Voluson Expert 22, Voluson Expert 20, Voluson Expert 18

The Voluson Expert 22, 20, and 18 systems are advanced general-purpose diagnostic ultrasound platforms with a distinct specialization in OB/GYN, utilizing a broad spectrum of imaging modalities and transducer technologies. These systems are significantly enhanced by new and improved AI-driven software for automated anatomical measurements, dynamic pelvic floor assessment, comprehensive fibroid mapping, and refined obstetric image analysis, aiming to improve diagnostic accuracy and workflow efficiency as demonstrated through rigorous verification on diverse clinical datasets.

The Voluson Expert Series, encompassing models 22, 20, and 18, are general-purpose ultrasound systems intended for use by qualified healthcare professionals in hospital or medical clinic environments. These systems feature a mobile console design and are particularly specialized for OB/GYN applications, offering real-time 3D/4D acquisition capabilities alongside a comprehensive suite of imaging modes such as B-mode, M-mode, various Doppler modalities (PW, CW, Color, Power), Harmonic Imaging, Coded Pulse, Elastography, and Shear Wave Elastography. They support a variety of transducers, including linear, curved linear, and matrix phased array types, and introduce the new RIC12-D probe. Key advancements include the integration and enhancement of several AI-driven software features: SonoAVC2.0 (Auto-Caliper) for automated measurements, SonoPelvicFloor2.0 for dynamic assessment and live tracking of the mid-sagittal plane (MHD) of the pelvic floor, Fibroid Mapping for automated segmentation of the uterus, endometrium, and fibroids, and substantial improvements to SonoLyst/SonoLyst Live for enhanced obstetric image recognition and grading, now including views for lower and upper limbs. Additional new software functionalities are BSI (Blood Speckle Imaging) and Ophthalmic Artery Measurement. The AI features have undergone specific verification using diverse, de-identified datasets collected from multiple geographical sites and system configurations. SonoPelvicFloor2.0 demonstrated high accuracy in identifying the MHD plane, achieving 96% on good image quality datasets and 93% on challenging ones, exceeding the predefined success criteria. Fibroid Mapping showed strong performance with qualitative success rates of 95-100% for uterus segmentation, 89-91% for endometrium, and 88% for fibroids; quantitative analysis yielded mean DICE scores of 0.89 for the uterus, and 0.70 for both endometrium and fibroids in successful segmentations. The improved SonoLyst/SonoLyst Live, with its expanded anatomical view recognition and user interface enhancements, maintained an average success rate of 80% or higher for its image sorting and grading functionalities. These Voluson Expert systems employ the same fundamental scientific ultrasound technology as their predicate devices and have been shown to meet applicable safety and performance standards, including biocompatibility and acoustic output limits, with substantial equivalence supported by non-clinical testing and detailed AI verification summaries.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the performance of the Fibroid Mapping AI vary with different fibroid sizes, types (e.g., intramural, submucosal, subserosal), and in the presence of co-existing uterine pathologies like adenomyosis?; What are the specific characteristics defining 'challenging image quality' for the SonoPelvicFloor2.0 AI, and what is the quantitative impact on the accuracy of MHD plane tracking under these conditions?; For the SonoLyst/SonoLyst Live enhancements, what were the most significant challenges in developing and validating the AI for the new 'Lower Limb' and 'Upper Limb' views, and how does its grading accuracy for these views compare to previously established fetal anatomical views?; Can the BSI (Blood Speckle Imaging) feature provide quantifiable advantages over traditional Color or Power Doppler in specific, challenging vascular assessments, such as low-flow states or the evaluation of placental microvasculature?; Considering the diverse geographical and multi-system data sources used for AI training, what specific methodologies were employed to ensure the robustness and generalizability of the AI models to patient populations and ultrasound system configurations not extensively represented in the initial training and validation sets?