Development of a practice-based score to predict extended duration of proton beam therapy session in pediatric patients

TL;DR

This study developed a score to predict long proton beam therapy sessions in children, helping improve treatment efficiency.

Contribution

The study introduces a new predictive score based on patient factors to estimate extended proton beam therapy session durations in pediatric patients.

Findings

A nomogram-based score with an area under the curve of 0.84 was developed to predict extended PBT session durations.

Age over 45 months was a strong predictor of extended session duration.

Factors like long-term venous access and craniospinal location also significantly increased session duration.

Abstract

Due to the labor intensity demanded by proton beam therapy (PBT) in pediatric patients, information on operational procedures related to efficiency is crucial to optimize quality and safety. We aimed to identify patient factors that affect the duration of the pediatric PBT session and to develop an easy-to-use predictive score of extended duration.

This is an observational retrospective cohort study in an academic medical centre, between May 2020 and February 2024. Seventy seven ASA III pediatric patients treated with PBT were recruited.

The mean age was 4.8 years [standard deviation (SD): 2.1] and 52% were women. The mean duration of the PBT session was 50 min (SD: 17). Extended duration of the PBT session (> 45 min) occurred in 39 patients (51%). Five predictors of extended duration were selected for the final prediction model. In the multivariable model, an age > 45 months showed a near eightfold increased odds of extended duration [Odds ratio (OR): 7.76, 95% confidence interval (95% CI) 1.63–36.99, P = 0.010]. The OR (95% CI) for long-term venous access, no recurrent tumors, hydrocephalus, and craniospinal location were 5.91 (1.47 to 23.79), 3.81 (0.67 to 21.69), 3.79 (0.90 to 15.97), and 2.59 (0.69 to 9.76), respectively. This five-variable model was used to build a nomogram-based score with an area under the receiver operating characteristic curve of 0.84 (95% CI 0.76–0.93).

A simple nomogram based on readily available pretreatment data has potential for planning pediatric PBT standard clinical expert practice.

Figures