Broncholab

Broncholab is a software medical device that provides quantitative analysis of pulmonary structures from CT scans to support clinical diagnosis, documentation, and follow-up examinations. It employs advanced image processing techniques on thoracic CT datasets to generate patient-specific 3D lung models and derive objective metrics such as volumetric data, lung density scores, airway dimensions, blood vessel density, and fissure integrity.

Broncholab is a Software as Medical Device (SaMD) designed to assist physicians by providing quantitative analysis of pulmonary tissue from CT thoracic datasets. It supports diagnosis, follow-up examinations, and the assessment of therapy effectiveness by transforming DICOM 3.0 compliant inspiratory CT scan images into patient-specific 3D lung models. The system utilizes a dedicated workflow that includes a web portal for uploading CT scans and patient data, and a client report accessory to convert the quantitative CT values into a digital report for physicians. Key image processing steps involve 3D segmentation, isolation of subcompartments, volumetric analysis, density evaluations, low density cluster analysis, and fissure evaluation, with both manual and automated quality checks implemented to ensure data integrity. The device generates several critical quantitative CT values to characterize lung structure and potential abnormalities. These include lung and lobar volumes, airway volume (defined from the trachea to segmental bronchi), and various lung density scores, such as low attenuation areas below -950 Hounsfield Units (LAA-950HU) and the 15th percentile of the density histogram (PD15), which are indicative of emphysema. Broncholab also quantifies blood vessel density through segmentation and 3D reconstruction from inspiratory CT scans and assesses fissure integrity by measuring the percentage of completeness. Non-clinical performance testing, which compared Broncholab to a predicate device and physician-corrected segmentations for airway volume, demonstrated that for parameters like lobe volume, emphysema, and blood vessel density (expressed as relative values), the mean differences were less than 5%, and for PD15, the mean difference was less than 1 HU. The Intraclass Correlation Coefficient for fissure integrity was 0.75, and airway volume measurements also showed minimal differences from the physician-derived ground truth, all performed using scans from a diverse range of CT scanners.

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 automated quality check algorithm for the 3D model generation differentiate between true anatomical variation and image artifacts, especially across different CT scanner protocols?; What is the specific methodology for "low density cluster analysis" mentioned, and how does its output clinically correlate with different phenotypes or severity of lung conditions beyond LAA-950HU and PD15?; Given that ground truth for airway volume was established by three physicians, what was the inter-observer variability among these physicians, and how does Broncholab's deviation from their mean compare to this variability?; While non-clinical data shows good agreement, what is the anticipated impact of real-world patient factors (e.g., motion artifacts, severe pathology distorting anatomy, comorbidities) on the accuracy and reproducibility of Broncholab's measurements?; How sensitive is the blood vessel density measurement to variations in CT contrast media timing and concentration, and are there specific CT acquisition protocols recommended to optimize this parameter's accuracy?