AUTHOR=Shi Zhenwei , Foley Kieran G. , Pablo de Mey Juan , Spezi Emiliano , Whybra Philip , Crosby Tom , Soest Johan van , Dekker Andre , Wee Leonard 

TITLE=External Validation of Radiation-Induced Dyspnea Models on Esophageal Cancer Radiotherapy Patients

JOURNAL=Frontiers in Oncology

VOLUME=Volume 9 - 2019

YEAR=2019

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2019.01411

DOI=10.3389/fonc.2019.01411

ISSN=2234-943X

ABSTRACT=Purpose: Radiation-induced lung disease (RILD), such as dyspnea, is a risk for patients receiving high-dose thoracic irradiation. This study is a TRIPOD Type 4 validation of previously-published lung toxicity models via secondary analysis of esophageal cancer SCOPE1 trial data. We quantify the predictive performance of these two models for predicting dyspnea 6 months after high-dose chemo-radiotherapy for primary esophageal cancer.
Material and methods: Lung cancer patients treated at MAASTRO Clinic (The Netherlands) were used to develop the previous dyspnea risk models. We tested the performance of the earlier models using baseline, treatment and follow-up data on 258 esophageal cancer patients in the UK enrolled into the SCOPE1 multi-centre trial. As some variables were missing randomly and cannot be imputed, 212 patients in the SCOPE1 were used for validation of model 1 and 255 patients were used for validation of model 2. The adverse event of interest was dyspnea ≥ Grade 2 within 6 months of the end of radiotherapy. The model parameter Forced Expiratory Volume in 1s (FEV1) was imputed using the WHO performance status. External validation was performed using an automated, decentralized approach, without exchange of individual patient data.
Results: Out of 258 patients with esophageal cancer in SCOPE1 trial data, 38 patients developed radiation-induced dyspnea (≥ Grade 2) within 6 months of the end of radiotherapy. The discrimination performance of the models in esophageal cancer patients treated with high-dose external beam radiotherapy was moderate, AUC of 0.68 (95% CI 0.55 – 0.76) and 0.70 (95% CI 0.58 - 0.77). The curves and AUCs derived by distributed learning were identical to the results from validation on a local host.
Conclusion: We have externally validated previously published dyspnea models using an esophageal cancer dataset. FEV1 that is not routinely measured for esophageal cancer was imputed using WHO performance status. Prediction performance was not statistically different from previous training and validation sets. Risk estimates were dominated by WHO score in Model 1 and baseline dyspnea in Model 2. The distributed learning approach gave the same answer as local processing, and could be performed without accessing a validation site’s individual patients-level data.