Characterization of PET/CT images using texture analysis: the past, the present… any future?

Abstract

After seminal papers over the period 2009 – 2011, the use of texture analysis of PET/CT images for quantification of intratumour uptake heterogeneity has received increasing attention in the last 4 years. Results are difficult to compare due to the heterogeneity of studies and lack of standardization. There are also numerous challenges to address. In this review we provide critical insights into the recent development of texture analysis for quantifying the heterogeneity in PET/CT images, identify issues and challenges, and offer recommendations for the use of texture analysis in clinical research. Numerous potentially confounding issues have been identified, related to the complex workflow for the calculation of textural features, and the dependency of features on various factors such as acquisition, image reconstruction, preprocessing, functional volume segmentation, and methods of establishing and quantifying correspondences with genomic and clinical metrics of interest. A lack of understanding of what the features may represent in terms of the underlying pathophysiological processes and the variability of technical implementation practices makes comparing results in the literature challenging, if not impossible. Since progress as a field requires pooling results, there is an urgent need for standardization and recommendations/guidelines to enable the field to move forward. We provide a list of correct formulae for usual features and recommendations regarding implementation. Studies on larger cohorts with robust statistical analysis and machine learning approaches are promising directions to evaluate the potential of this approach.

Hybrid Imaging: Instrumentation and Data Processing

“…Thanks to hybrid imaging lesion detection, tissue delineation and therapy monitoring have become more accurate [266]. Hybrid imaging has been key to cancer patient management [267], thanks also to its ability to depict tumor heterogeneity in a morphological and functional context [260]. Given that conventional indices, such as standard uptake value (SUV), maximum or metabolic tumor volume (MTV) are insufficient to characterize malignancies [268], several groups have started to investigate the feasibility of in vivo tumor characterization in the context of textural evaluation [269].…”

“…Textural parameters are strongly influenced by the choice of image reconstruction, voxel size and post processing steps (e.g., Gaussian filtering). Furthermore, the choice and definitions of the textural analysis metrics vary widely and the evaluation techniques differ substantially, thus, rendering a comparison between studies almost impossible [260,297]. The need for standardization efforts in the field of textural analyses has been highlighted previously [260,298,299], and efforts are currently under way to address this challenge [260,[300][301][302].…”

PET/MRI: Reliability/Reproducibility of SUV Measurements

“…Furthermore, volume-based PET metrics such as the total lesion glycolysis (TLG), calculated by multiplying SUV mean by the metabolic tumor volume (MTV) [16], radiomic and texture analysis [17,18], and parametric imaging have been suggested [14,15]. The TLG, used to assess global metabolic response of the whole lesion thus providing complementary information to SUV and it variants, was shown to be highly correlated with other PET response parameters and is reproducible [19].…”

“…Recent advances in PET image segmentation and delineation of lesion contours [20] combined with progress in partial volume correction techniques have enabled to automate the calculation procedure. More recently, radiomics and texture analysis emerged as new promising approaches enabling to circumvent the limitations of the above described oversimplified approaches by providing additional features including intratumoral heterogeneity through advanced image processing techniques and knowledge in systems biology [17,18]. An increasing number of pioneering studies support the underlying assumptions of these hypotheses; however, further research and development efforts using large clinical databases are still required before these approaches can translate to valuable and reliable tools that can be adopted in the clinic.…”

Volumetric and texture analysis of pretherapeutic 18F-FDG PET can predict overall survival in medullary thyroid cancer patients treated with Vandetanib

“…Of note, intratumoral heterogeneity derived from in-vivo PET reflects the heterogeneity of tracer uptake assessed by exvivo autoradiography [13]. Clinically, intratumoral heterogeneity (tumor texture) analysis is currently gaining ground to serve as a potential risk stratification tool in a variety of different cancer entities [14][15][16][17][18][19]. Hence, not surprisingly, combined 18 F-FDG PET-based approaches (i.e., volumetric plus tumor texture assessment) for outcome prediction have also been investigated, e.g.…”