Prognostic outcomes of surgery vs. non-surgery in initially unresectable HCC with successful conversion via ICI + TKI therapy

Abstract

Background

No consensus exists on optimal post-conversion therapy for initially unresectable hepatocellular carcinoma (uHCC) patients responding to a combination of immune checkpoint inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs). This study evaluated surgical benefits and prognostic factors in such patients.

Methods

Retrospective analysis of 67 uHCC patients receiving ICIs + TKIs conversion therapy. Log-rank test compared OS/PFS between surgical resection (SR, n = 30) and non-surgical (NSR, n = 37) groups; Cox regression identified prognostic factors.

Results

Baseline variables were balanced between groups. The SR group showed significantly higher PFS than the NSR group (P = 0.032), with 1-, 2-, and 3-year rates of 82.6%, 53.5%, and 31.2% versus 71.1%, 21.6%, and 14.4% for the NSR group. Partial response (PR) (HR = 6.0, P < 0.001) and AFP level ≥ 400 ng/ml (HR = 2.8, P = 0.021) emerged as independent prognostic factors associated with OS. For PFS, partial response (PR) (HR = 3.3, P = 0.002), number of tumors ≥ 3 (HR = 2.4, P = 0.020), and non-surgical treatment (HR = 2.2, P = 0.031) were relevant.

Conclusion

Surgical intervention resulted in better PFS than non-surgical approaches, while no significant difference in OS was observed.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04677-w.

Introduction

Hepatocellular carcinoma (HCC), the most common subtype of primary liver cancer globally, accounts for 75%–85% of all liver cancer cases worldwide [1, 2]. Most patients present with intermediate to advanced disease at the time of diagnosis, and initial clinical evaluation indicates unresectability, primarily due to large tumor size, multiple intrahepatic lesions, invasion of major blood vessels, or severe hepatic impairment [3]. With advancing medical technologies, conversion therapy-encompassing local treatments, systemic therapies, or multimodal combinations-enables some initially uHCC patients to achieve tumor downstaging and regain resectability [4, 5]. While conversion therapy has improved survival in other tumors by facilitating subsequent resection [6, 7]. Consensus is lacking on whether surgery or continued non-surgical regimens offer greater benefits for successfully converted uHCC patients. Surgery may achieve radical resection but carries risks of complications and recurrence, whereas non-surgical approaches (sustained targeted/immunotherapy or local minimally invasive treatments) avoid surgical trauma but have unproven long-term efficacy.

Current data show 5-year survival rates of 24.9%-57.0% for post-conversion surgery, comparable to early-stage HCC resection [8]. Before 2018, conversion therapy made limited progress due to stagnant systemic therapies and undefined regimens. However, recent advances in targeted-immunotherapy combinations have significantly enhanced downstaging and survival in uHCC [9]. A multicenter retrospective study of 1,244 advanced HCC cases [10] demonstrated that TACE plus targeted-immunotherapy was associated with enhanced OS, PFS, and objective response rate (ORR) with acceptable safety. However, it remains unclear whether post-conversion surgical resection translates to superior prognostic outcomes. To optimize clinical management of HCC and improve patient prognosis, we evaluated the prognosis of uHCC patients who received conversion therapy and were subsequently treated with either surgery or non-surgical treatment.

Methods

Study design and patient selection

A retrospective screening was conducted for patients initially diagnosed with uHCC at Hunan Provincial People’s Hospital from January 2020 to December 2024. A dedicated multidisciplinary team (MDT) collaboratively assessed and confirmed the unresectable status of HCC. Specifically, the definition of unresectability was based on the following criteria: (1) Future liver remnant (FLR) insufficiency: for cirrhotic patients, residual liver volume was considered insufficient if it represented < 40% of the standard liver volume. For patients without liver cirrhosis, a residual liver volume less than 30% of the standard liver volume was considered insufficient to tolerate resection. (2) Failure to achieve R0 resection: even if the FLR was sufficient, if complete tumor resection with a safe surgical margin could not be achieved, it was determined as unresectable. All patients underwent conversion therapy based on the combination of ICIs and TKIs. Camrelizumab, tislelizumab, and sintilimab were the ICIs used in the study; TKIs included lenvatinib and donafenib. Inclusion criteria: (1) Age between 18 and 80 years. (2) Diagnosed as uHCC. (3) Underwent conversion therapy comprising a combination of TKIs and ICIs, with optional concurrent transarterial chemoembolization during the conversion process. (4) Underwent R0 resection or continued TKIs combined with ICIs therapy after successful conversion. Exclusion criteria: (1) Combined with other malignant tumors. (2) Perioperative death. (3) Patients with incomplete clinical or follow-up data. Tumor response was assessed in accordance with the modified Response Evaluation Criteria in Solid Tumors (mRECIST) [11], details are provided in the supplementary materials. The study was approved by the Ethics Committee of the Hunan Provincial People’s Hospital (Ethics Review Number: [2023]-151).

