Survival Outcomes, Recurrence Patterns, and Its Corresponding Therapeutic Management after Conversion Therapy Followed by Curative Resection for Hepatocellular Carcinoma: A Real-World Study.

Introduction
Liver cancer is the sixth most commonly diagnosed cancer and the third leading cause of cancer-related mortality worldwide [1]. Hepatocellular carcinoma (HCC) is the predominant pathological type of primary liver cancer, accounting for 80–85% of all primary liver cancer. Treatment modalities such as liver resection, liver transplantation, and ablation offer the potential for cure [2]. However, in China, 60–70% of HCC patients are diagnosed at intermediate-advanced stages, when curative treatments are often no longer feasible [3]. For such patients with uHCC, clinical guidelines based on NCCN and BCLC staging systems recommend transarterial therapy (TAT) and systemic therapies including tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs) [4, 5]. During the course of these treatments, some patients experience tumor shrinkage or even downstaging, thereby becoming eligible for surgical resection.
Conversion therapy refers to a treatment approach to converting unresectable HCC into resectable HCC through locoregional therapies and/or systemic therapies. Locoregional therapies include transarterial chemoembolization (TACE), hepatic arterial infusion chemotherapy (HAIC), and radiotherapy, while systemic therapies consist mainly of TKIs, ICIs, and VEGF inhibitors [6]. Numerous previous studies have indicated that a considerable proportion of patients with uHCC achieved tumor reduction sufficient to undergo curative resection following transarterial interventional therapies, such as TACE, HAIC, or transarterial radioembolization, resulting in significantly improved OS [7–9]. With significant advances in systemic therapies, the objective response rate and conversion resection rate have markedly increased with the triple conversion regimens, referring to the combination of TAT, ICIs and TKIs or VEGF inhibitors [10–12]. Although numerous cases of conversion resection have been reported, the majority are based on small sample studies, with a notable lack of large-sample cohorts and long-term follow-up results. Consequently, the long-term clinical outcomes of conversion resection remain uncertain.
Tumor diameter, multifocality, vascular invasion, high histological grade, and higher serum α fetoprotein level are high-risk factors for postoperative recurrence in patients undergoing direct surgical resection [13], and they also represent potential indications for adjuvant therapy. The current consensus on adjuvant therapy after liver cancer surgery is that patients with these high-risk recurrence factors may require adjuvant treatment following resection. However, only a limited number of prospective randomized controlled trials have demonstrated positive outcomes. For example, the IMbrave050 study showed that adjuvant atezolizumab plus bevacizumab (T+A) significantly prolonged RFS [14], though its impact on OS required further investigation. A study by Li et al. [15] indicated that adjuvant HAIC significantly improved disease-free survival in postoperative patients with microvascular invasion (MVI). However, the aforementioned studies have focused exclusively on patients who underwent direct surgery. For those who achieved resection after conversion therapy, there is neither expert consensus nor prospective research to clarify whether adjuvant therapy is required or which specific regimen should be utilized.
The high postoperative recurrence rate, which exceeds 60% after liver resection [16], remains a critical factor affecting prognosis. Research showed that approximately 80% of postoperative recurrence occurred intrahepatically, which could be further categorized into solitary intrahepatic recurrence and multifocal intrahepatic recurrence [17]. Additional studies reported that over 30% of recurrence involve extrahepatic lesions, either as isolated extrahepatic lesion or combined intra- and extrahepatic recurrence [18–20]. Patients with multifocal intrahepatic recurrence or any extrahepatic recurrence generally have a poor prognosis due to the lack of curative treatment options. In contrast, those with solitary intrahepatic recurrence may still undergo curative interventions such as repeat resection or ablation and thus exhibit relatively better outcomes. These findings highlighted that the pattern of recurrence was a key determinant of survival after HCC recurrence. Therefore, investigating the recurrence patterns in patients who have undergone conversion resection and comparing treatment strategies across different recurrence patterns are essential for improving long-term outcomes in this specific and growing population.
Currently, there is a lack of large-scale, systematic studies on the long-term survival of patients after conversion resection. Moreover, no consensus exists regarding optimal treatment strategies for recurrent HCC after conversion resection. This retrospective study included 419 HCC patients who underwent conversion resection. We described the overall survival and recurrence situation among these patients and discussed factors influencing prognosis. Meanwhile, our study investigated whether postoperative adjuvant therapy could reduce the risk of recurrence and improve survival outcomes. Moreover, we outlined the recurrence patterns after conversion resection and discussed corresponding treatment strategies tailored to different types of recurrence, aiming to provide real-world evidence to inform clinical decision-making prior to the initiation of prospective studies.

Materials and Methods

Patients Collection
Patients received conversion resection from November 2014 to June 2023 were retrospectively collected in Sun Yat-sen University Cancer Center. The inclusion criteria were as follows: (1) diagnosed with HCC pathologically; (2) age of 18–80 years; (3) Child-Pugh class A or B; (4) a performance status (PS) score ≤ 2; (5) BCLC staging A to C; (6) received curative resection after conversion therapy, and (7) was considered unresectable by experienced hepatobiliary surgeons when initially diagnosed. The exclusion criteria were as follows: (1) had a history of other malignancies; (2) received palliative surgery, and (3) loss of follow-up. The detailed inclusion flowchart is shown in Figure 1.

