Automated prostate gland segmentation in challenging clinical cases: comparison of three artificial intelligence methods

Latrice A Johnson; Stephanie A Harmon; Enis C Yilmaz; Yue Lin; Mason J Belue; Katie M Merriman; Nathan S Lay; Thomas H Sanford; Karthik V Sarma; Corey W Arnold; Ziyue Xu; Holger R Roth; Dong Yang; Jesse Tetreault; Daguang Xu; Krishnan R Patel; Sandeep Gurram; Bradford J Wood; Deborah E Citrin; Peter A Pinto; Peter L Choyke; Baris Turkbey

doi:10.1007/s00261-024-04242-7

Automated prostate gland segmentation in challenging clinical cases: comparison of three artificial intelligence methods

Abdom Radiol (NY). 2024 Mar 21. doi: 10.1007/s00261-024-04242-7. Online ahead of print.

Authors

Latrice A Johnson¹, Stephanie A Harmon¹, Enis C Yilmaz¹, Yue Lin¹, Mason J Belue¹, Katie M Merriman¹, Nathan S Lay¹, Thomas H Sanford², Karthik V Sarma³, Corey W Arnold⁴, Ziyue Xu⁵, Holger R Roth⁵, Dong Yang⁵, Jesse Tetreault⁵, Daguang Xu⁵, Krishnan R Patel⁶, Sandeep Gurram⁷, Bradford J Wood^{8

9}, Deborah E Citrin⁶, Peter A Pinto⁷, Peter L Choyke¹, Baris Turkbey^{10

11}

Affiliations

¹ Molecular Imaging Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA.
² Department of Urology, Hawaii Pacific Health, Honolulu, HI, USA.
³ Department of Psychiatry and Behavioral Sciences, University of California, San Francisco, CA, USA.
⁴ Department of Radiology, University of California, Los Angeles, Los Angeles, CA, USA.
⁵ NVIDIA Corporation, Santa Clara, CA, USA.
⁶ Radiation Oncology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA.
⁷ Urologic Oncology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA.
⁸ Center for Interventional Oncology, National Cancer Institute, NIH, Bethesda, MD, USA.
⁹ Department of Radiology, Clinical Center, NIH, Bethesda, MD, USA.
¹⁰ Molecular Imaging Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. turkbeyi@mail.nih.gov.
¹¹ Molecular Imaging Branch (B.T.), National Cancer Institute, National Institutes of Health, 10 Center Dr., MSC 1182, Building 10, Room B3B85, Bethesda, MD, 20892, USA. turkbeyi@mail.nih.gov.

PMID: 38512516
DOI: 10.1007/s00261-024-04242-7

Abstract

Objective: Automated methods for prostate segmentation on MRI are typically developed under ideal scanning and anatomical conditions. This study evaluates three different prostate segmentation AI algorithms in a challenging population of patients with prior treatments, variable anatomic characteristics, complex clinical history, or atypical MRI acquisition parameters.

Materials and methods: A single institution retrospective database was queried for the following conditions at prostate MRI: prior prostate-specific oncologic treatment, transurethral resection of the prostate (TURP), abdominal perineal resection (APR), hip prosthesis (HP), diversity of prostate volumes (large ≥ 150 cc, small ≤ 25 cc), whole gland tumor burden, magnet strength, noted poor quality, and various scanners (outside/vendors). Final inclusion criteria required availability of axial T2-weighted (T2W) sequence and corresponding prostate organ segmentation from an expert radiologist. Three previously developed algorithms were evaluated: (1) deep learning (DL)-based model, (2) commercially available shape-based model, and (3) federated DL-based model. Dice Similarity Coefficient (DSC) was calculated compared to expert. DSC by model and scan factors were evaluated with Wilcox signed-rank test and linear mixed effects (LMER) model.

Results: 683 scans (651 patients) met inclusion criteria (mean prostate volume 60.1 cc [9.05-329 cc]). Overall DSC scores for models 1, 2, and 3 were 0.916 (0.707-0.971), 0.873 (0-0.997), and 0.894 (0.025-0.961), respectively, with DL-based models demonstrating significantly higher performance (p < 0.01). In sub-group analysis by factors, Model 1 outperformed Model 2 (all p < 0.05) and Model 3 (all p < 0.001). Performance of all models was negatively impacted by prostate volume and poor signal quality (p < 0.01). Shape-based factors influenced DL models (p < 0.001) while signal factors influenced all (p < 0.001).

Conclusion: Factors affecting anatomical and signal conditions of the prostate gland can adversely impact both DL and non-deep learning-based segmentation models.

Keywords: Deep learning; Machine learning; Magnetic resonance imaging; Prostate; Segmentation.