Conversion Therapy Based on TACE/HAIC-Based Treatment to Improve the Therapeutic Effect of Initially Unresectable Hepatocellular Carcinoma

Abstract

Objectives

The main objectives of this research are to evaluate the outcomes of patients with initially unresectable hepatocellular carcinoma (HCC) who received transcatheter arterial chemoembolization (TACE)/hepatic artery infusion chemotherapy (HAIC)-based combination therapy and to investigate the effects of liver resection following comprehensive conversion therapy on the short-term benefits and long-term survival of these patients.

Materials and Methods

A total of 301 initially unresectable HCC patients who received TACE/HAIC-based combination therapy between January 2019 and December 2021 were retrospectively reviewed. The study analyzed the conversion rate to resection, changes in tumor burden after treatment, and the survival outcomes.

Results

The study found that 20.9% (63/301) of initially unresectable HCC patients were able to undergo liver resection. The conversion resection rate among all patients was 38.2% (29/76) and 17.3% (23/132) for those with Barcelona Clinic Liver Cancer (BCLC) stage A and C. Patients who underwent surgery achieved promising outcomes with a pathological complete response (pCR) rate of 31.7% (20/63) and a 100% R0 resection rate. Kaplan-Meier survival analysis showed that patients who had successful surgery after conversion therapy had significantly longer median overall survival (OS) (not reached vs. 58.5 months) and progression-free survival (PFS) (42.83 months vs. 9.7 months) compared to those who did not (both p < 0.05). Additionally, patients achieving radiographic complete response (CR) had significantly better OS and PFS than those who did not. Multivariable logistic regression analysis showed that age (OR = 0.95, p < 0.001), positive HBsAg expression (OR = 0.34, p = 0.011), and alpha-fetoprotein levels ≥400 (OR = 0.49, p = 0.045), ECOG PS score of 1 (OR = 0.43, p = 0.038), BCLC stage B (OR = 0.23, p < 0.001) and stage C (OR = 0.44, p = 0.045), systemic inflammation response index (OR = 0.73, p = 0.018) were independent predictors for successful conversion surgery (all p < 0.05).

Conclusion

Patients with initially unresectable HCC can achieve promising curative effects and conversion resection rates with TACE/HAIC-based comprehensive therapy. More importantly, patients who undergo liver resection following conversion resection had significantly better long-term survival.

Introduction

Hepatocellular carcinoma (HCC) is the sixth most common cancer and the third leading cause of cancer-related death worldwide, accounting for approximately 4.3% of all new cases and 7.8% of all cancer-related deaths [1]. The development of HCC is frequently associated with hepatitis B virus infection, which affects over 350 million individuals worldwide [2]. In China, more than 85% of HCC patients also have a diagnosis of hepatitis B virus infection [3, 4].

Liver resection is considered one of the most effective treatments for HCC [5, 6]. Patients with early-stage-HCC have the potential to achieve a median survival time of over 5 years with radical treatments such as liver resection, ablation, or liver transplantation, according to the Barcelona Clinic Liver Cancer (BCLC) Guidelines [7] and the Chinese Guidelines for the diagnosis and treatment of Primary liver Cancer (2022 edition) [8]. Unfortunately, the BRIDGE study revealed that over 60% of HCC patients in China are diagnosed at intermediate or advanced stages (Chinese National Liver Cancer [CNLC] stage II or III), leading to a poor prognosis with a 5-year overall survival (OS) rate of only 3%, which makes curative treatments like hepatectomy or liver transplantation unfeasible for these patients [7, 9, 10]. As a result, systemic therapy has become the standard treatment for most HCC patients. While many novel anticancer agents have shown effectiveness in treating advanced or unresectable HCC in terms of tumor response and patient survival, the OS rate for HCC patients undergoing systemic therapy remains poor [11].

Conversion therapy is a novel therapeutic strategy for initially unresectable HCC that aims to decrease tumor size and achieve tumor downstaging through a combination of local and/or systemic therapies. This strategy also aims to shrink tumors and provide an opportunity for curative resection for patients with unresectable or borderline resectable malignancies [12–15]. In recent years, a combination of systemic and local treatments has allowed some patients with unresectable HCC to undergo successful liver resection. Transcatheter arterial chemoembolization (TACE) is a local therapy that combines embolization and chemotherapy to reduce tumor size and control tumor thrombus by selectively blocking tumor-feeding arteries and delivering high concentrations of chemotherapy drugs [16–19]. TACE has become a primary treatment option for patients with locally advanced liver malignancies who are not suitable for liver resection or other ablative treatments [7, 20–26].

Unlike TACE, hepatic artery infusion chemotherapy (HAIC) uses catheter technology to directly inject anticancer drugs into the main blood vessels supplying the tumor through the hepatic artery [27]. Theoretically, HAIC is more effective than systemic chemotherapy because hepatic artery perfusion of anticancer drugs can directly deliver large doses of drugs to HCC with high vascular density, including micrometastases that cannot be detected by imaging and tumors without obvious arterial supply [28]. The intrahepatic first-pass effect leads to a lower systemic drug level of HAIC than systemic administration, reducing toxic side effects and adverse reactions.

Numerous studies have demonstrated that combining TACE/HAIC with other therapies can improve the long-term survival of patients with unresectable HCC [19, 29–39]. TACE has been shown to effectively control tumor progression and promote future liver remnant (FLR) growth, potentially allowing for radical surgery in patients with large HCC [30]. The TACTICS trial has provided further evidence that combining TACE with sorafenib can significantly increase progression-free survival (PFS) and OS in unresectable HCC patients [34]. Additionally, the combination therapy of lenvatinib and TACE/HAIC has shown higher response rates and more favorable survival outcomes compared to TACE/HAIC alone or lenvatinib monotherapy in patients with a high tumor burden [36, 37, 40]. Similarly, in HCC patients with portal vein invasion, the median OS after HAIC combined with sorafenib was significantly longer than that with sorafenib monotherapy [19, 39, 41, 42]. Studies also pointed out that compared to TACE alone, HAIC combined with TACE further improved the objective response rate (ORR), OS, and safety in patients with unresectable HCC [43, 44]. Moreover, the combination of TACE/HAIC with targeted and immunotherapy achieved better ORR, OS, and PFS [38, 45]. Despite these advancements, there are still many controversies regarding the optimal timing of transarterial therapy, the most effective treatment combinations, the role of surgery after conversion, and the necessity of surgery after complete radiological response. In this study, we evaluated the short- and long-term outcomes of TACE/HAIC-based conversion therapy in initially unresectable HCC and tried to identify the specific patient population that could benefit from this treatment modality.

Methods

Patients

The present study retrospectively reviewed 301 patients with initially unresectable HCC who underwent combination therapy involving TACE/HAIC at the First Affiliated Hospital of Sun Yat-sen University between January 2019 and December 2021. Patients were enrolled according to the following inclusion criteria: (1) patients aged between 18 and 85 years; (2) confirmed diagnosis of initially unresectable HCC through imaging, histology, or cytology according to the Chinese guidelines [8]; (3) assessment of unresectable HCC by a multidisciplinary treatment (MDT) team as intermediate or advanced stage HCC (BCLC stage B/C, or CNLC stage II/II), or as BCLC stage A/CNLC stage Ib not suitable for curative liver resection due to inadequate FLR (cirrhosis patients <40%; patients with no cirrhosis <30%) [15]; (4) patients received at least one local interventional therapy (TACE/HAIC) with or without targeted therapy and immunotherapy; (5) presence of at least one evaluable lesion suitable for tumor burden assessment using modified Response Evaluation Criteria In Solid Tumors (mRECIST) [46] and Response Evaluation Criteria In Solid Tumors version 1.1 (RECIST v1.1) criteria [47]; (6) Eastern Cooperative Oncology Group performance status (ECOG-PS) of 0–1; (7) liver function categorized as Child-Pugh A or B.

