Surgery vs. Watch-and-Wait Strategy in Complete Responders for Hepatocellular Carcinoma

Improved systemic and locoregional treatments for hepatocellular carcinoma (HCC) have enabled some patients to achieve a radiologic complete response (rCR). Nevertheless, rCR does not necessarily indicate a pathological complete response because viable tumor cells may remain despite the absence of visible active tumor on imaging. The optimal management of patients with confirmed rCR who are candidates for curative-intent liver resection remains uncertain. Surgical consolidation may remove residual viable tumor and provide pathological information, but it also exposes patients to perioperative risks. Continuing systemic maintenance therapy may avoid unnecessary surgery but may leave clinically undetectable residual disease.

SWITCH-R is a multicenter, prospective, open-label, randomized, exploratory study comparing surgical resection with systemic maintenance therapy in patients with HCC who achieve and maintain rCR after anticancer treatment. The treatment leading to rCR must include a programmed cell death protein 1 or programmed death-ligand 1 (PD-1/PD-L1) inhibitor combined with a targeted or anti-angiogenic agent and may also include locoregional treatment, such as transarterial chemoembolization, hepatic arterial infusion chemotherapy, radiotherapy, or ablation.

After the first documentation of rCR according to modified Response Evaluation Criteria in Solid Tumors (mRECIST), patients continue their original systemic treatment and undergo confirmatory imaging using the same imaging method 4 to 8 weeks later. Patients with sustained rCR, no evidence of extrahepatic or other active HCC, and a multidisciplinary team assessment indicating that curative R0 liver resection is feasible are eligible for randomization.

Approximately 60 participants will be randomly assigned in a 1:1 ratio to the surgical resection group or the maintenance therapy group. Randomization will be stratified according to tumor resectability before the treatment that resulted in rCR (resectable versus unresectable). Because the two strategies are visibly different, participants and treating investigators will not be masked to treatment assignment.

Participants assigned to the surgical resection group will undergo curative-intent liver resection after an appropriate washout period based on their previous systemic treatment. Resected tissue will be assessed for margin status, pathological complete response, major pathological response, the proportion of residual viable tumor, microvascular invasion, and other pathological characteristics. Participants who recover adequately will resume their pre-randomization systemic treatment approximately 4 to 8 weeks after surgery.

Participants assigned to the maintenance therapy group will continue the same PD-1/PD-L1 inhibitor plus targeted or anti-angiogenic treatment they were receiving before randomization. Dose modifications and temporary treatment interruptions are permitted for treatment-related toxicity, but changing therapy for reasons other than disease progression is not planned.

In both groups, systemic treatment will continue until disease recurrence or progression, unacceptable toxicity, death, or fulfillment of the study-defined drug-off criteria. Systemic treatment may be stopped after all of the following conditions have been maintained for at least 24 weeks: no radiologic evidence of recurrence or metastasis after surgery in the surgical group, or continued mRECIST rCR in the maintenance group; no new tumor lesions; normalization and sustained stability of any tumor markers that were elevated before treatment, including alpha-fetoprotein or protein induced by vitamin K absence-II; and no other clinical evidence of active tumor. The decision to stop treatment will be made by a multidisciplinary team. Participants who stop treatment will remain in imaging, patient-reported outcome, and survival follow-up.

Tumor imaging will include contrast-enhanced computed tomography and/or magnetic resonance imaging of the chest, abdomen, and pelvis. Assessments are planned every 9 weeks through week 36 after randomization and every 12 weeks thereafter. Disease recurrence or progression will primarily be determined according to Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST v1.1), with mRECIST used as a supplementary assessment. Imaging will undergo blinded independent central review.

The primary endpoint is time to treatment failure as assessed by independent radiologic review, defined in this protocol as the time from randomization to the first occurrence of disease recurrence or progression according to RECIST v1.1 or death from any cause. Secondary and exploratory evaluations include overall survival, investigator- and mRECIST-assessed time to treatment failure, treatment discontinuation under the drug-off criteria, time to drug-off, drug-free survival, safety, surgical outcomes, patient-reported quality of life, pathological response, circulating tumor DNA, and tumor-marker dynamics.

The study uses a randomized Pick-the-Winner screening design. The planned sample size of 60 participants, with 30 participants in each group, is based on the exploratory purpose of the study and the feasibility of recruiting this selected patient population rather than on a formal superiority or noninferiority hypothesis. Treatment effects will therefore be described primarily using estimates and 95% confidence intervals to inform the design of a future confirmatory study.