Combined risk score enhances radiation pneumonitis prediction

Jul 23 2014

By Laura Cowen, medwireNews Reporter

Combining dose-volume histogram (DVH) parameters with age and baseline pulmonary fibrosis score creates a novel predictive risk score that improves prediction of radiation pneumonitis (RP) in patients receiving concurrent chemoradiotherapy for non-small-cell lung cancer (NSCLC), research shows.

The risk of RP, a potentially fatal dose-limiting adverse effect of concurrent chemoradiotherapy, is currently estimated using individual DVH parameters such as the percentage of lung volume receiving a dose in excess of 20 Gy (V20) and the mean lung dose (MLD), explain Kayoko Tsujino (Hyogo Cancer Center, Akashi, Japan) and colleagues.

In the present study, the researchers investigated whether severe RP (grade 3 or above) could be more accurately predicted by combining DVH metrics with clinical factors.

They reviewed the medical records of 122 patients (median age 63 years, 89% men) with locally advanced NSCLC treated with concurrent chemoradiotherapy and found that during a median 14.6 months of follow-up, 14 (11.5%) developed severe RP.

On univariate analysis, total lung volume, V5, V20, absolute lung volume spared from a 5 Gy dose (VS5), MLD, gender and pulmonary fibrosis on baseline computed tomography were significantly associated with severe RP. Age and smoking status showed a borderline significant association.

On multivariate analysis, a VS5 less than 1500 cc, baseline pulmonary fibrosis score of 2 or more, V20 of at least 26% and age 68 years or older were significant predictors of severe RP.

Based on the multivariate analysis beta coefficient, these four variables were assigned a score of 6, 5, 3 and 3, respectively, to create a combined predictive risk score (PRS).

Tsujino and team report that the cumulative incidence of severe RP at 12 months was 0%, 7.8%, 26.6% and 71.4% when the PRS was 0, 3–5, 6–8 and 9–14, respectively. No patient had a PRS score above 14.

In addition, the PRS predicted severe RP more accurately than V20 and VS5 combined, or V20 alone, with respective areas under the receiver operating characteristic curve of 0.888 versus 0.779 and 0.678.

“Currently, we recommend that the PRS should be less than eight and that the V20 dose constraint should be less than 30–35% when administering definitive [concurrent chemoradiotherapy] for NSCLC”, the researchers write in the Journal of Thoracic Oncology.

However, they conclude that further validation of the PRS is needed before it can be adapted for clinical use.

Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.