(84 days)
REX™ 1.0 is a software package intended for viewing and manipulating DICOM-compliant medical images acquired from CT scanners. REX™ 1.0 can be used for real-time image viewing, image manipulation, 3D volume rendering, virtual endoscopy, and issuance of reports.
REX™ 1.0 is a tool for 3D (three dimensional) and 2D (two dimensional) viewing and manipulation of DICOM compliant CT images. The proposed software provides real-time image viewing, image manipulation, 3D volume rendering, virtual endoscopy, and issuance of reports.
Here's an analysis of the provided text regarding the acceptance criteria and supporting study for the REX™ 1.0 device:
Important Note: The provided document is a 510(k) Premarket Notification Summary from 2002. At that time, the regulatory requirements and expectations for demonstrating substantial equivalence, particularly for software devices, were different from current standards, especially for AI/ML-driven devices. This document focuses on demonstrating equivalence to a predicate device rather than presenting extensive performance studies against defined acceptance criteria in the way a modern AI/ML device would.
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly present a table of quantitative acceptance criteria and corresponding reported device performance metrics in the way modern AI/ML device submissions typically do (e.g., sensitivity, specificity, AUC, etc.). Instead, the performance is framed in terms of substantial equivalence to a predicate device and adherence to functional requirements and standards.
| Acceptance Criteria (Inferred from Substantial Equivalence and Validation) | Reported Device Performance (REX™ 1.0) |
|---|---|
| Functional Equivalence to Predicate Device (Rapidia® V 2.0): | |
| - DICOM 3.0 Compliance | Conforms to DICOM Version 3.0 (Explicitly stated). |
| - Real-time image viewing | Yes (Matches predicate) |
| - Image manipulation | Yes (Matches predicate) |
| - 3D volume rendering | Yes (Matches predicate) |
| - Virtual endoscopy | Yes (Matches predicate) - specifically: instant access to lesions by single click, real-time display of endoscopic view, internal and external viewing of any hollow structures. |
| - Issuance of reports | Yes (Matches predicate) |
| - Operating System compatibility (Windows 2000) | Yes (Matches predicate) |
| - Patient demographics display | Yes (Matches predicate) |
| - TCP/IP networking | Yes (Matches predicate) |
| - PNG (Lossless) Image Compression | Yes (Feature present, predicate unspecified, but deemed acceptable for equivalence due to REX™ being a subset of features and no new safety risks) |
| - Annotations - marker | Yes (Matches predicate) |
| - Image Review (Still, Window, Level, Zoom, Pan, Flip) | Yes (Matches predicate) |
| - 2D Measurements (Length, Area) | Yes (Matches predicate) |
| - Image Source (CT only for REX™) | Yes (Supports CT, unlike predicate which also supports MR; REX™ is a subset of features and thus acceptable) |
| - Image Input (DICOM 3.0) | Yes (Matches predicate) |
| - Image Output (PNG lossless snapshots) | PNG (lossless snapshots) (Predicate outputs JPEG, BMP, DICOM; considered acceptable due to REX™ being a subset of features and no new safety risks) |
| - Use Standard Monitor | Yes (Matches predicate) |
| - Patient and Study Browser | Yes (Matches predicate) |
| - Measure CT Numbers (ROI) | Yes (Matches predicate) |
| - Standalone Software Type | Yes (Matches predicate) |
| - Local Image Storage | Yes (Matches predicate) |
| - True Color | Yes (Matches predicate) |
| - User Login | Yes (Predicate unspecified, but this is a security/access feature and does not introduce new safety risks that would preclude equivalence) |
| - Preset Window and Level | Yes (Matches predicate) |
| - Image Conversion (for viewing in browser) | Yes (Matches predicate) |
| - Trained Physicians as Users | Yes (Matches predicate) |
| - Volume Rendering algorithms | Yes (Matches predicate) |
| - Reporting algorithms | Yes (Matches predicate) |
| Adherence to Internal Requirements and Regulatory Practices: | |
| - Performance to functional requirements specified in SRS | Validation testing confirms REX™ 1.0 performs all input, output functions, and required actions according to the Software Requirements Specification. |
| - Adherence to Software Development Practices | Followed specified Software Development Practices and Validation and Verification Process. Procedures specify individuals responsible for developing/approving specs, coding/testing, validation, and maintenance. |
| - Hazard Control | Potential hazards identified in Hazard Analysis and controlled by designing controls, introducing protective measures, and warning users. |
| - No new potential safety risks | Concluded that REX™ 1.0 "does not result in any new potential safety risks and performs in accordance with its intended use as well as the Rapidia® V 2.0 device currently on the market." |
| - Substantial Equivalence to Predicate | PointDx considers features of REX™ 1.0 to be substantially equivalent to the subset of features in common with Rapidia® V 2.0. The FDA ultimately agreed and determined the device to be substantially equivalent. |
The "study" proving the device meets these criteria is the non-clinical performance data and substantial equivalence comparison presented in the 510(k) summary.
Detailed Breakdown of the Study/Evidence:
2. Sample Size Used for the Test Set and Data Provenance:
- Test Set Sample Size: Not explicitly stated as a separate "test set" with image counts. The validation testing mentioned refers to internal software validation. The comparison for substantial equivalence does not involve a specific clinical image dataset for performance metrics, but rather a feature-by-feature comparison against the predicate device.
- Data Provenance: Not applicable in the context of a clinical test set. The validation would have been performed on internally generated or standard DICOM test files to verify functionality.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications:
- Number of Experts/Qualifications: Not applicable. This submission predates the common requirement for human-in-the-loop or standalone clinical performance studies with expert ground truth for imaging software that provides basic viewing and manipulation functionalities. Substantial equivalence was primarily based on functional comparison.
4. Adjudication Method for the Test Set:
- Adjudication Method: Not applicable. There was no "test set" requiring adjudication in the way modern AI performance studies do. The equivalence was determined by comparing features and functionality to the predicate device.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done, and effect size:
- MRMC Study: No, an MRMC comparative effectiveness study was not done. The device is a "PACS / Image Processing Software" that provides viewing and manipulation tools, not an AI diagnostic aid requiring assessment of human reader improvement.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done:
- Standalone Performance: Not applicable in the context of a diagnostic algorithm. The device itself is a standalone software, but its "performance" is about its functionality (e.g., can it render 3D, can it measure length) rather than a diagnostic output that would have standalone metrics like sensitivity/specificity. The validation testing ensured the software functions as designed.
7. The Type of Ground Truth Used:
- Type of Ground Truth: For the functional validation, the "ground truth" would be the software requirements specification (SRS) and DICOM 3.0 standard compliance. For the substantial equivalence, the "ground truth" was the features and specifications of the predicate device (Rapidia® V 2.0). There was no clinical ground truth (like pathology or outcomes data) used as this device is a viewing/manipulation tool, not a diagnostic algorithm.
8. The Sample Size for the Training Set:
- Training Set Sample Size: Not applicable. REX™ 1.0 is described as PACS/Image Processing Software. It's a deterministic software program for viewing and manipulation, not an AI/ML algorithm that requires a "training set" in the modern sense.
9. How the Ground Truth for the Training Set was Established:
- Ground Truth for Training Set: Not applicable, as there is no "training set" for this type of software.
§ 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).