 2a (age < 75 years) was 70.8%. Contrastingly, the sequential therapy rate was 28.1% among patients with m-ALBI grade 2b who discontinued treatment due to AEs (Fig. 4).

Figure 4
figure 4

Profiles associated with sequential therapy. The pie graphs indicate the percentage of sequential therapy (white)/no sequential therapy (black) in each group.

Full size image

The results of univariate and multivariate analyses for transition to sequential therapy are summarized in Supplementary Table 6. Age, modified-ALBI grade, and discontinuation due to AEs were identified as independent factors for transition to sequential therapy (Supplementary Table 6).

Univariate and multivariate analyses of factors related to OS

Age, m-ALBI grade, maximum nodular diameter, macrovascular invasion, extrahepatic spread, AFP, DCP, discontinuation due to AEs, and post-progression treatment were included as variables in the univariate analysis. Multivariate analysis revealed that age, m-ALBI grade, macrovascular invasion, extrahepatic spread, discontinuation due to AEs, and sequential therapy were independent factors for OS (Table 4).

Table 4 Univariate and multivariate analyses of factors for OS.
Full size table

Discussion

This study evaluated the efficacy of sequential systemic therapy and investigated the clinical characteristics of patients, who were eligible for sequential systemic therapy. Systemic sequential therapy improved the prognosis of patients with advanced HCC. Multivariate analysis revealed that age, m-ALBI grade 1 or 2a, macrovascular invasion, extrahepatic spread, treatment discontinuation due to reasons other than AE, and sequential therapy were independent factors for OS. Additionally, decision tree analysis revealed that m-ALBI grade 1 or 2a and non-advanced age were clinically eligible characteristics for sequential systemic therapy.

Sequential therapy using MTAs has been recently considered an effective strategy for unresectable HCC in real-world clinical conditions11,18. During the era of mono-MTA, the MST of patients treated with SORA was 10.7 months1. Currently, various systemic agents, including REGO, RAM, CAB, and Atez/Beva are available. In our study, transition to later-line therapy prolonged OS compared with monotherapy. Compared with second-line therapy, third-line or later therapy drastically improved OS, with an MST of 35.6 months. Yengohag et al. reported that the median OS in the sequential treatment of SORA-REGO was 22.2 months19. Additionally, SORA as a first-line sequential systemic therapy for patients with unresectable HCC was found to prolong OS12. In our study, patients with HCC who successfully transited to third-line or later-line therapies showed the best prognosis. However, enforcement of the clinical trials regarding sequential systemic therapy is practically difficult to perform, therefore data on what is the best sequential strategy for unresectable HCC is lacking. Recently, the Markov model estimated the survival of different sequential strategies for unresectable HCC patients20. This model provided a strong rationale to support ongoing trials evaluating second-line after first-line ATezo/Beva therapy. Therefore, using these simulation models will be very important to determine the strategy of sequential therapy in real-world practice in the future.

The ALBI score and hepatic function are significant prognostic factors for sequential therapy for HCC16. Pre-treatment liver function based on the ALBI score was associated with discontinuation of MTA therapy due to AEs21,22,23. Some previous studies reported that ALBI scores gradually declines by sequential systemic therapy using MTAs11,24. We found that the ALBI score was significantly worse among patients with discontinuation due to AEs than among those without at the end of each line treatment and the liver function in pre-treatment was poorer in patients with discontinuation due to AEs compared to those without (ALBI score was − 2.36, and − 2.59, respectively). Moreover, the rate of transition to the next systemic therapy was lower among patients with discontinuation due to AEs than among those without. Hence, sequential systemic therapy should be introduced at the good liver function. Furthermore, we should be careful with hepatic decompensation during systemic chemotherapy. Hepatic decompensation hampers the uptake of subsequent lines of systemic treatment and it is associated with poorer survival after discontinuation of systemic chemotherapy25. Hence, the optimal management of liver function could improve the overall survival of HCC patients through long-term preservation of liver function26,27. In fact, the median post-first-line therapy survival time (PFST) was 5.3 months, 9.3 months, and 15.6 months among patients who had hepatic decompensation, adverse event, and HCC progression, respectively (Supplementary Fig. 3). Although there was no significant difference in PFST between hepatic decompensation and adverse event (PFST 5.3 months vs. 9.3 months, p = 0.010), we must promptly recognize hepatic decompensation and adequately manage it to allow the patient to continue the treatments.

Our findings demonstrated that sequential therapy and treatment discontinuation due to reasons other than AEs were among the independent factors for prolonged OS. Post-treatment progression is strongly correlated with post-progression survival (PPS); further, improved PPS is the most important factor for prolonging OS28. Ando et al.29 reported that sequential therapy with MTAs could improve the prognosis of patients who discontinued first-line therapy; further, good liver function is a favorable factor related to eligibility for second-line therapy. In this study, the median PPS was 6.2 months, 13.6 months, and 26.9 months in the non-transition to the second-line group, transitioned to the second-line group, and transitioned to the third-line or later therapies group, respectively (Supplementary Fig. 4). Thus, improved PPS is the most important factor for prolonging OS.

