• AI
  • Molecular Imaging
  • CT
  • X-Ray
  • Ultrasound
  • MRI
  • Facility Management
  • Mammography

Study: Adjunctive AI Imaging Software Enhances Contouring of Prostate Cancer


Artificial intelligence (AI) assisted contouring of prostate cancer demonstrated superior balanced accuracy than manual standard-of-care contouring and hemigland contouring with MRI, according to a new study.

Noting that accurate contouring can play a key role in the management and treatment of prostate cancer, the authors of a new study suggest that artificial intelligence (AI)-assisted contouring provides better accuracy than hemigland contouring and manual contouring with magnetic resonance imaging (MRI).

For the multireader retrospective study, recently published in the Journal of Urology, researchers compared AI-assisted contouring (Unfold AI, Avenda Health) — which incorporates MRI findings and clinical factors such as prostate-specific antigen (PSA) levels and biopsy results — with hemigland contouring and cognitive standard-of-care (SOC) contouring in 50 cases of patients who had a radical prostatectomy.

The 10 reviewing physicians (three radiologists and seven urologist) performed cognitive contouring and subsequently utilized the AI software for contouring on the same cases four weeks later, according to the study.

Study: Adjunctive AI Imaging Software Enhances Contouring of Prostate Cancer

In this case involving clinically significant prostate cancer (csPCa) and a comparison of standard-of-care (SOC) contours, hemigland cancer contours and AI-assisted contour, researchers noted a mean balanced accuracy of 64 percent for SOC, 77.7 percent for hemigland contours and 87.5 percent for AI-assisted cancer contours. (Images courtesy of the Journal of Urology.)

The researchers found that AI-assisted contouring had a balanced accuracy rate of 84.7 percent in contrast to 75.9 for hemigland contouring and 67.2 percent for cognitively defined contouring. The use of AI-assisted contouring had comparable sensitivity to hemigland contouring (97.4 percent vs. 98.4 percent but had a nearly 60 percent higher sensitivity rate than cognitive SOC contouring (38.2 percent), according to the study authors.

While cognitive SOC contouring offered the highest specificity rate (96.2 percent vs. 72.1 percent for adjunctive AI and 53.4 percent for hemigland contouring), it also had a significantly lower negative margin rate (1.6 percent) in contrast to adjunctive AI (72.8 percent) and hemigland contouring (86 percent).

“(Adjunctive AI) was significantly more accurate than hemigland and SOC contours, indicating a better balance between sensitivity and specificity. Furthermore, exposure to AI software encouraged physicians to define larger and more patient-specific contours, improving the negative margin rate. When AI software was used in our study population, the negative margin rate rose from 1.6% to 72.8% — a 45-fold increase,” wrote lead author Sakina Mohammed Mota, M.D,. a data scientist with Avenda Health, and colleagues.

(Editor’s note: For related content, see “Study Says AI Mapping More Effective than MRI for Assessing Extent of Prostate Cancer,” “FDA Clears AI ‘Contouring Assistant’ in MRI-Guided Ultrasound Ablation Procedures” and “MRI-Based Deep Learning Algorithm Shows Comparable Detection of csPCa to Radiologists.”)

Three Key Takeaways

1. Increased accuracy with AI-assisted contouring. AI-assisted contouring demonstrated higher accuracy (84.7 percent) compared to hemigland contouring (75.9 percent) and cognitive SOC contouring (67.2 percent). This suggests AI can improve the precision of prostate cancer treatment planning by better balancing sensitivity and specificity.

2. Improved negative margin rates. The use of AI software significantly increased the negative margin rate from 1.6 percent with cognitive SOC contouring to 72.8 percent. This indicates that AI-assisted contouring may lead to more effective and safer surgical outcomes by helping define more accurate tumor boundaries.

3. Enhanced treatment recommendations: The introduction of AI-assisted contouring led to changes in treatment recommendations in over 25 percent of cases. Notably, it resulted in a 9 percent increase in the recommendation for focal therapy, suggesting that AI can support more targeted and less invasive treatment options for prostate cancer.

The researchers also pointed out that the adjunctive AI software’s specificity advantage over hemigland contouring (72.1 percent vs. 53.4 percent) is a significant consideration with respect to targeted, efficient treatment of prostate cancer.

“Though the sensitivity of AI and hemigland contours were comparable, AI contours were smaller and more specific. It is therefore plausible that AI may facilitate faster treatment with fewer side effects, particularly because hemigland ablation entails lethal treatment in close proximity to the urethra, external sphincter, and ipsilateral nerve bundle,” added Mota and colleagues. “Avoidance of these structures, though common in clinical implementation, would reduce hemigland contour size and likely compromise their efficacy for a subset of cases.”

The researchers also noted that adjunctive AI changed treatment recommendations in over 25 percent of cases with a notable 9 percent change in urologists recommending focal therapy (mean of 21.5 percent with adjunctive AI vs. 12.5 percent based on SOC contouring).

“The cancer estimation map improved confidence in recommending focal therapy over whole gland therapy, with urologists trending towards a more targeted approach,” noted Mota and colleagues.

In regard to study limitations, the authors acknowledged the reviewed cases were derived from one institution and entirely drawn from patients who had radical prostatectomy procedures. The researchers noted the performance of hemigland contours may have been inflated due to limiting the cohort to those with apparently unilateral cancer.

Related Videos
Emerging Research at SNMMI Examines 18F-flotufolastat in Managing Primary and Recurrent Prostate Cancer
Could Pluvicto Have a Role in Taxane-Naïve mCRPC?: An Interview with Oliver Sartor, MD
Where the USPSTF Breast Cancer Screening Recommendations Fall Short: An Interview with Stacy Smith-Foley, MD
A Closer Look at MRI-Guided Transurethral Ultrasound Ablation for Intermediate Risk Prostate Cancer
Improving the Quality of Breast MRI Acquisition and Processing
Can Fiber Optic RealShape (FORS) Technology Provide a Viable Alternative to X-Rays for Aortic Procedures?
Can Diffusion Microstructural Imaging Provide Insights into Long Covid Beyond Conventional MRI?
Emerging MRI and PET Research Reveals Link Between Visceral Abdominal Fat and Early Signs of Alzheimer’s Disease
Nina Kottler, MD, MS
The Executive Order on AI: Promising Development for Radiology or ‘HIPAA for AI’?
Related Content
© 2024 MJH Life Sciences

All rights reserved.