Oncological Outcomes and Treatment Strategies for Borderline Resectable Hepatocellular Carcinoma Based on the Oncological Resectability Classification.

Introduction
Hepatocellular carcinoma (HCC) is a major cause of cancer-related deaths worldwide. Hepatic resection is still considered the standard treatment for selected patients [1–5]. However, in the absence of a standardized oncological framework to evaluate resectability, substantial variability in the prognosis has been observed, even among patients undergoing hepatic resection with curative intent. To address the existing lack of consensus, a novel oncological resectability classification was proposed in 2023 by the Japan Liver Cancer Association and the Japanese Society of Hepato-Biliary-Pancreatic Surgery. This classification stratifies patients into resectable (R), borderline resectable 1 (BR1), and borderline resectable 2 (BR2) categories [6]. Patients classified as BR are considered less likely to derive sufficient benefits from surgery alone, reflecting their unfavorable oncological characteristics.
Recent advances in systemic therapies, including immune checkpoint inhibitors and molecular-targeted agents, have resulted in significant improvements in tumor response rates and an expansion of treatment options for advanced HCC [7–12]. These developments have brought a paradigm shift in clinical practice, in which pharmacological treatment prior to surgery is being considered frequently to optimize long-term outcomes in high-risk patients.
The clinical appropriateness of surgery for borderline resectable HCC requires careful reappraisal. While both the BR1 and BR2 groups are defined by unfavorable factors, the clinical course may not be uniform, and treatment outcomes can be improved by tailoring strategies such as the selective use of upfront resection or the integration of preoperative systemic therapy [13–15].
The present study retrospectively analyzed initial hepatic resection cases for HCC at our institution based on the oncological resectability classification. The objective of this study was to elucidate which patients within the BR1 and BR2 categories would benefit from upfront surgery, which would require preoperative systemic therapy, thereby providing insights into optimal treatment strategies.

Methods

Patients
A retrospective analysis was conducted on consecutive patients who underwent initial hepatic resection for HCC at Nagasaki University Hospital between January 2014 and December 2024. Patients who had received prior local therapies, including radiofrequency ablation (RFA) or previous hepatic resection, were excluded from this study, as were those with recurrent HCC, combined hepatocellular-cholangiocarcinoma, or insufficient clinical records. Patients with adequate clinical, radiological, and follow-up data were included in the final analysis.

Classification of Oncological Resectability
Patients were stratified according to the oncological resectability criteria jointly proposed by the Japan Liver Cancer Association and the Japanese Society of Hepato-Biliary-Pancreatic Surgery in 2023. Patients were classified into three groups: R, the oncological status for which surgery alone may offer clearly better survival outcomes compared to the other treatment; BR1, the oncological status for which surgical intervention as a part of multidisciplinary treatment may offer a survival benefit; and BR2, the oncological status for which the efficacy of surgery is indeterminate and surgical indication should be carefully determined under the standard multidisciplinary management of HCC [6]. The various tumor factors constituting the oncological resectability criteria, including intrahepatic maximal tumor size, intrahepatic tumor number, presence of vascular invasion, and extrahepatic metastases, were evaluated using imaging studies.

Indications and Procedure of Hepatic Resection
The inclusion of hepatic resection of HCC was determined by the general condition tolerable to surgery, according to the Clinical Practice Guidelines for HCC proposed by the Japan Society of Hepatology [16], and by the liver function as measured by the indocyanine green retention rate at 15 min (ICG-R15) and liver scintigraphy findings using Tc-99m [17]. Hepatic resections were performed using open, laparoscopic, or robotic approaches. Robotic hepatic resection was initiated in May 2021 in our department and performed using the da Vinci Xi Surgical System (Intuitive Surgical, Inc., Sunnyvale, CA, USA). Hepatic transection was performed using an ultrasonic surgical aspirator (CUSA; Integra Life Sciences, Plainsboro, NJ, USA), an ultrasonic coagulating and cutting device, or a double bipolar technique in robots. The Pringle maneuver was routinely used to occlude blood inflow to the liver in all three procedures. The decision to perform conversion surgery for advanced HCC was determined at a multidisciplinary team meeting based on the oncological stability and technical feasibility of complete resection.

