Precise Image for Pediatrics

This is a retrospective, observational, multi-site, controlled, blinded, non-inferiority clinical investigation.

Precise Image is a reconstruction mode using a trained deep learning neural network to generate images with reduced noise and improved low contrast detectability at lower doses compared with standard FBP (filter back projection) reconstruction.

Pediatric subjects (from 18 months up to and including 18 years of age) scanned by the FDA-cleared Philips Incisive CT (K212441) or CT 5300 (K232491) systems.

The study will include reconstructed images of 168 CT scans in total. For each of the anatomies and resolutions, 42 scans will be used.

Inclusion criteria

• Subjects scanned by FDA-cleared and marketed Philips Incisive CT or CT 5300 systems, and according to the department standard of care clinical protocols for the applicable anatomies.
• Subject age: From 18 months up to and including 18 years of age.

Exclusion criteria

• Subject age: Below 18 months and above 18 years of age.
• Scans that were acquired using generic 'Head' and 'Body' scan types (as Precise Image does not support generic 'Head' and 'Body' scan types).
• Head scans (brain and sinus/facial/orbits) acquired with a kilovoltage peak (kVp) outside the range of 100-120 kVp (as Precise Image does not support kVp settings outside this range for these clinical scan types).
• Scans that were acquired using a kilovoltage peak of 70 kVp (as Precise Image does not support a kilovoltage peak lower than 80 kVp).
• Scans that the site radiologists have deemed as non-diagnostic image quality in standard practice (e.g. patient movement).
• Scans that failed internal technical checks (e.g., corrupt files, failure to reconstruct).