ESTRO 2025 - Abstract Book

S153

Invited Speaker

ESTRO 2025

4928

Speaker Abstracts Dose accumulation in radiotherapy of skin cancer Marta Szlag 1 , Magdalena Stankiewicz 2 , Sylwia Kellas-Ślęczka 2 , Piotr Wojcieszek 2

1 Radiotherapy Planning Department, Maria Sklodowska-Curie National Research Institute of Oncology Gliwice branch, Gliwice, Poland. 2 Brachytherapy Department, Maria Sklodowska-Curie National Research Institute of Oncology Gliwice branch, Gliwice, Poland

Abstract:

Treatment of multiple, non-simultaneous non-melanoma skin cancers (NMSCs) typically involves separate interstitial or superficial brachytherapy treatments, each with its own plan. Accurate dose accumulation for subsequent tumors requires transferring previous treatment dose maps to a common coordinate system, usually the most recent CT scan, to account for dose delivered to organs at risk and normal tissue. The choice of image registration algorithm (rigid [RIR], deformable [DIR], or structure-based deformable algorithm that uses delineated structure contours [SDIR]) and the dosimetric parameters for evaluating the cumulative dose are critical for accurate dose summation. Different algorithms will affect the accuracy of dose translation. Registration accuracy was assessed by comparing the maximal point dose Dmax in the original dose map to its value in the translated dose maps; significant discrepancies in Dmax indicate poor registration. While other dose volume histogram (DVH) parameters, such as the dose to 2 cm³ (D ₂ cm³) and dose to 0.1 cm³ (D ₀ . ₁ cm³ ) of the structure, may vary between their original and registered values due to substantial organ size changes, but they should remain relatively stable in regions such as the head and neck scalp, where significant volume changes are less likely. parameters. Alternatively, deformable image registration (DIR or SDIR) might be necessary. Analysis of a group of seven patients with multifocal non-melanoma skin cancer revealed that DIR significantly altered Dmax, D ₀ . ₁ cm³, and D ₂ cm³ likely due to poor skin surface recognition near the applicator. After deformation, Dmax reached 39.2 Gy, compared to its pre-translated value of 9.4 Gy. Concurrently, D ₀ . ₁ cm³ increased from 7.8 Gy to 16.4 Gy, and D ₂ cm³ increased from 5.7 Gy to 7.6 Gy. RIR only significantly affected Dmax (9.4 Gy vs 17.0 Gy). All values represent the median for the analyzed patient group. In contrast, SDIR maintained consistency in all three parameters, demonstrating its superiority in this context. Visual inspection and discrepancies in Dmax between pre- and post-translation dose maps revealed that DIR misidentified the applicator as skin, causing deformation and inaccurate skin contour representation. This resulted in artificially higher Dmax and overestimations of D ₀ . ₁ cm³, and D ₂ cm³ for the external contour as the deformed skin contour was closer to the source. In contrast, structure-based SDIR handled this deformation more effectively. Following image registration-based dose map translation, cumulative dose distributions can be displayed on the most recent CT scan. The presentation of spatial cumulated dose is extremely helpful in dose assessment as without this information it is difficult to draw conclusions about normal tissue exposure. The significance of high- and low-dose regions may depends on the distance between treated regions. Consecutive treatments in separate areas may increase the overall volume of tissue receiving lower doses, while overlapping target volumes may result in dose escalation within that region. Future directions in multifocal skin cancer brachytherapy require advances in the following areas: Robust 3D dose accumulation algorithms and standardised methodology and quality assurance; Biologically effective dose (BED) models that incorporate cumulative effects; and identification of key DVH parameters and development of predictive models that correlate dose with clinical outcomes. Focusing solely on the patient's external body contour and three dose-volume histogram (DVH) parameters (Dmax, D ₂ cm³, and D ₀ . ₁ cm³ ) the question is whether rigid registration accurately translates these dose