ESTRO 2025 - Abstract Book

S2750

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2025

1194

Digital Poster Dosimetric investigation of tiny metal implant impact on CyberKnife using DirectDensityTM CT reconstruction algorithm Tsz Ching Fok, Natalie Kar Wei Yip, Cheuk Lam Kelly Chung, Tin Lok Chiu, Siu Ki Yu Medical Physics Department, Hong Kong Sanatorium & Hospital, Hong Kong, Hong Kong Purpose/Objective: In current radiotherapy practice, handling CT artifacts with metallic implants requires manual contouring of the structure and, based on the implant information from the vendor, assigning an electron density (ED) or mass density (mD) in the treatment planning system (TPS) for correct dose calculation. Siemens Healthcare has developed the DirectDensity TM (DD) algorithm that offers direct ED information regardless of variation in kVp by correlating the ED using an analytical formula [1]. In CyberKnife (CK), some patients undergo craniotomy prior to cranium radiotherapy treatments, leading to tiny metallic implants in skulls where density override is unjustified due to their small dimensions. A previous study has shown that DD reconstruction eliminates metal shadows around metal inserts compared to conventional Hounsfield unit(HU)-ED reconstruction [2], highlighting the potential benefits of using DD. However, to date, no detailed dosimetric study has been conducted to examine the effects of DD algorithm implementation on CK patients with cranial metallic implants. Material/Methods: Five brain cancer patients who completed CK treatment were retrospectively recruited. The original sets of their image data were additionally reconstructed with the DD algorithm and imported to Precision TPS for re-optimized dose calculation with the same contours. Target and OAR criteria remained unchanged. DVH metrics of the targets and OARs were compared between the original plans. Gamma index analysis with a 2 mm distance-to-agreement and 3% dose difference was also performed on the datasets. Results: Since the DD-CT calculated plans were optimized to achieve similar target coverage as the original treatment plans, similarity was expected. The PTVs demonstrated a similar increase in Dmean and Dmax in plans generated with DD CT sets. Notable changes were observed in OARs, yet all doses were within the oncologist's criteria. The brainstem Dmax showed the largest dose difference, ranging from -4.51 Gy to +0.93 Gy. Nearly all the cases displayed visualized results of having large scale of dose difference >3% outside the target region and a relatively good agreement in the high dose gradient fall-off region in Gamma index analysis.