Efficacy and safety of anlotinib monotherapy for advanced hepatocellular carcinoma and clinical role of α-fetoprotein.

Introduction
Hepatocellular carcinoma (HCC) is the most common primary liver cancer in the world and ranks the third most common and second most fatal malignancy in China1,2. Chronic hepatitis B virus (HBV) infection is the key determinant of HCC development in Chinese patients3. The majority of patients with HCC are diagnosed with advanced disease, and only a few are eligible for potentially curative locoregional therapies. Due to the resistance to chemotherapy drugs, targeted therapy has been the standard care for advanced HCC since the first approval of sorafenib in 20074. However, the clinical benefit of sorafenib is limited with the median overall survival (mOS) of 8.5–14.7 months in patients worldwide and 6.5–8.9 months in Asian patients5–9. The first drug to get approved for the treatment of HCC in patients who had progressed after sorafenib therapy was regorafenib, a small molecule inhibitor with a broader inhibition of kinases than sorafenib10. To date, these two drugs are the standard of care for advanced HCC as first- and second-line monotherapy. With a lot of recent translational research and clinical trials, other targeted agents, such as lenvatinib, apatinib and donafenib, have been approved for first- or second-line therapy11–13. But no monotherapy has shown significantly better clinical effects than sorafenib, and drug resistance and tumor recurrence also remain inevitable issues. Thus, an unmet need for more effective targeted drugs remains among patients with advanced HCC.
Another clinical issue is to define biomarkers for targeted therapy and validate them in large-scale studies. HCC biomarkers are in urgent need in the following clinical fields: prognosis prediction; identification of a subgroup of patients for whom targeted therapy is more effective; early prediction of tumor response for targeted therapy. Alpha-fetoprotein (AFP) is a biomarker routinely assessed for the diagnosis and treatment of HCC. Some studies considered the pre-treatment AFP level as a prognostic factor and found its ability to screen patients who would benefit from targeted therapy14–16. Recently the post-treatment change of AFP level has been reported to predict tumor response although disagreement over the magnitude of the change still exists17–20. Other biomarkers such as blood-derived inflammatory markers and albumin-bilirubin (ALBI) grade have also been investigated for predicting survival of HCC patients21,22.
Anlotinib is a novel tyrosine kinase inhibitor (TKI) with potential anti-neoplastic and anti-angiogenic activities. Compared to other TKIs, it has more targets, including vascular endothelial growth factor receptor (VEGFR) 1–3, platelet-derived growth factor receptors (PDGFR) α/β, fibroblast growth factor receptors (FGFR) 1–4, c-Kit, c-FMS and discoidin domain receptor 1 (DDR1)23. Anlotinib has shown promising efficacy and tolerable toxicity in many malignancies, including advanced non-small-cell lung cancer, advanced soft tissue sarcoma, metastatic renal cell carcinoma and advanced medullary thyroid cancer24–27. Previous experimental and clinical data have already demonstrated the antitumor effects of anlotinib on HCC28,29. Our team’s previous single-arm clinical study demonstrated that anlotinib achieved a time to progression (TTP) of 5.9 months in first-line treatment and 4.6 months in second-line treatment30. However, the advantages of anlotinib compared to other targeted drugs and appropriate biomarkers are still unclear.
This multicenter retrospective study was conducted to explore the efficacy and safety of anlotinib for first- and second-line therapy. And patients receiving sorafenib treatment served as the control group. Predictive factors for disease control after anlotinib treatment were also analyzed in this study.

Materials and methods

Study design
Clinicopathological data of 210 patients diagnosed between May 2017 and July 2023 were retrieved from 3 tertiary medical research centers in China: National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College; and Qilu Hospital, Cheeloo College of Medicine, Shandong University. 55 patients received anlotinib monotherapy, while 155 patients received sorafenib monotherapy. Enrolled patients had a diagnosis of locally advanced or metastatic HCC. Diagnostic evaluation of HCC was based on liver biopsy or noninvasive measures combining imaging and blood tests. Eligible patients also conformed to the following inclusion criteria: age over 18 years old; Barcelona Clinic Liver Cancer (BCLC) stage B or C; Child–Pugh score < 8; an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0 or 1; at least one measurable lesion defined by the Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1); normal function of vital organs; complete medical records, including imaging and prognostic information. Exclusion criteria included diagnosis of cholangiocarcinoma, combined hepatocellular-cholangiocarcinoma (cHCC-CCA) or fibrolamellar hepatocellular carcinoma (FLHC); existence of serious comorbidities; treatment history of immunotherapy; local therapy combined during targeted therapy. The flow diagram of patients was shown as Fig. 1.

