Search

Additional file 1: Table S1. LifeX radiomic features description according to the Imaging Biomarker Standardization Initiative (IBSI) description (update 17 December 2019). LifeX version was 4.81. Figure S1. Box plots showing the distribution of SUVmax (panel A) and Volume (panel B). Figure S2. (A) Bar diagrams of intra-class correlation coefficient (ICC) values of RFs for robustness to SUVmax thresholding. Bars show the median ICC between the different segmentations for the absolute intensity rescale factor AR60. Range error bars (in black) encompass the lowest and highest values for different operators. (B) Boxplot of COVL for different threshold (20, 30, 40%) for each RFs, for the first operator (results superposable for the other operators). TLG (total lesion glycolysis) conventional parameter in our study corresponds to the TLSRE (total lesion somatostatin receptor expression). Figure S3. Radiomic features with moderate or poor consistency (ICC < 0.80), but high agreement (median COVL < 10%) to intensity discretization. The RFs were: GLCM_Entropy_log2, GLCM_Entropy_log10 (not shown), GLRLM_SRE, GLRLM_LRE and GLRLM_RP. Value of the RFs for each lesion are presented in the top row; boxplots of COVL for the first operator are presented in the bottom row. Figure S4. Radiomic features with high consistency (ICC > 0.90), but low agreement (median COVL > 10%) to SUVmax thresholds (0, 20, 30 and 40%). The RFs were: GLRLM_LGRE, GLRLM_SRLGE, GLZLM_LGZE and GLZLM_LZLGE. Value of the RFs for each lesion are presented in the top row; boxplots of COVL for the first operator are presented in the bottom row. Figure S5. Boxplot showing the distribution of RF value for each operator. The three RFs chosen are the most representative of the impact of segmentation on ICC. Segmentation did not affect SUVmean (A) and TLG (B) in terms of ICC, although TLG was characterized by not negligible dispersion (percentage of COVL) in our study. In contrast, segmentation had a high impact on GLZLM_SZLGE (C) in terms of both ICC and COVL. Mean COVL of SUVmean, TLG and GLZLM_SZLGE was 8.33±3.96, 13.38±8.52 and 30.67±27.29, respectively.

In prostate cancer (PCa), the use of new radiopharmaceuticals has improved the accuracy of diagnosis and staging, refined surveillance strategies, and introduced specific and personalized radioreceptor therapies. Nuclear medicine, therefore, holds great promise for improving the quality of life of PCa patients, through managing and processing a vast amount of molecular imaging data and beyond, using a multi-omics approach and improving patients’ risk-stratification for tailored medicine. Artificial intelligence (AI) and radiomics may allow clinicians to improve the overall efficiency and accuracy of using these “big data” in both the diagnostic and theragnostic field: from technical aspects (such as semi-automatization of tumor segmentation, image reconstruction, and interpretation) to clinical outcomes, improving a deeper understanding of the molecular environment of PCa, refining personalized treatment strategies, and increasing the ability to predict the outcome. This systematic review aims to describe the current literature on AI and radiomics applied to molecular imaging of prostate cancer.

Purpose: Medullary thyroid carcinoma (MTC) originates from thyroid parafollicular C-cells and represents