Conversion Therapy
The determination of unresectability was established through discussion by a multidisciplinary team (MDT), which included hepatobiliary surgeons, interventional radiologists, medical oncologists, radiation oncologists, and pathologists. Patients presenting with any of the following conditions were initially classified as having unresectable tumors and were thus candidates for conversion therapy: (1) inadequate future liver remnant volume (FLR <40% for cirrhotic patients or <30% for non-cirrhotic patients) (Fig. 2a, b); (2) multiple tumor lesions (Figure 2c and d); (3) presence of major vascular invasion (Figure 2e and f), and (4) tumor proximity to critical hepatic vasculature, preventing complete (R0) surgical removal (Figure 2g and h). The number of patients of each causes are shown in online suppl. Table S2 (for all online suppl. material, see https://doi.org/10.1159/000550954). Conversion therapy was performed to these patients, aiming to achieve resectability through tumor shrinkage, downstaging, or devitalization. Conversion therapy included locoregional therapy with or without the combination of systemic therapies. Locoregional therapy included FOLFOX-HAIC and TACE. FOLFOX-HAIC was done according to our previous study [9] and TACE was done according to Shi et al. [21] previously reported protocols. Systemic therapies included TKIs, ICIs, VEGF inhibitors, and their combination. TKIs in our study included lenvatinib (8 mg once daily), sorafenib (400 mg twice daily), and apatinib (250 mg once daily). ICIs in our study consisted mainly PD-1 inhibitors (sintilimab 200 mg, camrelizumab 200 mg, tislelizumab 200 mg) and PD-L1 inhibitors (atezolizumab 1200 mg), which were administered intravenously every 3 weeks. VEGF inhibitors referred to bevacizumab 15 mg/kg, administered intravenously every 3 weeks.

Conversion Resection
Tumor response was evaluated every 1 month after each cycle of conversion therapy by our MDT. The decision on surgical feasibility was made by our institutional MDT based on a thorough assessment. This decision primarily relied on the following factors: (1) the patient’ s overall condition was sufficient to tolerate surgery; (2) the FLR was greater than 30% (or greater than 40% in patients with cirrhosis), and (3) an R0 resection was deemed achievable. R0 resection was defined as the complete macroscopic removal of the tumor, with histopathologically confirmed negative surgical margins, ensuring the absence of microscopic residual disease and denoting a curative resection. Liver resection was done by experienced hepatobiliary surgeons with surgical experience more than 10 years. Anatomical resection or local resection is done according to the size, position, and liver function of patients.

Pathological Assessment
Each postoperative pathology underwent a dual evaluation by two independent, experienced pathologists. The surviving tumor activity and necrosis ratio were averaged from 2 independent assessments. Estimates exhibiting a divergence greater than 10% were re-evaluated and finalized by two senior pathologists to reach a consensus. Viable tumor cells ≤10% were defined as major pathological response (MPR), while pathological complete response was defined as the absence of viable tumor cells in any slice of excised tissue. Detailed pathological assessment procedures have already been described by the previous study of Ou-yang L-Y [22].

Postoperative Adjuvant Therapy
The decision regarding postoperative adjuvant therapy was made by our MDT during the first postoperative follow-up visit (at 1–2 months after resection) based on a comprehensive discussion of multiple factors. These included preoperative tumor burden and stage, response to conversion therapy, postoperative pathology, presence of MVI, preoperative and postoperative AFP level, surgical margin status, patients’ economic circumstances, personal preference, and anticipated treatment tolerance. Adjuvant therapy was initiated following patients’ first postoperative follow-up visit. The types of adjuvant therapy in our study included adjuvant FOLFOX-HAIC, TKIs, ICIs, VEGF inhibitors and their combination. The regimen of adjuvant FOLFOX-HAIC was in accordance with those reported by Li et al. [15] in their phase III clinical trial. The dosage and administration of TKIs, ICIs, and VEGF inhibitors were consistent with those previously described for the conversion therapy regimen. Patients received adjuvant therapy for approximately 6–8 months or until disease recurrence, the occurrence of unacceptable adverse events, the development of concurrent conditions that precluded further treatment, or patient noncompliance.

Postoperative Follow-Up
Patients were firstly monitored at the first month after liver resection, then every three to 4 months for the first 2 years, and every 6 months subsequently. Follow-up assessments included blood test (tumor markers including serum AFP and PIVKA-II levels, liver function tests, complete blood count, and HBV DNA levels for patients with positive HBV infection) along with contrast-enhanced MRI or CT. The diagnosis of recurrence was made on the basis of typical enhancement characteristics on imaging.