This study was approved by the Institutional Review Boards of National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (Ethical Approval Number: 21-197-2868) and Qilu Hospital of Shandong University (Ethical Approval Number: 2021-I2M-1-066). This study was conducted according to the Declaration of Helsinki. Because of the retrospective nature, this study was granted a full waiver of informed consent. All information provided by patients is maintained with confidentiality.

Targeted drugs
All patients in anlotinib group received treatment at a dose of 12 mg on day 1 through 14 of a 21-day cycle. Dosage reduction to 10 mg/day or even 8 mg/day would be proposed due to grade 3 or 4 hematologic toxicities. Sorafenib was taken 400 mg orally twice daily. Patients experiencing grade 4 hematologic toxicity or grade 2 non-hematologic toxicity require dose reduction. The dose reduction should not exceed two steps: the first reduction decreases the dose to 80% of the original dosage, and the second reduction decreases it to 50% of the original dosage. Patients received their assigned drugs until they were no longer benefiting from therapy or unacceptable toxic events occurred. Patients were regularly followed by the investigators.

Tumor response and outcomes
Tumor response was assessed using computed tomography (CT) scans or magnetic resonance imaging (MRI) examinations every 6 weeks by three radiologists independently and characterized by complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). Not to be evaluated (NE) denoted treatment effects which were not recorded. When disagreement happened on imaging evaluation, a third doctor was requested to reevaluate. The primary end point was overall survival (OS). OS was defined as the time from the first dose of drugs medication to death from any cause. The secondary endpoints included progression­free survival (PFS), objective response rate (ORR), disease control rate (DCR), survival rates at 12 or 24 months, and PFS rates at 6 or 12 months. PFS referred to the time from the first dose of drugs medication to disease progression according to RECIST 1.1, or death from any cause, whichever occurred first. ORR was defined as the proportion of confirmed CR or PR at the best response. DCR was defined as the percentage of confirmed CR, PR, or SD at the best response.

Follow‑up and data collection
Laboratory test results were collected from the time of drug assignment to 30 days following treatment discontinuation. Radiological imaging was done for all patients every 6 weeks after the first intervention dose. Patients or their family members were contacted every 1 month to obtain information on adverse events (AEs), survival status, and causes of death (if applicable) after treatment. Follow-up was conducted through outpatient visits, WeChat, and/or telephone interviews. AEs were graded according to the Common Terminology Criteria Adverse Events version 4.0 (CTCAE 4.0). The cutoff date for follow-up was July 1, 2024.
Percentage change of AFP, ALBI score, ALBI grade and three blood-derived inflammatory (neutrophil-to-lymphocyte ratio, NLR; platelet-to-lymphocyte ratio, PLR; prognostic nutritional index, PNI) markers were evaluated to predict disease control in HCC patients. Percentage change of AFP level was defined as (AFP level at 2, 4, 6 weeks after first dosage - AFP baseline level) / AFP baseline level. ALBI was a useful assessment tool for hepatic reserve function and ALBI score was calculated as follows: (log10 bilirubin (µmol/L) × 0.66) + (albumin (g/L) × − 0.085). The cutoff points for ALBI grades were as follows: ≤ − 2.60 (ALBI grade 1), more than − 2.60 to ≤ − 1.39 (ALBI grade2), and > − 1.39 (ALBI grade 3)31. NLR was calculated by dividing the total number of neutrophils by lymphocytes. PLR was calculated by dividing the total platelets count by lymphocytes count. PNI was calculated by multiplying albumin (g/L) by absolute lymphocyte count.

