Severe Radiation-Induced Lymphopenia during Concurrent Chemoradiotherapy for Lung Cancer: External Validation of Two Prediction Models

Abstract

<h3>Purpose/Objective(s)</h3> Severe radiation-induced lymphopenia in patients undergoing chemoradiotherapy (CRT) for non-small cell lung cancer (NSCLC) is associated with detrimental survival outcomes and decreased efficacy of subsequent immunotherapy. This impact provides a strong incentive to identify patients at high risk who could benefit from lymphopenia-mitigating strategies. At The Christie, a prediction model for grade ≥3 lymphopenia was developed in lung cancer patients. At MDACC, a model for grade 4 lymphopenia was developed in esophageal cancer patients. The aim of this study was to validate both models for predicting grade ≥3 and grade 4 lymphopenia during concurrent CRT in patients with stage III NSCLC. <h3>Materials/Methods</h3> Consecutive patients who underwent concurrent CRT for stage III NSCLC at our comprehensive cancer center between 2019 and 2020 were included. Mildly hypofractionated radiotherapy was administered (24 × 2.75 Gy to primary tumor and 24 × 2.42 Gy to involved lymph nodes). Chemotherapy consisted of daily low-dose cisplatin. The primary outcomes of grade ≥3 and grade 4 lymphopenia were defined as absolute lymphocyte count (ALC) nadirs during CRT of <0.5 and <0.2 K/mL, respectively. Predictors of the Christie model included age, baseline ALC, mean heart and lung doses, and thoracic vertebrae V20Gy. Predictors of the MDACC model were age, baseline ALC, and PTV in interaction with BMI. The prediction models' performance was assessed in terms of discrimination and calibration. <h3>Results</h3> A group of 68 patients was studied in which 51 patients (75%) developed grade ≥3 lymphopenia during CRT, with grade 4 lymphopenia in 9 patients (13%). For prediction of grade ≥3 lymphopenia, application of the Christie and MDACC models yielded c-statistics of 0.82 (95%CI: 0.71-0.94) and 0.85 (95%CI: 0.73-0.97), respectively. For prediction of grade 4 lymphopenia, the Christie and MDACC models yielded c-statistics of 0.68 (95%CI: 0.51-0.86) and 0.78 (95%CI: 0.62-0.94), respectively. Calibration (adjusted for the a-priori risk) demonstrated moderate agreement between the observed and predicted risk for grade ≥3 and grade 4 lymphopenia using the Christie model (Table 1). Good calibration was found using the MDACC model for both grade ≥3 and grade 4 risk predictions. <h3>Conclusion</h3> The MDACC prediction model for severe radiation-induced lymphopenia demonstrated superior external performance in the setting of concurrent CRT in patients with stage III NSCLC as compared with the Christie prediction model. As such, the prediction model may aid in identifying thoracic cancer patients at high risk for severe lymphopenia.