Treatments for Recurrent HCC
Management of recurrence during the follow-up period was determined by MDT according to the characteristics of tumors such as size, number of lesions, and location of recurrence. While the overall treatment strategy aligned with that for primary HCC, an exception was made for solitary extrahepatic recurrence. The therapeutic options for recurrent HCC mainly consisted of repeat resection, ablation, radiotherapy, locoregional intervention, systemic therapies, and the combination of the above. The protocols of resection, locoregional intervention, and systemic therapies have already been mentioned above. Ablation, including radiofrequency ablation (RFA) and microwave ablation, was performed by experienced surgeons under ultrasound guidance if the lesions located in the liver and were visible on ultrasound. If the lesions were not clearly visible on ultrasound or located extrahepatically, ablation was performed under the guidance of CT. Contrast-enhanced MRI or CT and blood test were done 1 month after ablation to check if complete ablation had been achieved. The detailed procedures of ablation and stereotactic body radiotherapy have already been described in our previous study by Xi et al. [23].

Outcomes
The primary endpoints include overall survival (OS) and overall survival after recurrence (rOS). OS is defined as the span from liver resection to death or last follow-up. rOS is defined as the span from tumor recurrence to death or last follow-up. The second endpoints include recurrence-free survival (RFS). RFS is defined as the duration from liver resection to tumor recurrence or death or last follow-up (online suppl. Fig. S1.). Tumor response is evaluated according to Response Evaluation Criteria in Solid Tumors 1.1 (RECIST 1.1) and the modified Response Evaluation Criteria in Solid Tumors 1.1 (mRECIST) [24, 25]. Tumor response is classified as complete response (CR), partial response (PR), stable disease (SD), and progression disease (PD).

Statistical Analysis
All statistical analyses were carried out with R software (version 4.4.2, R Foundation, Vienna, Austria) and SPSS (version 25.0, SPSS Inc., Chicago, USA). Categorical data were analyzed with Pearson chi-square test or Fisher’s exact test to compare the frequencies and percentages of events between different groups. Continuous variables were described by median (IQR) and analyzed using Student’s t test when the variables distribute normally and were analyzed using non-parametric test when the variables were not normally distributed. Bonferroni method was used to correct p values for multiple comparisons. Survival curves were performed by Kaplan-Meier analysis, and log-rank test was used to compare the OS, PFS, and RFS between groups. Clinically relevant variables showing a potential association (p < 0.10) with the survival outcome (OS or RFS) in the univariable Cox regression analysis were included in the initial multivariable Cox proportional hazards model. Collinearity diagnostics was conducted to exclude the variables that were strongly relevant before performing the multiple Cox regression test. Propensity score methodologies were applied to balance baseline characteristics. Primary analysis involved propensity score matching (PSM) using the MatchIt package in R. Sensitivity analysis was conducted through inverse probability of treatment weighting (IPTW) with the WeightIt package, using the same variables as PSM. A two-tailed p value <0.05 was considered statistically significant.

Results

Baseline Characteristics
Between November 2014 and June 2023, 419 patients who underwent conversion therapy followed by curative liver resection in Sun Yat-sen University Cancer Center were enrolled. The median age of these patients was 54 years (IQR, 43–61), with 365 patients (87.1%) being male and the vast majority (380, 90.7%) being hepatitis B virus carriers. At initial diagnosis, the median maximum tumor diameter was 84.0 mm (IQR, 56.0–109.0). According to the BCLC staging system, 143 patients (34.1%) were classified as BCLC A, 140 (33.4%) as BCLC B, and 136 (32.5%) as BCLC C. Among these patients who received conversion therapy, 207 (49.4%) underwent TAT alone, 48 (11.5%) received dual therapy (TAT combined with TKIs or ICIs), and 162 (38.7%) received triple therapy (TAT combined with TKIs and ICIs). The median time from the initiation of conversion therapy to curative resection was 3.6 months (IQR, 2.3–5.0) (Table 1.).

Description of OS in Conversion Resection Patients
The median follow-up time for the study was 44.3 months (95% CI, 26.4–62.3). During the follow-up period, recurrence occurred in 261 patients (62.3%) until the end of follow-up. A total of 120 patients (28.6%) died during follow-up, the majority due to tumor recurrence. Overall, the 1-, 3-, and 5-year OS rates for the 419 patients who underwent conversion resection were 94.5%, 76.3%, and 63.5%, respectively, with a median OS not reached (online suppl. Fig. S2a). The corresponding 1-, 3-, and 5-year RFS rates were 58.2%, 40.4%, and 34.1%, respectively, with a median RFS of 18.9 months (95% CI, 13.6–25.4) (online suppl. Fig. S2b).

