Dose-volume histogram parameters for predicting radiation pneumonitis using receiver operating characteristic curve

Abstract

Objective: To assess the predictability of dose-volume histogram (DVH) parameters for radiation pneumonitis (RP) using receiver operating characteristic (ROC) curve.

Methods: One hundred and thirty-five cases of locally advanced non-small cell lung cancer patients treated with three-dimensional radiotherapy and chemotherapy were analyzed retrospectively. The end point of follow-up was ≥2 grade RP defined according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0. The ROC curve was used to explore the predictive sensitivity (SEN), specificity (SPE), positive predictive value (PPV), and negative predictive value (NPV) for potential DVH parameters associated with RP.

Results: Relative volumes of total lungs receiving ≥5 Gy (V(5)), ≥10 Gy (V(10)), ≥13 Gy (V(13)), ≥20 Gy (V(20)), and mean lung dose (MLD), were all correlated to the development of RP (p < 0.05), among which V 5 and V 20 were the most important factors (p = 0.045 and 0.037; OR = 3.166 and 3.030). However, collinearity was found between V(5) and V(20) (Spearman's rho 0.771, p < 0.01). The area under the ROC curve was 0.643 and 0.648 for using V(5) and V(20) as predictors. If predictive cut-off values were established as follows: V(5) = 0.8 and V(20) = 0.3, the parameters could provide predictive SEN, SPE, PPV and NPV were 0.387 and 0.581, 0.882 and 0.701, 0.444 and 0.321, and 0.855 and 0.873, respectively.

Conclusions: V(5) and V(20) could act as predictors for RP; however, single DVH metrics did not appear to have high predictive power for RP.