Minimum Lung Volume Spared for Lung Stereotactic Ablative Therapy: A Useful Metric or Outdated Endpoint?

Abstract

Stereotactic body radiation therapy (SBRT) is the preferred treatment for malignant lesions of the lung in patients unable to tolerate an anatomic resection. While this modality is typically well-tolerated, there is potential for morbidity or even mortality in this population. Radiation pneumonitis (RP) is one such toxicity, and ongoing NRG protocols recommend Lung V20 < 10-15% and minimum volume of lung spared (MVLS) from 12.4 or 13.5 Gy for four and five fraction SBRT as the only two lung parameters for planning. There is some controversy on whether to include additional lung factors besides V20 in the treatment planning process. We sought to assess the performance of MVLS compared with cutoffs associated with RP from our team's previous study [1] using data from another institution.Inclusion criteria from our previous study [1] were applied to a new dataset. Patients receiving SBRT to the lung between 2010 and 2018 with tumoricidal (BED10 > 100 Gy) fractionation schemes (e.g., 50-55/5, 48-50/4, 60/8) were included. A minimum clinical follow-up of 6 months or documentation of RP was required. Symptomatic RP was defined as imaging consistent with RP and indication for corticosteroids (CTCAE Grade 2-3). Various dosimetric parameters and patient factors were recorded. The univariate logistic regression and dosimetric constraints from Institution #1 to predict probabilities of symptomatic RP have been published [1]. The accuracy of these predictive probabilities was assessed by applying derived dosimetric values predicting a 20%, 33% and 50% probability of developing symptomatic RP to Institution #2.For the sample from Institution #2, 25.7% (n = 17/66) developed symptomatic RP, which is higher than the published rate of 8.6% (n = 8/93) observed in the Institution #1. Minimum volume of lung spared (MVLS) from 12.4 or 13.5 Gy > 1000 ccs was achieved by all but one patient (a pediatric patient with small lung volumes). Multiple DVH thresholds predicting a 20% rate of symptomatic RP from Institution #1 correlated well with Institution #2 (Table 1). However, the model was not reliable for the 33% and 50% predicted probability cutoffs.This study of symptomatic RP identified several dosimetric parameters that accurately captured RP between two cohorts. MVLS appears to have limited utility in the clinical setting. In addition to Lung V20, future SBRT consortium trials may benefit from more correlative dose constraints.