Statistical analysis
In this observational study, a 2:1 nearest neighbor propensity score matching (PSM) was used to achieve a balanced covariate distribution between the two treatment groups. Variables included in the propensity score calculation were provided in Table S1. Statistical computation of PSM method was finished by the R package MatchIt (version 4.3.2)32. A caliper for matching process was set at 0.1. The Student’s t-test or the Wilcoxon rank-sum test was conducted to assess baseline comparability for continuous variables between the two groups. The mean and standard deviation (SD) were used to describe normally distributed continuous data, while the median and interquartile range (IQR) were used to describe non-normally distributed continuous data. Categorical variables were calculated using both frequencies and percentages. And they were compared by the Chi-square test or Fisher’s exact test. The OS and the PFS curves were obtained using the Kaplan-Meier (KM) method and compared by the log-rank test. The best cut-off values for hematological biomarkers were determined by the receiver operating characteristic (ROC) curve. Univariate and multivariate analyses were conducted using the Cox proportional hazards model or logistic regression analysis to analyze factors associated with survival or disease control. Variables with a P value < 0.05 in the univariate analyses were included in the multivariate analyses. Tests within each subgroup used an unstratified Cox proportional hazards model and the forest plots illustrated the test results. All statistical analyses were performed using RStudio version 1.4.1717 (Integrated Development Environment for R, Boston, MA) and SPSS version 26.0 (IBM Corp., Armonk, NY, USA). All tests were two-tailed, and P values < 0.05 were considered statistically significant.

Results

Baseline clinical characteristics
Initially, a total of 210 patients with locally advanced or metastatic HCC were included in this study according to the inclusion and exclusion criteria (Table S1). Compared to anlotinib group, patients in sorafenib group had incomparable composition ratio of Child–Pugh class (P = 0.074). PSM analysis generated an assembly of 158 patients with a mean age of 52.7 years (SD = 9.7). 54 patients received anlotinib and 104 received sorafenib. Imbalances in patients’ characteristics between two groups were minimized (Table 1). Among the 158 patients, 141 (89.2%) were males, 127 (80.4%) had HBV infection, 66 (41.8%) had a baseline AFP level of more than 400 ng/mL, 70 (44.3%) had an ECOG PS score of 1, 122 (77.2%) were BCLC stage C, 16 (10.1%) were Child–Pugh class B (scores 7) and 107 (67.7%) were CNLC stage III. The median tumor size was 6.4 cm (IQR [4.3, 10.4]). The presence of distant metastasis happened in 73 (46.2%) patients, including 40 (25.3%) lung metastasis and 16 (10.1%) bone metastasis. 75 (47.5%) patients received targeted therapy for first-line therapy, and 83 (52.5%) patients received targeted therapy for second-line therapy. 80 patients underwent surgery during first or second-line treatment, accounting for 50.6% of the total population.

In anlotinib group, 24 (44.4%) patients received anlotinib as first-line therapy, whereas 30 (55.6%) patients received anlotinib as second-line therapy. 10 (18.5%) patients underwent surgery after achieving disease control with anlotinib. In sorafenib group, 51 (49.0%) patients received sorafenib as first-line therapy, whereas 53 (51.0%) patients received sorafenib as second-line therapy. 18 (17.3%) patients underwent surgery after achieving disease control with sorafenib.

Overall prognosis analysis
As shown in Figure S1, the median follow-up time for the whole patient (n = 158) was 13.0 months (range: 1.0–30.0 months). For patients (n = 75) receiving TKI therapy as first-line treatment, overall median PFS was 4.0 months (95% confidence interval [CI]: 3.0–5.0 months; Figure S1A) and overall median OS was 15.0 months (95% CI: 14.0–16.0 months; Figure S1B). For patients (n = 83) receiving TKI therapy as second-line treatment, overall median PFS was 5.0 months (95% CI: 4.0–6.0 months; Figure S1C) and overall median OS was 13.0 months (95% CI: 10.0–16.0 months; Figure S1D).
Univariate Cox regression analyses were used to explore prognostic factors in HCC patients receiving TKI therapy as first-line (Table S2) or second-line (Table S3) treatment. For patients receiving TKI therapy as first-line treatment, China Liver Cancer (CNLC) stage and macrovascular invasion (MaVI) were strongly associated with both PFS (CNLC, P = 0.001; MaVI, P = 0.019) and OS (CNLC, P = 0.026; MaVI, P = 0.004). Large tumor size and distant metastasis were found to have obviously negative effects on PFS (tumor size, P = 0.032; distant metastasis, P = 0.004) while they had no significant influences on OS (tumor size, P = 0.767; distant metastasis, P = 0.932). AFP level > 400 ng/mL was a risk factor only for OS (P = 0.005). For patients receiving TKI therapy as second-line treatment, Child–Pugh class B, MaVI and distant metastasis were related with worse PFS (Child–Pugh class, P = 0.011; MaVI, P = 0.021; distant metastasis, P = 0.005), while AFP level > 400 ng/mL was associated with worse OS (P = 0.011). Moreover, CNLC stage III was associated with both worse PFS (P = 0.011) and OS (P = 0.026).
The results of multivariate analyses were presented in Tables 2 and 3. The presence of distant metastasis was an independent risk factor associated with PFS both in HCC patients receiving TKI therapy as first-line treatment (HR [hazard ratio]: 1.542, 95% CI: 1.003–2.561, P = 0.023) and those receiving TKI therapy as second-line treatment (HR: 1.595, 95% CI: 1.182–1.982, P = 0.022). And high AFP (> 400 ng/mL) was an independent risk factor for poorer OS (first-line, HR: 2.645, 95% CI: 1.822–3.172, P = 0.016; second-line, HR: 1.827, 95% CI: 1.127–2.963, P = 0.014). Moreover, CNLC stage III was a risk factor independently related to PFS only in patients receiving TKI therapy as second-line treatment (HR: 1.081; 95% CI: 1.013–1.177; P = 0.042).

