QLAB Advanced Quantification Software

The QLAB Advanced Quantification Software System, through its new 3D Auto MV application, intends to improve cardiac diagnostics by offering a semi-automatic tool for detailed mitral valve quantification from Philips ultrasound image data. This is accomplished by integrating a machine-learning based segmentation engine with established morphological assessment algorithms, which enables automated 3D model generation and the reporting of a comprehensive suite of mitral valve parameters.

The Philips QLAB Advanced Quantification Software System is a software application package designed for viewing and quantifying image data obtained from Philips ultrasound systems. The latest iteration, version 15.0, introduces a significant enhancement: the 3D Auto MV cardiac quantification application, a new Q-App which aims to provide clinicians with a dynamic method for clinical quantification of the mitral valve by integrating existing technologies from QLAB HeartModel and TomTec-Arena 4D MV Assessment. The 3D Auto MV application operates as a semi-automatic tool, combining a machine-learning derived segmentation engine with morphological and functional assessment algorithms to analyze the mitral valve. This integration facilitates the semi-automatic creation of a 3D surface model of the mitral valve, with the system proposing initial anatomical landmarks that users can review, edit, or accept, thereby aiming to improve workflow. The underlying machine learning algorithm is specifically trained for the mitral valve and supports the introduction of several new measurement parameters (such as D-Shaped Annulus Perimeter, Maximum Prolapse Height, D-Shaped Annulus Area, and Total Open Coaptation Area) for a more comprehensive assessment. Non-clinical verification and validation, including algorithm training and a validation study, were performed to confirm its performance specifications.

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 machine learning algorithm in the 3D Auto MV application differentiate between normal and pathological mitral valve morphologies, and what was the diversity of the dataset used for its training and validation?; What is the inter-observer and intra-observer variability for the new mitral valve quantification parameters introduced by the 3D Auto MV application (e.g., D-Shaped Annulus Perimeter, Maximum Prolapse Height)?; Can the 3D Auto MV software be integrated with non-Philips ultrasound systems, or are there specific data acquisition protocols from Philips systems required for optimal performance of its quantification features?; How does the semi-automatic landmark proposal feature in the 3D Auto MV application impact workflow efficiency and reduce user input time compared to fully manual methods or previous semi-automatic tools?; What are the known limitations or specific patient populations and image quality scenarios where the 3D Auto MV application's accuracy or performance might be suboptimal?