Biograph Horizon

The Biograph Horizon PET/CT system integrates Positron Emission Tomography and X-ray Computed Tomography to provide fused metabolic and anatomical images for oncological, neurological, and cardiological diagnostics, utilizing radiopharmaceuticals and X-ray transmission data. It employs advanced software methodologies, including AI-driven motion correction (OncoFreeze AI), dynamic multi-parametric PET (FlowMotion AI), and enhanced scatter correction, to improve image quality, quantitative accuracy, and workflow efficiency as validated by performance testing.

The Biograph Horizon PET/CT system is a combined multi-slice X-Ray Computed Tomography (CT) and Positron Emission Tomography (PET) scanner designed for whole-body oncology, neurology, and cardiology examinations. It provides registration and fusion of high-resolution metabolic information from PET (which images and measures the distribution of PET radiopharmaceuticals) and anatomic information from CT (which produces cross-sectional images via X-ray transmission). The CT component also provides attenuation correction maps for PET studies and anatomical references for fused images, with both modalities capable of independent operation. The system is intended for use by trained healthcare professionals to aid in detecting, diagnosing, staging, and restaging various diseases, including cardiovascular disease, neurological disorders, and cancer, and can also assist in radiotherapy planning, interventional radiology, and low-dose lung cancer screening. This iteration of the Biograph Horizon PET/CT system introduces software modifications including corrections and new features such as OncoFreeze AI (Data Driven Gating), CardioFreeze, FlowMotion Multi-Parametric PET AI, PET FAST Planning, QualityGuard, and Whole Body Scatter Correction, alongside minor hardware updates. Performance testing according to NEMA NU2-2018 on 3-ring and 4-ring versions demonstrated adherence to acceptance criteria for resolution, count rate, scatter, sensitivity, image quality, and co-registration accuracy. Specific evaluations showed OncoFreeze AI (deviceless gating) achieved SUV and volume measurements comparable to Anzai-based gating; OncoFreeze/CardioFreeze enabled significantly more counts per gate; FlowMotion Multi-Parametric PET AI improved noise and delineation with Patlak transformation during reconstruction; PET FAST Planning accurately defined bed ranges (87.1%-100%); QualityGuard offered substantial time savings and reduced staff radiation exposure; and Whole Body Scatter Correction significantly improved ROI accuracy over single bed methods.

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 'OncoFreeze AI' deviceless gating specifically compensate for respiratory motion artifacts compared to traditional Anzai-based gating, and what are the implications for lesion quantification accuracy in different anatomical regions (e.g., lung base vs. liver dome)?; Can the 'FlowMotion Multi-Parametric PET AI' performing Patlak transformation during reconstruction be adapted for other kinetic models, and what is the computational overhead increase compared to static reconstruction?; Regarding the 'Whole Body Scatter Correction,' what are the primary algorithmic differences from 'Single Bed Scatter Correction' that account for the substantial improvement in ROI accuracy, particularly in areas distant from the primary activity concentration?; While 'QualityGuard' improves efficiency and reduces staff exposure, does its automated calibration process maintain the same level of long-term system stability and image consistency as manual calibration, particularly across different system configurations (3-ring vs. 4-ring TrueV)?; For 'PET FAST Planning (FlowMotion AI),' what are the most common failure modes or anatomical scenarios leading to the lower end of the bed range identification accuracy (87.1%), and are there manual override or correction capabilities for the user?