Clarius Prostate AI

Clarius Prostate AI is a machine learning algorithm integrated into the Clarius ultrasound app that performs semi-automatic measurement of prostate volume on ultrasound images. The device automatically detects and measures prostate length, width, and height, then calculates volume; users can manually adjust measurements as needed. It is intended for use by trained healthcare professionals in adult male patients only.

Clarius Prostate AI uses a Deep Neural Network for image segmentation and border detection of the prostate gland, with automatic view classification. It operates on ultrasound data from Clarius curvilinear and endo-cavitary scanners running on iOS and Android platforms. The predicate device (Clarius Bladder AI) uses the same non-adaptive machine learning approach but segments the bladder instead; both employ identical architectural principles for segmentation and measurement.

IEC 62304:2006 + A1:2015 (medical device software lifecycle), ISO 14971:2019 (risk management), NEMA PS 3.1–3.20 (DICOM), IEC 62366-1:2015 + A1:2020 (usability engineering), ISO 15223-1:2021 (medical device symbols). Testing followed FDA guidance on General Principles of Software Validation (2002), Device Software Functions (2023), and Cybersecurity in Medical Devices (2023).

Both subject and predicate devices are ultrasound image processing software using non-adaptive ML algorithms trained on clinical data for automated segmentation, measurement, and volume calculation of anatomical structures. Clinical verification on 139 independent cases showed Clarius Prostate AI prostate volume measurements are non-inferior to expert clinician measurements (p-value 1.146e-5, equivalence margin 22%), with strong Bland-Altman agreement (ICC 0.87 for endo-cavitary, 0.67 for curvilinear probes). The only material difference is the anatomical target (prostate vs. bladder), which does not raise safety or effectiveness concerns.