AI Takes Over the First Step in Abdominal Ultrasound Measurement: Philips Alturion Receives U.S. Marketing Clearance

Ultrasound examinations depend heavily on the operator not only to find the appropriate imaging plane, but also to place measurement calipers in the correct positions. Philips has launched the Alturion ultrasound system in an effort to use artificial intelligence to handle these repetitive steps and reduce manual work in high-volume imaging departments. The system has received U.S. Food and Drug Administration (FDA) 510(k) marketing clearance; Philips also said that the product is available in the United States and Europe after receiving the CE mark.

Alturion’s AI Auto Measure Abdomen identifies the contours of abdominal organs and semi-automatically measures the kidney’s longitudinal length, transverse width, and height, as well as the spleen’s length. The caliper positions proposed by the system are not final; healthcare professionals can still adjust them or perform separate manual measurements. The feature is currently limited to use with the C5-1 and C9-2 transducers and requires software version 14.0 or later.

FDA documents show that Philips conducted a performance analysis using existing ultrasound images from 150 adults, comparing the algorithm’s results with the average measurements of three experienced clinical sonographers. The data came from three clinical sites, and the validation set was separated from the training data by site, time period, and equipment source. Participants ranged in age from 22 to 85, with equal numbers of men and women, and body mass index values spanning underweight to obese populations.

Bland–Altman agreement analyses for all four measurements met the prespecified acceptance criteria. Based on mean differences, the AI and manual reference values differed by approximately 0.26% to 2.34%. However, this validation was classified as nonclinical performance testing and did not demonstrate that the system can improve diagnostic accuracy, examination time, or patient outcomes. The FDA also explicitly stated that no additional clinical data were required for this substantial equivalence determination.

The composition of the dataset also limits interpretation of the results. Among the kidney images, 22% were classified as abnormal, but abnormal spleens accounted for only 1.33%. The available documents are therefore insufficient to determine whether the algorithm is equally stable when handling rare lesions, pronounced organ deformation, or more complex cases. Although the average of the three experts’ measurements provides a consistent comparison benchmark, it is not equivalent to pathology findings or long-term clinical outcomes.

Alturion also offers 2D Auto Cine, which continuously saves two-dimensional images during scanning so users can select suitable frames afterward. The system also supports strain and shear-wave elastography and can quantify tissue stiffness. The FDA classifies Alturion as a Class II medical device, with clearance covering multiple imaging applications, including abdominal, cardiac, obstetric and gynecological, vascular, musculoskeletal, and pulmonary uses. However, the core purpose of the 510(k) process is to confirm that it is substantially equivalent to an existing legally marketed device; it should not be interpreted as evidence that the FDA has confirmed that every automated feature provides clinical benefit.