The exclusion criteria were as follows: (1) patients diagnosed with intrahepatic cholangiocarcinoma or combined HCC; (2) patients with primary HCC and extrahepatic metastasis; (3) no interventional therapy performed or only salvage TACE received as initial therapy; (4) patients with a history of other types of malignant tumors; (5) patients with severe complications, cardiovascular disease, renal dysfunction, or severe encephalopathy, and (6) patients with insufficient preoperative laboratory or pathological data, including unknown initial treatment in other hospitals.

The study protocol adhered to the ethical guidelines of the World Medical Association Declaration of Helsinki. The Institutional Review Board of the First Affiliated Hospital of Sun Yat-sen University reviewed and approved the study protocol (ethical approval number: LunShen [2023] No. 399), and the requirement for written informed consent was waived due to the retrospective nature of the study.

Data collection during the perioperative phase included a comprehensive clinical examination, laboratory evaluations comprising complete blood count, liver and kidney function tests, and alpha-fetoprotein (AFP) levels, along with radiological assessments such as abdominal ultrasonography and triphasic computed tomography (CT) and/or magnetic resonance imaging (MRI).

CNLC Stage System

The CNLC system classifies HCC into seven stages: CNLC stage Ia, Ib, IIa, IIb, IIIa, IIIb, and IV, based on a combination of performance status (PS), Child-Pugh liver function class, tumor size and number, vascular invasion, and extrahepatic spread. The specific stages are as follows: CNLC stage Ia: PS 0–2, Child-Pugh A/B, single tumor, diameter ≤5 cm, no vascular tumor thrombus or extrahepatic metastasis. CNLC stage Ib: PS 0–2, Child-Pugh A/B, single tumor, diameter >5 cm or 2–3 tumors, maximum diameter ≤3 cm, no vascular tumor thrombus or extrahepatic metastasis. CNLC stage IIa: PS 0–2, Child-Pugh A/B, 2–3 tumors, maximum diameter >3 cm, no vascular tumor thrombus or extrahepatic metastasis. CNLC stage IIb: PS 0–2, Child-Pugh A/B, tumor number ≥4, no matter the diameter of the tumor, with no vascular tumor thrombus or extrahepatic metastasis. CNLC stage IIIa: PS 0–2, Child-Pugh A/B, regardless of tumor size and number, imaging showed vascular tumor thrombus without extrahepatic metastasis. CNLC stage IIIb: PS 0–2, Child-Pugh A/B, regardless of the size and number of tumors, regardless of the presence or absence of vascular tumor thrombus on imaging, but with extrahepatic metastasis. CNLC stage IV: PS 3–4 and Child-Pugh grade C, regardless of tumor size and number, regardless of the presence or absence of vascular tumor thrombus on imaging, regardless of extrahepatic metastasis [8].

Conversion Therapy Strategy and Patient Selection

All patients in this study initially received at least one local interventional therapy, either TACE or HAIC, with or without systemic therapy (e.g., targeted therapy or immunotherapy). The comprehensive treatment regimens included: (1) TACE/HAIC monotherapy (TACE/HAIC), (2) TACE/HAIC combined with targeted therapy (mainly tyrosine kinase inhibitors [TKIs]) (TACE/HAIC+TKIs), and (3) TACE/HAIC combined with both TKIs and immunotherapy (immune checkpoint inhibitors [ICIs]) (TACE/HAIC+TKIs+ICIs). The treatment decisions were determined by MDT discussions after thorough evaluation.

All patients were initially deemed unresectable, and MDT-based assessment determined whether they were eligible to initiate TACE/HAIC-based conversion therapy, according to the following criteria: (1) presence of large tumor burden, multifocal disease, or macrovascular invasion that precluded curative resection at diagnosis; (2) preserved liver function (Child-Pugh A or selected B), and ECOG PS of 0–1; (3) absence of extensive extrahepatic metastases, or presence of limited lesions considered controllable or resectable, and (4) potential for achieving resectability after tumor shrinkage or vascular involvement regression following locoregional and/or systemic therapy.

The primary intent of initiating conversion therapy was tumor downstaging, with the goal of achieving R0 surgical resection. For patients who failed to meet resection criteria after treatment, the strategy will followed the treatment recommendation from BCLC/CNLC staging, aiming to stabilize disease, improve quality of life, and extend survival.

Transarterial Interventional and Systemic Therapies

The transarterial interventional therapy includes TACE and HAIC. Transarterial intervention was performed every 3–4 weeks, with evaluation by CT or MRI before each session.

TACE included both conventional TACE and drug-eluting bead TACE, following previously reported protocols [48]. HAIC was performed using the FOLFOX regimen through percutaneous catheterization without an implantable port system [19, 49–52].

Systemic therapy included a range of TKIs (e.g., sorafenib, lenvatinib, apatinib, regorafenib) and ICIs (e.g., camrelizumab, atezolizumab, sintilimab, toripalimab). Detailed drug information is provided in online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000547725).

Hepatectomy and Follow-Up

Liver resection was performed after meeting the criteria for resectable HCC and obtaining informed consent. The criteria for curative resection included the following factors: (a) R0 resection could be achieved with adequate residual liver volume and function (patients without liver cirrhosis having a remnant liver volume ≥30% of standard liver volume, or patients with liver cirrhosis having a remnant liver volume ≥40% of standard liver volume); (b) no macrovascular invasion (no tumor thrombus in the main portal vein or the first branch of portal vein), or portal vein tumor thrombus disappeared or reduced to resectable range; (c) single or few (≤3) tumor lesions, or multiple lesions confined to one liver lobe, and (d) no other contraindications to partial hepatectomy were found [8]. In addition, patients who did not meet surgical criteria continued to receive combination therapy until the disease has progressed and toxicity is no longer tolerated.

Patients were followed 1 month after hospital discharge and regularly followed every 3 months for the first 2 years post-surgery, and every 6 months after that. At each follow-up, a physical examination, abdominal contrast-enhanced CT or MRI, and AFP examination were performed.

Pathological Response

Pathological response was evaluated in all patients who underwent conversion surgery. Resected specimens were fixed in formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. The extent of tumor necrosis was assessed as the proportion of necrotic tissue relative to the total tumor area.

Pathological response was defined by the percentage of nonviable tumor cells in the resected specimens. Pathological complete response (pCR) was defined as the absence of any viable tumor cells in all assessable lesions, corresponding to 100% tumor necrosis, in accordance with recently established criteria for HCC conversion therapy studies. Major pathological response (MPR) was defined as ≥70% tumor necrosis [53].

To minimize interobserver variability, standardized reporting standards were used for pathology reports, and all pathological assessments were performed using a standardized reporting protocol. Each case was independently reviewed by at least two experienced hepatopathologists. In cases of discrepancy, a consensus diagnosis was reached through joint discussion. Although formal inter-observer agreement analysis was not conducted in this study, prior multicenter research has demonstrated high concordance among expert pathologists when using standardized criteria [53, 54], providing indirect support for the reliability of our pathological evaluation.

