Continuation of First-Line Systemic Therapy with Progression-Directed Radiotherapy in Patients with Oligoprogressive Hepatocellular Carcinoma: A Multicenter, Single-Arm, Phase II Trial

Abstract

Purpose:

Evidence for radiotherapy (RT) in oligoprogressive hepatocellular carcinoma (OP-HCC) is limited. We evaluated the efficacy and safety of progression-directed RT (PDRT) alongside ongoing first-line systemic therapy (FLST) in patients with OP-HCC.

Patients and Methods:

Patients who developed OP-HCC during FLST were enrolled and received PDRT with a biologically effective dose of at least 60 Gy while continuing FLST. The primary endpoint was progression-free survival (PFS); secondary endpoints were overall survival (OS), objective response rate (ORR), disease control rate (DCR), duration of response (DOR), toxicities, and quality of life (QoL).

Results:

From March 2024 through May 2025, 36 patients were enrolled from 10 cancer centers. At a median follow-up time of 10.9 months, median PFS time was 7.0 months (95% confidence interval 4.9–9.7), with 3-, 6-, and 9-month PFS rates of 73.7%, 64%, and 38.8%, respectively. Type of FLST and albumin–bilirubin (ALBI) grade at oligoprogression were independently associated with PFS. Median OS and DOR times were not reached; 1-year OS rates were 86.4%, and 3-, 6-, and 9-month DOR rates were 84.6%, 79.6%, and 70.8%, respectively. ORR and DCR were 64.7% and 98.0%, respectively. QoL measures generally remained stable, except for transient increases in fatigue and pain scores 1 month after PDRT. RT-related toxicities (mostly grades 1–2) occurred in 16 patients (44.4%), including grade ≥3 events in four patients (11.1%).

Conclusions:

Maintaining FLST with PDRT was effective, safe, and preserved QoL, supporting its feasibility for OP-HCC. FLST type and baseline ALBI grade may provide risk stratification and prognosis for PFS.

Translational Relevance.

Given the limited options and low efficacy of current second-line systemic therapy (SLST), an effective and well-tolerated treatment strategy is needed for patients who develop oligoprogressive hepatocellular carcinoma (OP-HCC) during first-line systemic therapy (FLST) to avoid premature transition to SLST. This prospective phase II trial in OP-HCC demonstrated that maintaining FLST with progression-directed radiotherapy (RT) resulted in favorable disease control and safety outcomes. These findings support the use of RT as a “1.5-line” treatment strategy to bridge FLST and SLST for OP-HCC.

Introduction

Hepatocellular carcinoma (HCC) is often diagnosed at advanced stages (1), and only 50%–60% of patients are eligible for systemic therapy at the time of diagnosis (2). Although immunotherapy-based first-line systemic therapy (FLST) has improved survival in advanced HCC (3), effective second-line systemic therapy (SLST) options remain limited (4), with modest response rates (3, 5), underscoring the need for strategies to bridge FLST and SLST. This issue is particularly relevant in patients who develop oligoprogression (defined as ≤5 lesions in ≤3 organs) during FLST. Although progression has traditionally prompted a switch to SLST, emerging evidence, including that from the IMbrave150 trial (6, 7), suggests that continuing systemic therapy beyond progression, often in combination with locoregional treatments such as radiotherapy (RT), is becoming more common.

RT has been widely investigated in oligometastatic and oligoprogressive (OP) disease across tumor types (8–12). However, in oligoprogressive HCC (OP-HCC), its role remains insufficiently defined. Seong and colleagues (13) demonstrated the feasibility of stereotactic ablative RT (SABR) in oligometastatic HCC; however, oligoprogression was not specifically addressed in that study. In our prior retrospective multicenter study (14), FLST maintenance combined with progression-directed RT (PDRT) was associated with longer median progression-free survival (PFS) than switching to SLST (8.6 vs. 3.1 months). However, the retrospective design limited causal interpretation. Therefore, we conducted this prospective trial to evaluate the efficacy and safety of continuing FLST in combination with PDRT in patients with OP-HCC.

