Updated assessment of the six-minute walk test as predictor of acute radiation-induced pneumonitis

Purpose: To assess the utility of the 6-minute walk test (6MWT) as a predictor of symptomatic radiation-induced pneumonitis (RP). Methods: As part of a prospective trial to study radiation-induced lung injury, 53 patients receiving thoracic radiotherapy (RT) underwent a pre-RT 6MWT, pulmonary function tests (PFTs), and had ≥3-month follow-up for prospective assessment of Grade 2 or worse RP (requiring medications or worse). Dosimetric parameters (e.g., the percentage of lung receiving ≥30 Gy) were extracted from the lung dose-volume histogram. The correlations between the 6MWT and PFT results were assessed using Pearson's correlation. The receiver operating characteristic technique was used in patient subgroups to evaluate the predictive capacities for RP of the dosimetric parameters, 6MWT results, and PFT results, or the combination (using discriminant analysis) of all three metrics. ROCKIT software was used to compare the receiver operating characteristic areas between each predictive model. The association of the decline in 6MWT with the development of RP was evaluated using Fisher's exact test. Results: The pre-RT PFT and 6MWT results correlated weakly (r = 0.44-0.57, p ≤ 0.001), suggesting that they measure somewhat different physiologic functions. Of the 53 patients, 9 (17%) developed RP. The dose-volume histogram-based dosimetric parameters were the best single-metric model for predicting RP (e.g., percentage of lung receiving ≥30 Gy, receiver operating characteristic area 0.73, p = 0.03). Including the PFT or 6MWT results with the percentage of lung receiving ≥30 Gy did not improve the predictions. The predictive abilities of dosimetric-based models improved when the analysis was restricted to those patients whose tumors were not causing regional lung dysfunction. No correlation was found between the decline in the 6MWT result and the RP rate (p = 0.6). Conclusion: Although the PFTs and 6MWT are related to each other, the correlation coefficients were weak, suggesting that they could be measuring different physiologic functions. In the present data set, the addition of the PFTs or 6MWT did not increase the ability of the dosimetric parameters to predict for acute symptomatic RP. Additional work is needed to better understand the interaction among the PFT results, exercise tolerance (6MWT), and the risk of RT-induced lung dysfunction.