qCT LN Quant

qCT LN Quant is a software device utilizing deep learning to provide quantitative characterization, including volumetric and diameter measurements, of user-identified solid pulmonary nodules from CT scans. Its methodology involves semi-automatic, user-assisted segmentation and integrates decision support tools like the Brock Score and Lung-RADS calculators to aid clinicians in the assessment and tracking of these nodules.

The qCT LN Quant is a software device designed for the tracking, assessment, and quantitative characterization of user-detected solid pulmonary nodules (≥ 6mm) from non-contrast chest CT scans. Its core methodology involves a deep-learning algorithm that processes these scans after a user performs semi-automatic segmentation of nodules using double seed points, with options for interactive fine-tuning of boundaries. The system provides volumetric and diameter measurements, alongside computer analysis based on morphological characteristics, either for a single study or longitudinally across several thoracic studies. Results, including 2D and 3D reconstructed views of nodules with their respective measurements, can be integrated into existing workflows via PACS as a secondary capture series or viewed on a dedicated web application. The device is intended for use by trained medical specialists such as radiologists and pulmonologists. Key quantitative outputs include average diameter and volume, with additional features such as long and short axis diameters, effective diameter, Mean/Minimum/Maximum Hounsfield Units (HU), volume change over time, and Volume Doubling Time (VDT). The software also incorporates a Brock Score - Risk Calculator, which uses the device-measured nodule diameter and clinician-provided inputs to estimate malignancy probability, and a Lung-RADS™ calculator based on ACR guidelines to assist in decision-making. Furthermore, it can provide recommendations based on Fleischner Society guidelines. Clinical performance testing on 216 solid nodules demonstrated median absolute normalized errors of 11.1% for long axis diameter, 14.3% for short axis diameter, and 20.7% for volume. The device is explicitly not intended for clinical diagnosis or to replace a full patient evaluation but serves as an informational tool.

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 semi-automatic segmentation process using double seed points compare in terms of user time and accuracy to fully automated or purely manual segmentation methods for pulmonary nodules?; What is the impact of variability in CT acquisition parameters (e.g., slice thickness, reconstruction kernel) on the accuracy of the qCT LN Quant's quantitative measurements, especially for volume doubling time calculations?; How was the deep learning algorithm trained and validated, particularly concerning the diversity of nodule morphologies and patient populations, to ensure robustness and minimize bias beyond the provided study of 216 nodules?; Could the integration of the Brock Score and Lung-RADS calculators within the software potentially influence clinician decision-making, and are there studies planned or underway to evaluate this impact on patient management pathways and outcomes?; What are the specific performance characteristics of the qCT LN Quant when analyzing nodules at the lower limit of its indicated size (≥ 6mm) or those with subtle morphological features?