Efficacy Assessment and Follow-Up

Patients were regularly followed up every 3 months for 2 years post-surgery and then every 6 months after that. At each follow-up, physical examination, abdominal contrast-enhanced CT or MRI, and AFP examination were performed. Recurrence is diagnosed based on AFP, radiology, and/or pathological diagnosis (tissue obtained through ultrasound-guided fine-needle aspiration). Tumor response was assessed according to mRECIST and RECIST v1.1 guidelines [46, 47], which categorize response as complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). The best tumor response during treatment was recorded and reassessed at the end of therapy. The primary outcomes of this study were OS and PFS. OS was defined as the date of the initial transarterial interventional therapy to the date of death. PFS was calculated from the time of receiving treatment to disease progression or death, whichever occurred first, and progression was determined based on contrast-enhanced CT or MRI scans following mRECIST criteria [47]. Other endpoints included surgical conversion rate and pathological CR (pCR) (with a definition of the complete absence of residual viable tumor cells after HE staining from the completely resected specimen) rate. The second endpoints included ORR and disease control rate (DCR). ORR represented the percentage of patients achieving CR or PR lasting more than 4 weeks from the initial radiological confirmation. DCR included patients with SD in addition to those with ORR.

Safety was assessed by the physical examination, vital signs, hematologic and biochemical laboratory tests, and the incidence of adverse events (AEs) from the start of the study treatment to after the last dose or the end of the study. The treatment-related AEs (TRAEs) were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 5.0, and the postoperative complications were graded according to the Clavien-Dindo classification system [55, 56].

Biomarkers of Peripheral Blood Inflammatory Response

The biomarkers of peripheral blood inflammatory response included neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), aspartate aminotransferase-to-neutrophil ratio (ANRI), and systemic inflammation response index (SIRI). Perioperative data, including laboratory evaluation (including complete blood count, liver functions, kidney functions, and AFP, and radiological assessment (including abdominal ultrasonography and triphasic CT and/or MRI), were collected from electronic medical records within 7 days before the first TACE/HAIC treatment. Neutrophil-to-lymphocyte ratio was calculated as the neutrophil count/lymphocyte count (10⁹/L); platelet-to-lymphocyte ratio was calculated as the platelet count/lymphocyte count (10⁹/L); aminotransferase-to-the neutrophil ratio was calculated as AST/neutrophil count (10⁹/L); SIRI was calculated as the neutrophil count × monocyte count/lymphocyte count (10⁹/L) [57].

Propensity Score Matching

Propensity score matching (PSM) was conducted using R software (version 4.0.5; R Foundation, Vienna, Austria) to address the selection bias between patients in the non-surgery and surgery groups. In this study, the nearest neighbor matching (NNM) method was utilized for 1:4 ratio with a caliper value of 0.15, a commonly used value in previous studies, to improve matching accuracy and maximize case inclusion. The variables considered in the PSM included the following factors: sex, age, clinically significant portal hypertension (CSPH), ALT, AST, Child-Pugh grade, pre-conversion tumor size, pre-conversion BCLC stage, interventional therapy cycle, and conversion regimens.

Statistical Analysis

In this study, continuous variables were presented as either the median with interquartile range or the mean with standard deviation. Differences between groups were compared using Student’s t test or the Mann-Whitney U test as appropriate. Categorical variables were described using numbers and percentages, and differences were assessed using the chi-square test or Fisher’s exact test if necessary. A significant level of p < 0.05 was used for all tests. Logistic regression was employed for univariate and multivariate analysis to identify factors influencing prognosis. The Kaplan-Meier survival analysis method was used to generate survival curves, with the Log-Rank test used for comparison. All statistical analyses were performed using R software (version 4.0.5; R Foundation, Vienna, Austria).

Results

Clinicopathologic Characteristics of Patients

The patient characteristics of the non-surgery and surgery groups are summarized in Table 1. The study enrolled a total of 301 patients with unresectable HCC who met the specified inclusion criteria (Fig. 1). Upon initial diagnosis, all patients were considered unresectable and underwent TACE/HAIC-based comprehensive therapy. Following this treatment, 63 out of 301 (20.9%) patients proceeded to hepatectomy after comprehensive conversion therapy. Of the patients, the median age was 55.38 years, with 92.0% (277/301) being male. According to the CNLC stage classification, the distribution of patients was as follows: 6.3% at stage Ia, 18.6% at stage Ib, 6.3% at stage IIa, 24.6% at stage IIb, and 44.2% at stage IIIa. The majority of patients (85.7%, 258/301) had Child-Pugh grade A liver function, and BCLC staging showed that 25.2% (76/301) were at stage A, 30.9% (93/301) were at stage B, and 43.9% (132/301) were at stage C, respectively. Approximately 41.2% (124/301) of patients had tumors measuring 10 cm or larger. Additionally, we also found that the proportion of patients in the non-surgery group who received four or more interventional therapy cycles was significantly higher than that in the surgery group (40.8% vs. 15.9%, p < 0.001) (details shown in Table 1).

Table 1.

