Influence of filter choice on 18F-FDG PET segmentation accuracy determined using generalized estimating equations.

This study aims to quantify how filter choice affects several fluoro-deoxy-glucose (FDG)-positron emission tomography (PET) segmentation methods and present the use of model fitting via generalized estimating equations (GEEs) to appropriately account for the properties of a common segmentation quality metric (Dice similarity coefficient). Spherical and irregularly shaped 'hot' objects filled with 18F-FDG were placed in a medium with background activity and imaged for 1, 2 and 5 min durations at low and high contrasts. Images were filtered with Gaussian and bilateral filters of 5 and 7 mm full-width half maximum (FWHM), with and without 3 mm FWHM Gaussian pre-smoothing. Four segmentation methods were used: 40% thresholding, adaptive thresholding, k-means clustering and seeded region-growing. Segmentation accuracy was quantified by overlap (using Dice similarity coefficient (DSC)) and distance between surfaces (using symmetric-mean-absolute-surface-distance (SMASD)) of the ground truth and segmented volumes. All segmentation methods showed mean DSC values between 0.71-0.87 and mean SMASD values between 0.72-2.10 mm across filters. The bilateral filter with 3 mm FWHM Gaussian pre-smoothing had mean DSC 0.80 ± 0.17 and mean SMASD 1.17 ± 1.51 mm displaying approximately equal performance to a 5 mm Gaussian filter with mean DSC 0.79 ± 0.18 and mean SMASD 1.27 ± 1.52 mm. Results from models fit using GEE with a binomial distribution and exchangeable correlation structure estimated the correlation between DSC values as 0.118 and 0.290 for spheres and irregular objects, respectively. The GEE approach accounts for several factors specific to the DSC metric that simpler statistical approaches do not, providing more accurate estimations of experimental effects commonly associated with nuclear medicine segmentation studies.

Combining multiple FDG-PET radiotherapy target segmentation methods to reduce the effect of variable performance of individual segmentation methods.

PURPOSE: Many approaches have been proposed to segment high uptake objects in 18F-fluoro-deoxy-glucose positron emission tomography images but none provides consistent performance across the large variety of imaging situations. This study investigates the use of two methods of combining individual segmentation methods to reduce the impact of inconsistent performance of the individual methods: simple majority voting and probabilistic estimation. METHODS: The National Electrical Manufacturers Association image quality phantom containing five glass spheres with diameters 13-37 mm and two irregularly shaped volumes (16 and 32 cc) formed by deforming high-density polyethylene bottles in a hot water bath were filled with 18-fluoro-deoxyglucose and iodine contrast agent. Repeated 5-min positron emission tomography (PET) images were acquired at 4:1 and 8:1 object-to-background contrasts for spherical objects and 4.5:1 and 9:1 for irregular objects. Five individual methods were used to segment each object: 40% thresholding, adaptive thresholding, k-means clustering, seeded region-growing, and a gradient based method. Volumes were combined using a majority vote (MJV) or Simultaneous Truth And Performance Level Estimate (STAPLE) method. Accuracy of segmentations relative to CT ground truth volumes were assessed using the Dice similarity coefficient (DSC) and the symmetric mean absolute surface distances (SMASDs). RESULTS: MJV had median DSC values of 0.886 and 0.875; and SMASD of 0.52 and 0.71 mm for spheres and irregular shapes, respectively. STAPLE provided similar results with median DSC of 0.886 and 0.871; and median SMASD of 0.50 and 0.72 mm for spheres and irregular shapes, respectively. STAPLE had significantly higher DSC and lower SMASD values than MJV for spheres (DSC, p < 0.0001; SMASD, p = 0.0101) but MJV had significantly higher DSC and lower SMASD values compared to STAPLE for irregular shapes (DSC, p < 0.0001; SMASD, p = 0.0027). DSC was not significantly different between 128 × 128 and 256 × 256 grid sizes for either method (MJV, p = 0.0519; STAPLE, p = 0.5672) but was for SMASD values (MJV, p < 0.0001; STAPLE, p = 0.0164). The best individual method varied depending on object characteristics. However, both MJV and STAPLE provided essentially equivalent accuracy to using the best independent method in every situation, with mean differences in DSC of 0.01-0.03, and 0.05-0.12 mm for SMASD. CONCLUSIONS: Combining segmentations offers a robust approach to object segmentation in PET. Both MJV and STAPLE improved accuracy and were robust against the widely varying performance of individual segmentation methods. Differences between MJV and STAPLE are such that either offers good performance when combining volumes. Neither method requires a training dataset but MJV is simpler to interpret, easy to implement and fast.