Clinical efficacy and safety
54 patients in anlotinib group and 104 patients in sorafenib group were included in efficacy and safety analysis. We separately analyzed the differences in the impact of anlotinib and sorafenib on clinical outcomes in first- and second-line treatments. Among all patients, 24 (44.4%) patients in anlotinib group and 51 (49.0%) in sorafenib group received targeted drugs as first-line therapy (Table 4). 2 (8.3%) patients achieved PR, 16 (66.7%) exhibited SD and no one achieved CR in anlotinib subgroup, yielding an ORR of 8.3% and DCR of 75.0%. And for sorafenib subgroup, 1 (2.0%) patient reached CR, 2 (3.9%) obtained PR and 35 (68.6%) achieved SD. The ORR was 5.9%, and the DCR was 74.5%, with no significant difference compared to the anlotinib subgroup (ORR, P = 0.921; DCR, P = 0.964). Median PFS was 5.0 months (95% CI: 2.6–7.4 months) with anlotinib and 4.0 months (95% CI: 3.4–4.6 months) with sorafenib (HR: 1.465; 95% CI: 0.885–2.425; P = 0.144) (Fig. 2A). Median OS was improved in anlotinib subgroup versus sorafenib subgroup (20.0 vs. 15.0 months; HR: 1.588; 95% CI: 0.927–2.719; P = 0.077) (Fig. 2B).

A total of 30 (55.6%) patients in anlotinib group and 53 (51.0%) in sorafenib group received targeted drugs as second-line therapy (Table 4). 3 (10.0%) patients achieved PR, and 17 (56.7%) exhibited SD in anlotinib subgroup, whereas 5 (9.4%) patients achieved PR, and 24 (45.3%) exhibited SD in sorafenib subgroup. No one achieved CR in either subgroup. The ORR was 10.0% or 9.4% (P = 0.762), and the DCR was 66.7% or 54.7% (P = 0.288), in the anlotinib and sorafenib subgroup, respectively. Median PFS was 5.0 months (95% CI: 4.3–5.7 months) versus 5.0 months (95% CI: 4.2–5.8 months) in anlotinib and sorafenib subgroup, respectively (HR: 0.969; 95% CI: 0.599–1.565; P = 0.948; Fig. 2C). Median OS (13.0 vs. 11.0 months) was significantly improved with anlotinib versus sorafenib (HR: 1.966; 95% CI: 1.189–3.250; P = 0.010; Fig. 2D).
Drug-related adverse effects (AEs) in patients were summarized in Table S4 and Table 5. 46 (85.2%) or 95 (91.3%) patients with anlotinib or sorafenib experienced drug-related AEs (Table S4). Drug-related grade ≥ 3 AEs occurred in 15 (27.8%) or 39 (37.5%) patients with anlotinib versus sorafenib (P = 0.296). The most common AE was diarrhea both in anlotinib (28 patients, 51.9%) and sorafenib (56 patients, 53.8%) group. Specifically, in first-line treatment, the incidence of AEs did not differ significantly between HCC patients treated with anlotinib and those treated with sorafenib (Table 5). Among HCC patients receiving TKI therapy as second-line treatment, the incidence of drug-related grade ≥ 3 AEs was significantly lower in the anlotinib group (16.7% vs. 41.5%, P = 0.020).