Indication and Treatment of Systemic Chemotherapy
Systemic chemotherapy was administered in accordance with the following eligibility criteria: HCC not amenable to locoregional treatment and Child-Pugh class A or B liver function. The atezolizumab plus bevacizumab (Ate/Bev) regimen consisted of intravenous administration of a fixed dose of 1,200 mg of atezolizumab combined with 15 mg/kg of bevacizumab, administered every 3 weeks. Lenvatinib (LEN) was administered orally at a dosage of 8 mg/day to patients with a body weight <60 kg and 12 mg/day to those with a body weight ≥60 kg.

Follow-Up Protocol
In practice, after curative hepatic resection, we perform regular imaging and laboratory surveillance to enable early detection of recurrence. This typically includes dynamic CT or MRI and tumor marker tests every 3 months during the first 2 years, followed by continued monitoring every 6 months up to 5 years, in line with standard guidelines.

Data Collection
A comprehensive array of demographic, clinical, and tumor-related variables was retrieved from institutional databases. These included age, sex, hepatitis B surface antigen, hepatitis C antibody status, tumor markers (alpha-fetoprotein [AFP] and protein induced by vitamin K absence or antagonist-II [PIVKA- II]), liver function reserve (albumin, bilirubin, platelet count, prothrombin time, Child-Pugh score, and modified albumin-bilirubin [mALBI] grade), and tumor characteristics (size, number, vascular or biliary invasion). The perioperative variables included the surgical approach (open vs. minimally invasive), the type and extent of resection, operative time, blood loss, transfusion requirements, and postoperative complications, which were classified according to the Clavien-Dindo (CD) classification. Postoperative mortality was defined as death within 90 days of surgery. The utilization of preoperative systemic therapy, including Ate/Bev or LEN, has been documented considering its evolving role in conversion surgery strategies. Subsequent datasets encompassed recurrence patterns, treatment modalities for recurrent cases (including repeated hepatectomy, RFA, systemic therapy, or transarterial chemoembolization [TACE]), the recurrence-free survival (RFS) which was defined from the time of curative hepatic resection to the first tumor recurrence, and patients who died without recurrence were censored at their time of death, and the overall survival (OS). For patients classified as BR1 or BR2, additional subgroup analyses were conducted according to the receipt of preoperative systemic therapy and vascular invasion status. The study’s primary endpoint was the OS, compared across the R, BR1, and BR2 groups.

Statistical Analyses
Continuous variables are expressed as medians with interquartile ranges and compared using the Mann-Whitney U test. Categorical variables were analyzed using the chi-square test or Fisher’s exact test, as appropriate. Survival curves for the RFS and OS were generated using the Kaplan-Meier method, and intergroup differences were assessed using the log-rank test. Statistical significance was determined using a p value <0.05. All statistical analyses were performed using the JMP software program (SAS Institute, Cary, NC, USA).

Results

Patient Characteristics
A total of 184 patients were included in the analysis, comprising 149 in the R group, 21 in the BR1 group, and 14 patients in the BR2 group, respectively (Table 1). The median age was comparable across the 3 groups (75 years old in R, 70 years old in BR1, and 70 years old in BR2), and the proportion of female patients ranged from 4.7% in BR1 to 28.6% in BR2, without significant differences. The prevalence of hepatitis C virus infection and hepatitis B surface antigen positivity was comparable among the groups. Preoperative systemic therapy was administered significantly more often in the BR1 (33.3%) and BR2 (14.3%) groups than in the R group (3.4%). The median AFP levels did not differ markedly among the groups; however, the median PIVKA-II levels were markedly elevated in BR2 (1,790 mAU/mL) compared with R (112 mAU/mL) and BR1 (136 mAU/mL) (p < 0.01). The liver function, as indicated by ICG-R15, the liver-to-heart-plus-liver uptake ratio at 15 min, platelet count, and prothrombin time-international normalized ratio (PT-INR), was found to be comparable across the three groups. However, a statistically significant difference was observed in the proportion of patients with mALBI grade ≥2a, with BR2 demonstrating a higher percentage (71.4%), followed by R (34.2%), and BR1 (33.3%) (p < 0.01). With respect to perioperative variables, the median operative time was found to be longer in the BR groups (432 min in BR1 and 485 min in BR2) than in the R group (375 min). A statistically significant increase in blood loss was observed in BR2 (median 1,437 g) compared to that in R (533 g) and BR1 (551 g) (p < 0.01). The R group exhibited a significantly higher prevalence of minimally invasive surgery (59.1%) than the BR1 (19.1%) and BR2 (7.1%) groups (p < 0.01). A higher frequency of perioperative transfusions was observed in BR2 (50.0%) than in R (24.2%) or BR1 (23.1%). Conversely, the incidence of major postoperative complications (CD grade ≥3a) was significantly higher in BR2 (57.1%) than in R (16.8%) or BR1 (19.0%) (p < 0.01). The median postoperative hospital stay was also the longest in BR2 (25.5 days), followed by BR1 (15 days) and R (13 days).

