Translating Live EPID based Inspiration Level Assessment (LEILA) into clinical practice

Abstract

Purpose

Deep inspiration breath hold (DIBH) minimises patient motion and reduces radiation exposure to organs at risk during breast radiotherapy. An effective DIBH strategy relies on patients maintaining consistent and reproducible breathing patterns, which is typically guided by monitoring external surrogates. The LEILA system is a real-time verification system, utilising electronic portal imaging device (EPID) images to monitor internal anatomy during DIBH breast radiotherapy. This work describes the development and implementation of LEILA as a clinic-ready application.

Methods

For the treatment planning phase, the LEILA system includes a fluence model to automatically predict EPID images, enabling the estimation of lung depths (LDs) and skin distances (SDs) for each control point. During treatment delivery, EPID images are acquired and LDs and SDs differences are quantified.

A pilot study was conducted to validate the LEILA system’s feasibility, by monitoring DIBH alongside existing motion management strategy in our clinic.

Results

The LEILA system was deployed on two Varian TrueBeam linear accelerators with aS1200 EPIDs. It showed low latency, with an average image processing time of 74.0 ms (StDev = 9.4 ms). Differences in mid LDs and SDs, categorised by breath-holds, for the 17 monitored beams resulted in averages and standard deviations of −1.2 mm ± 3.1 mm and 1.5 mm ± 4.1 mm respectively.

Conclusions

The LEILA system’s streamlined and automated workflow meets clinical needs at minimal extra workload. LEILA assesses the accuracy and reproducibility of DIBH, providing immediate feedback with deviations and retrospective insights to refine DIBH monitoring strategies.