Session Item

The aim of this study was 1) to assess  the prevalence of tube feeding dependence (TFD) during the course of radiotherapy in head and neck cancer (HNC) patients treated over a time span of 13 years of technological improvements and improved organ sparing capabilities and 2) to build and validate NTCP models to estimate the benefit of proton therapy (IMPT) over VMAT with regard to reducing TFD.

The study population was composed of 1181 HNC patients who underwent definitive (chemo)radiation. Patients were treated with parotid sparing (ParRT: 2007-2010), swallowing sparing (SwRT: 2011-2017, including additional sparing of pharyngeal constrictor muscle (PCM), supraglottic larynx and oesophageal inlet muscle) or multi-organ sparing (MoRT: 2018-2020, including additional sparing of oral cavity and submandibular glands) radiotherapy (Figure 1). All patients were subjected to a prospective data registration protocol in which TFD (CTCAEv4.0 grade 4 dysphagia) was scored at baseline (BSL), weekly during treatment and at week 12. Dose-volume data for a wide range of OAR were collected for the clinical plan and (in the case of IMPT) the backup VMAT plans. First, the prevalence of TFD was studied for the aforementioned three subsequent time periods. Subsequently, multivariable NTCP modelling was performed in development (2007-2016) and validation (2017-2020) cohorts using recently reported methods (R&O 148, p.151; R&O 157, p.147). NTCP models were used for model-based clinical validation and for estimating the clinical benefit of proton therapy in 135 patients receiving IMPT.

The prevalence of TFD was 40.7%, 41.3% and 31.7% (last week of RT, W6/7) and 34.9%, 30.6% and 19.2% (week 12) during ParRT, SwRT and MoRT, respectively (Figure 1). The predictors for the final models were: Oral cavity Dmean; Superior PCM Dmean; Accelerated radiotherapy; Chemoradiation; and Bioradiation. With MoRT, TFD in the last week of RT was 31.7%, which was lower than predicted (34.5%, 95%CI: 30.6-38.3%). In week 12, the prevalence of TFD was 19.2%, also lower than predicted (24.9%, 95%CI: 20.7-29.1%). The NTCP of TFD with the clinical Mo-IMPT plans was reduced with 5.1 (46.9-41.8) and 6.6 (36.9-30.3) percent point in the last week of RT and in week 12 as compared with the backup Mo-VMAT plans. The observed prevalence with IMPT was 4.2 (46.9-42.7) and 13.7 (36.9-23.2) percent point lower than the NTCP predicted for the backup Mo-VMAT plans in the same weeks, respectively (Figure 1).

A marked decline of TFD during treatment for head and neck cancer was observed during 13 years of continuous technological improvements and improved organ sparing capabilities. In a recent cohort of patients treated with multi-organ sparing VMAT or IMPT the prevalence of TFD was even lower than expected. IMPT further reduces the NTCP of TFD with 5-7 percent point compared with VMAT, and with IMPT recovery of TFD seems even better than expected in the weeks following RT.