ESTRO 2024 - Abstract Book

S4083

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

Accuracy of seed- and grey value match in Elekta XVI software for prostate and mamma radiotherapy

Chin Loon Ong, Marieke Overdevest, Eric M Franken

HagaZiekenhuis, Radiotherapy, The Hague, Netherlands

Purpose/Objective:

Image guidance (IGRT) plays a vital role in modern radiotherapy. With increasing hyperfractionation regimes and highly conformal dose delivery, the treatment outcome is profoundly dependent upon the consistency and precision of IGRT in matching patient’s daily treatment position. Using Elekta CBCT and XVI software, daily online registration can be carried out using bone match, seed/marker match, or gray value match. The purpose of this study is to evaluate the 3D registration’s accuracy of seed match vs gray value match (GV) in prostate and mamma radiotherapy.

Material/Methods:

15 accelerated partial breast (APBI) and 15 prostate patients were included in this study. For each patients, the registration of 3 treatment fractions (beginning, middle and end) were analyzed. Both groups of patients had 3-4 implanted gold markers in the target area. Prior to treatment, daily CBCT was acquired and dual registration was performed using Elekta XVI software. Firstly, a rigid registration of the bony anatomy within a large region of interest (ROI) around the target area was performed, followed by second matching on a smaller ROI (usually created by expanding the target area with 0-5mm margin). The second match was executed using both seed and grey value matching method. Seed match focuses only in the registration of the high density marker within the ROI meanwhile GV registers voxel gray scale intensity value throughout the entire target volume within the ROI. All matches underwent clinical review by a senior RTT for acceptability. The match results for all 6 directions (translations and rotations) for both methods were compared.

Results:

The average differences and standard deviations for the seed- and GV match in all 6 different directions for both groups of patients are presented in table 1. The differences in most translational axes were <1mm (except in vertical direction for APBI). The differences between the 2 techniques were larger in rotational axes, with standard deviations of up to 8⁰ in the pitch axis for prostate patients. All GV registrations were clinically acceptable. 4 and 13 fractions of the seed registration for ABPI and prostate patients, respectively, resulted in rotational correction of >10⁰. These fractions were not clinically acceptable. When these fractions were excluded from the analysis, the rotational standard deviations decreases to <3⁰. The large rotation is caused by image artifacts or the present of calcification within the target area. In some fractions, the large rotational correction in seed match can be avoided by repeating the match (Figure 1). However, this workaround diminishes the reliability and consistency of the seed match.

Table 1. Average differences and standard deviations between seed matcn and grey values match for accelerated partial breast irradiation (n = 15) and prostate irradiation (n = 15)

Longitudinal, cm

Lateral,cm

Vertical, cm

Pitch, º

Roll, º

Yaw, º