Patients and Methods

This prospective, multicenter, single-arm phase II trial (NCT06261047) was conducted at 10 academic hospitals in China. The trial was conducted in accordance with the Declaration of Helsinki and the Good Clinical Practice guidelines of the International Council for Harmonization. The study protocol was approved by the Institutional Review Board at each participating center. All eligible participants provided written informed consent before any study-related procedures.

Eligible patients were ages 18 to 75 years and had a histologically, cytologically, or clinically confirmed diagnosis of primary HCC. All patients had received FLST, subsequently developed oligoprogression, and had Barcelona Clinic Liver Cancer (BCLC) stage B or C disease at enrollment. Lesions were required to be measurable according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 and suitable for RT. Additional criteria included a Child–Pugh score of ≤7, an Eastern Cooperative Oncology Group performance status of 0 to 1, stable disease (SD) on FLST for at least 3 months, and a life expectancy of at least 6 months. Detailed eligibility criteria and definition of oligoprogression are provided in the protocol (Supplementary Materials S1).

Systemic therapy and RT

All standard regimens for advanced HCC were permitted and administered according to the National Comprehensive Cancer Network Guidelines (version 1.2024) and the China Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2022 edition; ref. 15). The pre-oligoprogression FLST regimen was managed in a risk-adapted manner during RT. For patients treated with stereotactic body RT (SBRT) or other hypofractionated regimens, systemic therapy was generally continued without interruption. For those receiving conventionally fractionated RT, nonimmunotherapy systemic agents were temporarily withheld when the irradiated volume included the primary tumor, whereas systemic therapy was continued when the primary tumor was not included in the RT target volume.

All OP lesions received RT to a biologically effective dose (BED; LQ, α/β = 10) of at least 60 Gy during continued FLST. The RT technique, dose, and fractionation were individualized based on the location and characteristics of each lesion. Treatment was delivered as intensity-modulated RT (45–60 Gy in 25–30 fractions) or SBRT (30–50 Gy in 3–5 fractions), as appropriate. Dose constraints were guided by the QUANTEC consensus (16), with additional reference to AAPM TG-101 (17) when applicable. All treatment plans underwent a quality assurance review to ensure adherence to protocol standards.

Outcome measures

The primary endpoint was PFS, defined as the time from RT initiation to disease progression or death. Secondary endpoints included overall survival (OS), defined from OP to death; objective response rate (ORR), the proportion achieving complete response (CR) or partial response (PR); disease control rate (DCR), the proportion achieving CR, PR, or SD; and duration of response (DOR), defined as the time from the first documented CR or PR to progression or death.

Patients were monitored weekly during RT for acute adverse events (AE). After RT, follow-up visits with clinical evaluation and imaging were scheduled at 1, 3, 6, 9, and 12 months during the first year and every 6 months thereafter until disease progression, death, or study completion (±14-day window). At each visit, assessments included survival status, disease evaluation per RECIST v1.1, AEs, and laboratory tests. Laboratory abnormalities were recorded as AEs when meeting Common Terminology Criteria for Adverse Events (CTCAE) criteria. All AEs were graded according to CTCAE v5.0 (18). Attribution of AEs to RT was determined prospectively by the treating physicians at each visit based on clinical assessment and temporal relationship to RT. Events occurring during RT or within 90 days were classified as acute, and those thereafter as late. Survival status was confirmed at clinic visits or by telephone, with censoring at the last contact.

Quality of life

Patient quality of life (QoL) was assessed with the EORTC's Quality of Life Questionnaire-Core 30 (QLQ-C30; refs. 19–21), which measures global health status, functional domains, and symptom domains. Scores were linearly transformed to a 0 to 100 scale per the scoring manual; higher functional scores indicate better functioning, whereas higher symptom scores indicate greater symptom burden. Assessments were performed at baseline (within 1 month before RT) and at 1, 2, and 3 months after RT.

Statistical analysis

The primary endpoint was PFS. This interval reflects disease control following local intervention for prior progression and is conceptually analogous to a postprogression, PFS2-type endpoint.

This single-arm, phase II trial was powered to detect an improvement in median PFS over the historic median of 3.1 months derived from our retrospective cohort of patients with OP-HCC who received SLST (14). The trial was also supported by the regorafenib arm of the RESORCE trial (22). In our previous retrospective study (14), continued SLST plus PDRT was associated with a median PFS of 8.6 months compared with 3.1 months when switching to SLST alone. Therefore, a conservative target median PFS of 5 months was assumed for this prospective trial.