Data collection

The following data of enrolled patients were collected: gender, age, HBV infection status, presence of liver cirrhosis, albumin (ALB), total bilirubin (TBil), alanine transaminase (ALT), aspartate transaminase (AST), alpha-fetoprotein (AFP) levels before conversion therapy, tumor size and number, tumor response, and duration of conversion therapy before hepatectomy.

Post-conversion evaluation and decision-making

Tumors were evaluated every 2–4 cycles using computed tomography (CT) or magnetic resonance imaging (MRI), with positron emission tomography (PET) selectively used when necessary. All patients were regularly re-evaluated by the same dedicated multidisciplinary team (MDT) following conversion therapy. Patients were defined as having achieved “successful conversion” only when they met the pre-specified criteria for potentially curative R0 resection, as confirmed by cross-sectional imaging and MDT consensus. The feasibility of R0 resection is determined by at least two senior physicians (holding senior professional titles) in our team. Neither the specific conversion therapy regimens nor the patients’ clinical information were disclosed to the evaluating physicians during the assessment, ensuring the blinding of the evaluation process. In case of discrepancies between the two physicians, a third senior physician was consulted for joint discussion to reach a consensus. The Cohen’s kappa (κ) statistic for inter-rater reliability between the two independent observers was calculated as 0.819, indicating strong agreement. Following successful conversion, patients and their families were fully counseled on the potential benefits and risks of both surgical resection and continued systemic therapy. As a result, some patients opted for surgical treatment, while others chose to continue systemic therapy. The final decision to forgo surgery was primarily based on the patients’ informed personal preferences, with common considerations including concerns about perioperative risks, the desire to maintain quality of life, and financial constraints.

Follow-Up

We followed up with all the patients to record their disease progression, recurrence, and survival. In the first two years after treatment, follow-up was arranged every three months, every six months from the second year to the fifth year. Follow-up included AFP level detection, abdominal enhanced CT or MRI. All CT/MRI scans were carried out in our institution in a standardized way. PFS: The time from successful conversion to recurrence, progression, final follow-up or death. OS: The duration from successful conversion to death or final follow-up. In the case that the results of simple radiology were uncertain whether there was recurrence, MDT held a meeting to discuss and determine whether tissue biopsy was needed for further confirmation.

Statistical analysis methods

The differences between groups were analyzed by T test, χ² test, and Fisher’s exact test. Survival curves were generated using the Kaplan-Meier method. Cox regression was used for univariate and multivariate analysis, and P < 0.1 was included in univariate analysis. P value < 0.05 was considered statistically significant. The forest map was generated using R.

Results

Baseline characteristics of patients

A cohort of 67 patients with initially uHCC was recruited for the present study, with the enrollment process detailed in the flow chart (Fig. 1). Among them, 30 patients received surgical treatment (SR) and 37 patients continued to receive the conversion regimen (NSR). The median interval from successful conversion to surgical resection was 16 days. No statistically significant disparities were detected between the two groups with respect to gender distribution, age, hepatitis status, liver function parameters, tumor size, proportion of multiple tumors, presence of cirrhosis, vascular invasion, history of interventional therapy, or proportion of complete remission. Among the SR and NSR groups, 13 and 23 patients had an AFP level ≥ 400 ng/ml, respectively; 18 and 23 patients had multiple tumors, while the median tumor sizes were 68 mm and 65 mm in each group. Vascular invasion was present in 9 patients in the SR group and 14 in the NSR group, and 11 and 13 patients in the two groups achieved complete remission following conversion therapy, respectively (Table 1). To further mitigate the potential selection bias inherent to the retrospective study design, we performed 1:1 propensity score matching (PSM) to balance the baseline characteristics between the two groups. A total of 29 patients were successfully matched in each group, and the detailed baseline data are presented in Supplementary Table 1.

Fig. 1.

Flow diagram for patient selection. uHCC: unresectable hepatocellular carcinoma; SD: stable disease; PD: progressive disease

Table 1.

