# Detector‐specific correction factors for small‐field photon dosimetry in magnetic resonance‐guided radiation therapy: A systematic review and meta‐analysis

## TL;DR

This study reviews and analyzes correction factors for small-field dosimetry in magnetic resonance-guided radiation therapy, showing that a single correction factor is insufficient due to high variability.

## Contribution

The study provides a systematic review and meta-analysis of detector-specific correction factors in MRgRT, revealing the sources of variability and the need for specific correction factors.

## Key findings

- High heterogeneity (I² > 92%) was found across correction factors, making a single global correction factor statistically inappropriate.
- Correction factors depend on magnetic field strength, detector model, and orientation.
- Field output factors showed a strong, statistically significant dependence on field size.

## Abstract

Accurate small‐field (SF) dosimetry is essential for Magnetic Resonance‐guided Radiation Therapy (MRgRT), particularly for stereotactic treatments. The static magnetic field (B‐field) alters detector response, necessitating detector‐specific correction factors kB,Qf,f to ensure dosimetric accuracy. While numerous studies have reported such factors, the published data shows considerable variability, creating uncertainty for clinical practice.

This study was undertaken to (1) systematically review the literature on SF photon dosimetry in MRgRT and (2) perform a meta‐analysis to pool quantitative data and investigate the sources of this variability.

A systematic search of four databases was conducted following PRISMA 2020 guidelines. A qualitative synthesis was performed on all included studies (n = 86). A random‐effects meta‐analysis was performed on studies providing correction factors (beam quality correction factor, k
msr; field output correction factor, k
clin) or uncorrected output factors (FOF) for fields ≤ 4 × 4 cm2 with associated uncertainties (n = 13 studies, 441 data points). Heterogeneity (I
2) was quantified, and subgroup analyses (B‐field, detector model, and orientation) and meta‐regression (field size) were used to investigate its sources.

The qualitative review identified key themes in foundational physics, detector‐specific characteristics, and clinical quality assurance. The meta‐analysis revealed extremely high heterogeneity (I2
 > 92%) across all pooled data, confirming that a single global‐average correction factor is statistically inappropriate. Subgroup analyses demonstrated that this heterogeneity is systematically driven by three main factors: magnetic field strength (e.g., 0.35 T vs. 1.5 T), specific detector model (with significant variations observed even among models of the same type), and detector orientation relative to the beam and B‐field. Meta‐regression showed that correction factors k
msr and k
clin demonstrated no statistically significant dependence on field size in this SF regime, which suggests that the theoretical variation expected due to beam property changes is statistically masked by the high heterogeneity (I2 > 92) observed in the literature; conversely, FOFs showed a strong, statistically significant field‐size dependence.

The findings confirm that detector response in MRgRT is a complex function of the measurement setup. A single correction factor is insufficient for clinical use. Dosimetric accuracy in MRgRT requires the use of correction factors that are specific to the MR‐Linac's B‐field strength, the exact detector model, and the measurement orientation. This analysis provides pooled, stratified data to support clinical physicists and inform the development of future dosimetry protocols.