Sample size was calculated using a one-sample log-rank test in PASS version 15.0, based on the following assumptions: a one-sided α of 0.05, 80% power, an H₀ median PFS of 3.1 months, an H₁ median PFS of 5 months, and an estimated dropout rate of 10%. Under these assumptions, a total of 36 patients were required.

Survival endpoints were estimated using the Kaplan–Meier method and compared using the log-rank test. Cox proportional hazards models were used for univariate and multivariate analyses, and the results were reported as hazard ratios (HR) and 95% confidence intervals (CI). Continuous variables were summarized as the mean ± standard deviation (SD) or the median [interquartile range (IQR)], and categorical variables were summarized as frequencies (%). Changes in QoL were assessed using the Friedman test. When significant, post hoc pairwise comparisons with Bonferroni correction were performed. Statistical analyses were performed using SPSS Statistics version 27.0 (IBM Corp.; RRID: SCR_002865), and figures were generated using GraphPad Prism version 10.0 (GraphPad Software; RRID: SCR_002798).

Results

Patient characteristics

Between March 2024 and May 2025, 50 patients with OP during FLST were screened, of whom 14 were excluded (12 before treatment assignment and 2 after assignment because of missing critical follow-up data). Figure 1 provides the specific reasons for exclusion. The final analysis included 36 patients. Baseline characteristics are summarized in Table 1. 77.8% received combination therapy [47.2% tyrosine kinase inhibitor plus immune checkpoint inhibitor (TKI + ICI); 30.6% received anti–vascular endothelial growth factor agent plus ICI (anti-VEGF + ICI)], and 22.2% received TKI monotherapy. The median time from HCC diagnosis to OP was 22.1 months (IQR 10.8–57.9), and that from FLST initiation to OP was 10.2 months (IQR 4.3–17.3). Most patients (72.2%) had a single OP lesion, predominantly intrahepatic (25.5%) or macrovascular (23.5%), followed by pulmonary metastases (21.6%). The median BED was 70.6 Gy. The representativeness of the study participants is shown in Supplementary Table S1.

Figure 1.

CONSORT diagram.

Table 1.

Patient characteristics.

Characteristics	Value or number (%)
Age, years, median (IQR)	59 (52–65)
Sex	​
Male	34 (94.4)
Female	2 (5.6)
Etiology	​
Nonviral	8 (22.2)
HBV	26 (72.2)
HCV	1 (2.8)
HBV + HCV	1 (2.8)
FLST types	​
TKI + ICI	17 (47.2)
Anti-VEGF + ICI	11 (30.6)
TKI	8 (22.2)
Time to oligoprogression from HCC diagnosis, mo., median (IQR)	22.1 (10.8–57.9)
Time to oligoprogression from FLST start, mo., median (IQR)	10.2 (4.3–17.3)
Number of oligoprogressive lesions per patient, median (IQR)	1 (1–2)
Number of metastatic lesions per patient, median (IQR)	3 (3–4)
Number of oligoprogressive lesions per patient	​
1	26 (72.2)
2	6 (16.7)
3	4 (11.1)
Number of metastatic lesions per patient after oligoprogression	​
1	4 (11.1)
2	5 (13.9)
3	15 (41.7)
≥4	12 (33.3)
Site of oligoprogression (per lesion)	​
Intrahepatic progression	13 (25.5)
Macrovascular invasion	12 (23.5)
Lung	11 (21.6)
Lymph node	9 (17.6)
Adrenal gland	4 (7.6)
Bone	2 (3.9)
Oligoprogression subtype	​
Metachronous	7 (19.4)
Repeat	24 (66.7)
Induced	5 (13.9)
ECOG performance status at the time of oligoprogression	​
0	3 (8.3)
1	33 (91.7)
AFP at the time of oligoprogression, ng/mL	​
<200	22 (61.1)
≥200	14 (38.9)
Child–Pugh score at the time of oligoprogression	​
5	30 (83.3)
6	4 (11.1)
7	2 (5.6)
ALBI grade at the time of oligoprogression	​
1	20 (55.6)
2	15 (41.7)
3	1 (2.8)
Fractionation scheme	​
IMRT	17 (47.2)
SBRT	19 (52.8)
BED of oligoprogressive lesions (α/β = 10), median (IQR)	70.6 (59.9–75.5)

Abbreviations: AFP, α-fetoprotein; ECOG, Eastern Cooperative Oncology Group; HBV, hepatitis B virus; HCV, hepatitis C virus; IMRT, intensity-modulated RT.

