Incorporating dose\u2013volume histogram parameters of swallowing organs at risk in a\xa0videofluoroscopy-based predictive model of radiation-induced dysphagia after head and neck cancer intensity-modulated radiation therapy

Stefano Ursino,Emanuele Neri,Riccardo Morganti,Alessia Giuliano,Fabiola Paiar,Patrizia Giusti,Antonio Stefanelli,Claudio Traino,Giacomo Aringhieri,Fabio Di Martino,Paola Cocuzza,Alessandro Molinari

doi:10.1007/s00066-020-01697-7

Abstract

PurposeTo develop a videofluoroscopy-based predictive model of radiation-induced dysphagia (RID) by incorporating DVH parameters of swallowing organs at risk (SWOARs) in a machine learning analysis.MethodsVideofluoroscopy (VF) was performed to assess the penetration-aspiration score (P/A) at baseline and at 6 and 12 months after RT. An RID predictive model was developed using dose to nine SWOARs and P/A-VF data at 6 and 12 months after treatment. A total of 72 dosimetric features for each patient were extracted from DVH and analyzed with linear support vector machine classification (SVC), logistic regression classification (LRC), and random forest classification (RFC).Results38 patients were evaluable. The relevance of SWOARs DVH features emerged both at 6 months (AUC 0.82 with SVC; 0.80 with LRC; and 0.83 with RFC) and at 12 months (AUC 0.85 with SVC; 0.82 with LRC; and 0.94 with RFC). The SWOARs and the corresponding features with the highest relevance at 6 months resulted as the base of tongue (V65 and Dmean), the superior (Dmean) and medium constrictor muscle (V45, V55; V65; Dmp; Dmean; Dmax and Dmin), and the parotid glands (Dmean and Dmp). On the contrary, the features with the highest relevance at 12 months were the medium (V55; Dmin and Dmean) and inferior constrictor muscles (V55, V65 Dmin and Dmax), the glottis (V55 and Dmax), the cricopharyngeal muscle (Dmax), and the cervical esophagus (Dmax).ConclusionWe trained and cross-validated an RID predictive model with high discriminative ability at both 6 and 12 months after RT. We expect to improve the predictive power of this model by enlarging the number of training datasets.

Highlights

Radiation-induced dysphagia (RID) represents the real Achille’s heel of radiation-based organ preservation treatment in head and neck cancer (HNC)
We discovered that the dose received by the Swallowing Organs at Risk (SWOARs) located in the upper HN region (SPCM, BOT, MPCM, and parotid glands (PGs)) was extremely important for predicting RID at 6 months, whereas the dose received by the SWOARs located in the lower region (IPCM, glottic larynx (GL), cricopharyngeal muscle (CPM), and EC) was extremely important for predicting RID at 12 months after treatment
We strongly suggest that clinicians and physicists maximally enforce the dosimetric constraints of the lower SWOARs in the IMRT plan optimization process for this specific subset of HNC patients in order to reduce the risk of late RID

Summary

Introduction

Radiation-induced dysphagia (RID) represents the real Achille’s heel of radiation-based organ preservation treatment in head and neck cancer (HNC). A multivariable clinical normal tissue complication probability model (NTCP) of swallowing dysfunction at 6 months following RT was externally validated by correlating the grade of dysphagia according to the Radiation Therapy Oncology Group (RTOG) and European Organization for Research and Treatment of Cancer (EORTC) late radiation morbidity score (a clinical primary endpoint) and average mean dose (Dmean) to the Swallowing Organs at Risk (SWOARs) and salivary glands, reporting a good overall performance, discrimination, and goodness of fit [4] In this regard, five patterns of clinical swallowing dysfunction related to RT doses to the upper and lower pharyngeal region as well as to the larynx and salivary glands were revealed (low, intermediate, and severe persistent; transient and progressive), suggesting different underlying radiobiological mechanisms [5]

Objectives

Methods

Results

Discussion

Conclusion