Subgroup analysis
Subgroup analysis was performed to identify which part of patients could benefit from anlotinib therapy as first-line or second-line treatment. As shown in Figure S2, the HR values for PFS continued to favour anlotinib versus sorafenib across most prespecified subgroups in the first-line treatment of HCC, and statistically significant differences were only observed in the high AFP subgroup (AFP > 400 ng/mL, HR: 0.442, 95% CI: 0.179–0.992, P = 0.047). Subgroup comparisons of OS showed that anlotinib brought improvement in patients’ survival in the first-line treatment, and significant differences were noted in the subgroup with a baseline AFP concentration of more than 400 ng/mL (HR: 0.161, 95% CI: 0.046–0.561, P = 0.004) and BCLC stage C (HR: 0.490, 95% CI: 0.258–0.927, P = 0.028; Figure S3).
As shown in Fig. 3, anlotinib did not demonstrate a clear subgroup advantage in promoting PFS in the second-line treatment setting except for the high AFP subgroup (AFP > 400 ng/mL, HR: 0.479, 95% CI: 0.224–0.802, P = 0.027). Subgroup analyses of OS demonstrated that anlotinib provided clear survival benefits in the second-line therapy, with statistically significant differences observed in most subgroups (Fig. 4). Better OS was achieved regardless of whether the patients’ AFP levels were above or below 400 ng/mL (AFP ≤ 400 ng/mL, HR: 0.330, 95% CI: 0.150–0.724, P = 0.006; AFP > 400 ng/mL, HR: 0.422, 95% CI: 0.202–0.879, P = 0.021). Moreover, anlotinib exhibited a significant advantage in HBV-positive patients (HR: 0.386, 95% CI: 0.215–0.694, P = 0.001). However, this effect was not observed in HBV-negative patients.

The clinical significance of anlotinib in the high AFP subgroup (AFP > 400 ng/mL) was illustrated using KM curves. For patients receiving TKI therapy as first-line treatment, median PFS was 9.0 months (95% CI: 4.0–13.5 months) in anlotinib subgroup versus 4.0 months (95% CI: 2.0–5.0 months) for sorafenib subgroup (P = 0.001; Fig. 5A). Median OS was 20.0 months (95% CI: 10.0–19.5 months) for anlotinib subgroup versus 11.0 months (95% CI: 8.0–16.0 months) for sorafenib subgroup (P = 0.027; Fig. 5B). For patients receiving TKI therapy as second-line treatment, median PFS was 7.0 months (95% CI: 5.0–10.0 months) in anlotinib subgroup versus 3.0 months (95% CI: 2.0–6.0 months) for sorafenib subgroup (P = 0.039; Fig. 5C). Median OS was 12.0 months (95% CI: 10.0–15.5 months) for anlotinib subgroup versus 8.0 months (95% CI: 7.0–17.0 months) for sorafenib subgroup (P = 0.016; Fig. 5D).

Predictive factors for disease control in anlotinib group
For patients in anlotinib group, in order to identify early predictive factors which can predict disease control after anlotinib therapy, risk factors in Tables 2 and 3 and several hematological biomarkers (NLR, PLR, PNI, ALBI score and grade, rate of change in AFP level at 2, 4, 6 weeks after first dosage) were included in predictive factors analysis. ROC analysis was conducted to obtain cutoff values of hematological biomarkers. Results of ROC analysis were summarized in Table S5. The optimal cutoff values of NLR, PLR, PNI and ALBI score for disease control were determined as 3.0, 115, 55 and − 2.5. The result of ROC analysis showed that patients with disease control were well differentiated from those with disease progression when AFP decreased by ≥ 25% at 4 weeks after first dosage, or by ≥ 27% at 6 weeks after first dosage. The 4-week time window was selected as it offers an early assessment of treatment response, enabling timely evaluation of therapeutic efficacy. Finally, the cutoff used for the rate of change in AFP level was − 0.25 with specificity and sensitivity levels of 93.8% and 78.9% (Figure S4). According to the results of previous studies, other time windows (2 weeks, 4 weeks and 6 weeks) and cutoffs for rate of change in AFP levels (20%, 50%, and 75%) were also included in logistic regression analysis to analyze factors associated with disease control17.
The results were shown in Table 6 and NLR, PLR and ALBI grade were found to be significantly correlated with disease control in univariate analysis (P = 0.033, 0.003, 0.025). Decrease rate of AFP in 2 or 4 weeks by 20% or 25% were also predictors for disease control (P = 0.022, 0.005, 0.031, < 0.001). Multivariate analysis showed only decrease rate of AFP in 4 weeks by 25% was an independent predictor for disease control in anlotinib group (OR [odds ratio]: 1.854, 95% CI: 1.364–2.476, P = 0.001). Accordingly, AFP response was defined as a reduction of ≥ 25% from the baseline level.