Clinical outcomes

The median follow-up was 10.9 months (IQR 5.7–13.8). A total of 51 OP lesions were evaluated for response. Overall, 26 patients (72.2%) reached the primary endpoint (25 progression events and 1 death without prior documented progression). The median PFS time was 7 months (95% CI, 4.9–9.7), with 3-, 6-, and 9-month PFS rates of 73.7%, 64%, and 38.8%, respectively (Fig. 2A).

Figure 2.

Survival outcomes in enrolled patients. Kaplan–Meier curves show (A) PFS, (B) OS, and (C) DOR.

Five patients died during follow-up (one without prior progression and four after progression). Median OS was not reached; 6-, 12-, and 18-month OS rates were 90.5%, 86.4%, and 69.1%, respectively (Fig. 2B). Median DOR was not reached, with 3-, 6-, and 9-month rates of 84.6%, 79.6%, and 70.8%, respectively (Fig. 2C). The lesion-level ORR was 64.7% (9 CR, 24 PR), with 17 additional lesions achieving SD, yielding a DCR of 98%. A swimmer plot is shown in Supplementary Fig. S1.

In exploratory analysis, we assessed the duration of FLST, defined as the time from initiation of FLST to its discontinuation due to disease progression after RT. The median duration of FLST was 18.3 months (95% CI 14.1–23.9). In univariate analyses, FLST type, albumin–bilirubin (ALBI) score, and ALBI grade at oligoprogression were associated with PFS. Owing to collinearity, ALBI score was excluded from multivariate analysis; FLST type and ALBI grade remained independent prognostic factors (Supplementary Table S2). Stratified analyses showed longer median PFS in ALBI-1 versus ALBI-2/3 patients (9.7 vs. 4.9 months; HR 0.318; 95% CI, 0.129–0.780; P < 0.001; Supplementary Fig. S2A). Patients receiving TKI + ICI had longer median PFS than those receiving TKI alone (9.2 vs. 4.5 months; HR 0.272; 95% CI, 0.074–0.994; P = 0.002; Supplementary Fig. S2B). No significant differences were observed between TKI + ICI and anti-VEGF + ICI or between anti-VEGF + ICI and TKI-only groups (Supplementary Fig. S2C and S2D). By the time of the last follow-up, 25 of 36 patients had experienced disease progression (Supplementary Fig. S3). The most common patterns were new intrahepatic metastasis (32%) and new distant metastasis (24%). The majority of recurrences occurred outside the irradiated volume (outfield failure).

Safety

RT-related AEs (Table 2) occurred in 16 patients (44.4%), including RT-related symptomatic AEs (10 patients, 27.8%) and RT-related laboratory abnormalities (11 patients, 30.6%). Some patients experienced both symptomatic and laboratory RT-related AEs. Most events were grade 1 to 2; grade ≥3 RT-related AEs were observed in four patients (11.1%). No cases of RT-induced liver disease or RT-related deaths were reported. The most common RT-related AEs of any grade were abdominal pain, leukopenia, and thrombocytopenia. A comprehensive summary of all recorded AEs, regardless of attribution, is provided in Supplementary Table S3.

Table 2.

AEs clinically attributable to RT.