Patient characteristics

Characteristics	SR	NSR	P value
n	30	37
Sex, n (%)			0.659
Male	23 (76.7%)	30 (81.1%)
Female	7 (23.3%)	7 (18.9%)
Age(years), mean ± sd	55.1 ± 12.8	56.7 ± 9.8	0.564
HBV, n (%)			0.988
No	5 (16.7%)	5 (13.5%)
Yes	25 (83.3%)	32 (86.5%)
ALB(g/l), median (IQR)	39.5 (38.0, 41.8)	40.0 (37.0, 43.0)	0.834
ALT(u/l), median (IQR)	29.0 (24.0, 31.0)	27.0 (25.0, 31.0)	0.874
AST(u/l, median (IQR)	30.0 (26.3, 37.0)	32.0 (26.0, 37.0)	0.766
TBIL(µmol/l), median (IQR)	15.3 (13.6, 17.7)	16.4 (14.4, 18.0)	0.256
AFP(ng/ml), n (%)			0.383
<400	17 (56.7%)	17 (45.9%)
≥ 400	13 (43.3%)	20 (54.1%)
Tumor number, n (%)			0.857
<3	12 (40.0%)	14 (37.8%)
≥ 3	18 (60.0%)	23 (62.2%)
Tumor size(mm), median (IQR)	68.0 (52.5, 95.0)	65.0 (49.0, 93.0)	0.609
Vascular invasion, n (%)			0.502
No	21 (70.0%)	23 (32.2%)
Yes	9 (30.0%)	14 (37.8%)
Cirrhosis, n (%)			0.625
No	12 (40.0%)	17 (45.9%)
Yes	18 (60.0%)	20 (54.1%)
TACE, n (%)			0.457
≥ 2	22 (73.3%)	24 (64.5%)
<2	8 (36.7%)	13 (35.1%)
Treatment response, n (%)			0.897
CR	11 (36.7%)	13 (35.1%)
PR	19 (63.3%)	24 (64.5%)

Survival analysis

In the SR group, the median follow-up duration was 23.5 months, with a range of 6.0 to 48.0 months; for the NSR group, this was 22.0 months (range, 8.0–38.0 months), and the difference between the two groups was not statistically significant (P = 0.489). The 1-, 2-, and 3-year OS rates in the SR group and NSR group were 96.4%, 88.4%, 56.8% and 94.1%, 70.3%, 31.3%, respectively (P = 0.245) (Fig. 2A). Kaplan-Meier curves showed that PFS was significantly prolonged in the surgical group (82.6%, 53.5%, 31.2% vs. 71.1%, 21.6%, 14.4%; P = 0.032) (Fig. 2B). After PSM, PFS in the SR group remained superior to that in the NSR group, whereas there was no statistically significant difference in OS between the two groups. The median OS was 37 months in the SR group and 28 months in the NSR group (P = 0.156); the median PFS was 25 months in the SR group and 16 months in the NSR group (P = 0.045) (Supplementary Fig. 1).

Fig. 2.

Kaplan–Meier survival analysis for OS and PFS. A and (B) showed OS and PFS curves

Handling of tumor recurrence or progression in both groups

During the follow-up period, 14 patients in the SR group developed recurrence, while 24 patients in the NSR group experienced recurrence or progression. Following discussions by the MDT and taking into account the wishes of patients and their families, in the SR group, 2 patients underwent re-resection, 3 received microwave ablation, 7 were treated with TACE or HAIC, and 2 initially received systemic therapy. In the NSR group: 3 patients underwent surgical resection, 5 received microwave ablation, 7 were treated with TACE or HAIC, 7 switched to an alternative systemic therapy regimen, and 2 received best supportive care(Table 2). There was no significant difference in treatment regimens between the two groups (P = 0.598).

Table 2.

Treatments for tumor recurrence or progression

	SR	NSR	P
Recurrence or progression	14	24	0.598
Surgical resection	2	3
Microwave ablation	3	5
TACE or HAIC	7	7
Systemic therapy	2	7
Best supportive care	0	2