Characteristics of patients with initially unresectable HCC

Variable	N = 301	Non-surgery group (n = 238)	Surgery group (n = 63)	p value
Age, years	55.38 (48.00, 64.05)	55.84 (49.06, 64.73)	52.25 (43.39, 57.28)	0.007
Female sex, n (%)	24 (8.0)	16 (6.7)	8 (12.7)	0.123
HBsAg, n (%)	256 (85.0)	210 (88.2)	46 (73.0)	0.005
ALT, U/L	37.00 (24.00, 58.00)	39.50 (26.00, 61.75)	29.00 (19.50, 46.00)	0.002
AST, U/L	51.00 (33.00, 82.00)	56.50 (37.00, 90.75)	35.00 (25.50, 43.00)	<0.001
GGT, U/L	122.00 (70.00, 248.00)	131.50 (76.00, 264.00)	98.00 (54.50, 170.00)	0.009
ALB, g/L	35.10 (32.40, 38.30)	35.00 (32.30, 38.10)	35.40 (33.25, 39.10)	0.138
PLT, ×109/L	186.00 (121.00, 242.00)	179.00 (111.25, 240.00)	202.00 (140.00, 266.50)	0.058
PT, s	12.40 (11.60, 13.40)	12.40 (11.50, 13.50)	12.20 (11.60, 12.95)	0.129
INR	1.06 (0.99, 1.15)	1.06 (0.99, 1.16)	1.05 (0.99, 1.12)	0.215
AFP ≥400, n (%)	131 (43.5)	114 (47.9)	17 (27.0)	0.004
NLR	2.38 (1.75, 3.53)	2.47 (1.83, 3.76)	2.05 (1.58, 2.68)	0.004
PLR	119.25 (80.69, 175.65)	116.57 (79.57, 172.98)	129.39 (84.94, 197.63)	0.192
ANRI	14.60 (8.86, 26.69)	15.44 (9.54, 28.63)	11.49 (7.29, 19.83)	0.009
SIRI	1.23 (0.78,2.24)	1.30 (0.84,2.52)	0.98 (0.65,1.75)	0.005
CSPH, n (%)	86 (28.6)	82 (34.5)	4 (6.3)	<0.001
Child-Pugh, n (%)	​	​	​	0.044
Grade A	258 (85.7)	199 (83.6)	59 (93.7)	​
Grade B	43 (14.3)	39 (16.4)	4 (6.3)	​
Pre-conversion tumor size, cm	8.60 (5.37, 12.00)	8.40 (5.30, 12.20)	8.70 (6.15, 10.75)	0.843
Pre-conversion tumor size ≥ 10 cm, n (%)	124 (41.2)	100 (42.0)	24 (38.1)	0.666
Pre-conversion tumor number, n (%)	​	​	​	<0.001
1	129 (42.9)	89 (37.4)	40 (63.5)	​
2	39 (13.0)	29 (12.2)	10 (15.9)	​
3	8 (2.7)	6 (2.5)	2 (3.2)	​
≥4	125 (41.5)	114 (47.9)	11 (17.5)	​
Conversion regimens, n (%)	​	​	​	0.651
TACE/HAIC	130 (43.2)	106 (44.5)	24 (38.1)	0.457
TACE	77 (25.6)	62 (26.1)	15 (23.8)	​
HAIC	6 (2.0)	4 (1.7)	2 (3.2)	​
TACE+HAIC	47 (15.6)	40 (16.8)	7 (11.1)	​
TACE/HAIC+TKIs	99 (32.9)	77 (32.4)	22 (34.9)	0.174
TACE+TKIs	58 (19.3)	48 (20.2)	10 (15.9)	​
HAIC+TKIs	2 (0.7)	1 (0.4)	1 (1.6)	​
TACE+HAIC+TKIs	39 (13.0)	28 (11.8)	11 (17.5)	​
TACE/HAIC+TKIs+ICIs	72 (23.9)	55 (23.1)	17 (27.0)	0.443
TACE+TKIs+ICIs	20 (6.6)	15 (6.3)	5 (7.9)	​
HAIC+TKIs+ICIs	7 (2.3)	4 (1.7)	3 (4.8)	​
TACE+HAIC+TKIs+ICIs	45 (15.0)	36 (15.1)	9 (14.3)	​
Interventional therapy, n (%)	​	​	​	0.194
TACE	155 (51.5)	125 (52.5)	30 (47.6)	​
HAIC	15 (5.0)	9 (3.8)	6 (9.5)	​
TACE+HAIC	131 (43.5)	104 (43.7)	27 (42.9)	​
Interventional therapy cycle, n (%)	​	​	​	<0.001
1 cycle	62 (20.6)	39 (16.4)	23 (36.5)	​
2 cycles	86 (28.6)	66 (27.7)	20 (31.7)	​
3 cycles	46 (15.3)	36 (15.1)	10 (15.9)	​
≥4 cycles	107 (35.5)	97 (40.8)	10 (15.9)	​
Pre-conversion BCLC stage, n (%)	​	​	​	<0.001
Stage A	76 (25.2)	47 (19.7)	29 (46.0)	​
Stage B	93 (30.9)	82 (34.5)	11 (17.5)	​
Stage C	132 (43.9)	109 (45.8)	23 (36.5)	​
Pre-conversion CNLC stage, n (%)	​	​	​	<0.001
Stage Ia	19 (6.3)	14 (5.9)	5 (7.9)	​
Stage Ib	56 (18.6)	32 (13.4)	24 (38.1)	​
Stage IIa	19 (6.3)	14 (5.9)	5 (7.9)	​
Stage IIb	74 (24.6)	68 (28.6)	6 (9.5)	​
Stage IIIa	133 (44.2)	110 (46.2)	23 (36.5)	​

NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; ANRI, aspartate aminotransferase-neutrophil ratio; SIRI, systemic inflammatory response index; CSPH, clinically significant portal hypertension; TKIs, tyrosine kinase inhibitors; ICIs, immune checkpoint inhibitors; BCLC, Barcelona Clinical Liver Cancer; CNLC, Chinese National Liver Cancer.

Fig. 1.

Flowchart of patient selection for conversion therapy in patients with initially unresectable primary hepatic carcinoma. A total of 886 patients were initially assessed. After exclusions, a total of 301 patients (n = 301) were included in this study, of whom 63 patients have successfully received conversion surgery (n = 63) following conversion therapy, and 238 patients did not receive conversion surgery (n = 238).

Among 301 patients receiving conversion therapy, 90.7% (273/301) experienced at least one TRAE. The most frequent events included elevated ALT/AST (69.4%), increased white blood cell/neutrophil count (71.8%), gastrointestinal symptoms (23.6%), and decreased platelet count (15.3%). Grade ≥3 TRAEs were observed in 20.6% (62/301) of patients. In the surgical cohort (n = 63), 66.7% (42/63) experienced postoperative complications, while 19.0% (12/63) had events of Clavien-Dindo grade ≥3. The most common grade ≥3 complication was abdominal/pleural effusion requiring drainage (15.9%). Importantly, no postoperative treatment-related deaths occurred (details shown in online suppl. Tables 2, 3).

Conversion Therapy for Unresectable HCC

ORR and DCR

According to the mRECIST criteria, the surgery group showed favorable outcomes, with 27.0% (17/63) of patients achieving CR, 57.1% (36/63) showing PR, and 14.3% (9/63) experiencing SD. The ORR in the surgery group was 84.1% (53/63), and the DCR reached 98.4% (62/63). In contrast, the non-surgery group had inferior outcomes, with a significantly lower ORR (69.3% vs. 84.1%) and DCR (88.2% vs. 98.4%) (both p < 0.05). According to the RECIST v1.1 criteria, the patient’s best responses were 0 CR, 19.0% (12/63) PR, 76.2% (48/63) SD, and 4.8% (3/63) PD in the surgery group. And in the non-surgery group, there were 0 CR, and 20.2% (48/238), 60.9% (145/238), and 18.9% (45/238) showed PR, SD, and PD. The ORR and DCR evaluated by RECIST v1.1 showed that the DCR rate of the surgical group was relatively higher compared with the non-surgical group (95.2% vs. 81.1%, p = 0.006), while there was no significant difference in the ORR rate (19.0% vs. 20.2%, p = 1.000) (Table 2).

Table 2.

Efficacy evaluated by mRECIST and RECIST v1.1 criteria after conversion therapy

Variable	N = 301	mRECIST criteria			N = 301	RECIST v1.1 criteria
		non-surgery (n = 238)	surgery (n = 63)	p value		non-surgery (n = 238)	surgery (n = 63)	p value
ORR	218 (72.4)	165 (69.3)	53 (84.1)	0.026	60 (19.9)	48 (20.2)	12 (19.0)	1.000
DCR	272 (90.4)	210 (88.2)	62 (98.4)	0.014	253 (84.1)	193 (81.1)	60 (95.2)	0.006
CR	69 (22.9)	52 (21.8)	17 (27.0)	0.402	0 (0.0)	0 (0.0)	0 (0.0)	–
PR	149 (49.5)	113 (47.5)	36 (57.1)	0.203	60 (19.9)	48 (20.2)	12 (19.0)	1.000
SD	54 (17.9)	45 (18.9)	9 (14.3)	0.464	193 (64.1)	145 (60.9)	48 (76.2)	0.027
PD	26 (8.6)	25 (10.5)	1 (1.6)	0.022	48 (15.9)	45 (18.9)	3 (4.8)	0.006
NE	3 (1.0)	3 (1.3)	0 (0.0)	–	–	–	–	–

ORR, objective response rate; DCR, disease control rate; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, non-evaluated.

Prognosis after Successful Conversion Therapy

For all patients in the study, the median follow-up time for OS was 20.47 months (95% CI: 15.80–24.13) (online suppl. Fig. 1A), the median OS was not reached (95% CI: 58.5-not reached) (Fig. 2a). And the median follow-up time for PFS was 26.33 months (95% CI: 16.83–36.17) (online suppl. Fig. 1B), the median PFS was 13.3 months (95% CI: 10.6–18.4) (Fig. 2b). Kaplan-Meier survival analysis showed that patients who successfully received conversion surgery had significantly longer median OS (not reached vs. 58.5 months) and PFS (42.83 months vs. 9.7 months) compared to those who did not receive conversion surgery (both p < 0.05) (Fig. 3a, b).

