ESTRO 2024 - Abstract Book

S1136

Clinical - Haematology

ESTRO 2024

745

Poster Discussion

whole body MRI for contouring of total marrow and lymph node irradiation

Damiano Dei 1,2 , Pierina Navarria 1 , Luisa Bellu 1 , Elena Clerici 1 , Nicola Lambri 2,1 , Ricardo Coimbra Brioso 3 , Leonardo Crespi 3 , Daniele Loiacono 3 , Stefania Bramanti 4 , Marta Scorsetti 2,1 , Pietro Mancosu 1 1 IRCCS Humanitas Research Hospital, Radiotherapy and Radiosurgery Department, Milan, Italy. 2 Humanitas University, Department of Biomedical Sciences, Milan, Italy. 3 Politecnico di Milano, Dipartimento di Elettronica, Milan, Italy. 4 Humanitas Clinical and Research Hospital, Bone Marrow Transplantation Unit, Milan, Italy

Purpose/Objective:

Total Marrow and Lymph node Irradiation (TMLI) is a modulated radiotherapy technique delivered as conditioning regimen for bone marrow transplantation, which allows to spare dose from organs at risk (1) . The TMLI target includes bones, lymphatic system, and spleen (2) . In particular, accurate lymph-node delineation is fundamental to provide optimal plan. Intravenous contrast use during CT planning can enhance tissue contrast optimizing target definition. Unfortunately long acquisition time of the total body CT does not allow for optimal administration of contrast agents. Our aim is to evaluate the potential of whole body simulation MRI in optimizing lymph node delineation.

Material/Methods:

Since 2010, 126 patients have been treated in our center using volumetric modulated arc therapy (VMAT). Since March 2022, whole body MRI simulation without contrast agent has been implemented in the workflow for TMLI planning on 15 patients. A dedicated all body frame immobilization system was used only CT procedure. Raystation treatment planning system was used to coregister simulation MRI-CT Imaging and define the lymph node CTV (CTV_LN) exclusively on CT (CTV_LN_CT) and on MRI (CTV_LN_MRI). Deformable registration was performed using various specific ROIs for better registration accordance. MRI-based segmentation was mapped on planning CT. A customized script for data extraction was used. CTV_LN_CT and CTV_LN_MRI geometrical similarity, volumes comparison and dosimetric coverage were assessed.

Results:

The CTV_LN_MRI for a standart patient is reported in figure 1. Similar results were obtained in all patients. A specific analysis was performed on 3 randomly selected patients. Mean difference between CTV_LN_CT and CTV_LN_MRI was 175 cc with larger volumes for MRI approach. Mean Dice Similarity Coefficient (DSC) for CTV_LN was 0.76 (range 0.68-0.81), while Hausdorff Distance (HD) was 3 mm (range 2.2-3.8 mm). The best similarity results were found for pelvic area with a DSC of 0.81 (range 0.79-0.82) and HD of 1.55 mm (range 1.4-1.6 mm), and the worst for abdominal region with a DSC of 0.72 (range 0.55-0.83) and HD of 2.85 mm (range 1.6-3.8 mm). Mean V95%