Impact of segmentation and discretization on radiomic features in 68Ga-DOTA-TOC PET/CT images of neuroendocrine tumor

Virginia Liberini; Bruno De Santi; Osvaldo Rampado; Elena Gallio; Beatrice Dionisi; Francesco Ceci; Giulia Polverari; Philippe Thuillier; Filippo Molinari; Désirée Deandreis

doi:10.1186/s40658-021-00367-6

Impact of segmentation and discretization on radiomic features in ⁶⁸Ga-DOTA-TOC PET/CT images of neuroendocrine tumor

EJNMMI Phys. 2021 Feb 27;8(1):21. doi: 10.1186/s40658-021-00367-6.

Authors

Affiliations

¹ Nuclear Medicine Unit, Department of Medical Sciences, University of Turin, Corso Dogliotti 14, 10126, Turin, Italy. virginia.liberini@unito.it.
² Biolab, Department of Electronics and Telecomunications, Politecnico di Torino, Turin, Italy.
³ Medical Physics Unit, AOU Città della Salute e della Scienza, Turin, Italy.
⁴ Nuclear Medicine Unit, Department of Medical Sciences, University of Turin, Corso Dogliotti 14, 10126, Turin, Italy.
⁵ Department of Endocrinology, University Hospital of Brest, Politecnico di Torino Brest, Turin, France.

Abstract

Objective: To identify the impact of segmentation methods and intensity discretization on radiomic features (RFs) extraction from ⁶⁸Ga-DOTA-TOC PET images in patients with neuroendocrine tumors.

Methods: Forty-nine patients were retrospectively analyzed. Tumor contouring was performed manually by four different operators and with a semi-automatic edge-based segmentation (SAEB) algorithm. Three SUV_max fixed thresholds (20, 30, 40%) were applied. Fifty-one RFs were extracted applying two different intensity rescale factors for gray-level discretization: one absolute (AR60 = SUV from 0 to 60) and one relative (RR = min-max of the VOI SUV). Dice similarity coefficient (DSC) was calculated to quantify segmentation agreement between different segmentation methods. The impact of segmentation and discretization on RFs was assessed by intra-class correlation coefficients (ICC) and the coefficient of variance (COV^L). The RFs' correlation with volume and SUV_max was analyzed by calculating Pearson's correlation coefficients.

Results: DSC mean value was 0.75 ± 0.11 (0.45-0.92) between SAEB and operators and 0.78 ± 0.09 (0.36-0.97), among the four manual segmentations. The study showed high robustness (ICC > 0.9): (a) in 64.7% of RFs for segmentation methods using AR60, improved by applying SUV_max threshold of 40% (86.5%); (b) in 50.9% of RFs for different SUV_max thresholds using AR60; and (c) in 37% of RFs for discretization settings using different segmentation methods. Several RFs were not correlated with volume and SUV_max.

Conclusions: RFs robustness to manual segmentation resulted higher in NET ⁶⁸Ga-DOTA-TOC images compared to ¹⁸F-FDG PET/CT images. Forty percent SUV_max thresholds yield superior RFs stability among operators, however leading to a possible loss of biological information. SAEB segmentation appears to be an optimal alternative to manual segmentation, but further validations are needed. Finally, discretization settings highly impacted on RFs robustness and should always be stated.

Keywords: 68Ga-DOTATOC PET/CT; Neuroendocrine tumor; Radiomics; Robustness; Semi-automatic segmentation; Texture analysis.