Inter-Observer Delineation Differences in Esophageal Cancer GTV on MR, CT, and PET/CT Protokol šečerov Ermenc 28dec19

Overview

Esophageal cancer radiotherapy requires accurate delineation of the gross tumor volume (GTV) to ensure adequate tumor coverage and to minimize radiation dose to surrounding organs at risk. Computed tomography (CT) is routinely used for target delineation, while positron emission tomography-computed tomography (PET/CT) may provide additional functional information. However, uncertainty in tumor boundaries and inter-observer variation remain clinically relevant issues. Magnetic resonance imaging (MRI) provides improved soft tissue contrast and diffusion-weighted imaging (DWI) information, which may improve visualization of esophageal tumors. This study evaluates whether MRI-based imaging, alone or fused with CT or PET/CT, reduces inter-observer variation in GTV delineation compared with standard CT and PET/CT-based delineation. A total of 21 patients with locally advanced esophageal cancer undergoing standard radiotherapy preparation will undergo MRI simulation in addition to routine CT and PET/CT simulation. Five experienced radiation oncologists will independently delineate the primary tumor GTV on CT, MRI, PET/CT, CT-MRI fusion, and PET/CT-MRI fusion datasets. Inter-observer variation will be assessed using conformity indices and spatial analysis tools.

Accurate target volume delineation is essential in radiotherapy for esophageal cancer, as contouring uncertainty may lead to geographic miss or increased dose to organs at risk. CT imaging is the standard modality used for radiotherapy planning; however, CT has limited soft tissue contrast, which can make delineation of the proximal and distal tumor extent challenging. PET/CT may assist in staging and may provide functional tumor information, but its value for precise target delineation remains uncertain. MRI has recently gained importance in radiotherapy planning due to improved anatomical soft tissue visualization and functional imaging such as diffusion-weighted imaging (DWI). Technical advances have improved thoracic MRI feasibility, and MRI may allow improved identification of tumor borders in the esophagus. The clinical role of MRI in reducing inter-observer delineation variability in esophageal cancer radiotherapy planning is not yet well established. This prospective study compares inter-observer variation in gross tumor volume (GTV) delineation across different imaging modalities used in radiotherapy preparation. Patients with locally advanced esophageal cancer will undergo standard PET/CT simulation according to institutional protocol and additional MRI simulation performed with identical immobilization. MRI will include T2-weighted sequences and DWI sequences without intravenous contrast. Five experienced radiation oncologists will independently delineate the primary tumor GTV on anonymized datasets using the Eclipse treatment planning system. Contouring will be performed separately on the following image sets: CT, MRI, PET/CT, CT fused with MRI, and PET/CT fused with MRI. Pathologic lymph nodes will not be included in the GTV. A washout period of at least 14 days will be maintained between contouring sessions for the same patient on different imaging modalities. Inter-observer variation will be assessed using the generalized conformity index (CIgen). Additionally, reference contours will be generated using the STAPLE algorithm. The Contour Analysis Tool (CAT) will be used to calculate planar and volumetric conformity indices (PCI and VCI) and to perform spatial analysis based on inter-delineation distance measurements. Contouring time, image quality, and perceived delineation difficulty will also be recorded by observers. The primary objective is to determine whether MRI-based imaging reduces inter-observer variability in GTV delineation compared with standard CT and PET/CT imaging in esophageal cancer radiotherapy planning.