Why This Study Matters
Locally advanced, unresectable non-small cell lung cancer (NSCLC) is typically treated with concurrent chemoradiation, but uniform dose escalation — as tested in the RTOG 0617 trial — failed to improve survival and instead increased toxicity, showing that a one-size-fits-all radiation dose does not benefit every patient equally. That result has driven interest in biologically guided, risk-adaptive radiotherapy: using mid-treatment imaging response to decide which patients actually need a higher dose, intensifying treatment only where it is likely to help rather than exposing everyone to added toxicity.
The Phase II FLARE-RT trial (Functional Lung Avoidance and Response-adaptive Escalation Radiation Therapy) tested exactly this approach in 49 patients at a single institution: using mid-treatment FDG-PET response to selectively boost radiation dose only in tumors that were not responding, while sparing perfused lung in everyone to limit toxicity. This report presents the trial's mature, long-term outcomes.
Study Design
FLARE-RT was a single-institution, non-randomized, risk-stratified Phase II trial (NCT02773238) conducted from 2016 to 2021. Patients underwent baseline FDG-PET imaging before treatment and a second FDG-PET scan at week 3, after 24 Gy. Response status was determined prospectively by multidisciplinary consensus using multifactorial radiological criteria assessing changes in FDG avidity and metabolic tumor volume.
| Group | Regimen |
|---|---|
| PET responders (n=33, 67.3%) | Continued standard chemoradiation: 60 Gy in 30 fractions. |
| PET non-responders (n=16, 32.7%) | Concomitant integral boost of 14 Gy in 15 fractions to residual metabolically active disease (total integral dose 74 Gy; peak tumor doses exceeding 90 Gy), spatially conformed to and non-uniformly scaled by mid-treatment FDG-PET uptake. |
The primary endpoint was overall survival. Secondary outcomes included progression-free survival, locoregional progression and distant metastasis cumulative incidence (estimated with competing-risks methods), and predictors of treatment failure identified through Fine-Gray regression.
Patient Population
The analysis included 49 patients with AJCC version 7 stage IIB-IIIB NSCLC or N2 locoregional recurrence who both enrolled and initiated chemoradiation. Most had adenocarcinoma histology (61.2%), N2 nodal disease (63.3%), a smoking history (83.6%), and received carboplatin-paclitaxel chemotherapy (63.3%). Half the cohort — 25 of 49 patients (51%) — received consolidation durvalumab following the PACIFIC regimen, and just over half, 26 of 49 (53.1%), were treated with proton rather than photon radiotherapy.
| Characteristic | All patients (n=49) | PET responders (n=33) | PET non-responders (n=16) |
|---|---|---|---|
| Adenocarcinoma histology | 30 (61.2%) | 18 (54.5%) | 12 (75.0%) |
| Squamous cell carcinoma | 17 (34.7%) | 13 (39.4%) | 4 (25.0%) |
| N2 nodal disease | 31 (63.3%) | 19 (57.6%) | 12 (75.0%) |
| Stage IIIA | 24 (49.0%) | 14 (42.4%) | 10 (62.5%) |
| Former/current smoker | 41 (83.6%) | 28 (84.9%) | 13 (81.3%) |
| Proton radiotherapy | 26 (53.1%) | 19 (57.6%) | 7 (43.8%) |
Dosimetric parameters for target volumes and organs at risk were otherwise similar between the standard-dose and boost-dose groups; only esophageal V20 differed significantly (46.6% vs. 39.3%, P=.04).
Primary Endpoint Results
At a median follow-up of 52.3 months, 1-year and 2-year overall survival were 81.6% (95% CI, 71.5%-93.2%) and 54.2% (95% CI, 41.8%-70.4%), respectively; median overall survival was 29.1 months (95% CI, 17.3-NA). Progression-free survival at 1 and 2 years was 53.1% (95% CI, 40.9%-69.0%) and 40.5% (95% CI, 28.8%-57.0%), respectively; median progression-free survival was 12.3 months (95% CI, 9.2-NA).
Locoregional progression occurred in 9 patients (3 nodal-only, 6 with local progression), while distant metastasis was the dominant mode of failure. PET non-responders who received the adaptive dose boost showed a non-significant trend toward lower locoregional progression than responders (HR 0.23); at 1 year, locoregional progression was 6.3% in boosted non-responders versus 18.2% in standard-dose responders.