AE	No. of patients (%)
	Grade 1	Grade 2	Grade ≥3
Abdominal distension	​	​	​
Acute	3 (8.3)	0	0
Late	0	1 (2.8)	0
Abdominal pain	​	​	​
Acute	2 (5.6)	2 (5.6)	0
Late	0	2 (5.6)	1 (2.8)
Nausea	​	​	​
Acute	0	0	0
Late	1 (2.8)	0	0
Diarrhea	​	​	​
Acute	0	0	0
Late	0	2 (5.6)	0
Fatigue	​	​	​
Acute	1 (2.8)	0	0
Late	2 (5.6)	0	0
Gastrointestinal hemorrhage	​	​	​
Acute	0	0	0
Late	0	1 (2.8)	0
Bone pain	​	​	​
Acute	1 (2.8)	0	0
Late	1 (2.8)	0	0
Leukopenia	​	​	​
Acute	3 (8.3)	2 (5.6)	2 (5.6)
Late	0	0	0
Thrombocytopenia	​	​	​
Acute	4 (11.1)	1 (2.8)	1 (2.8)
Late	0	0	0
Decreased albumin	​	​	​
Acute	2 (5.6)	2 (5.6)	0
Late	0	0	0
Bilirubin increase	​	​	​
Acute	3 (8.3)	2 (5.6)	0
Late	0	0	0
Creatinine increased	​	​	​
Acute	1 (2.8)	0	0
Late	0	0	0
INR increased/prothrombin time prolonged	​	​	​
Acute	1 (2.8)	0	0
Late	0	0	0

Note: Percentages are calculated based on 36 patients. Patients may have experienced multiple events.

Abbreviation: INR, international normalized ratio.

QoL variations

QoL scores (mean ± SD) at four time points were presented in Supplementary Table S4. Overall, global QoL remained stable, with transient increases in fatigue and pain. At 1 month after PDRT, fatigue and pain scores were significantly higher than baseline (P = 0.037 and P = 0.010, respectively; Supplementary Fig. S4A and S4B), with no significant differences at other time points. No longitudinal changes were observed in other QoL domains.

Discussion

In this uncontrolled trial, PDRT combined with continued FLST achieved meaningful local control with a favorable safety profile. These results imply that combining PDRT with continued FLST is feasible for patients with OP-HCC and could delay SLST while potentially prolonging survival. In current practice, later-line TKI-based therapies after progression are associated with modest response rates and increased toxicity (22, 23). In this context, PDRT offers a pragmatic strategy by controlling resistant sites while preserving a first-line regimen that maintains control of remaining disease. By deferring SLST, this approach may sustain the clinical benefit of FLST and postpone exposure to the higher toxicity burden of later-line therapy.

The phase III RTOG 1112 trial (24) evaluated the addition of SBRT to sorafenib in patients with locally advanced HCC characterized by extensive intrahepatic disease and frequent macrovascular invasion. In this randomized study, SBRT followed by sorafenib was compared with sorafenib alone to evaluate differences in local control and survival. In contrast, our study addresses OP-HCC arising during FLST. Rather than upfront intensification, RT was introduced after prior benefit from FLST to target focal acquired resistance manifesting as oligoprogression. Thus, RT functioned as a PDRT to control resistant lesions while preserving the existing systemic regimen and delaying subsequent disease failure.

Interpretation of our findings should consider the distinct clinical profile of the enrolled cohort. Patients experienced a relatively prolonged period of disease control before oligoprogression, with median intervals of 22.1 months from diagnosis and 10.2 months from FLST initiation. In addition, restricting eligibility to oligoprogression and excluding patients with markedly elevated α-fetoprotein further enriched the cohort with favorable prognostic features. Collectively, these characteristics suggest a biologically selected population that may be more likely to benefit from PDRT. Future studies should further investigate additional clinical and biological predictors—such as FLST type, dynamic liver function parameters, advanced imaging features, and circulating tumor DNA—to refine patient selection and identify subgroups most likely to derive durable benefit.

After systemic therapy, oligometastatic disease may present as oligopersistence, characterized by radiographically stable residual lesions despite overall disease control (25). However, radiographic stability does not imply biological quiescence. Genomic and evolutionary evidence suggests that residual lesions—including primary and metastatic sites—may harbor resistant clones and retain metastatic potential under therapeutic pressure (26, 27). From this perspective, comprehensive RT to all residual sites may seem biologically justified. However, this must be balanced against organ-specific toxicity and overall treatment tolerability. In advanced HCC, high tumor burden and impaired baseline liver function make extensive RT—particularly to intrahepatic disease—concerning because of radiation-induced liver toxicity and QoL deterioration. In line with this reasoning, our exploratory analysis identified ALBI score as an independent determinant of PFS, underscoring the importance of hepatic reserve when defining RT extent. Thus, restricting RT to OP lesions represents a pragmatic compromise between disease control and toxicity mitigation, consistent with prior studies (28, 29).

