Real World Rates and Predictive Factors of Pneumonitis in Advanced, Non-Resectable NSCLC Treated with Concurrent Chemoradiation and Durvalumab

Purpose/Objective(s)

Real world outcomes related to pneumonitis in patients with locally advanced (LA) non-resectable non-small cell lung cancer (NSCLC) treated with concurrent chemoradiation (chemoRT) and consolidative durvalumab (D) may differ from those seen in prospective trials given selection biases and other variables in clinical trials. We aimed to assess the real-world incidence of pneumonitis for LA NSCLC patients and identify clinical or dosimetric predictors of pneumonitis.

Materials/Methods

We retrospectively identified consecutive patients with locally advanced NSCLC treated with chemoRT and D at a single institution from 4/2017 to 6/2019 from our prospectively maintained institutional database. The actuarial incidence of pneumonitis was estimated using the Kaplan-Meier method and compared using log-rank tests. Toxicity was evaluated via retrospective examination of clinical documents and prescription records (i.e., steroids or oxygen) and graded according to Common Terminology Criteria for Adverse Events, version 5.0. Logistic regression was used to assess associations between pneumonitis and relevant clinical and dosimetric variables. Patients were followed with clinical examination and surveillance computed tomography scans every 3 months for 2 years and every 6 months afterwards.

Results

We identified 61 patients with locally advanced NSCLC that were consecutively treated with a median RT dose of 66 Gy (range 60-70Gy). Median follow up was 29 months (range 4-55). Intensity-modulated radiation therapy (IMRT) was used in 36 (59%) patients, and 25 (41%) patients received 3D-conformal RT (3DCRT). Twenty-three (37.7%) patients developed pneumonitis. Highest grade pneumonitis was 2 (n=17, 27.9%), and 3 (n=5, 8.2%). Of patients who developed pneumonitis, the median time to onset was 3.2 months from completion of RT (range 1.2-9.1), with a 1-year actuarial rate of 43.5% for all grade pneumonitis. On logistic regression, significant predictors of any grade pneumonitis were higher mean lung dose, V5, V10, V15, V20 and V25. There was no association between time to pneumonitis and the volume of the planning target volume (PTV) or use of 3DCRT vs IMRT (p > 0.05).

Conclusion

In real-world patients with locally advanced NSCLC treated with concurrent chemoRT and consolidative durvalumab, the incidence of any grade and grade 3 pneumonitis may be higher than previously observed in prospective clinical trials. Several dosimetric parameters (MLD, V5, V10, V15, V20 and V25) predicted development of pneumonitis, emphasizing the global importance of minimizing dose to normal lung tissue.