Quality assurance of craniospinal irradiation in helical tomotherapy using a dose reconstruction tool based on leaf open time

TL;DR

This study evaluates a dose reconstruction tool for quality assurance in craniospinal irradiation using helical tomotherapy, showing it can effectively identify delivery issues.

Contribution

The study introduces a novel method using leaf open time data to assess dose delivery accuracy in craniospinal irradiation.

Findings

Delivery Analysis showed high accuracy in predicting dose quality assurance outcomes with area under the curve values of 0.858 and 0.714 for cranial and spine plans.

Modulation factors and beam parameters were identified as key risk factors for dose delivery failure.

Reoptimizing failed plans with a higher modulation factor significantly improved passing rates in both DA and MatriXX assessments.

Abstract

The delivery quality assurance (DQA) of craniospinal irradiation (CSI) due to the target length results in no ideal verification devices. Delivery Analysis (DA) could calculate the dose distribution based on the measured multi leaf open time in helical tomotherapy (HT). This study aimed to evaluate the efficacy of DA for DQA of CSI in HT.

32 CSI plans were classified into two groups based on γ analysis of the PTV‐cranial and PTV‐spine plans using a 2D ionization chamber matrix (MatriXX). Plans with γ passing rates ≥ 95% at 3%/2 mm were classified as the passed group, while those < 95% were classified as the failed group. Receiver operating characteristic (ROC) curves identified optimal passing rate threshold for DA in HT. Logistic regression analyzed risk factors for DQA failure, and failed plans were reoptimized according to the adjusted parameter.

For PTV‐cranial plans, 30 passed and two failed; for PTV‐spine plans, 21 passed and 11 failed. ROC analysis revealed areas under the curve of 0.858 (PTV‐cranial, threshold: 89.0%) and 0.714 (PTV‐spine, threshold: 86.0%). Logistic regression identified planned modulation factor (MF‐plan; p = 0.046; p = 0.023) and actual modulation factor (MF‐actual; p = 0.027; p = 0.008) as independent risk factors for DQA failure in both MatriXX and DA. Additionally, beam on time (p = 0.043), gantry period (p = 0.007) and maximum leaf open time (p = 0.007) were identified as independent risk factors for DA. Reoptimization of failed plans with MF‐plan = 2.6 significantly improved passing rates in DA (73.70% ± 13.30% vs. 88.20% ± 12.30%; p = 0.010) and MatriXX (91.20% ± 2.60% vs. 96.10% ± 1.40%; p < 0.001).

Delivery Analysis could be a feasible tool for DQA of CSI in HT. Increasing the MF‐plan is recommended to enhance the passing rate.

Figures