Improving survival prediction of oesophageal cancer patients treated with external beam radiotherapy for dysphagia

Introduction The recent POLDER trial investigated the effects of external beam radiotherapy (EBRT) on dysphagia caused by incurable oesophageal cancer. An estimated life expectancy of minimally three months was required for inclusion. However, nearly one-third of the included patients died within three months. The aim of this study was to investigate if the use of prediction models could have improved the physician's estimation of the patient's survival. Methods Data from the POLDER trial (N = 110) were linked to the Netherlands Cancer Registry to retrieve patient, tumour, and treatment characteristics. Two published prediction models (the SOURCE model and Steyerberg model) were used to predict three-month survival for all patients included in the POLDER trial. Predicted survival probabilities were dichotomised and the accuracy, sensitivity, specificity, and the area under the curve (AUC) were used to evaluate the predictive performance. Results The SOURCE and Steyerberg model had an accuracy of 79% and 64%, and an AUC of 0.76 and 0.60 (p = .017), respectively. The SOURCE model had higher specificity across survival cut-off probabilities, the Steyerberg model had a higher sensitivity beyond the survival probability cut-off of 0.7. Using optimal cut-off probabilities, SOURCE would have wrongfully included 16/110 patients into the POLDER and Steyerberg 34/110. Conclusion The SOURCE model was found to be a more useful decision aid than the Steyerberg model. Results showed that the SOURCE model could be used for three-month survival predictions for patients that are considered for palliative treatment of dysphagia caused by oesophageal cancer in addition to clinicians' judgement.

Automated summary

This study investigated the use of prediction models to improve the estimation of patient survival in dysphagia caused by incurable esophageal cancer. The results showed that the SOURCE model was a more useful decision aid than the Steyerberg model, with higher accuracy and specificity.