The ability of AFP response as a predictor for disease control was validated in different treatment lines (Table S6). Logistic regression analysis within each treatment line supported that AFP response was an early predictor of disease control both in first-line treatment (OR: 2.378, 95% CI: 1.648–3.475, P = 0.001) and second-line treatment (OR: 1.872, 95% CI: 1.378–2.655, P = 0.001). Moreover, patients in anlotinib group were divided into high-AFP (baseline AFP ≥ 400 ng/mL) and low-AFP (baseline AFP < 400 ng/mL) subgroup. Logistic regression analysis within each subgroup supported that AFP response was an early predictor of disease control both in high-AFP subgroup (OR: 1.745, 95% CI: 1.087–2.665, P = 0.002) and low-AFP subgroup (OR: 1.475, 95% CI: 1.211–1.851, P = 0.018).

Discussion
This multicenter retrospective real-world study revealed that the anlotinib showed promising efficacy and acceptable safety as monotherapy for patients with advanced HCC. Treatment with anlotinib or sorafenib was generally well-tolerated, and the most frequently recorded drug-related AE was diarrhea for both. Anlotinib exhibited a moderately lower incidence of grade ≥ 3 AEs compared to sorafenib (27.8% vs. 37.5%). The better efficacy profile of anlotinib versus sorafenib was mainly reflected by statistically significant extension in OS. Subgroup analysis demonstrated significant OS benefits of anlotinib in most subgroups, including those reflective of the characteristics of the Chinese HCC population, such as high AFP levels, positive HBV infection, and a higher proportion of patients with BCLC stage C. In comparison to western countries, patients presented with more advanced baseline disease status in China, which further underscores the positive therapeutic response of anlotinib.
The REFLECT study demonstrated that lenvatinib as a first-line treatment exhibited clinical benefits comparable to sorafenib, with a median OS increase of 1.3 months (13.6 vs. 12.3 months, P > 0.05)13. A large-scale phase III clinical trial indicated that donafenib, a drug independently developed in China, was the only first-line TKI to show superior OS compared to sorafenib (12.1 vs. 10.3 months, P = 0.025)12. In this study, patients treated with anlotinib as first-line therapy achieved a longer median OS compared to those receiving sorafenib (20.0 vs. 15.0 months, P = 0.077). Currently, both donafenib and lenvatinib have been included in the standard for diagnosis and treatment of primary liver cancer in China, and large-sample, double-arm clinical trials are still needed to further validate the clinical therapeutic advantages of anlotinib33.
Our team’s prior research has demonstrated that anlotinib is a viable option for HCC refractory to first-line therapy, with a median TTP of 4.6 months30. Compared to the placebo control group, both regorafenib and apatinib, as second-line standard treatment regimens, have been shown to improve OS in patients (10.6 vs. 7.8 months, P < 0.001; 8.7 vs. 6.8 months, P = 0.048)10,11. Anlotinib also showed trend of OS improvement in this study (13.0 vs. 11.0 months, P = 0.010). The recurrence or progression of HCC is highly dependent on angiogenesis, as the formation of new blood vessels provides essential nutrients to support tumor cell proliferation. Compelling evidence has revealed that FGFR signaling plays an important role in HCC progression, angiogenesis and therapy-resistance34. For example, FGFR1/AKT/mTOR signaling pathway and FGF19/FGFR4 axis are responsible for sorafenib resistance35,36. On one hand, the pan-FGFR inhibition of anlotinib offers a potential solution to overcome HCC resistance to anti-angiogenic agents; On the other hand, its favorable safety profile enables sustained anti-angiogenic effects. Taken together, anlotinib shows promise as a potential standard second-line therapeutic option in the future.
AFP has remained the main diagnostic and prognostic biomarker for HCC for decades. Many studies showed that high baseline AFP level appeared to be associated with worse patient survival16,37–39. This study revealed that baseline AFP level was identified as the only independent prognostic factor for patients’ OS. Moreover, Subgroup analysis revealed that significant PFS improvement was only observed for patients with a baseline AFP of 400 ng/mL or greater. Andrew et al. analyzed patients’ data from REACH-1 and REACH-2 phase 3 trial, and the results showed ramucirumab, an angiogenesis inhibitor, met its primary endpoint in the subgroup of patients with AFP levels above 400 ng/mL15,16. Another clinical trial reported that the magnitude of OS benefit brought by lenvatinib was greatest in patients with high AFP level (≥ 200 ng/mL)13. These data indicated AFP functioned as a biomarker for identifying subgroups of HCC patients and guiding adjustments in treatment strategies.