Survival Outcomes
The median follow-up duration was 39 months (range, 0.5–60). Kaplan-Meier analysis revealed substantial disparities in the RFS among the three groups (Fig. 1a). The 5-year RFS rates were 43.9% in the R group, 39.3% in BR1, and 20.0% in BR2 (p < 0.001). Patients in the R group exhibited the most favorable recurrence profile, while those in the BR2 group experienced early and frequent recurrences, with the majority relapsing within 2 years after resection. The BR1 group demonstrated an intermediate recurrence pattern, with RFS inferior to that of R, but still significantly superior to that of BR2. Furthermore, the OS varied substantially among the groups (Fig. 1b). The 5-year OS rates were 74.6% in the R group, 76.9% in BR1, and 33.8% in BR2 (p < 0.001).

Subgroup Analyses of BR1
A comprehensive analysis was conducted in the BR1 subgroup, stratifying patients based on their vascular invasion status and the utilization of preoperative systemic therapy (Fig. 2). Patients with vascular invasion who underwent upfront surgery without preoperative therapy had suboptimal outcomes. One patient experienced early recurrence, leading to cancer-related death, while several others developed rapid intrahepatic recurrence or extrahepatic metastasis, resulting in unfavorable post-hepatic resection survival. Conversely, patients with BR1 who exhibited vascular invasion and underwent preoperative systemic therapy showed enhanced outcomes. Despite uniformly high PIVKA-II levels and evidence of vascular invasion, the postoperative survival proved to be relatively favorable, with a single death attributable to cancer progression and further fatality resulting from liver failure during the follow-up period. Meanwhile, BR1 patients without vascular invasion exhibited a favorable long-term survival following upfront resection. Despite the prevalence of intrahepatic recurrences, these are largely amenable to management through repeated local treatments, including RFA, TACE, stereotactic radiotherapy, and repeat hepatic resection. Notably, no cancer-related mortalities occurred in this subgroup, and the majority of patients exhibited a prolonged survival, with deaths primarily attributed to noncancerous causes.

Subgroup Analyses of BR2
Among the 14 patients classified as having BR2, outcomes were uniformly poor regardless of the presence or absence of vascular invasion. The median tumor size was substantial, measuring 6.3 cm in patients with documented vascular invasion and 7.5 cm in those without such invasion. Furthermore, most patients had multiple tumors, with a median count of two to three nodules. As indicated in Table 2, elevated levels of both AFP and PIVKA-II were observed, with ≥70% of patients exhibiting an impaired liver function, as defined by mALBI grade ≥2a. A Kaplan-Meier analysis confirmed that there was no significant survival difference between BR2 patients with and without vascular invasion (5-year OS: 28.6% vs. 35.7%, p = 0.88). Both groups exhibited a dismal RFS (Fig. 3a).
When the data were stratified according to preoperative systemic therapy in patients with vascular invasion, a trend toward improved outcomes was observed. However, statistical significance was not achieved because of the small cohort size (Table 3). Patients who underwent upfront surgery without systemic therapy (n = 5) exhibited early recurrence and a poor survival, whereas those who responded to preoperative treatment (n = 2) achieved durable disease control (Fig. 3b).

