Prognostic value of automatically acquired biomarkers using artificial intelligence in 18F-Choline PET/CT in high-risk prostate cancer.

Conclusions: Automated deep learning-based measurements of 18F-choline uptake in the prostate gland were significantly associated with prostate cancer specific survival in patients with hormone-naive prostate cancer. This type of deep learning-based methods could be applied to other prostate cancer PET tracers as well, for example PSMA. C-index and Univariate Cox proportional hazards regression model.Variable ClinicalC-index95% CIHazard ratio95% CIp-valAge0.610.46 - 0.760.970.88 - 1.060.50PSA0.650.50 - 0.791.021.00 - 1.050.034Gleason0.690.47 - 0.921.700.87 - 3.330.12T stage0.740.51 - 0.972.100.79 â" 5.650.14PET/CTProstate Vol0.730.61 - 0.851.031.00 - 1.060.016Lesion Vol0.760.62 - 0.911.051.02 - 1.080.001TLU0.730.58 - 0.882.371.22 â" 4.590.011SUVmax0.580.44 - 0.721.140.92 - 1.410.24
Source: Journal of Nuclear Medicine - Category: Nuclear Medicine Authors: Tags: Technical Advances Posters Source Type: research