Comparison of Survival in Patients with Different Tumor Response after Conversion Therapy
According to mRECIST criteria, patients were categorized based on the best observed response during conversion therapy into the following 4 groups: complete response (n = 57, 13.8%), partial response (n = 201, 48.6%), stable disease (n = 144, 34.8%), and progressive disease (n = 12, 2.9%). The median OS was not reached in the CR and PR groups, while the SD group had a median OS of 69.6 months (95% CI, 54.9–NA). A significant difference in OS was observed among the four groups (p = 0.015, Fig. 3a.). Multiple comparisons demonstrated significantly superior OS in patients achieving CR compared to the other three groups. The median RFS was 83.8 months (95% CI, 43.0–NA) in the CR group, 16.8 months (95% CI, 11.3–29.6) in the PR group, 13.6 months (95% CI, 10.5–22.5) in the SD group, and 8.6 months (95% CI, 4.1–NA) in the PD group, showing a progressively decreasing trend with statistical significance (p < 0.001, Fig. 3b). Similar analysis using RECIST 1.1 criteria were shown in online suppl. Figure S3a and S3B.
We next assessed the impact of changes in tumor markers on the prognosis of patients who underwent conversion resection. A reduction of ≥75% in AFP and PIVKA-II from pre-conversion to post-conversion therapy (before surgery) was defined as the AFP decrease group (n = 226, 53.9%) and PIVKA-II decrease group (n = 260, 62.1%), respectively, while a reduction of <75% was defined as the AFP stable group (n = 193, 46.1%) and PIVKA-II stable group (n = 140, 33.9%). The AFP decrease group exhibited a significantly longer RFS compared to the stable group (27.3 months [95% CI, 15.5–56.9] vs. 12.8 months [95% CI, 9.8–20.8], p < 0.001) (Fig. 3d). Similarly, the PIVKA-II decrease group showed a significantly longer RFS than its corresponding stable group (23.4 months [95% CI, 15.4–42.3] vs. 12.5 months [95% CI, 8.6–24.5], p = 0.005) (online suppl. Fig. S3d). Notably, although no significant differences in OS were observed between the decrease and stable groups for either AFP (p = 0.094, Fig. 3c) or PIVKA-II (p = 0.43, online suppl. Fig. S3c), the decrease groups demonstrated a trend toward superior OS compared to their corresponding stable groups.
Subsequent analysis compared the prognostic impact of the pathological response in patients following conversion therapy. Pathological assessment revealed a MPR in 110 patients (26.2%) and the absence of MPR in 319 patients (73.8%). Patients who achieved MPR demonstrated significantly longer OS (NA vs. 74.6 months [95% CI, 59.9-NA], p < 0.001) and RFS (83.8 months [95% CI, 43.0–NA] vs. 12.7 months [95% CI, 10.3–16.4], p < 0.001) compared to those who did not, as indicated in Figure 3e and f.
Univariate and multivariate Cox regression analysis were conducted to identify independent risk factors related to OS and RFS for primary HCC (Table 2). Baseline multiple tumor lesions and postoperative AFP >40 ng/mL were independent risk factors for both OS (HR: 1.80, 95% CI: 1.09–2.97, p = 0.02; HR: 2.02, 95% CI: 1.35–3.04, p = 0.01) and RFS (HR: 1.90, 95% CI: 1.32–2.73, p < 0.001; HR: 2.78, 95% CI: 1.89–4.10, p < 0.001), whereas the presence of MPR was an independent protective factor for OS (HR: 0.42, 95% CI: 0.20–0.87, p = 0.02) and RFS (HR: 0.63, 95% CI: 0.42–0.94, p = 0.025). In addition, cycles of TAT exceeding 4 (HR: 1.58, 95% CI: 1.06–2.36, p = 0.026) and the presence of microvascular invasion (HR: 1.63, 95% CI: 1.20–2.21, p = 0.002) suggested a higher risk of tumor recurrence.

Adjuvant Therapy after Conversion Resection
The necessity of adjuvant therapy following curative resection remains controversial, and whether it improves survival after conversion therapy and subsequent radical resection is still unclear. To investigate this, we divided the patients into adjuvant therapy group (n = 88, 20.8%) and active surveillance group (n = 331, 79.2%) and compared their survival and recurrence outcomes. Although baseline variables were generally comparable between the two groups, the adjuvant group versus surveillance group demonstrated significantly higher percentages of patients with BCLC C stage (37 [42.0%] vs. 99 [29.9%], p = 0.023), CNLC III stage (37 [42.0%] vs. 99 [29.9%], p = 0.017), and microvascular invasion (27.3%] vs. 55 [16.6%], p = 0.034), as indicated in online suppl. Table S1. The adjuvant group and active surveillance group demonstrated similar OS (65.4 months [95% CI, 59.9–NA] vs. NA, p = 0.66, Figure 4a) and similar RFS (16.8 months [95% CI, 13.4–43.0] vs. 19.6 months [95% CI, 12.7–26.9], p = 0.94, Figure 4b). Moreover, there were 59.0% patients in the adjuvant group and 63.1% patients in the surveillance group experienced recurrence, respectively, with no significant difference (p = 0.549).
Due to inherent baseline imbalances, a 2:1 PSM was performed using BCLC stage, CNLC stage, and presentation of MVI as covariates. The post-matching baseline characteristics are presented in online suppl. Table S1, which demonstrated that the two groups were well balanced. Subsequently, we compared OS and RFS between the matched cohorts. After PSM, there was no statistically significant difference in OS between the adjuvant therapy group and the active surveillance group (65.4 months [95% CI, 59.9–NA] vs. NA, p = 0.84, Figure 4c). Furthermore, the median RFS was 16.8 months (95% CI, 13.4–43.0) in the adjuvant therapy group and 15.1 months (95% CI, 10.3–32.4) in the active surveillance group, which also did not reach statistical significance (p = 0.82, Fig. 4d). Similar results were obtained in the analyses of OS and RFS after IPTW adjustment (Fig. 4e and f).
To further validate our findings, we performed subgroup analyses for both OS and RFS by incorporating multiple baseline factors. Consistent with our primary findings, adjuvant therapy did not confer a significant benefit in either OS or RFS across any of the predefined subgroups (online suppl. Fig. S4a and b). As indicated in the univariate and multivariate analysis in Table 2, adjuvant therapy was not an independent risk or protective factor for OS (HR: 1.12, 95% CI: 0.69–1.81, p = 0.66) or RFS (HR: 0.99, 95% CI: 0.73–1.34, p = 0.94), either. Furthermore, no significant difference in RFS can be observed in each adjuvant treatment type when compared to active surveillance (online suppl. Fig. S5b, d, and f), while adjuvant HAIC demonstrated worse OS compared to active surveillance (HR: 2.59, 95% CI: 1.18–6.14, p = 0.014, online suppl. Figure S5a).

