ScreenDx

ScreenDx is a software-only device designed to analyze lung computed tomography (CT) imaging data, providing qualitative adjunctive information for findings related to interstitial lung disease as part of a clinician referral pathway. Its methodology involves leveraging proprietary algorithms validated through a retrospective multicenter study, demonstrating high sensitivity and specificity in identifying these lung findings.

Imvaria, Inc.'s ScreenDx, a software-only device, has undergone review and been determined substantially equivalent to a legally marketed predicate device. This clearance allows the device, classified as Class II, to be marketed. ScreenDx is designed to receive and analyze lung computed tomography (CT) imaging data, providing a qualitative output of findings related to interstitial lung disease. Its purpose is to offer adjunctive information as part of a referral pathway to qualified clinicians, supplementing existing standard-of-care workflows. However, the device does not replace current diagnostic methods for lung fibrosis, and its results are not intended to rule out or rule in the condition. The device's review included an assessment of its Predetermined Change Control Plan (PCCP), which outlines specific modifications to the software's algorithm that can be made without requiring further review, provided they adhere to predefined protocols for data management, re-training, and performance evaluation. Non-clinical testing ensured the device's safety, and a retrospective multicenter study demonstrated its performance, achieving a sensitivity of 91.4% and specificity of 95.2% in identifying interstitial lung findings. An additional independent validation study further confirmed strong performance. The company must also comply with ongoing regulatory requirements, including quality system regulations, labeling, and adverse event reporting.

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. What specific features or patterns does ScreenDx identify in CT images to produce its qualitative output, and how are these features categorized or weighed?; Given its role as an adjunctive information tool, how is ScreenDx's output integrated into existing clinical workflows, and what studies are planned or completed to assess its impact on patient outcomes or diagnostic efficiency?; The PCCP allows for algorithm modifications without further review; what are the specific thresholds or criteria for performance degradation that would trigger a mandatory re-review, even under the PCCP?; The device demonstrated 91.4% sensitivity and 95.2% specificity; how do these performance metrics compare to human expert readers or other existing automated ILD detection tools, particularly for early-stage disease?; Since ScreenDx does not 'rule out or rule in' lung fibrosis, what are the potential risks of misinterpretation or over-reliance on its qualitative output by clinicians, and how are these mitigated through labeling or training?