Unexpected Lack of Survival Benefit in Liver Cancer Treated With Proton Therapy

Proton therapy failed to improve survival in locally advanced hepatocellular carcinoma (HCC) compared with photons, surprising investigators in the multicenter randomized NRG-GI003 trial.

Patients treated with photon radiation therapy lived more than twice as long as patients randomized to proton therapy (54.9 vs 26.6 months). A subgroup analysis showed no convincing effect of either modality in most groups with the exception of tumor size, as photon therapy dramatically outperformed protons for treating larger tumors (≥5 cm), showing a greater than twofold impact on overall survival (OS). Progression-free survival (PFS) and rates of local progression did not differ between treatment groups.

Protons did have a dosimetric advantage, significantly reducing exposure of uninvolved tissue, which translated into a numerically low rate of severe toxicity. Additionally, photon therapy performed far better than hypothesized when the trial was designed, reported Theodore Hong, MD, of Dana-Farber Cancer Institute in Boston, at the American Society for Radiation Oncology (ASTRO) meeting in Boston.

“Both arms performed favorably with respect to overall survival, with the photon arm performing noticeably better than expected,” said Hong. “The post hoc tumor size interaction suggests the value of protons may not be uniform across all patients. Further work is needed to determine whether tumor burden, liver reserve, or other factors can identify patients most likely to benefit from proton therapy.”

Although photon therapy demonstrated an OS advantage in the NRG-GI003 trial, questions about the role of proton therapy in oncology and performance versus photons remain open, according to two ASTRO discussants.

The trial had an ambitious goal for survival improvement with proton therapy, a hazard ratio of 0.58, noted Erqi Pollom, MD, of Stanford Medicine in California. The unexpected performance of photons was likely driven by improvements in systemic therapy, particularly immune checkpoint combinations.

The trial had an “isotoxic” design, wherein the delivered tumor dose of radiation was determined by normal liver constraints, she continued. Theoretically, greater proton dose escalation should have been feasible, but only 58% of patients randomized to proton therapy received the highest dose level. The 2-year local progression rates were about 20% in both arms, inferior to published results with proton therapy. The dose-escalation issue warrants further study to explore opportunities for further dose escalation.

“Ultimately, this trial shows that modern liver radiation, whether using photons or protons, is effective and safe, and it challenges the current trend in liver radiation, often considered as a last resort behind other liver-directed therapies,” said Pollom. “Instead of evaluating if protons can improve the already excellent outcomes of photons for these patients, could we use protons to expand the clinical boundaries of liver radiation and reach those high-risk cohorts who may be comfortable being treated with protons?”

The most unexpected result was the superiority of photons for larger tumors, when the opposite might have been predicted, said Marco Schwarz, PhD, of the University of Washington in Seattle.

“The trial design had to define PTV [planning target volume] doses, liver doses, and tumor volumes in such a way that all patients could be treated safely at sufficiently high doses for both photons and protons,” he said. “The trade-offs between these three variables led to a situation where the superiority of protons ended up sparing the liver instead of increasing two more doses. The reduction in mean liver dose suggests that protons could indeed deliver a higher PTV dose.”

“In my opinion, the trial question remains open, remains sensible, and very much worth asking,” he added.

By way of introduction, Hong noted that the NRG-RTOG 1112 trial established stereotactic body radiotherapy as standard of care for locally advanced HCC. Proton therapy may improve survival compared with photon therapy as a result of decreased liver compensation owing to favorable dose distributions and the ability to spare nontarget liver tissue. NRG-GI003 tested that hypothesis in a population of patients with advanced HCC.

Eligible patients had measurable disease, up to three single or multinodular lesions, and Child-Turcotte-Pugh (CTP) score of A or B7. Prior chemotherapy, targeted therapy, surgery, chemoembolization, and ablation were allowed.

Treatment could be delivered in five or 15 fractions. The trial design assumed a 14-month median OS for the photon control arm, improved to 24 months by use of proton therapy.

Data analysis included 115 randomized patients who had a median age of 72. Men accounted for 79% of the study population. More than 90% of patients had CTP A disease, 79% had a single tumor, and 59% of patients received treatment in 15 fractions.

Grade ≥3 treatment-related adverse events (TRAEs) occurred in 24% of the photon arm and 11% of the proton arm (P=0.093). The most common TRAE was decreased lymphocytes (10 in the photon group and three in the proton arm).

The trial had an efficacy stopping point of log-rank P value ≤0.011 and a futility stopping point of hazard ratio >1. After a median follow-up of 20.4 months, an interim analysis showed a hazard of 1.3 (95% CI 0.83-2.04), meeting criteria for futility. The subgroup analysis identified three factors with OS predictive value: CTP grade A5 vs B7 (P=0.0009), tumor vascular thrombus (P=0.0092), and treatment/tumor size interaction (P=0.046).

Median PFS was 16.9 months in the proton arm and 13.7 months in the photon arm (HR 1.02, 95% CI 0.65-1.62). The rate of local progression and first site of treatment failure did not differ significantly between groups.