Comparison between Different Recurrence Stages according to BCLC Staging
The median follow-up time after recurrence was 36.0 months (95% CI, 33.4–44.9). Median time to recurrence was 7.2 months. Among the 261 patients who experienced recurrence, the median rOS was 40.2 months (95% CI, 33.4–48.7), with a 1-year, 2-year, 3-year rOS rates of 85.6%, 69.1%, 53.0%, respectively (online suppl. Fig. S6a). Most recurrent tumors occurred intrahepatically (n = 181, 69.3%), while 22.6% (n = 59) of them occurred extrahepatically and 8.0% (n = 21) of them occurred both intra- and extrahepatically. The majority of recurrences (67.0%) were observed within the first year after resection. Using BCLC staging system for recurrent tumors (rBCLC staging), the patient distribution was as follows: rBCLC 0, 68 patients (26.1%); rBCLC A, 61 (23.4%); rBCLC B, 41 (15.7%); and rBCLC C, 91 (34.9%). Other clinical variables during postoperative recurrence are shown in Table 3.
The median rOS was 48.7 months (95% CI, 43.6–NA) for patients with rBCLC 0/A, 34.9 months (95% CI, 22.0–NA) for those with rBCLC B, and 29.6 months (95% CI, 21.6–40.2) for those with rBCLC C. Noteworthy improvements in rOS were seen in patients with rBCLC 0/A in contrast to patients with rBCLC B (HR: 1.75, 95% CI: 1.05–2.91, p = 0.032) and rBCLC C(HR: 1.90, 95% CI: 1.26–2.85, p = 0.002), while no significant difference was observed between the latter 2 groups (online suppl. Fig. S6b.). The 3-year survival rates for patients with rBCLC 0/A, B, C were 64.5%, 48.6%, 39.6%, respectively.
Univariate and multivariate analysis demonstrated that a higher ALBI score at recurrence (HR: 1.84, 95% CI: 1.12–3.02, p = 0.017), AFP at recurrence >400 ng/mL (HR: 2.05, 95% CI: 1.33–3.18, p = 0.001), and treatment patterns were independent risk factors for rOS, while time to recurrence>1 year (HR: 0.48, 95% CI: 0.27–0.84, p = 0.010) was independent protective factor for rOS (Table 4).