Treatment response is generally evaluated by radiological technology. Although imaging techniques such as CT and MRI can visually reflect changes in tumor size following treatment, they also have certain limitations, including the subjectivity of radiologists, interference from tissue changes induced by pharmacological treatments, and the lack of standardized imaging guidelines for the evaluation of systemic therapy for HCC. Most importantly, imaging methods cannot predict treatment efficacy at an early stage. In addition to the diagnostic and prognostic significance of baseline AFP levels, numerous studies have proposed that changes in AFP levels during or following treatment may serve as surrogate biomarkers for assessing response to systemic and locoregional therapies17–20. However, there is considerable variability among studies in the criteria used to define AFP response, including the time window for AFP measurement after treatment and the magnitude of AFP level reduction. In this study, we systematically evaluated multiple cutoffs at multiple time points and found that a 25% reduction at 4 weeks had the strongest predictive performance based on ROC and logistic regression analyses. This threshold offers a balance of clinical utility and early assessment, though further validation is needed across different cohorts and treatment settings.
In recent years, with the widespread clinical application of immune checkpoint inhibitors (ICIs), a variety of immunotherapy regimens represented by ICIs have emerged in the field of hepatocellular carcinoma treatment. Results from IMbrave150 study demonstrated that the combination of atezolizumab and bevacizumab significantly prolonged mOS and mPFS compared with sorafenib40,41. Subgroup analysis indicated that the combination therapy also provided significant clinical benefits for Chinese patients. Several studies have shown that anlotinib, in addition to inhibiting tumor proliferation and angiogenesis, can enhance the infiltration of natural killer (NK) cells, antigen-presenting (APC) cells, and CD8+ T cells while inducing apoptosis of cancer-associated fibroblasts (CAFs)42–45. This process facilitates the transformation of an immunosuppressive microenvironment into an immune-activated microenvironment, thereby activating and initiating T cell recognition of tumor antigens to elicit antitumor immune responses. Considering the efficacy and safety profile of anlotinib, we believe that anlotinib, either as monotherapy or in combination with ICIs, holds significant clinical potential for patients who develop resistance to immunotherapy or experience immune-related adverse events (irAEs). Furthermore, clinical trials investigating the safety and efficacy of anlotinib in combination with ICIs are currently underway (ClinicalTrials.gov identifier: NCT05453383, NCT03825705, NCT04052152, and NCT06031480).
In summary, anlotinib acts as an effective monotherapy agent for patients with unresectable or metastatic HCC. It also showed acceptable safety and tolerability. However, this study has some limitations. Firstly, because of the retrospective nature, this study utilized PSM analysis to filter the original data and minimize potential biases in outcome analysis. But clinical trials with a large sample size are still needed to further validate the clinical therapeutic advantages of anlotinib compared to sorafenib or other TKIs. Secondly, this study included only Chinese patients. Thus, future studies are required to evaluate the efficacy and safety of anlotinib in Western populations.

Conclusion
Anlotinib demonstrated encouraging efficacy and tolerability in Chinese patients with advanced hepatocellular carcinoma in both first- and second-line settings. AFP response may serve as an early predictor of disease control during anlotinib treatment. And these findings warrant confirmation in future prospective clinical studies.

Supplementary Information
Below is the link to the electronic supplementary material.