SaMD for Detection of Pulmonary Lung Nodules

Case Studies >>

Regulatory and Market Authorization Support for SaMD for Detection of Pulmonary Lung Nodules from CT scans

An AI-based Software Tool was designed for the automated assessment of multiple lung pathologies on CT scans by a Start-up company. The tool detects lung nodules for the assessment of lung cancer and estimates the 25-point CTSS (Computed Tomography Severity Score) for Covid-19 from CT scans of patients diagnosed with or suspected to have lung cancer and Covid-19 patients. The subject device aims to support radiologists in improving their workflow management in screening lung nodules and in estimating the CTSS in suspected Covid-19 cases.

The company behind this innovative tool sought Nextvel’s assistance in conducting a retrospective analysis of CTs to establish the sensitivity and specificity of the tool. Nextvel designed and conducted the study as per the Clinical Investigation Plan after obtaining the necessary approvals. The study was first performed for lung nodule detection and later internally validated for estimating the CTSS in suspected Covid-19 cases.

During the study, 244 scans were annotated, with 165 scans meeting the eligibility criteria and included for analysis. The AI achieved an overall average sensitivity of 85.71%, which was found to be sufficient to fulfil the intended purpose.

The successful validation of this AI-based tool has the potential to revolutionize the detection of lung nodules and the assessment of Covid-19 severity, enabling earlier intervention and improved patient outcomes.