ALBI grade has been shown to provide superior prognostic discrimination compared with Child–Pugh class, with higher grades associated with poorer survival (19). In our prior multicenter study (13), ALBI remained an independent prognostic factor in both univariable and multivariable analyses. Among ALBI-1 patients, switching to SLST yielded a median PFS of 5.6 months, whereas PDRT plus continued FLST achieved 16.7 months. Although outcomes were shorter in ALBI-2 patients, the addition of RT still conferred a PFS advantage over SLST alone. These findings were reproduced in the present trial, in which median PFS was 9.7 months for ALBI-1 versus 4.9 months for ALBI-2/3 (P < 0.001). Although most patients had preserved Child–Pugh class A liver function (34/36, 94.4%), limiting its discriminatory value, ALBI demonstrated a more balanced distribution (55.6% ALBI-1; 44.4% ALBI-2), enabling clearer risk stratification. Importantly, ALBI should be interpreted as a prognostic marker rather than a determinant of treatment eligibility. Patients with higher ALBI grades may still benefit from PDRT, provided that careful treatment planning minimizes irradiation of uninvolved liver and hepatic function is closely monitored. Future studies should evaluate whether optimizing hepatic reserve—through antiviral therapy, nutritional support, sarcopenia management, ascites and portal hypertension control, and medication adjustment—can improve ALBI grade and potentially prolong PFS.

We acknowledge several limitations. First, the absence of a control group, small sample size, and limited follow-up warrant cautious interpretation of the findings. Second, heterogeneity in systemic regimens and RT techniques may have influenced outcomes. Third, systemic therapy management during RT was not entirely consistent with current consensus recommendations (30), including guidance on interruption of VEGF(R) inhibitors or multitargeted TKIs. Although the observed safety profile was manageable, extrapolation to contemporary practice should be undertaken with caution. At the time of study design, prospective data on SLST in OP-HCC were limited. Accordingly, sample size estimation relied on retrospective and unselected SLST trial data, which may not fully reflect the prognosis of this specific population. As newer second-line agents demonstrate improved efficacy, the historic reference PFS may underestimate contemporary outcomes and influence interpretation of the observed benefit. Finally, exploratory analyses suggested shorter PFS among patients receiving TKI monotherapy. These findings should be interpreted cautiously, as the study was powered for the overall cohort and subgroup analyses were exploratory. The heterogeneity in systemic therapy reflected the broad eligibility criteria and real-world treatment patterns. Future phase III trials should prospectively account for FLST type, for example through stratification or prespecified subgroup analyses.

Overall, the findings of this nationwide, multicenter, prospective study suggest that PDRT combined with continued FLST is a safe and feasible treatment strategy for patients with OP-HCC. These results provide new evidence to optimize treatment and guide clinical decision-making. Furthermore, the type of FLST and ALBI grade may be valuable prognostic markers. These observations may be further validated in our ongoing phase III randomized trial (NCT06841172).

Supplementary Material

Supplementary Materials 1

Supplementary tables, figures and protocol.

Data Availability

The datasets generated and/or analyzed in this study are not publicly available because of the risk of compromising patient privacy and consent. However, they are available from the corresponding author upon reasonable request.

Authors’ Disclosures

J. Yue reports grants from Noncommunicable Chronic Diseases–National Science and Technology Major Project and grants from National Natural Science Foundation of China during the conduct of the study. No disclosures were reported by the other authors.

Authors’ Contributions

F. Shi: Conceptualization. H. Wang: Data curation, formal analysis, writing–original draft, writing–review and editing. Y. Yang: Resources, data curation. L. Ding: Resources, data curation. Y. Ding: Resources, data curation. Y. Zhang: Resources, data curation. M. Xi: Resources, data curation. Y. Qu: Resources, data curation. Z. Chen: Resources, data curation. Y. Lu: Resources, data curation. Y. Yang: Resources, data curation. L. Liu: Resources, data curation. J. Yue: Conceptualization, supervision, funding acquisition.