AnX Robotica said that FDA submissions for its NaviCam ProScan reading technology and for a recently cleared comparator AI-assisted capsule endoscopy platform both reported reductions in mean review time, while stating that the two studies were conducted independently and should not be read as a head-to-head comparison.
According to the FDA's De Novo summary for NaviCam ProScan (DEN230027, decision Dec. 12, 2023), the prospective multicenter ARTIC study covered 133 patients and found mean small-bowel reading time fell from 33 minutes 42 seconds with standard reading to 3 minutes 50 seconds with AI-assisted reading (p < 0.001). Diagnostic yield was reported as 62.4% with standard reading and 73.7% with AI-assisted reading.
For the comparator, the FDA's 510(k) summary (K254230, decision Sept. 25, 2026) reports a reader study of 111 patients in which mean review time fell from 45.7 minutes to 34.4 minutes with AI assistance, a mean difference of 11.3 minutes, with AI-assisted patient-level sensitivity of 92.3% and diagnostic yield of 87.6%. The same clearance identifies NaviCam ProScan as the primary predicate device.
The company said the two evaluations used different capsule systems, patient populations, study designs, readers, reference standards and validation datasets, so the figures describe what was reported in each product's FDA submission. Both products are intended to assist rather than replace clinician interpretation, and the ProScan De Novo summary states the clinician remains responsible for assessment and that ProScan is not intended as a stand-alone diagnostic device. Tim Thomas, the company's vice president of regulatory and clinical affairs, said performance claims should be considered in the context of each product's cleared indication and study design.
