Dosiomics: Extracting 3D Spatial Features From Dose Distribution to Predict Incidence of Radiation Pneumonitis

In conclusion, the spatial features of dose distribution extracted by the dosiomics method effectively improves the prediction ability. Introduction Radiation pneumonitis (RP) is one of the major toxicities of thoracic radiation therapy. The clinical symptoms range from fever, cough to pulmonary function failure, which may occur during the first 6 months after irradiation. Reducing the prescription dose could lower the risk of RP incidence, but also impairs tumor control. An accurate RP predictor (or prediction model) is desired to “safely” irradiate the tumor target without increasing the risk of RP incidence. RP incidence is directly associated with the dose distribution within lung volume. Dosimetric factors, such as mean lung dose (MLD) and the lung volume within which the dose is greater than xGy (Vx), are widely used for RP prediction. Boonyawan et al. reported that RP incidence increases with V10 and V20 (1). Ramella et al. found that RP incidence is associated with V20 and V30 (2). Briere et al. reported that RP incidence significantly increases if the sparing lung volume is <1852cc (receiving dose≤40Gy) (3). Pinnix et al. reported that V5 has better prediction capability than V10, V15, and V20 (4). Palma et al. analyzed 836 patient cases from international institutions and concluded that symptomatic RP is associated with V20, and fatal RP associated with the mean dose per day during treatment (5). Those studies demonstrate tha...
Source: Frontiers in Oncology - Category: Cancer & Oncology Source Type: research