Fig. 2.

Overall survival (OS) and progression-free survival (PFS) curves of the entire cohort. a OS. b PFS.

Fig. 3.

Kaplan-Meier survival analysis comparing overall survival (OS) and progression-free survival (PFS) between patients who underwent surgery and those who did not. a, b OS and PFS before propensity score matching (PSM) between surgery and non-surgery groups. c, d OS and PFS before PSM between patients who underwent surgery and those who did not in patients who meet surgical criteria. e, f OS and PFS after PSM between patients who underwent surgery and those who did not in patients who meet surgical criteria.

In this study, the median interval between initiation of conversion therapy and surgery was 2.90 months (95% CI: 1.80–6.03 months). Notably, of the 301 patients, 90 (29.9%) were assessed as meeting surgical resection criteria after conversion therapy. However, only 63 (70%) of these patients ultimately underwent liver resection, while the remaining 27 (30%) did not. Baseline analysis showed that the incidence of CSPH was significantly higher in the non-surgery group than in the surgery group (48.1% vs. 6.3%, p < 0.001). Additionally, the surgery group had larger tumor diameters both before and after conversion therapy (pre-conversion: 8.70 cm vs. 6.70 cm, p = 0.018; post-conversion: 7.20 cm vs. 4.60 cm, p = 0.001), and a higher proportion of tumors ≥10 cm in size after conversion therapy (28.6% vs. 3.7%, p = 0.010) (online suppl. Table 4). Upon further analysis, following PSM, the two groups were well balanced in terms of sex, age, tumor size and number, presence of CSPH, Child-Pugh grade, number of interventional therapy cycles, and conversion therapy regimens (online suppl. Table 5). Kaplan-Meier analysis revealed that among patients who met resection criteria, those who underwent surgery had significantly improved OS and PFS compared to those who did not before and after PSM (all p < 0.05) (Fig. 3c–f), highlighting the critical role of surgery in maximizing long-term outcomes once resectability is achieved.

Notably, among patients who underwent hepatectomy after conversion therapy, the majority were classified as BCLC stage A and C, with conversion resection rates of 38.2% (29/76) and 17.3% (23/132), respectively. Regarding the CNLC stage classification, most patients who had hepatectomy after conversion were primarily concentrated in stages Ia, Ib, and IIa, with conversion resection rates of 26.3% (5/19), 42.9% (24/56), and 21.1% (4/19), respectively. Moreover, despite being earlier in their disease progression, patients classified as BCLC stage B and CNLC stage IIb had the lowest conversion resection rates. Specifically, only 11.8% (11/93) of BCLC stage B and 8.1% (6/74) of CNLC stage IIb patients underwent hepatectomy after conversion therapy. Upon further analysis, we found that 79.6% of BCLC stage B patients presented with ≥4 lesions. In contrast, only 38.6% of BCLC stage C patients had ≥4 lesions (online suppl. Table 6). Additionally, these rates represented the lowest conversion resection rates among the three conversion regimens studied, with BCLC stage B patients showing rates of 7.5%, 16.7%, 13.0%, and 3.0%, 13.6%, and 10.5% for CNLC stage IIb patients, respectively (Table 3).

Table 3.

Conversion resection rates of three treatment regimens at each tumor stage

​	TACE/HAIC	TACE/HAIC+TKIs	TACE/HAIC+TKIs+ICIs	Total
BCLC stage
Stage A	13/43 (30.2)	10/21 (47.6)	6/12 (50.0)	29/76 (38.2)
Stage B	3/40 (7.5)	5/30 (16.7)	3/23 (13.0)	11/93 (11.8)
Stage C	8/47 (17.0)	7/48 (14.6)	8/37 (21.6)	23/132 (17.4)
CNLC stage
Stage Ia	3/14 (21.4)	0/2 (0)	2/3 (66.7)	5/19 (26.3)
Stage Ib	10/29 (34.5)	10/18 (55.6)	4/9 (44.4)	24/56 (42.9)
Stage IIa	2/7 (28.6)	2/8 (25.0)	1/4 (25.0)	4/19 (21.1)
Stage IIb	1/33 (3.0)	3/22 (13.6)	2/19 (10.5)	6/74 (8.1)
Stage IIIa	8/47 (17.0)	7/49 (14.3)	8/37 (21.6)	23/133 (17.3)
Total	24/130 (18.5)	22/99 (22.2)	17/72 (23.6)	63/301 (20.9)

TKIs, tyrosine kinase inhibitors; ICIs, immune checkpoint inhibitors; BCLC, Barcelona Clinical Liver Cancer; CNLC, Chinese National Liver Cancer.

Patients Achieving CR

Monitoring the treatment progress of patients who achieved CR after conversion therapy was crucial. To address this, we conducted a separate analysis focusing on patients in the cohort who attained CR. We found that patients achieving CR had significantly better OS and PFS than those who did not (both p < 0.05) (Fig. 4a, b). Furthermore, we also found that in the non-surgery group, patients who achieved CR had significantly improved OS and PFS compared to those who did not (both p < 0.05) (Fig. 4c, d). However, these differences were not statistically significant in the surgery group (Fig. 4e, f).

Fig. 4.

Kaplan-Meier survival curves comparing overall survival (OS) and progression-free survival (PFS) based on treatment response. a, b OS and PFS between complete response (CR) and non-CR groups before propensity score matching (PSM) in the total group. c, d Subgroup analysis: OS and PFS between CR and non-CR groups before PSM in the surgery group. e, f OS and PFS between CR and non-CR groups before PSM in the non-surgery group.

Additionally, we compared the outcomes of patients who attained pCR or MPR after conversion therapy. Among the 63 patients who underwent surgical resection after conversion therapy, 20 (31.7%) patients achieved pCR, and 24 (38.1%) patients achieved MPR (≥70% necrosis) (Table 4). Kaplan-Meier analysis showed that patients who achieved pCR had significantly better PFS compared to those with residual viable tumor (p = 0.0065), although no difference was observed in OS (p = 0.55) (Fig. 5a, b). Similarly, patients with MPR also showed superior PFS (p = 0.0014), but not OS (p = 0.43) compared to non-MPR patients (Fig. 5c, d). These findings suggest that a higher degree of pathological tumor necrosis is associated with better long-term outcomes after conversion surgery and is consistent with the previous research results [53]. Notably, among the 17 patients with radiographic CR, 11 patients (64.7%, 11/17) achieved a pCR confirmed by pathology after surgical intervention, demonstrating that radiographic CR does not always indicate true pathological CR (Table 4), and the Kaplan-Meier survival analysis also showed that patients who achieved radiographic CR and pCR had significantly better PFS than those who did not, while the OS difference was insignificant (Fig. 5e, f). The images of 3 representative patients before and after conversion in surgery groups are shown in Figure 6.

Table 4.

The rate of patients achieving CR and pathology response

N = 63	Complete response (CR)
	no (n = 46)	yes (n = 17)	p value
pCR, n (%)	​	​	0.002
No (n = 43/63)	37 (80.4)	6 (35.3)	​
Yes (n = 20/63)	9 (19.6)	11 (64.7)	​
MPR, n (%)	​	​	0.018
No (n = 39/63)	33 (71.7)	6 (35.3)	​
Yes (n = 24/63)	13 (28.3)	11 (64.7)	​

pCR, pathology complete response; MPR, major pathological response.

Fig. 5.