Analysis of prognostic factors and forest plots

Cox regression was used for univariate analyses of factors influencing OS and PFS, with forest plots generated to visualize the hazard ratios of each factor (Fig. 3A-B, Supplementary Tables 2–3). Subsequently, to further clarify independent risk factors for OS and PFS, multivariate analyses were conducted (Supplementary Tables 2–3). For OS: PR (HR = 6.0; 95%CI: 2.0-17.7, P < 0.001) and AFP ≥ 400 ng/ml (HR = 2.8, 95%CI: 1.2–6.6, P = 0.021) were identified as risk factors. In contrast, other factors—including surgical resection—were not associated with improved OS in the multivariate analysis. For PFS: PR (HR = 3.3; 95%CI: 1.6–6.9, P = 0.002), tumor number ≥ 3 (HR = 2.4, 95%CI: 1.1-5.0, P = 0.020), and non-surgical treatment (HR = 2.2, 95%CI: 1.1–4.4, P = 0.031) were significantly associated. Other factors were not identified as risk factors for PFS. We also conducted univariate and multivariate analyses on the post-PSM data (Supplementary Tables 4–5). The final results showed that multiple satellite lesions were also an independent risk factor affecting OS in the surgical group, while all other findings were consistent with those before PSM.

Fig. 3.

Forest plots of univariate analyses. A and (B) represent the forest plot of univariate analysis for OS and PFS, respectively

Discussion

This study investigated whether surgical resection (SR) improves survival over non-surgical treatment (NSR) in initially uHCC patients after successful conversion therapy. Retrospective analysis of 67 converted patients showed SR significantly prolonged PFS. We also analyzed OS and PFS prognostic factors, providing evidence for individualized treatment.

For unresectable intermediate-advanced HCC, downstaging via systemic therapies plus local treatments (e.g., TACE) can enhance prognosis. Recent application of next-generation targeted agents (e.g., lenvatinib) and ICIs has enabled breakthroughs in treating intermediate-advanced HCC. Studies show the atezolizumab plus bevacizumab regimen achieves a 36% objective response rate (ORR) with significantly prolonged median OS and PFS [12]. Moreover, triple combination therapy integrating TACE/HAIC with molecular targeted therapy and immunotherapy has been adopted and yielded promising outcomes [13, 14]. However, it remains unclear whether radical surgery or continued original treatment yields a better prognosis for successfully converted patients. Surgeons typically recommend prompt resection after successful conversion to achieve curative effects. In contrast, internists may prefer conservative management, recommending surgery only when serum biomarkers increase or new lesions emerge [15, 16]. A single-center retrospective analysis showed radical resection post-conversion significantly improved PFS but not OS [5]. Currently, conversion rates reported in multiple studies vary considerably (5%–60%) [17–19], likely stemming from inconsistent definitions of unresectability and non-uniform conversion therapy regimens. In our study, 20% of patients achieved remission following conversion treatment. As response rates to systemic therapy improve, a key clinical question arises: Does surgery after successful conversion translate to better patient prognosis? Evidence indicates that systemic or local therapy can downstage uHCC, thereby increasing resectability rates and improving overall survival [20, 21].

In our current study, both the surgical (SR) and non-surgical (NSR) groups met established resectability criteria after treatment initiation, thereby enabling a more balanced and unbiased comparison. Our findings revealed that the SR group achieved significantly superior PFS. Although no statistically significant difference in OS was observed between the SR and NSR groups in this study, the Kaplan-Meier survival curves showed a distinct separation trend after 24 months, suggesting that prolonged follow-up may reveal the long-term OS benefit of surgical intervention. However, the lack of OS benefit with surgery in this study may be closely associated with extensive salvage therapies administered after disease progression/recurrence, which effectively narrowed the survival disparity between the two groups.

Accumulating evidence indicates that VEGF/VEGFR-targeted angiogenesis inhibition enhances ICIs’ anti-tumor effects [22, 23]. Emerging data highlight the value of combining locoregional therapies with high-efficacy TKIs and ICIs for downstaging or reducing tumor size in hepatobiliary malignancies [24]. These studies suggest synergistic effects from multimodal combinations. Our center’s results show that most successfully converted patients received early local therapy combined with TKIs and ICIs, enhancing tumor responsiveness while promoting long-term survival. Although such therapeutic regimens show potential in conversion therapy, well-conducted trials are needed to deliver high-level clinical evidence.