Treatment Patterns for Recurrent HCC after Conversion Resection
For patients who experience recurrence after conversion resection, multiple therapeutic options are available, primarily including repeat resection, ablation, radiotherapy, TACE, systemic therapy (TKIs/ICIs), and supportive care. The selection among these modalities is guided mainly by the size, location, number of lesions, and tumor stages. In this section, treatment patterns are classified into four categories: curative treatment, locoregional TAT, systemic therapies, and combined therapies.
For early-stage recurrence (68 at rBCLC stage 0 and 61 at rBCLC stage A), curative treatment was the primary approach (n = 78, 60.4%). This included repeat hepatectomy (n = 15), radiofrequency/microwave ablation (n = 50), and curative radiotherapy (n = 13). Locoregional TAT was administered to 20 patients (15.5%), comprising 18 TACE and 2 HAIC. Systemic therapies were administered to only 4 patients (3.1%). Combined therapies were given to 20 patients (15.5%), with 17 receiving TACE combined with systemic therapies and 3 receiving HAIC combined with systemic therapy. An additional 7 patients had undocumented therapy. The treatment patterns are listed in Figure 1 in detail. Patients who experienced early-stage recurrence were characterized by relatively small tumors, with a median maximum diameter of 14.0 mm (IQR, 10.0–20.0). The majority presented with a solitary tumor (n = 93, 72.1%) and an AFP level ≤400 μg/L at recurrence (n = 108, 83.7%). Median rOS for patients received curative treatments, locoregional TAT, systemic therapies, and combined therapies were NA (95% CI, 46.3–NA), 32.5 months (95% CI, 15.8–NA), NA (95% CI, 7.43–NA), 24.9 months (95% CI, 16.5–NA), respectively, with significant difference (p = 0.0007, Figure 5a).
For patients with rBCLC stage B disease (n = 41), TAT combined with systemic therapies (n = 28, 68.2%) served as the primary treatment approach. Additionally, 2 patients remained eligible for curative treatments, both of whom underwent TACE-RFA. Six patients received TACE alone, 4 received systemic therapy alone, and 1 had an unspecified treatment regimen (Fig. 1). Median rOS for patients received curative treatments, locoregional TAT, systemic therapies, and combined therapies were NA, 54.6 months (95% CI, 19.8–NA), 16.7 months (95% CI, 13.6–NA), 33.4 months (95% CI, 20.0–NA), respectively, with no significant difference (p = 0.22, Figure 5b). HR was not available in this group because of the small sample size.
For patients had recurrent HCC staging rBCLC C (n = 91). Treatment modalities based on systemic therapies represented the predominant approach, including systemic therapies alone (n = 40) or locoregional TAT combined with systemic therapy (n = 18). Additionally, 20 patients still underwent curative treatment, consisting mainly of ablation or resection of extrahepatic lesion (Fig. 1). The remaining 10 patients had undocumented therapy. In this group with rBCLC C recurrence, the median tumor size was 14.5 mm (IQR, 9.0–22.3). Most patients (n = 63, 69.2%) presented with multifocal recurrence and extrahepatic lesion (n = 75, 82.4%), 71.4% (n = 65) had AFP ≤400 μg/L at recurrence, and 4.4% (n = 4) exhibited vascular invasion at the time of recurrence. Median rOS for patients received curative treatment, locoregional TAT, systemic therapies, and combined therapies were NA (95% CI, 30.8–NA), 35.2 months (95% CI, 5.3–NA), 34.9 months (95% CI, 22.7–NA), and 20.8 months (95% CI, 12.9–NA), respectively, with significant difference (p = 0.04, Figure 5c).