Discussion
In this study, we conducted a retrospective analysis of patients who underwent initial hepatic resection for HCC at our institution. The present analysis was based on the oncological resectability classification proposed in 2023 [6]. We examined the validity of surgery in borderline resectable categories. Consequently, the patients in the R group exhibited the most favorable outcomes and served as the reference group. Conversely, both the BR1 and BR2 groups demonstrated elevated recurrence rates, with notably poor survival outcomes in the BR2 group. It is noteworthy that while BR1 patients exhibited a significantly inferior RFS, their OS was comparable to that of the R group, whereas BR2 patients demonstrated poor outcomes with regard to both the RFS and OS.
Patients classified as BR1 exhibited an inferior RFS compared to those classified as R (5-year RFS 39.3% vs. 43.9%), indicative of a heightened propensity for postoperative recurrence. However, this does not result in a corresponding OS disadvantage. The 5-year OS rate in the BR1 group (76.9%) was nearly equivalent to that in the R group (74.6%). This paradox can be explained by the existence of effective post-resection treatments. In the present cohort, the majority of BR1 patients who developed intrahepatic recurrences were able to undergo additional therapies, including repeat hepatic resection, RFA, TACE, or stereotactic radiotherapy, to control the disease. The aggressive surveillance and intervention for recurrence observed in this study contributed to the prolonged survival of the patients. Indeed, Shindoh et al. [18] also reported that despite nearly 80% recurrence at 5 years in BR1, the 5-year OS still approached 60%, underscoring the critical role of post-recurrence interventions in prolonging the survival. In summary, while BR1 tumors exhibited a higher propensity for recurrence, effective management of these recurrences enabled numerous BR1 patients to attain long-term survival outcomes comparable to those of truly resectable HCC.
A subgroup analysis of BR1 cases indicated the significance of microvascular invasion as a prognostic factor. BR1 patients without vascular invasion exhibit commendable long-term prognoses after upfront surgery. Despite the occurrence of tumor recurrence in a significant proportion of cases, these were predominantly amenable to locoregional therapeutic interventions. No cancer-related mortality was observed in this subgroup during the follow-up period. This finding suggests that for BR1 tumors without macrovascular invasion, hepatic resection provides a significant survival benefit and recurrences can be managed with curative or semi-curative intent, thereby enabling patients to achieve prolonged survival. In contrast, BR1 patients with vascular invasion exhibited a distinctly more aggressive disease course when managed solely with surgery. In several cases, early and diffuse recurrence occurred shortly after hepatic resection, resulting in a poor survival. This finding is consistent with the established adverse prognostic impact of microvascular invasion in HCC, as it frequently indicates micrometastatic dissemination and aggressive tumor biology [19]. However, the findings of this study suggest that the adverse consequences of vascular invasion in BR1 might be mitigated, at least in part, by preoperative systemic therapy. In the present series, this approach was associated with favorable disease control; patients who received preoperative systemic therapy demonstrated at least comparable survival outcomes despite a higher tumor burden, such as elevated tumor marker levels, and only one cancer-related death occurred during follow-up. Despite the limited sample size, these observations suggest a possible advantage of integrating systemic therapy prior to resection in high-risk BR1 patients. However, this potential benefit should be interpreted with caution and confirmed in larger studies. Kawamura et al. [20] recently demonstrated that BR1 patients, despite having a significantly worse oncological status than resectable cases, achieved a comparable overall survival when systemic therapy was integrated, especially in combination with multidisciplinary approaches such as locoregional therapy and R0 resection. A rational treatment algorithm can be proposed from the available data. For BR1 patients without vascular invasion, upfront surgery is recommended as it yields favorable survival rates with manageable recurrences. For BR1 patients with vascular invasion, systemic therapy prior to surgery should be considered to potentially improve tumor biology and outcomes. This bespoke strategy has the potential to optimize the benefits of surgery in patients with borderline resectable tumors.
Patients classified as BR2 exhibited uniformly poor outcomes in the present study, thereby emphasizing the aggressive nature of the tumors in this category. The BR2 designation encompasses cases with very large tumors, multifocal disease, and often elevated tumor markers (AFP and PIVKA-II) alongside an impaired liver function – an overall profile of advanced disease that approaches traditional “unresectable” criteria. Despite undergoing curative intent resection, the majority of patients experienced relapses within a short time frame. The majority of BR2 patients in the present series experienced recurrence within the first year or two following surgeries, with a 5-year RFS rate of 20.0%. The OS outcomes were similar; the 5-year OS for BR2 was only 33.8%, and the 1-year OS was approximately 50%, indicating that nearly half of the BR2 patients succumbed within a year of surgery despite the absence of gross residual disease at resection. The present findings are consistent with those reported by Komatsu et al. [21] who demonstrated that the median survival time for BR2 patients following hepatic resection was 20.1 months. It is noteworthy that, in contrast to the findings in BR1, the presence of vascular invasion did not significantly stratify outcomes within BR2. Once a tumor fulfills the BR2 criteria, such as very high tumor burden or marginally resectable extrahepatic spread, the disease course is suggested to be dominated by its aggressive biology, and traditional adverse features, such as microvascular invasion, simply add to an already dire prognosis. From a clinical perspective, these findings indicate that surgery alone is largely inadequate for BR2 HCC and that alternative or additional therapies are crucial to improve outcomes in this group.
While the overall prognosis for BR2 patients was poor, our experience provides a note of cautious optimism, as a subset of BR2 patients achieved unexpectedly favorable outcomes when systemic therapy was incorporated as part of their treatment. Notably, 2 patients in the BR2 group who received preoperative systemic therapy exhibited notably improved outcomes. We report the case of a BR2 patient with elevated tumor marker levels who underwent preoperative therapy with Ate/Bev. Following this treatment, the tumor markers normalized and imaging revealed tumor shrinkage. At the time of resection, both the primary tumor and the initially present portal vein tumor thrombus exhibited complete pathological necrosis. As a result, the patient remained disease-free for over 3 years following hepatic resection, which represents an exceptionally favorable outcome considering the prognosis of BR2 disease [22]. Another BR2 patient in our cohort who underwent systemic therapy prior to surgery, although recurrence eventually occurred, achieved a long-term survival after resection. The findings of the present study are consistent with emerging clinical reports on conversion surgery in advanced HCC. Xu et al. reported 3-year OS rates of approximately 80% after successful conversion surgery [23]. Furthermore, patients with a complete pathological response after conversion surgery have demonstrated favorable survival outcomes [24]. The presented cases combined with these previous findings, underscore the necessity of a multidisciplinary multimodal approach for BR2 patients. Rather than performing surgery in all cases of BR2, an ideal strategy would be to individualize the timing and necessity of resection based on the treatment response. In the event of a robust response that may be induced by significant tumor shrinkage or necrosis, conversion surgery has the potential to be used as curative treatment. Conversely, patients whose tumors progress on systemic therapy are spared the morbidity of extensive surgery that would likely have been futile. This approach ensures that surgery, when undertaken in BR2 cases, is more likely to provide benefits and leverages advances in systemic therapy to improve overall outcomes.
As with any retrospective study, the present analysis has certain limitations. The sample size of BR2 patients was relatively small, and treatments were not randomized, which may introduce selection bias in terms of patients who received preoperative therapy or aggressive recurrence treatments. Furthermore, the results for BR1 were not derived from the statistically powered subgroup analyses. Indeed, the visual representations of the data, which were presented as swimmer plots of the selected cases, should be interpreted with caution. In addition, the therapeutic landscape of HCC underwent significant alterations during the study period (2014–2024). Notably, LEN and Ate/Bev became clinically available only after 2018, signifying that the follow-up period for patients treated with these agents remains relatively brief. Consequently, the influence of time period effects on outcomes is a potential concern.