For patients with resectable tumors post-conversion therapy, optimal radical surgery timing is debated in oncology. Some advocate immediate surgery after tumor remission (1–9-month variable observation) [9], with considerations including drug onset and withdrawal timing. In this study, among 30 surgical patients, the 12-month postoperative recurrence rate was 16.6% (5/30). Among 37 non-surgical patients (remission post-conversion but no surgery), the 1-year recurrence rate was 21.6% (8/37), significantly higher than the surgical group. One SR patient achieved successful conversion following 9 months of targeted therapy combined with immunotherapy. During the preoperative drug withdrawal period, the patient’s AFP level increased; however, the patient still elected to undergo surgical resection. Recurrence was documented at 6 months postoperatively, and the patient succumbed to the disease several months later. This may be because: (1) tumor volume shrank post-conversion, but serum markers rose/remained elevated after drug withdrawal (suggesting residual tumor activity); (2) surgical destruction of the tumor microenvironment may increase postoperative recurrence risk. Current studies report 12–16.8-month postoperative recurrence rates of up to 36.6% in successfully converted patients [25, 26]. Thus, surgical timing depends on detailed tumor response assessment (radiological, serological) and dynamic observation of response stability. Based on our research, we believe that surgical resection can be actively considered for patients who meet the resectable standard after conversion therapy.

In this study, Cox multivariate analysis revealed that AFP ≥ 400 ng/ml and partial response were independent prognostic risk factors affecting patients’ OS. As a common HCC tumor marker, AFP levels correlate with tumor activity and prognosis—high AFP usually indicates high malignancy and poor prognosis [27, 28]. The prognosis of patients achieving only PR after conversion therapy is poor, which may be due to high baseline risk, such as late stage, high tumor load or unfavorable molecular subtypes. The potential mechanism underlying this result may be as follows: Tumor cells in patients achieving CR exhibit higher sensitivity to conversion therapy, leading to maximal reduction or even complete elimination of tumor burden, thus resulting in a better prognosis. In contrast, patients with PR still have residual viable tumor tissue, which embodies an inherent risk of disease recurrence and progression. This suggests that more aggressive treatment strategies (such as intensified postoperative adjuvant therapy and close follow-up monitoring) should be adopted for these patients to reduce the risk of tumor progression. Our results also found that multiple tumor foci (tumor number ≥ 3) were independent risk factors for PFS. Studies have shown that multifocal tumors may be closely related to the biological behavior and invasiveness of tumors [29–31]. Compared with single-tumor patients, patients with multifocal HCC show more biological behaviors of malignant tumors, and have a higher possibility of intrahepatic metastasis or residual tumor cells, which leads to a significantly higher postoperative recurrence rate and mortality. These findings provide an important reference for clinical prognosis evaluation and individualized treatment plans. Studies have also shown that accurate tumor counting in clinical practice is not only helpful for prognosis judgment, but also can guide the choice of adjuvant therapy [32, 33].

However, our research has some limitations. First of all, this is a single-center retrospective study with a relatively small sample size, which may be affected by selection bias and confounding factors. Second, the follow-up duration was insufficiently long, which may have limited our ability to capture the long-term survival benefit of surgical intervention and the late recurrence risk of tumors. Third, the diversity of patients’ conversion treatment and non-surgical treatment schemes in the study may interfere with the research results. Future research should include multi-center, large-sample, prospective randomized controlled trials with extended follow-up periods to further verify our conclusions, explore the impact of different conversion/non-surgical schemes on prognosis, and formulate more accurate and effective treatment strategies.

In conclusion, for uHCC patients who attain resectable status following conversion therapy, surgical intervention is superior to non-surgical management in improving PFS outcomes. However, no statistically significant difference was observed in OS among patients, which may be partially attributed to the relatively small sample size of our study. In addition, tumor number and AFP level are independent risk factors for PFS and OS, respectively. Clinicians should make individualized treatment strategies according to the specific conditions of patients and comprehensively consider various factors to improve the survival rate and quality of life.

Authors’ contributions

Zengpeng Sun and Zhiguo Tan contributed to the study concepts and design. Yutao Wang, Jia Zhou, and Xu Chen acquired the data. Ou Li and Chuang Peng were responsible for quality control of data and algorithms, and manuscript review. Zengpeng Sun and Zhiguo Tan analyzed and interpreted the data, with Zhiguo Tan conducting statistical analysis, preparing and editing the manuscript, and Yutao Wang assisting in manuscript preparation.

Funding

This work was supported by the Leading Talent Project of Hunan Provincial People’s Hospital (20230327-1012); Hunan Provincial Natural Science Foundation (2024JJ6276); Hunan Provincial Department of Science and Technology Project (2024JK2110); Cancer-Free Initiative of Sichuan Xinxin Charity Foundation (SCF251028).

Data availability

Data and materials were included in the manuscript.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Hunan Provincial People’s Hospital (Ethics Review Number: [2023]-151) in accordance with the Declaration of Helsinki. All patients voluntarily participated and provided written informed consent.

Consent for publication

The authors give consent for publication of the manuscript.

Competing interests

The authors declare no competing interests.