Comparative methods for PET image segmentation in pharyngolaryngeal squamous cell carcinoma

Abstract

Purpose Several methods have been proposed for the segmentation of 18 F-FDG uptake in PET. In this study, we assessed the performance of four categories of 18 F-FDG PET image segmentation techniques in pharyngolaryngeal squamous cell carcinoma using clinical studies where the surgical specimen served as the benchmark. Methods Nine PET image segmentation techniques were compared including: five thresholding methods; the level set technique (active contour); the stochastic expectationmaximization approach; fuzzy clustering-based segmentation (FCM); and a variant of FCM, the spatial wavelet-based algorithm (FCM-SW) which incorporates spatial information during the segmentation process, thus allowing the handling of uptake in heterogeneous lesions. These algorithms were evaluated using clinical studies in which the segmentation results were compared to the 3-D biological tumour volume (BTV) defined by histology in PET images of seven patients with T3-T4 laryngeal squamous cell carcinoma who underwent a total laryngectomy. The macroscopic tumour specimens were collected "en bloc", frozen and cut into 1.7-to 2-mm thick slices, then digitized for use as reference.Results The clinical results suggested that four of the thresholding methods and expectation-maximization overestimated the average tumour volume, while a contrast-oriented thresholding method, the level set technique and the FCM-SW algorithm underestimated it, with the FCM-SW algorithm providing relatively the highest accuracy in terms of volume determination (−5.9±11.9%) and overlap index. The mean overlap index varied between 0.27 and 0.54 for the different image segmentation techniques. The FCM-SW segmentation technique showed the best compromise in terms of 3-D overlap index and statistical analysis results with values of 0.54 (0.26-0.72) for the overlap index. Conclusion The BTVs delineated using the FCM-SW segmentation technique were seemingly the most accurate and approximated closely the 3-D BTVs defined using the surgical specimens. Adaptive thresholding techniques need to be calibrated for each PET scanner and acquisition/processing protocol, and should not be used without optimization.

Automatic method for tumor segmentation from 3-points dynamic PET acquisitions

“…Our results turns to be competitive with those reported in [3], where a set of segmentation techniques are compared in terms of SOI using a dataset of 7 patients suffering from pharyngolaryngeal squamous cell carcinoma. In particular, in [3] it is found that one of the best performing method, termed FCM-SW, yields an average SOI of 0.54.…”

“…However, PET segmentation remains an open problem mainly because of the limited image resolution and presence of acquisition noise [3].…”

“…In particular, in [3] it is found that one of the best performing method, termed FCM-SW, yields an average SOI of 0.54. FCM-SW employs a more complex segmentation approach as opposed to our simple 3-points RX Detector: it is based on fuzzy c-Means algorithm coupled with anisotropic diffusion filter to incorporate spatial information and wavelet transform.…”

“…However validation of accuracy and precision of these algorithms is still unresolved, due also to the lack of standard approaches and guidelines by nuclear medicine and radiation oncology professional societies [3].…”

Role of fluorine-18 fluorodeoxyglucose PET/CT in head and neck oncology: the point of view of the radiation oncologist

“…5,31,[72][73][74]76,80 However, few groups have validated delineation process using different imaging modalities against surgical resection specimens. 31,[81][82][83][84][85] In general, all imaging modalities overestimated the tumour extension compared with surgical specimen. Nevertheless, none of the image modalities (CT, MRI or PET) completely encompassed the surgical specimen volume because of an underestimation of superficial tumour extension in the mucosa, 31 as also reported by Ng et al 81 According to Daisne et al, 31 the GTV delineated from 18 F-FDG-PET applying an adaptive signal-to-background method was significantly smaller than GTV delineated by CT or MRI.…”