Currently, it has been reorted thah OS was significantly shorter among patients with discontinuation due to AEs than among those without30,31. Generally, fatigue and appetite loss are common AEs that cause discontinuation of systemic therapy; additionally, these AEs significantly affect the worsening ALBI scores15,22,32 and are associated with shortened treatment duration. In fact, the change in ALBI score was significantly worse in patients with discontinuation of treatment due to fatigue and appetite loss than in patients with discontinuation of treatment due to reasons other than fatigue and appetite loss in this study [Fatigue; (from − 2.34 to − 1.89 vs. from − 2.53 to − 2.19, p < 0.001: first-line therapy, from − 2.06 to − 1.49 vs. from − 2.45 to − 2.01, p < 0.001: second-line therapy)], [Appetite loss; (from − 2.46 to − 2.19 vs. from − 2.49 to − 2.13, n.s: first-line therapy, from − 2.39 to − 1.93 vs. from − 2.41 to − 2.01, p = 0.03: second-line therapy)]. In addition, time is required to recover from discontinuation due to AEs, with the tumor enlarging during treatment interruption33. Therefore, it is important to avoid discontinuation of sequential systemic treatment due to AEs. However, this study enrolled more patients who received LEN treatment in initial systemic chemotherapy. It has been reported that appetite loss and fatigue lead to discontinuation of treatment in LEN treatment15,33,34. In patients with discontinuation of initial systemic chemotherapy, the development of fatigue > G2 and appetite loss > G2 was significantly higher in the LEN group than in the SORA or Atezo/Beva groups [Fatigue > G2; LEN (35.9%) vs. SORA (3.3%), p =  < 0.001, LEN (35.9%) vs. Atezo/Beva (10.5%), p = 0.015, appetite loss > G2; LEN (29.4%) vs. SORA (6.6%), p =  < 0.001, LEN (29.4%) vs. Atezo/Beva (0.0%), p =  < 0.001, respectively]. Although we cannot deny that the results in the study were affected by high ratio of patients who were initially treated with LEN, the study showed that which factors are important for administration of sequential systemic therapy. Thus, we consider that it is important to investigate the clinically eligible characteristics for sequential systemic therapy.

Additionally, advanced age is a critical factor that causes discontinuation of systemic therapy due to AEs15. This could be attributed to patients with advanced age being vulnerable to the toxicity of anti-cancer agents35. Although some previous reports revealed that MTA-monotherapy could be used safely in elderly patients with advanced HCC36,37,38, there are few reportson whether the impact of advanced age is how to influence in sequential systemic therapy. In this study, we found that the incidence of fatigue, appetite loss were significantly higher in the ≥ 75 years group compared to in the < 75 years group (p =  < 0.001) in the first-line, and second line therapy, respectively (Supplementary Tables 1 and 2). Moreover, we found that the incidence of progressive disease was siginicantly higher in the < 75 years group compared to in the ≥ 75 years group. We have previously reported that advanced age is associated with the discontinuation of systemic therapy due to severe AEs15. It suggests that patients with advanced age are vulnerable to the toxicity of anti-cancer agents 35. In contrast, patients with non-advanced age are tolerable to the toxicity of the drug. Therefore, prevalence of progressive disease was relatively higher in the < 75 years group. Although clinical trials mostly include patients of non-advanced age, in real-world clinical settings, numerous patients with advanced age are treated with systemic therapy. It is important to promptly detect symptoms of AE in patients with advanced age during sequential systemic treatment.

To prevent worsening liver function and discontinuation due to AEs, we should actively introduce the nursing and pharmaceutical department intervention for education regarding self-monitoring and AEs management, and telephone follow-up in every 2–4 weeks is also important in proper management39. Moreover, we recently reported a useful protocol for LEN involving a 5 days-on/2 days-off administration schedule (the weekends-off protocol)40. However, a robust strategy to prevent worsening liver function and discontinuation due to AEs has not been established yet. Thus, additional clinical trials and real-world evidence could facilitate the establishment of a strategy for systemic sequential therapy.

This study has several limitations. First, this was a retrospective study. Second, there was a selection bias in drug selection. Especially, the selection of first-line drugs contributes to having a great effect on sequential systemic therapy41. Third, second-line and later-line therapies were performed at the oncologist’s discretion. Fourth, the follow-up time of Atezo/Beva was short. Fifth, we did not evaluate the dose modification and relative dose intensity of MTAs, the time of occurrence of AE, and oesophageal varices. Sixth, the factors that occurred after starting therapies, such as sequential therapy used in the analysis cause bias for OS. Nonetheless, this is the first study to clarify the relationship between sequential systemic therapy and m-ALBI 1 or 2a or non-advanced age. Thus, it is necessary to accumulate clinical data and establish beneficial sequential systemic therapy20. Future studies are warranted to establish an evidence-based strategy for an improved order of sequential systemic therapy.

