乌鸦传媒

TORONTO -- Prophylactic cranial irradiation (PCI) failed to improve survival in patients with locally advanced non-small cell lung cancer (NSCLC) treated with curative therapy, 10-year follow-up of a randomized trial showed.

Patients treated with PCI had a 10-year survival of 17% versus 13.3% for patients who underwent observation after curative treatment for NSCLC. Consistent with a previous analysis, PCI did improve disease-free survival (DFS) and reduced the incidence of brain metastases by almost 50%.

A subgroup analysis showed that patients who did not have surgery as primary treatment had significantly fewer brain metastases and significantly better DFS and overall survival (OS), Alex Sun, MD, of the University of Toronto, reported here at the .

"The early accrual closure failed to provide adequate power to detect the hypothesized difference in overall survival, and the survival rates were not significantly different between PCI and observation," said Sun. "Subgroup analyses based on stratification factors suggest that PCI may improve survival among nonsurgical patients."

Invited discussant John Armstrong, MD, of St. Luke's Radiation Oncology Network in Dublin, Ireland, cited several factors in the trial's failure demonstrate improved survival with PCI. Investigators enrolled about a third of the patients specified in the trial design (356 of 1,038). The nonsurgical group that appeared to benefit from PCI was not prespecified for subset. The reduction in brain metastases nonsurgical group was fairly modest as compared with the entire study population.

"There is no reason why PCI improved the primary endpoint in this subset," said Armstrong.

Citing a of biologically effective doses for NSCLC, Armstrong said the protocol of 30 Gy in 15 fractions used in the trial translated into a total dose of 45.38 Gy, substantially lower than calculated doses needed to achieve tumor control.

Whether PCI has a future in NSCLC remains to be seen, he added. Factors to consider include the potential for identifying an "ultra-high-risk" group of patients, development of a genomics-based test to help guide therapeutic decisions, application of translational research involving trials of locally advanced NSCLC that do not test PCI, and the potential value of aggressive surveillance with volumetric MRI for high-risk patients in complete remission.

Sun reported 10-year follow-up data from the NRG/RTOG 0214 randomized clinical trial. The rationale for the study came from improved survival survival in locally advanced NSCLC as a result of better locoregional and systemic disease control. However, several previous studies showed that as survival improves, the incidence of central nervous system (CNS) failure increases.

The trial involved patients with stage IIIa/b NSCLC who had completed locoregional therapy within the previous 16 weeks, no acute or subacute toxicity ≤ grade 2, stable disease or better result with treatment, and no CNS metastases on MRI or brain CT. Patients were randomized to PCI (30 Gy, 20 fractions) or observation, and the primary endpoint was overall survival. Secondary endpoints included DFS, CNS metastases, neurocognitive function, and quality of life.

Previous analyses showed no difference in OS or DFS after or of follow-up. The incidence of brain metastases was significantly lower in the PCI group.

The 10-year analysis included 340 patients who had a median follow-up of 2.1 years, including 9.2 years for patients who remained alive. With longer follow-up, the results still showed no difference in OS, although a statistically significant difference in DFS had emerged in favor of PCI (12.6% vs 7.5%, P=0.0286). Analysis of CNS metastasis continued to show a significant lower incidence in the PCI group (16.7% vs 28.3%, P=0.0033).

Of the 340 evaluable patients, 225 had no surgery. In that subgroup, PCI was associated with significantly fewer brain metastases (13.0% vs 31.0%, P=0.0030), significantly better DFS (11.7% vs 4.9%, P=0.0148), and significantly better OS (16.6% vs 8.9%, P=0.0263).

The study population included 157 evaluable patients in the PCI group. Worst acute toxicity was grade 1 in 14.6%, grade 2 in 35.0%, grade 3 in 3.8%, and one patient (0.6%) had grade 4 clinical depression. Worst late toxicity was grade 1 in 12.7% of patients, grade 2 in 8.7%, and grade 3 in 3.3%.

Data on neurocognitive function and quality of life were published previously. The longer follow-up yielded insufficient data for additional analyses, said Sun.

Comment