Discussion
This study included 419 uHCC patients, demonstrating that initially uHCC patients can achieve favorable survival outcomes if successful conversion resection was achieved. Previous study compared survival between patients who underwent conversion resection and those who received non-surgical therapy after conversion, showing significantly longer median OS and PFS in the resection group, underscoring the critical role of surgical resection in HCC management [26]. Our findings further supported that conversion resection was a vital treatment strategy for patients with locally advanced, initially unresectable HCC. Additionally, we provided a detailed analysis of patients who experienced recurrence after conversion resection, which had not been reported. Treatment strategies for different recurrent stages based on BCLC staging were discussed, offering insights for managing recurrence after conversion resection.
Our study indicated that MPR identified in postoperative pathology was an independent protective factor for both OS and RFS, which was consistent with the findings reported by Ou-yang L-Y et al. [22]. Additionally, analysis based on mRECIST criteria showed that patients achieving radiological CR had significantly better OS and RFS. And among these patients, over 80% and nearly 70% demonstrated MPR and pathological complete response in postoperative pathology, respectively, indicating good consistency between radiological and pathological assessments regarding tumor response. Notably, no significant survival differences were observed between patients classified as PR and SD according to RECIST 1.1. Furthermore, a reduction in tumor markers such as AFP and PIVKA-II before and after conversion therapy was associated with significantly improved RFS. These findings suggested that tumor activity, rather than tumor size alone, may be a more important indicator of prognosis, and clinicians should place greater emphasis on tumor activity when assessing treatment response.
The necessity of adjuvant therapy following surgery has long been a controversial topic in the management of HCC. Furthermore, whether adjuvant treatment provides survival benefit after successful conversion resection for uHCC remains an area with limited evidence and has been scarcely investigated. Our study found no survival benefit from adjuvant therapy after conversion resection, though this result needs prospective validation. Similar results were found after PSM and IPTW. Subgroup analysis revealed that adjuvant therapy did not confer a significant benefit compared to active surveillance in either OS or RFS across any of the predefined subgroups. Therefore, we preliminarily concluded that routine administration of postoperative adjuvant therapy might offer limited benefit to patients undergoing conversion resection. Many studies have already discussed adjuvant therapy for high recurrence risk HCC, but few exhibited positive outcomes in OS. IMbrave050 demonstrated that in patients at high risk of recurrence after tumor resection or ablation, adjuvant therapy with the A+T regimen (atezolizumab plus bevacizumab) significantly improved RFS compared with active surveillance. However, due to the relatively short follow-up time for OS, meaningful conclusions regarding OS cannot yet be drawn [14]. Similarly, a study by Wang et al. [27]. reported consistent findings, showing that adjuvant treatment with sintilimab significantly improved RFS in MVI-positive patients after curative surgery, though no significant difference was observed in OS Additionally, a multicenter prospective study by Li et al. [15] indicated that postoperative adjuvant HAIC significantly prolonged disease-free survival in MVI-positive HCC patients compared with active surveillance. Nevertheless, there were no significant differences in 1-, 2-, or 3-year OS rates between the two groups. Thus, the results from these studies consistently suggested that while postoperative adjuvant therapy significantly prolonged RFS compared with no adjuvant therapy, it did not lead to a significant improvement in OS. Our study shared similarity with these prospective trials in that postoperative adjuvant therapy did not extend OS. This might be attributed to the fact that OS is influenced not only by adjuvant treatment but also by patterns of recurrence and subsequent therapies. Additionally, even the most effective currently available systemic regimen for uHCC achieves an objective response rate no greater than 40% [28]. Consequently, the subset of patients who are both responsive to adjuvant therapy and derive clinical benefit from it may constitute a very small proportion of the overall cohort, likely resulting in insufficient statistical power to detect a significant difference. On the other hand, our finding also shared distinction with these trials in that adjuvant therapy failed to even prolong RFS or reduce recurrence rates. An important contributing factor is that our study population exclusively comprised patients who had undergone conversion therapy prior to surgery, whereas the aforementioned prospective trials enrolled only those who proceeded directly to resection. Mechanistically, HCC exhibits high genetic heterogeneity. While conversion therapy induces tumor necrosis or regression, it may also select for and expand resistant clones [29, 30]. Moreover, residual micrometastases after conversion resection often harbor resistance mutations (e.g., VEGFR2/PDGFR mutations associated with TKI resistance), thereby diminishing the efficacy of adjuvant treatments [31]. In summary, our study indicated that adjuvant therapy after conversion resection did not reduce recurrence rates or prolong survival. Instead, it may pose potential risks of liver function damage and impose additional financial burdens on patients. Therefore, we recommend that patients adhere to regular antiviral medication, undergo active surveillance, and focus on preserving liver function. Even in cases of recurrence, long-term survival may still be achievable through curative treatments such as surgery or ablation.
Recurrence is a pivotal factor adversely affecting long-term postoperative survival, irrespective of whether resection follows conversion therapy or is performed directly. Our data demonstrated that recurrence was associated with a more than 50% reduction in the 5-year survival rate. Consequently, understanding the characteristics and management of recurrence after conversion surgery is crucial. The 5-year recurrence rate in our series of 419 patients was 63.9%, which is comparable to rates reported for patients undergoing primary resection [32]. However, 67% of recurrences in our cohort occurred within the first postoperative year, with a median time to recurrence of only 7.2 months, a duration shorter than previously reported for patients undergoing primary resection [32]. This higher rate and shorter duration suggested a predisposition for early recurrence (within 1 year) following conversion resection. This might be attributed to the fact that patients requiring conversion therapy typically presented with a higher baseline tumor burden, more advanced disease stage, and inherently more aggressive tumor biology, all of which predisposed them to the presence of radiographically occult micrometastases. While conversion therapy effectively reduced the size or activity of the primary tumor, its efficacy against these slowly proliferating micrometastatic deposits was often limited. Subsequent surgical resection, accompanied by inflammatory responses, ischemic injury, and oxidative stress, might then create a favorable environment for the growth of residual micrometastases and the implantation of circulating tumor cells, precipitating early recurrence [33]. Given that our study found no evidence of benefit from postoperative adjuvant therapy, intensive surveillance within the first postoperative year is particularly critical. While international guidelines generally recommend follow-up every 3–4 months for the first 2 years post-resection [4, 34, 35], the pronounced risk of early recurrence in the conversion resection population leads us to specifically recommend a more intensive schedule of every 3 months during the first postoperative year.
We also analyzed the location of recurrence in this cohort. Isolated intrahepatic recurrence remained the most common pattern (69.3%). However, in contrast to patients with initially resectable disease, the proportion of isolated extrahepatic recurrence was 22.6%, significantly higher than the 12% reported by Tabrizian et al. [32] and the 6.7% reported by Qi et al. [36]. This spatial distribution pattern reveals that, when extrahepatic recurrence does occur in patients undergoing conversion resection, it is more likely to present as an isolated extrahepatic event without concurrent intrahepatic lesion. Isolated extrahepatic recurrence retains the potential for curative-intent treatment and rarely impairs liver function, resulting in a more favorable prognosis compared to combined intra- and extrahepatic recurrence. This interesting finding underlies the lack of survival difference between rBCLC stage B and C patients in our study. Consequently, in postoperative surveillance, we recommend alternating MRI of the liver with CT scans that include chest sections to ensure extrahepatic recurrences, particularly lung metastases, are not overlooked.
How to treat recurrence after conversion resection has rarely been reported in previous studies. Our findings demonstrated a median rOS of 40.2 months in these patients, suggesting that effective treatment can still yield significant survival benefit despite recurrence. Parrisa et al. [32] in 2014 reported that the median OS after HCC recurrence was 21 months, with a 5-year survival rate of only 23%. This comparison suggested that the prognosis for recurrent HCC has significantly improved over the past decade, likely due to advancements in interventional techniques, targeted drugs, and immunotherapies. Given the considerable diversity in treatments for recurrent patients and based on our center’s evidence showing similar efficacy of surgical resection, ablation, and stereotactic body radiotherapy for recurrent small HCC [23, 37], these three methods were collectively classified and discussed as curative treatment. For recurrent cases classified as rBCLC 0 or rBCLC A stages, curative treatment remained the primary and most effective approach for achieving long-term survival. However, within this patient cohort, a subset received TAT with or without systemic therapy due to factors such as multiple lesions (2–3 lesions), tumors being too small for ultrasonographic detection, high surgical complexity, or patient refusal of surgery. Their prognosis was significantly worse compared to patients who underwent curative treatment. Furthermore, no survival difference was observed between patients who received TAT alone and those who received combined TAT and systemic therapy. This suggested that for this category of early-stage recurrence patients, there is currently no evidence supporting additional survival benefit or improved disease control from more aggressive treatment strategies, a conclusion that warrants further validation through prospective clinical trials. In patients with rBCLC stage B disease, curative treatment is often challenging due to the presence of multiple lesions. Nevertheless, a small subset of patients with few and small lesions may still qualify for RFA following TACE. In this study, only 2 such patients were identified, both of whom remained alive at the end of follow-up. In contrast, over 20% of rBCLC stage C patients were able to undergo curative treatment and achieve long-term survival, which was a characteristic distinct from primary HCC. These patients typically presented with limited, small extrahepatic metastases that did not compromise liver function, enabling them to receive radical approaches such as metastasis ablation, resection, or radiotherapy, ultimately attaining survival rates comparable to those in rBCLC stages 0 and A. Furthermore, treatment strategies based on systemic therapy represented the predominant approach for rBCLC C patients. Notably, combination therapy did not demonstrate superior OS benefits compared to systemic therapy alone. Therefore, irrespective of the disease stage at recurrence, pursuing curative treatment remains a critical pathway to long-term survival. Additionally, for patients in whom curative treatment is not feasible, the combination of locoregional and systemic treatments does not appear to provide a survival advantage over either modality alone. Therefore, in the management of recurrence after conversion resection, simple and focused treatment strategies might be preferable, as aggressive combined regimens require careful consideration of their potential impact on liver function.
This study had several limitations: (1) as a single-center retrospective analysis, the sample size was relatively limited and may introduce selection bias. (2) The patients included in this study spanned over a decade, during which the treatment landscape for HCC has evolved substantially. Over this period, systemic therapy transitioned from the dominance of sorafenib to the emergence of diverse targeted and immune agents, while HAIC has been validated by multiple prospective studies demonstrating superior efficacy in uHCC. These developments have introduced considerable heterogeneity into our analysis. (3) Over 90% of the HCC patients in this study were HBV-related. Whether the efficacy of conversion resection and long-term postoperative management for non-HBV-related HCC, which is known to be less responsive to immunotherapy, remains comparable to the outcomes reported herein requires further investigation.