In conclusion, we demonstrated that sequential systemic therapy improved the prognosis of patients with unresectable HCC. Additionally, we showed that discontinuation due to AEs resulting from deterioration of liver function and advanced age could disturb sequential systemic therapy. Finally, good hepatic function and non-advanced age are characteristics of clinical eligibility for sequential systemic therapy.

Material and methods

Study design and patients recruitment

A total of 1003 consecutive patients with unresectable HCC underwent systemic therapy as a first-line treatment from 2009 to November 30, 2021 and were registered at eight independent institutions in Japan. The exclusion criteria were as follows: treatment with systemic therapy before 2017 (n = 562), Child–Pugh class B or C (n = 50), Barcelona Clinic Liver Cancer (BCLC) stage A (n = 13), lost to follow-up (n = 10), or clinical trial (n = 3). Accordingly, we included 365 patients (Supplement Fig. 5). The study was conducted in accordance with the Helsinki Declaration and was approved by the Ethical Committee of Kurume University School of Medicine (Approval Code: 21006). Informed consent was obtained using the opt-out approach.

Diagnosis of HCC

HCC was diagnosed using a combination of following serum markers and imaging: AFP and DCP, and imaging procedures, including ultrasonography, enhanced CT and MRI scans. Macrovascular invasion was radiologically diagnosed by enhanced CT or MRI. In the study, macrovascular invasion included invasion into the portal vein and hepatic vein.

Treatment protocol and safety evaluation

Sequential systemic therapy was selected based on the evidence-based clinical practice guidelines for HCC of BCLC staging and treatment strategy8,42. Based on the industrial recommendations, the patients underwent first-line therapy comprising SORA (Nexval; Bayel Co., Ltd., Osaka, Japan), LEN (Eisai Co., Ltd, Tokyo, Japan), and Atezo/Beva (Chugai Pharmaceutical Co. LTD, Tokyo, Japan). SORA introduced as first-line therapy until LEN was approved in Japan. After LEN was approved in 2018, LEN introduced as first-line therapy. Moreover, since Atezo/Beva was approved, we introduced Atezo/Beva as first-line therapy. Depending on the approval time, the observation starting point of SORA started in 2017, because regorafenib was approved as second-line in Japan. Moreover, the observation starting point of LEN and Atezo/Beva was started in March 2018 and October 2020, respectively. Regarding contraindications of Atezo/Beva, we did not administer Atezo/Beva therapy in patients with autoimmune diseases. AEs were evaluated based on the National Cancer Institute Common Terminology Criteria for Adverse Events. Treatment was discontinued in case of development of any unacceptable or serious AE, or observation of clinical tumor progression. Systemic treatment using the second-line or subsequent therapies was performed depending on the physician’s decision.

Assessment of hepatic functional reserve

Hepatic functional reserve was evaluated using the albumin-bilirubin (ALBI) score. ALBI scores were calculated as follows: [Log10 bilirubin level (μmol/L) × 0.66] + [albumin level (g/L) ×  − 0.085]43. The modified ALBI (m-ALBI) grade was defined based on the ALBI score as follows: m-ALBI grade 1, ≤ -2.60; m-ALBI grade 2a, > − 2.60 to ≤ − 2.27; m-ALBI grade 2b, >  − 2.27 to ≤ − 1.39; m-ALBI grade 3, > − 1.3943.

Evaluation of treatment response and follow up

The therapeutic response of HCC was evaluated based on the modified Response Evaluation Criteria in Solid Tumors44 using computed tomography or magnetic resonance imaging at 4–6 weeks after the initial treatment. Subsequently, evaluation was performed at 3 months intervals until death or the study endpoint (November 31, 2021).

Statistical analysis

All statistical analyses were performed using JMP Pro version 15 (SAS Institute Inc., Cary, NC). All data are presented as the number or median (range). OS was calculated using the Kaplan–Meier method and analyzed using the log-rank test or Bonferroni method45. Factors related to OS were evaluated through univariate and multivariate analyses using the Cox proportional hazards model. Additionally, we performed a decision tree analysis to identify factors related to the possibility of sequential therapy, as previously described15. Statistical significance was set at a two-tailed p-value < 0.05.

Decision-tree algorithm

A decision-tree algorithm was constructed to reveal profiles associated with the sequential systemic therapy according to the instructions provided with the R software package as previously described46. The following variables were used for the decision-tree analysis for sequential systemic therapy: we used the same variables in the Cox proportional hazard model analysis.

Ethics approval statement

This study was approved by the Ethical Committee of Kurume University School of Medicine (Approval Code: 21074) and was conducted according to the Helsinki Declaration.