Artificial intelligence for reduced dose 18F-FDG PET examinations: a real-world deployment through a standardized framework and business case assessment

Abstract Background To determine whether artificial intelligence (AI) processed PET/CT images of reduced by one-third 18-F-FDG activity compared to the standard injected dose, were non-inferior native scans and if so assess potential impact commercialization. Materials methods SubtlePET™ AI was introduced in a center Italy. Eligible patients referred for 18F-FDG prospectively enrolled. Administered two-thirds dose. Patients underwent one low-dose CT two sequential PET scans; “PET-processed” with dose acquisition time, “PET-native” an elapsed time simulate PET-processed reconstructed using SubtlePET™. PET-native defined as reference. The datasets anonymized independently evaluated random order four blinded readers. evaluation included subjective image quality (IQ) assessment, lesion detectability, assessment business benefits. Results From February April 2020, 61 Subjective IQ not significantly different between (4.62±0.23, p =0.237) all scanner models, “almost perfect” inter-reader agreement. There no significant difference lesions’ target mean SUV max value, liver value (182.75/181.75 [SD:0.71], 9.8/11.4 [SD:1.13], 2.1/1.9 [SD:0.14] respectively). No false-positive lesions reported examinations. Agreed price per examination 15-20% FDG savings. Conclusion This is first real-world study demonstrate non-inferiority examinations obtained 66% methodology define solution price.