(225 days)
SoftView is intended to generate an enhanced, secondary digital radiographic image of the chest. The enhanced AP or PA image of the chest provides improved visibility of the lung parenchyma through bone suppression and tissue equalization, and may facilitate discerning the presence or absence of nodules. The SoftView image provides adjunctive information and is not a substitute for the original PA/AP image. This device is intended to be used by trained professionals, such as physicians, radiologists, and technicians on patients with risk of having lung nodules and is not intended to be used on pediatric patients.
SoftView is a dedicated post-processing application which suppresses bone structures from digital radiographic image of the chest.
Here's a breakdown of the acceptance criteria and study information for the SoftView™ device, extracted from the provided text:
SoftView™ Acceptance Criteria and Study Information
The acceptance criteria for SoftView™ were demonstrated through a reader study and a comparative analysis of the contrast-to-noise ratio (CNR) against a predicate device.
1. Table of Acceptance Criteria and Reported Device Performance
| Acceptance Criteria Category | Specific Metric | Acceptance Criteria Description | Reported Device Performance |
|---|---|---|---|
| Reader Study | Area Under the Localization Receiver-Operating Characteristic (LROC) Curve | A statistically significant improvement in the LROC curve area when using SoftView™ assistance for detecting actionable lung nodules, compared to without DeepView™ assistance. | The mean difference in the area under the LROC curve was -0.098 (95% CI: -0.116 to -0.080), indicating a statistically significant improvement with SoftView™. |
| Sensitivity for Actionable Lung Nodules | No explicit acceptance criterion for a minimum sensitivity value is stated, but the study aimed to demonstrate improvement with SoftView™. | Sensitivity was 49.5% (95% CI: 45.9-53.0) without SoftView and 66.3% (95% CI: 63.1-69.7) with the addition of the SoftView image. | |
| Specificity | No explicit acceptance criterion for a minimum specificity value is stated, but the study aimed to demonstrate performance with SoftView™. | Specificity was 96.1% (95% CI: 95.0-97.1) with the standard image and 91.8% (95% CI: 89.5-93.5) with the SoftView image. | |
| Comparative Analysis | Contrast-to-Noise Ratio (CNR) of Residual Bone (Rib and Clavicle Regions) | SoftView™ must be substantially equivalent to the predicate device's hardware/software process (GE Medical Systems Dual Energy and Tissue Equalization Software Options) in terms of CNR of residual bone in soft tissue images. Equivalency and non-inferiority tests were performed, with an "ideal" CNR of residual ribs being 0.0. The analysis stratified data by modality (CR and DR) and lung regions (pleural, mid-lung, and hilum). | For indicated strata, SoftView™ was demonstrated to be equivalent to DES soft tissue images. The exception was the middle area of the lung for the DR modality (GE DES), where SoftView™ was found to be better (closer to an "ideal" CNR of 0.0). Furthermore, SoftView™ was non-inferior to DR and CR dual energy devices' soft tissue images across all strata. Descriptive statistics of means and standard deviations of CNRs across different modalities and regions of the lungs were in agreement. |
2. Sample Size Used for the Test Set and Data Provenance
The document does not explicitly state the exact sample size for the test set used in the reader study or the comparative analysis. It mentions "an independent dataset" for the comparative analysis of CNR.
The data provenance is not specified (e.g., country of origin, retrospective or prospective).
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications
This information is not provided in the given text. The document refers to "detecting actionable lung nodules" in the reader study, implying a clinical assessment, but does not detail how the ground truth for these nodules was established or by whom.
4. Adjudication Method for the Test Set
The adjudication method for establishing ground truth for the test set is not provided in the document.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and Effect Size
Yes, a multi-reader study was conducted. The study assessed the benefit of SoftView™ to radiologists for detecting actionable lung nodules.
Effect Size:
The mean difference in the area under the LROC curve was -0.098 (95% CI: -0.116 to -0.080), which is described as a statistically significant improvement for human readers with the aid of SoftView™. The sensitivity for detecting nodules increased from 49.5% without SoftView to 66.3% with SoftView.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
A standalone performance study of the algorithm without human intervention was not explicitly described in terms of diagnostic effectiveness for nodules. The "Comparative Analysis" focused on the algorithm's output (soft tissue image) against a predicate device's output based on CNR, which is a technical image quality metric rather than direct standalone diagnostic performance. The primary effectiveness study was a reader study with human readers.
7. The Type of Ground Truth Used
The type of ground truth used for the reader study is not explicitly stated. It refers to "actionable lung nodules," implying clinical diagnosis or follow-up, but the method (e.g., expert consensus, pathology, outcome data) is not detailed.
For the comparative analysis, the ground truth was the soft tissue images produced by the predicate DES devices (GE and Fuji), used to compare the CNR of residual bone.
8. The Sample Size for the Training Set
The document does not provide information regarding the sample size used for the training set. It mentions that the model is "built from DES data by using simple image features extracted from the standard PA, along with target values derived from a DES soft tissue image," but the size of this training data is not specified.
9. How the Ground Truth for the Training Set Was Established
The ground truth for the training set (from which the SoftView™ model was built) was established by using target values derived from a DES soft tissue image. This means that commercially available Dual Energy Subtraction (DES) systems were used as the reference standard to create the "ideal" bone-suppressed images that the SoftView™ algorithm was trained to replicate. The exact process of "deriving target values" is not detailed further.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).