Conclusion
This study provided a comprehensive view of clinical outcomes for patients with uHCC undergoing curative resection after conversion therapy, demonstrating favorable overall prognosis in this population. Prognosis was significantly associated with radiological response during conversion therapy, changes in tumor markers, and pathological response. Current evidence does not support the routine use of postoperative adjuvant therapy. Given the high risk of recurrence within the first postoperative year, active surveillance is essential. In cases of recurrence, curative-intent treatment should be prioritized whenever possible. When curative options are not feasible, simple and less aggressive therapeutic approaches are preferable.

Statement of Ethics
This study was approved by the Ethics Committee of Sun Yat-sen University Cancer Center (B2023-673-01) and was performed in full accordance with the Declaration of Helsinki. The data were collected, reviewed, de-identified, and anonymized prior to analysis, and the Ethics Committee waived the requirement for informed consent.

Conflict of Interest Statement
The authors have no relevant financial or nonfinancial interests to disclose.

Funding Sources
This work is funded by the National Natural Science Foundation of China (82103566 to D.D. Hu), the Postdoctoral Fellowship Program of CPSF (No: GZC20251523 to Z.Y. Yang), and the China Postdoctoral Science Foundation (No: 2025M782360 to Z.Y. Yang).

Author Contributions
Rui Xiao drafted the manuscript. Rui Xiao, Zhenyun Yang, Zehao Zheng, and Na Liu participated in the research design. Rui Xiao, Wei Yu, and Qingyang Lin collected data and analyzed data. Jinbin Chen and Juncheng Wang helped draw figures and tables. Zhongguo Zhou and Yaojun Zhang helped a lot in academic decision-making and content guidance. Minshan Chen and Dandan Hu provided the financial support and supervised the study. All authors read and approved the final manuscript.