PMID- 32737518 OWN - NLM STAT- MEDLINE DCOM- 20210528 LR - 20210528 IS - 1619-7089 (Electronic) IS - 1619-7070 (Print) IS - 1619-7070 (Linking) VI - 48 IP - 2 DP - 2021 Feb TI - Machine learning-based analysis of [(18)F]DCFPyL PET radiomics for risk stratification in primary prostate cancer. PG - 340-349 LID - 10.1007/s00259-020-04971-z [doi] AB - PURPOSE: Quantitative prostate-specific membrane antigen (PSMA) PET analysis may provide for non-invasive and objective risk stratification of primary prostate cancer (PCa) patients. We determined the ability of machine learning-based analysis of quantitative [(18)F]DCFPyL PET metrics to predict metastatic disease or high-risk pathological tumor features. METHODS: In a prospective cohort study, 76 patients with intermediate- to high-risk PCa scheduled for robot-assisted radical prostatectomy with extended pelvic lymph node dissection underwent pre-operative [(18)F]DCFPyL PET-CT. Primary tumors were delineated using 50-70% peak isocontour thresholds on images with and without partial-volume correction (PVC). Four hundred and eighty standardized radiomic features were extracted per tumor. Random forest models were trained to predict lymph node involvement (LNI), presence of any metastasis, Gleason score >/= 8, and presence of extracapsular extension (ECE). For comparison, models were also trained using standard PET features (SUVs, volume, total PSMA uptake). Model performance was validated using 50 times repeated 5-fold cross-validation yielding the mean receiver-operator characteristic curve AUC. RESULTS: The radiomics-based machine learning models predicted LNI (AUC 0.86 +/- 0.15, p < 0.01), nodal or distant metastasis (AUC 0.86 +/- 0.14, p < 0.01), Gleason score (0.81 +/- 0.16, p < 0.01), and ECE (0.76 +/- 0.12, p < 0.01). The highest AUCs reached using standard PET metrics were lower than those of radiomics-based models. For LNI and metastasis prediction, PVC and a higher delineation threshold improved model stability. Machine learning pre-processing methods had a minor impact on model performance. CONCLUSION: Machine learning-based analysis of quantitative [(18)F]DCFPyL PET metrics can predict LNI and high-risk pathological tumor features in primary PCa patients. These findings indicate that PSMA expression detected on PET is related to both primary tumor histopathology and metastatic tendency. Multicenter external validation is needed to determine the benefits of using radiomics versus standard PET metrics in clinical practice. FAU - Cysouw, Matthijs C F AU - Cysouw MCF AUID- ORCID: 0000-0001-5946-0173 AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. m.cysouw@amsterdamumc.nl. FAU - Jansen, Bernard H E AU - Jansen BHE AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Urology, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - van de Brug, Tim AU - van de Brug T AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Epidemiology and Biostatistics, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - Oprea-Lager, Daniela E AU - Oprea-Lager DE AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - Pfaehler, Elisabeth AU - Pfaehler E AD - Department of Nuclear Medicine and Molecular Imaging, Medical Imaging Center, University of Groningen, Groningen, the Netherlands. FAU - de Vries, Bart M AU - de Vries BM AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - van Moorselaar, Reindert J A AU - van Moorselaar RJA AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Urology, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - Hoekstra, Otto S AU - Hoekstra OS AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - Vis, Andre N AU - Vis AN AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Urology, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. FAU - Boellaard, Ronald AU - Boellaard R AD - Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, De Boelelaan, 1117, Amsterdam, the Netherlands. LA - eng PT - Journal Article PT - Multicenter Study DEP - 20200731 PL - Germany TA - Eur J Nucl Med Mol Imaging JT - European journal of nuclear medicine and molecular imaging JID - 101140988 SB - IM MH - Humans MH - Machine Learning MH - Male MH - *Positron Emission Tomography Computed Tomography MH - Prospective Studies MH - *Prostatic Neoplasms/diagnostic imaging MH - Risk Assessment PMC - PMC7835295 OTO - NOTNLM OT - Machine learning OT - PSMA PET-CT OT - Prostate cancer OT - Radiomics COIS- R. Boellaard has a scientific collaboration with Philips Healthcare, Netherlands, not pertaining to the present study. No other potential conflicts of interest relevant to this article exist. EDAT- 2020/08/02 06:00 MHDA- 2021/05/29 06:00 PMCR- 2020/07/31 CRDT- 2020/08/02 06:00 PHST- 2020/04/21 00:00 [received] PHST- 2020/07/22 00:00 [accepted] PHST- 2020/08/02 06:00 [pubmed] PHST- 2021/05/29 06:00 [medline] PHST- 2020/08/02 06:00 [entrez] PHST- 2020/07/31 00:00 [pmc-release] AID - 10.1007/s00259-020-04971-z [pii] AID - 4971 [pii] AID - 10.1007/s00259-020-04971-z [doi] PST - ppublish SO - Eur J Nucl Med Mol Imaging. 2021 Feb;48(2):340-349. doi: 10.1007/s00259-020-04971-z. Epub 2020 Jul 31.