Kaplan-Meier survival curves comparing overall survival (OS) and progression-free survival (PFS) based on treatment response among those who successfully underwent conversion surgery. a, b OS and PFS between patients with pathological complete response (pCR) and non-pCR among those who successfully underwent conversion surgery (n = 63). c, d OS and PFS between patients with major pathological CR (MPR) and non-pCR among those who successfully underwent conversion surgery (n = 63). e, f OS and PFS between patients with pathological complete response (pCR) and non-pCR among those who have achieved radiographic CR and successfully underwent conversion surgery (n = 17).

Fig. 6.

Images of 3 representative patients who received conversion therapy. a Case 1: the patient was diagnosed with primary initially unresectable HCC (S8, 40 mm × 33 mm × 30 mm), accompanied by tumor thrombus formation in the right anterior branch of the portal vein (BCLC stage C, CNLC stage IIIa). The patient received conversion therapy with a regimen of TACE/HAIC+TKIs+ICIs. After conversion therapy, no significant blood supply to the primary tumor lesion was observed on enhanced MR, and the portal vein thrombus had resolved. mRECIST evaluation indicated a complete response (CR). Subsequently, the patient underwent radical hepatectomy, and postoperative pathology revealed no viable tumor cells in the lesion, confirming a pathological complete response (pCR). b Case 2: the patient was diagnosed with primary HCC in the right lobe of the liver with tumor thrombus formation in the right branch of the portal vein (112 mm × 88 mm) (BCLC stage C, CNLC stage IIIa). The patient received conversion therapy with a regimen of TACE/HAIC+TKIs+ICIs. After conversion therapy, enhanced MR indicated significant necrosis and liquefaction of the primary tumor lesion, with no definite blood supply to the tumor. mRECIST evaluation showed a CR. The patient subsequently underwent radical hepatectomy, and postoperative pathology revealed extensive necrosis of the tumor tissue, consistent with primary HCC, though pCR was not achieved. c Case 3: the patient was diagnosed with primary HCC (S8, 74 × 56 mm) (BCLC stage A, CNLC stage Ib). Due to insufficient estimated residual liver volume, the patient received conversion therapy with a regimen of TACE/HAIC+TKIs+ICIs. After conversion therapy, the primary tumor lesion reduced in size (68 × 60 mm), and mRECIST evaluation indicated a partial response (PR). The patient subsequently underwent radical hepatectomy, and postoperative pathology revealed no viable tumor cells in the lesion, confirming a pCR.

Prognostic Factors Associated with Survival

To identify prognostic factors influencing OS, we conducted a multivariable time-dependent Cox regression analysis, which better accounts for immortal time bias and more accurately assesses the impact of surgery and other covariates on survival outcomes. The results showed that both the successful surgery (HR: 0.17, 95% CI: 0.07–0.40, p < 0.001) and achieving radiographic CR during treatment (HR: 0.24, 95% CI: 0.11–0.52, p < 0.001) were independent favorable prognostic factors for OS, and older age (HR: 1.03, 95% CI: 1.00–1.06, p = 0.031) and pre-conversion tumor size ≥10 cm (HR: 2.11, 95% CI: 1.16–3.84, p = 0.014) were associated with worse survival (Fig. 7a). Additionally, the successful surgery (HR: 0.37, 95% CI: 0.22–0.61, p < 0.001) and interventional therapy cycle during treatment (HR: 0.80, 95% CI: 0.69–0.92, p = 0.002) were associated with favorable PFS. Although radiographic CR did not reach statistical significance (HR: 0.65, 95% CI: 0.41–1.02, p = 0.063), it showed a strong trend toward improved PFS. Similarly, tumor size shrinkage showed borderline significance (HR: 1.00, 95% CI: 1.00–1.01, p = 0.049) (Fig. 7b) (Table 5).

Fig. 7.

Multivariable time-dependent Cox regression analysis of prognostic factors associated with overall survival (OS) and progression-free survival (PFS): Forest plot displaying the hazard ratios (HRs) and 95% confidence intervals (CIs) for key clinical variables included in the multivariable time-dependent Cox model. a OS. b PFS.

Table 5.

Univariable and multivariable analysis for predictors of OS and PFS in patients with initially unresectable HCC

Variables	OS			PFS
	univariable p value	multivariable HR (95% CI)	p value	univariable p value	multivariable HR (95% CI)	p value
Age	0.927	1.03 (1.00–1.06)	0.031	0.076	0.99 (0.98–1.01)	0.375
Female sex	0.297	0.52 (0.16–1.74)	0.290	0.675	1.06 (0.59–1.92)	0.849
HBsAg (positive)	0.420	–	–	0.496	–	–
AFP ≥400	0.134	–	–	0.002	1.33 (0.94–1.88)	0.113
Tumor size ≥10 cm	0.002	2.11 (1.16–3.84)	0.014	0.017	1.32 (0.91–1.92)	0.138
Tumor size shrinkage	0.605	–	–	0.033	1.00 (1.00–1.01)	0.049
Tumor number	​	–	–	​	​	​
1	Ref	–	–	Ref	Ref	​
2	0.977	–	–	0.380	0.99 (0.54–1.79)	0.965
3	0.557	–	–	0.803	0.75 (0.27–2.13)	0.594
≥4	0.265	–	–	0.011	1.08 (0.66–1.74)	0.766
Vascular invasion	<0.001	6.65 (0.68–64.48)	0.102	0.003	7.04 (0.70–70.39)	0.097
CSPH positive	0.194	1.17 (0.61–2.23)	0.632	0.046	1.17 (0.80–1.71)	0.417
Child-Pugh grade B	0.305	0.42 (0.12–1.40)	0.157	0.329	0.95 (0.57–1.58)	0.848
BCLC stage	​	​	​	​	​	​
Stage A	Ref	​	​	Ref	Ref	​
Stage B	0.364	0.89 (0.39–2.02)	0.777	0.015	1.41 (0.76–2.64)	0.276
Stage C	0.002	0.36 (0.04–3.28)	0.368	<0.001	0.26 (0.03–2.44)	0.236
Successful surgery	0.005	0.17 (0.07–0.40)	<0.001	<0.001	0.37 (0.22–0.61)	<0.001
Radiographic CR	0.007	0.24 (0.11–0.52)	<0.001	0.180	0.65 (0.41–1.02)	0.063
Interventional therapy cycle	0.080	0.81 (0.63–1.05)	0.108	<0.0001	0.80 (0.69–0.92)	0.002
Conversion regimens	​	–	–	​	–	–
TACE/HAIC	Ref	–	–	Ref	–	–
TACE/HAIC+TKIs	0.200	–	–	0.074	–	–
TACE/HAIC+TKIs+ICIs	0.291	–	–	0.876	–	–

HR, hazard ratio; CI, confidence interval; AFP, Alpha-fetoprotein; CSPH, clinically significant portal hypertension; BCLC, Barcelona clinic liver cancer; CR, complete response; TKIs, tyrosine kinase inhibitors; ICIs, immune checkpoint inhibitors.

Prognostic Factors Associated with the Successful Conversion Therapy

To further investigate the factors that contribute to successful conversion therapy, a comprehensive analysis was performed to identify the determinants of successful conversion surgery in patients initially diagnosed with unresectable HCC. The results of the univariable and multivariable regression analysis based on preoperative data were presented in Table 6. Multivariable analysis identified age (OR = 0.95, p < 0.001), positive HBsAg expression (OR = 0.34, p = 0.011), and AFP levels ≥ 400 (OR = 0.49, p = 0.045), ECOG-PS score 1 (OR = 0.43, p = 0.038), BCLC stage B (OR = 0.23, p < 0.001) and stage C (OR = 0.44, p = 0.045), SIRI (OR = 0.73, p = 0.018) as independent predictors for successful conversion surgery (all p < 0.05) (Table 6).