Subgroup and Risk-Stratification Analyses
Beyond the overall trial results, the investigators used 5-fold cross-validation to test whether imaging and treatment features could stratify patients into risk groups for locoregional progression and distant metastasis.
| Stratifying feature | Outcome | Low-risk group | High-risk group | P-value |
|---|---|---|---|---|
| Baseline metabolic burden | 1-yr locoregional progression | 7.7% | 17.4% | .056 |
| Mid-treatment metabolic burden | 1-yr distant metastasis | 25.0% | 47.6% | .027 |
| Durvalumab use | 1-yr distant metastasis | 20.0% (durvalumab) | 50.0% (no durvalumab) | .032 |
| Combined: baseline burden + histology | 1-yr locoregional progression | 0.0% | 24.0% | .012 |
| Combined: durvalumab + mid-treatment burden | 1-yr distant metastasis | 20.0% | 50.0% | .011 |
In the multivariable model for locoregional progression, adenocarcinoma histology carried a hazard ratio of 11.67 versus squamous cell carcinoma or other histology (95% CI, 1.80-75.85); the model's concordance index was 0.82 (95% CI, 0.68-0.95). The distant-metastasis model, built on durvalumab use and mid-treatment metabolic burden, had a concordance index of 0.66 (95% CI, 0.53-0.77).
Safety Profile
Among all 49 patients, grade 2 or higher dermatitis occurred in 28.6%, grade 2 or higher pneumonitis in 40.8%, and grade 2 or higher esophagitis in 59.2%. Only one grade 3 or higher esophagitis event was reported, and it occurred in a PET responder receiving standard-dose radiotherapy — no grade 3 or higher esophagitis occurred among boosted PET non-responders. No significant cardiotoxicity was observed under the strict heart dosimetric constraints applied in response to prior RTOG 0617 findings.
| Adverse event (grade 2+) | PET responders (60 Gy, n=33) | PET non-responders (74-90 Gy, n=16) |
|---|---|---|
| Pneumonitis | 15/33 (45.5%) | 5/16 (31.2%) |
| Dermatitis | 11/33 (33.3%) | 3/16 (18.8%) |
| Esophagitis | 16/33 (48.5%) | 13/16 (81.3%) |
Pneumonitis also varied by systemic therapy: it occurred in 12 of 25 (48.0%) patients who received durvalumab versus 8 of 20 (40.0%) who did not. Although differences did not reach statistical significance, PET responders generally experienced more dermatitis and pneumonitis than non-responders — a pattern the authors attribute to responders' greater intrinsic radiosensitivity — while esophagitis followed the opposite trend, likely reflecting the higher esophageal dose delivered to the boosted non-responder group.
Interpretation and Broader Context
In summary, the biological image-guided and response-adaptive FLARE-RT trial achieved high locoregional control with consistent overall survival relative to RTOG 0617 (60 Gy arm) / RTOG 1106 (adaptive arm) and comparable to PACIFIC (durvalumab arm).
— Discussion, Clinical Cancer Research
The authors report that FLARE-RT's 2-year locoregional progression rate of 18.5% compares favorably with 33.7% in the adaptive arm of RTOG 1106 and 31.8% in a University of Michigan trial. Grade 3+ pneumonitis in FLARE-RT was 8.2%, versus 6.9% in RTOG 1106 and 2.1% in the Michigan trial — though the trials differed in durvalumab use (51% in FLARE-RT vs. 0% in RTOG 1106 vs. 43% in the Michigan trial) and other patient characteristics, so these are cross-trial historical comparisons rather than head-to-head results.
Compared with the PACIFIC trial's durvalumab arm, 2-year overall survival was numerically higher in PACIFIC (66.3%) than in FLARE-RT (54.2%), though the FLARE-RT estimate falls within PACIFIC's reported confidence interval. The authors note that only half of FLARE-RT patients received consolidation durvalumab and the trial enrolled patients by intent to treat — including some who did not complete the full chemoradiation course — whereas PACIFIC enrolled only patients who successfully completed chemoradiation, a more selected population.
Despite achieving strong local control, distant metastasis remained the dominant failure pattern in FLARE-RT, underscoring that risk-adaptive radiotherapy targeting local disease does not by itself solve systemic disease control. The authors highlight consolidation durvalumab and metabolic-burden-guided systemic therapy intensification as complementary strategies worth exploring. Because this was a single-institution, non-randomized study of 49 patients, the authors emphasize that risk-adaptive dose escalation guided by FDG-PET requires validation in larger, independent, and ideally randomized cohorts before it could inform routine practice. What's next: the investigators call for validating pretreatment and mid-treatment metabolic tumor burden as predictors of locoregional progression and distant metastasis, respectively, in larger independent cohorts, and for next-generation trials that combine risk-adaptive radiotherapy with systemic therapy intensification guided by integrated imaging, tissue, and blood biomarkers.
In this single-institution Phase II trial, using mid-treatment FDG-PET response to selectively boost radiation dose only in non-responding tumors achieved strong locoregional control (2-year locoregional progression of 18.5%) with manageable toxicity — no grade 3+ esophagitis among boosted patients — and 2-year overall survival of 54.2% that compares reasonably with prior adaptive-radiotherapy and durvalumab-consolidation trials. But distant metastasis, not local failure, drove most treatment failures, and higher pretreatment and mid-treatment tumor metabolic burden separately predicted locoregional and distant relapse, pointing toward combining PET-adaptive local therapy with systemic therapy intensification. With only 49 patients treated at a single center without randomization, these results need confirmation in larger, ideally randomized, multi-institutional trials before PET-adaptive dose escalation can be considered standard practice.