A systematic review on artificial intelligence evaluating PSMA PET scan for intraprostatic cancer

Area of Science:

• Oncology
• Radiology
• Artificial Intelligence

Background:

• Prostate cancer (PCa) evaluation before treatment relies on accurate imaging.
• Prostate-specific membrane antigen positron emission tomography (PSMA PET) is increasingly used for PCa staging.
• Assessing intraprostatic PCa on PSMA PET scans is crucial for treatment decisions.

Approach:

• A systematic review was conducted searching major databases for studies on AI evaluating intraprostatic PCa on PSMA PET scans.
• Included studies focused on AI's ability to assess lesions prior to active treatment (radiotherapy or prostatectomy).
• The Prediction model Risk Of Bias Assessment Tool (PROBAST) was used for quality assessment.

Key Points:

• AI demonstrated high accuracy in differentiating high-grade PCa (ISUP Grade Group ≥3 and ≥4).
• AI showed capability in identifying non-PSMA-avid lesions and detecting extraprostatic extensions.
• AI can automate lesion segmentation and gross tumor volume measurement on PSMA PET scans.

Conclusions:

• Current AI applications for evaluating intraprostatic PCa on PSMA PET scans are promising but experimental.
• AI offers potential benefits in local staging, identifying occult lesions, and standardizing reporting.
• Further large-scale, prospective, multicenter studies are necessary to establish AI for routine clinical use.