# Benchmarking Modern Day Pencil Beam Scanning Proton Therapy Treatment Times: Insights From Real Time Location Service Treatment Time Data

## TL;DR

This paper uses real-time tracking data to analyze and benchmark proton therapy treatment times across different patient groups and equipment.

## Contribution

The study introduces a novel use of Real-Time Location System (RTLS) data to benchmark treatment times in proton therapy.

## Key findings

- Treatment times ranged from 26 minutes for vertebrae to 52 minutes for craniospinal and gynecologic cases.
- High-volume sites like prostate and head and neck had consistent treatment times but with outliers due to complex setups.
- Gantry differences in treatment times were observed, likely due to staff familiarity and case distribution.

## Abstract

Collecting consistent and accurate treatment time data is challenging, particularly when relying on manual records or predictive models. Our center uses a Real-Time Location System (RTLS), which provides precise, automated tracking of patient movement throughout the clinic. This study leverages RTLS data to benchmark treatment times in a high-volume pencil beam scanning proton facility where ancillary systems like cone beam CT, surface imaging, and ultrasound bladder scans are used to aid in setup.

We analyzed RTLS data from 12,551 fractions delivered between January 2023 and July 2024 across three gantries of an IBA Proteus Plus system. Treatment time was defined as the duration between a patient entering and exiting a gantry zone, recorded by infrared and radio frequency badges. Outliers and erroneous entries were excluded using clinical and statistical criteria. Mean and median treatment times were calculated by disease sites, subsites, and gantries. Imaging protocols and setup factors were provided to give context to the data.

Treatment times ranged from 26 minutes (vertebrae) to 52 minutes (craniospinal and gynecologic), with an overall average of 35.5 minutes. High-volume sites such as prostate (4479 fractions), head and neck (2153), and central nervous system (CNS) (1954) showed relatively tight distributions, but had many outliers due to complex setups or anatomical variability. Discrepancies in treatment time across gantries were observed, with certain gantries consistently faster, likely due to staff familiarity and case concentration on specific gantries.

RTLS provides a scalable method for tracking treatment times in proton therapy. These data offer an unbiased, data-driven foundation for clinical and business planners to develop more accurate and sustainable proton center models as well as serve as benchmark data for other proton centers seeking to optimize performance.