Conclusion
Although BR1 HCC is associated with a higher risk of recurrence than resectable disease, appropriate surgical resection combined with intensive surveillance and effective post-recurrence or preoperative systemic therapies may allow selected patients to achieve favorable long-term survival. In particular, treatment strategies tailored to vascular invasion status appear essential to optimize outcomes in BR1 patients. In contrast, BR2 HCC is generally associated with poor outcomes, and surgery alone is often insufficient to achieve durable disease control. Therefore, multidisciplinary treatment strategies incorporating systemic and locoregional therapies should be considered, with the role of surgery carefully evaluated on an individual basis. The management of HCC should be individualized, with therapeutic strategies tailored to the distinct oncological profiles defined by classification.

Statement of Ethics
The study was approved by the Institutional Ethics Committee of Nagasaki University (Approval No. 19102143-3) and conducted in accordance with the Declaration of Helsinki (1964) and its later amendments. As this was a retrospective study, the requirement for informed consent was waived by the Institutional Ethics Committee of Nagasaki University. Study details were disclosed on the hospital website.

Conflict of Interest Statement
The authors have no conflicts of interest to declare.

Funding Sources
This work was supported by a Health and Labour Sciences Research Grant from the Ministry of Health, Labour and Welfare of Japan (Grant Number: 24HB2002).

Author Contributions
T.H., A.S., H.M., A.K., K.M., H.I., S.N., A.S., K.K., T.A., R.S., H.M., and S.E. contributed to conception and design; T.H. and H.M. collected the data; T.H., A.S., and H.M. performed the statistical analysis; T.H., A.S., H.M., and S.E. drafted the manuscript. T.H., A.S., H.M., A.K., K.M., H.I., S.N., A.S., K.K., T.A., R.S., H.M., and S.E. read and approved the final manuscript.