Table 6.

Factors associated with successful conversion surgery in patients who received conversion therapy for initially unresectable HCC

Variables	Univariable p value	Multivariable OR (95% CI)	p value
Age	0.008	0.95 (0.92, 0.98)	<0.001
Female sex	0.126	​	​
HBsAg (positive)	0.003	0.34 (0.15, 0.78)	0.011
AFP ≥400	0.003	0.49 (0.24, 0.99)	0.045
Tumor Size ≥ 10 cm	0.574	​	​
Tumor number	​	​	​
1	Ref	​	​
2	0.521	​	​
3	0.721	​	​
≥4	<0.001	​	​
ECOG PS: 1 vs. 0	0.003	0.43 (0.19, 0.95)	0.038
Child-Pugh grade B	0.052	​	​
BCLC stage	​	​	0.002
Stage A	Ref	Ref	​
Stage B	<0.001	0.23 (0.10, 0.55)	<0.001
Stage C	0.001	0.44 (0.20, 0.98)	0.045
Radiographic CR	0.389	​	​
Conversion regimens	​	​	​
TACE/HAIC	Ref	​	​
TACE/HAIC+TKIs	0.482	​	​
TACE/HAIC+TKIs+ICIs	0.384	​	​
SIRI	0.007	0.73 (0.56, 0.95)	0.018

OR, odds ratio; CI, confidence interval; AFP, alpha-fetoprotein; ECOG PS, Eastern cooperative oncology group performance status; BCLC, Barcelona clinic liver cancer; CR, complete response; TKIs, tyrosine kinase inhibitors; ICIs, immune checkpoint inhibitors; SIRI, systemic inflammation response index.

The above seven independent predictors of successful conversion surgery were incorporated into a nomogram prediction model to predict the probability of successful conversion (Fig. 8a). Each factor was assigned a score based on its OR. The accuracy of the nomogram was evaluated using ROC curve analysis, which yielded an AUC of 0.802 (95% CI: 0.743–0.860) (Fig. 8b). Additionally, the calibration curve, generated using 100 bootstraps resamples, showed strong alignment between predicted and actual conversion outcomes (Fig. 8c), and the decision curve analysis further demonstrated that the new models offered greater clinical benefit (Fig. 8d).

Fig. 8.

Nomogram and validation plots for predicting the probability of successful conversion to resectable HCC. a The nomogram model for estimation of the risk of successful conversion. b Performance of the nomogram in predicting successful conversion. c The calibration curve of the nomogram for estimation of the risk of successful conversion. d The decision curve analysis (DCA) of the nomogram for estimation of the risk of successful conversion.

Discussion

Patients with initially unresectable HCC generally face a poor prognosis due to limited curative options. Recent advancements in locoregional and systemic therapies, such as targeted therapy and immunotherapy made conversion therapy possible in unresectable HCC treatment. In this study, we found that TACE/HAIC-based comprehensive conversion therapy effectively controlled disease progression in initially unresectable HCC patients, achieving a conversion resection rate of 20.9%, with a pCR rate of 31.7% (20/63). And patients who underwent radical liver resection after conversion therapy showed a better OS and PFS. In addition, patients who achieved CR or pCR showed superior survival outcomes. This study indicates that TACE/HAIC-based conversion therapy has the potential to increase the conversion resection rate and improve the long-term survival of patients.

Conversion therapy, targeting tumor downstaging to enable curative resection, has emerged as a key strategy for initially unresectable HCC. In recent years, TACE has been recommended as a treatment option for intermediate-stage HCC in various guidelines, with its efficacy and safety being well-established in clinical practice [7]. Several studies have further explored the potential benefits of combining therapies for initially unresectable HCC. For example, a phase III randomized controlled clinical [48] reported that combining lenvatinib with TACE significantly improved PFS and OS compared with lenvatinib monotherapy, and also increased the ORR (54.1% vs. 25.0%, p < 0.001). Another study found that adding pembrolizumab to TACE plus lenvatinib significantly improved the ORR (47.1% vs. 27.8%, p = 0.017) and prognosis, with a longer PFS and OS in the combination group [58]. In addition, studies have shown that the combination of TACE/HAIC+TKIs+ICIs achieves better ORR, OS, and PFS than TACE/HAIC alone [38]. Consistent with these findings, in our study, we observed that the TACE/HAIC+TKIs and TACE/HAIC+TKIs+ICIs groups showed higher ORR (73.7% and 84.7%) compared to the TACE/HAIC monotherapy group (64.6%) (p = 0.008) (online suppl. Table 7), suggesting that a multimodal combination therapy involving interventional therapy, targeted therapy, and immunotherapy might offer superior effectiveness and disease control for patients with initially unresectable HCC.

The effectiveness of liver resection after conversion therapy for unresectable HCC remains a topic of debate. Recent studies have evaluated various combination therapies as conversion therapy for unresectable HCC and have reported a surgical conversion rate of 15.9–37.5% [15, 59, 60]. However, the sample size of the included studies was generally small. Our research included 301 initially unresectable HCC patients, of which 63 patients underwent hepatectomy after successful conversion therapy, resulting in a conversion resection rate of 20.9% (63/301). In parallel, patients with surgery showed promising pathological outcomes, with a pCR rate of 31.7% (20/63) and a 100% rate of achieving R0 resection. In the survival analysis, patients who underwent surgery had significantly longer median PFS and OS compared to those who did not undergo surgery (p < 0.05) (Fig. 3a, b). Furthermore, we found that, before and after adjustment for confounding factors, the Kaplan-Meier survival analysis showed that among patients who met the surgical resection criteria, the prognosis of those who underwent surgery was significantly better than that of those who did not (Fig. 3c, f), indicating that liver resection was an effective treatment option for patients who achieve successful conversion therapy, which highlights the importance of liver resection in improving survival outcomes, especially the long-term survival benefits of surgery.

It has been reported that 50–75% of HCC patients are diagnosed with multiple lesions [61, 62]. Multifocal HCC, which consists of multiple lesions, poses a greater challenge in treatment due to its complex tumor heterogeneity [61, 63]. Different lesions within the same patient may have significant molecular and biological differences [61, 64–67], which can influence treatment response and reduce the effectiveness of conversion therapy [61, 68–70]. In our study, we found that only a small proportion of patients with multiple lesions were able to undergo liver resection following conversion therapy (11.8% in BCLC stage B and 8.1% in CNLC stage IIb), which indicates that multiple lesions may limit the success rate of conversion resection and diminish the effectiveness of treatment strategies. In contrast, some BCLC C patients, although initially presenting with vascular invasion, had more localized tumors (e.g., solitary lesions with PVTT), and thus could become resectable following vascular shrinkage or necrosis after combination therapy. In summary, the lower resection rate in conversion therapy among patients with multifocal HCC suggests that the number of lesions plays a crucial role in the success of conversion therapy. Therefore, to improve the success rate of conversion therapy in patients with multifocal HCC, future therapeutic strategies should focus on more individualized assessments of each lesion. This study underscores the importance of enhancing the assessment of lesion number and heterogeneity in treatment decision-making.

