Assessment of various strategies for 18F-FET PET-guided delineation of target volumes in high-grade glioma patients
- 630 Downloads
The purpose of the study is to assess the contribution of 18F-fluoro-ethyl-tyrosine (18F-FET) positron emission tomography (PET) in the delineation of gross tumor volume (GTV) in patients with high-grade gliomas compared with magnetic resonance imaging (MRI) alone.
Materials and methods
The study population consisted of 18 patients with high-grade gliomas. Seven image segmentation techniques were used to delineate 18F-FET PET GTVs, and the results were compared to the manual MRI-derived GTV (GTVMRI). PET image segmentation techniques included manual delineation of contours (GTVman), a 2.5 standardized uptake value (SUV) cutoff (GTV2.5), a fixed threshold of 40% and 50% of the maximum signal intensity (GTV40% and GTV50%), signal-to-background ratio (SBR)-based adaptive thresholding (GTVSBR), gradient find (GTVGF), and region growing (GTVRG). Overlap analysis was also conducted to assess geographic mismatch between the GTVs delineated using the different techniques.
Contours defined using GTV2.5 failed to provide successful delineation technically in three patients (18% of cases) as SUVmax < 2.5 and clinically in 14 patients (78% of cases). Overall, the majority of GTVs defined on PET-based techniques were usually smaller than GTVMRI (67% of cases). Yet, PET detected frequently tumors that are not visible on MRI and added substantially tumor extension outside the GTVMRI in six patients (33% of cases).
The selection of the most appropriate 18F-FET PET-based segmentation algorithm is crucial, since it impacts both the volume and shape of the resulting GTV. The 2.5 SUV isocontour and GF segmentation techniques performed poorly and should not be used for GTV delineation. With adequate setting, the SBR-based PET technique may add considerably to conventional MRI-guided GTV delineation.
KeywordsPET/CT 18F-FET Image segmentation Gross tumor volume High-grade glioma
This work was supported by the Swiss National Science Foundation under grant No. 3152A0-102143, the Indo Swiss Bilateral Research Initiative (ISBRI) supported by the Swiss State Secretariat for Education and Research under grant No. AP24, and the foundation Cellex International. This paper is dedicated to the memory of Prof. Bruce Hasegawa (Department of Radiology, UCSF), a brilliant scientist and true friend who sadly passed away last summer.
- 17.Erdi Y, Mawlawi O, Larson S, et al. Segmentation of lung lesion volume by adaptive positron emission tomography image thresholding. radioimmunodetection and radioimmunotherapy of cancer, Princeton, New Yersey, 1997:2505–2509.Google Scholar