Our study underscores the importance of achieving CR and pCR after conversion therapy. The prognosis of patients who achieved CR during conversion therapy was significantly better than that of those who did not, especially in the non-surgery group, where CR was associated with significantly better survival (p < 0.05) (Fig. 4a, b). There was no significant difference in survival based on CR status in the surgery group, which might be explained by the fact that all patients underwent liver resection, resulting in a considerable improvement in their long-term survival. It is also noteworthy that some patients initially diagnosed with insufficient FLR showed an increase in FLR after receiving comprehensive treatment involving TACE/HAIC, while also achieving tumor control. Some patients, although classified as SD, had sufficient FLR and were able to undergo hepatectomy after evaluation by the MDT. Conversely, there were patients who chose to continue comprehensive treatment and refused surgery during conversion therapy based on personal preference. Therefore, not all patients with CR/PR underwent liver resection, and some patients who did undergo surgery had a tumor response of SD before surgery.

Additionally, our study further supports the emerging evidence that pathological response after conversion therapy is closely associated with long-term survival in patients with initially unresectable HCC. In our cohort, patients who achieved pCR and MPR had significantly improved PFS following resection. This finding is consistent with results from the recent multicenter global collaborative consortium (NeoHCC) analysis, which demonstrated that pCR and MPR were strongly correlated with relapse-free survival (HR 0.19 for pCR compared to <100% necrosis, and 0.26 for MPR, compared to <70% necrosis) [53]. These data suggest that pathological response, particularly pCR, may serve as a robust surrogate endpoint for clinical benefit in conversion therapy. Although no statistically significant difference in OS was observed between pCR and non-pCR patients in our study, this may be partially attributed to the fact that tumor resection provides substantial disease control even in non-pCR patients, thereby prolonging the OS. Moreover, heterogeneity in post-recurrence treatment strategies and individual responses may lead to survival compensation in non-pCR patients, potentially attenuating OS differences between groups. Importantly, the relatively high pCR rate observed in our cohort (31.7%) underscores the potential efficacy of TACE/HAIC-based combination regimens and further highlights the value of incorporating pathological response assessment in postoperative evaluation. Therefore, we recommend surgical resection in patients who achieve CR. Unlike colorectal liver metastases, where a “Wait and Watch” approach may be recommended due to chemotherapy sensitivity and the possibility of needing anal resection, the scenario for HCC is different. Patients with CR in HCC who undergo surgery have several advantages: (1) imaging-based CR does not always correlate with pCR, with only a 64.7% (11/17) concordance rate in our study; (2) after liver resection, patients can potentially reduce or discontinue their treatments, while non-surgery patients require lifelong medication, and (3) surgery improves survival and overall quality of life, with some patients achieving long-term tumor-free outcomes. Considering these benefits, surgical resection for CR in HCC provides accurate assessment, reduces treatment burden, and leads to better long-term outcomes.

Numerous studies have shown a strong connection between cancer and inflammation, with inflammation having been associated with increased cancer risk and poorer outcomes for tumor patients. Various inflammation-related markers have been identified as predictors of cancer survival [71–75]. Previous studies have shown that chronic inflammation caused by viral infections can alter the liver microenvironment and increase the risk of HCC [76]. Additionally, higher AFP levels, a marker of advanced HCC, have been identified as a predictor of conversion therapy outcomes, possibly contributing to therapy failure [77].Consistent with these findings, our study identified several independent risk factors for predicting successful conversion therapy, including age, positive HBsAg expression, AFP levels ≥400, ECOG PS score of 1, BCLC stage B or C, SIRI. To improve the accuracy of predicting successful conversion surgery, we developed a nomogram prediction model incorporating these six independent predictors. The nomogram visually represents the probability of achieving successful conversion (Fig. 8a). The model showed strong discriminative power, as reflected by the high AUC in the ROC analysis. Furthermore, the calibration curve showed good alignment between predicted and observed conversion rates, and the decision curve analysis indicated the model’s potential clinical utility in guiding treatment decisions for patients considering conversion surgery.

In recent years, the development of innovative anticancer drugs has significantly improved the survival and quality of life of cancer patients [78]. However, access to these therapies is often limited by high costs, uneven health resources, and socioeconomic disparities. Previous studies have identified socioeconomic status as an independent prognostic factor in HCC patients [79], and the prognosis of low-income patients is significantly worse than that of high-income patients [80]. In addition, lower education and household income have also been associated with a later BCLC stage at diagnosis, fewer treatment opportunities, and worse long-term survival [81]. In our study, we similarly took into account that the choice of systemic therapy during conversion therapy is often influenced by patient preferences and economic situations. Especially in the case of the high cost of targeted therapy and immunotherapy, patients with limited financial resources may be more inclined to receive TACE/HAIC alone, while those with better means were more likely to receive TACE/HAIC combined with systemic therapy (TKIs/ICIs), potentially leading to higher ORR and conversion resection rates. Unfortunately, due to the retrospective nature of our study, we could not directly assess which patients gave up systemic therapy for financial reasons. As a potential confounding factor affecting treatment choice and prognosis, the specific impact of economic status on survival requires further studied.

There are several limitations to this study. First, it is a single-center retrospective study, which may introduce inherent bias. Second, although no Grade 3 or higher AEs were reported, the lack of detailed documentation and evaluation may have led to an underestimation of these events due to the retrospective nature of the study. Third, as a retrospective study, this research was unable to completely eliminate potential unmeasured confounding factors, which may to some extent affect the research conclusions. Lastly, the treatment regimens were not completely standardized, potentially leading to differences in drug efficacy. Therefore, future prospective studies, especially large-scale multicenter trials, are necessary to validate these findings under more consistent conditions.

Conclusions

Patients with initially unresectable HCC can achiever promising curative effects and conversion resection rates with TACE/HAIC-based comprehensive therapy. More importantly, for patients who met the criteria for resection after conversion therapy, those who ultimately underwent surgery had significantly longer OS and PFS both before and after propensity score matching, highlighting the crucial role of surgery in maximizing long-term benefits after achieving resectability.

Acknowledgments

We gratefully thank all patients who participated in this study.

Statement of Ethics

The study protocol adhered to the ethical guidelines of the World Medical Association Declaration of Helsinki. The Institutional Review Board of the First Affiliated Hospital of Sun Yat-sen University reviewed and approved the study protocol (ethical approval No. LunShen [2023] No. 399), and the requirement for written informed consent was waived due to the retrospective nature of the study.

Conflict of Interest Statement

The authors declare that there is no conflict of interest or disclosures in this study.

Funding Sources

The authors greatly acknowledge the financial support from the Guangdong Basic and Applied Basic Research Foundation (2023B1515230006, 2022A1515010537), Guangdong Provincial Key Areas Research and Development Program (2023B1111020007), and the Natural Science Foundation of Guangdong Province (2022A1515010862).

Author Contributions

S.S. and M.K. designed the research and served as guarantors for the overall content, and Y.H., S.S., and M.K. obtained funding; S.L., Z.S., P.C., and X.Y. performed data collection and data evaluation. S.L. performed data analysis and drafted the manuscript, Z.S. and X.Y. participated in case follow-up. W.X., Y.H., S.L., S.S., and M.K. supervised the study, developed the protocol, and coordinated tissue collection. All authors revised and approved the final manuscript.

Funding Statement

The authors greatly acknowledge the financial support from the Guangdong Basic and Applied Basic Research Foundation (2023B1515230006, 2022A1515010537), Guangdong Provincial Key Areas Research and Development Program (2023B1111020007), and the Natural Science Foundation of Guangdong Province (2022A1515010862).

Data Availability Statement

The data that support the findings of this study are not publicly available due to privacy reasons but are available from the corresponding author upon reasonable request.