Search Results

MDxNet is intended for use under physician's orders and/or oversight and management.

MDxNet is indicated for acquisition of related sets of 2D ultrasound images and 3D reconstruction of ultrasound images in the adult and pediatric patient utilizing existing ultrasound systems. It is intended to acquire, analyze, store and retrieve digital ultrasound images for computerized 2D and 3D image processing. MDxNet is indicated to acquire, digitize, archive and retrieve single or sequences of 2D ultrasound images.

MDxNet is intended as a general purpose digital image processing and archiving tool for use in abdominal, pelvic, fetal, cardiac, peripheral vascular and neurovascular imaging.

MDxNet add-on indications are for the assessment of pulmonary function to acquire, analyze, store and retrieve this assessment in male/female adult and pediatric patients utilizing FDA approved Spirometry. MDxNet additional add-on will provide 12-lead resting electrocardiogram (ECG) which permits the detection of abnormalities in the transmission of the cardiac impulse through the heart muscle and serves as an important aid in the diagnosis of heart ailments through utilizing 12-lead ECG.

MDxNet consists of MDxPAC, MDxStation (PC or Notebook, laptop computer), MDxServer, and the proprietary software component called MDxView. MDxPAC consists of limited hardware, that bundle and power FDA approved Commercial Off-The-Shelf devices (COTS). MDxNet features an integrated 2D/3D streaming engine which permits PC's or notebooks to control the server and to review 2D and 3D reconstructions. MDxNet is capable of image review, archiving, data collection, database maintenance, reporting and basic 3D capabilities including color rendering. Upon receipt of FDA approval to market MDxNet, finished devices will be distributed only under direction and/or orders of physicians.

MDxNct inherits and integrates concepts first pioneered in the National Aeronautics and Space Administration (NASA) sponsored Telemedicine Instrument Pack (TIP) project and the Defense Advanced Research Projects Agency (DARPA) sponsored Medical Ultrasound Three-dimensional, Portable with Advanced Communication (MUSTPAC) project, and its predecessors. It expands capability to include a server component, augments data acquisition to include wireless options, redesigns the Graphical User Interface (GUI) for more intuitive interaction and provides improved packaging for better usability.

MDxNet has the capability to record ultrasound transducer spatial position in six degrees of freedom during use. Coordinate tracking is achieved with a miniature magnetic field sensor within a transmitted pulsed magnetic field. This is done by attaching a plastic holding plate to the FDA approved probe of the host Ultrasound system, to which the receiver of an electromagnetic sensor device is attached.

2D ultrasound images are acquired sequentially in a series of steps as the ultrasound transducer is moved across the patient scan site. The resulting set of digitized 2D images is then converted into a 3D data volume.

IQmark Digital ECG and Spirometer both FDA approved devices have been integrated into MDxNet as requested by the end user of the validation study. The Midmark software for each of these devices and indications of use have not been altered, though permit the end user as demonstrated in both clinical studies to capture additional potentially key clinical results within MDxNet for data collection, review, comparison, reporting and archiving.

Here's a breakdown of the acceptance criteria and study information for the CyberMDx MDxNet device, based on the provided FDA 510(k) Premarket Notification:

The provided document K060035 for CyberMDx, Inc. MDxNet is a 510(k) Premarket Notification. It explicitly states that MDxNet is an "Image Processing Server" and a "System, Image Processing, Radiological." This type of device primarily functions as a tool for displaying, processing, archiving, and retrieving medical images. It's important to note that this document does not describe a device that performs AI-driven interpretation or diagnosis, nor does it present "acceptance criteria" in the sense of performance metrics (like sensitivity, specificity, or AUC) that an AI algorithm would need to meet.

Instead, the submission focuses on demonstrating substantial equivalence to existing legally marketed predicate devices by comparing features and intended uses. The "studies" mentioned are primarily about verifying the device's functionality and safety as an image processing system rather than validating the diagnostic accuracy of an AI component.

Therefore, the requested information, particularly regarding AI-specific performance metrics, ground truth, expert adjudication, sample sizes for AI training/test sets, and MRMC studies, is largely not applicable (N/A) in the context of this 2006 510(k) for an image processing server. The device predates the widespread adoption and regulatory requirements for AI in medical imaging.

However, I will extract and organize the relevant information that is present in the document.

Acceptance Criteria and Reported Device Performance

Given that this is an image processing server, the "acceptance criteria" discussed are primarily related to its functional capabilities and conformity to design intent and quality standards, rather than diagnostic accuracy metrics. The document emphasizes functional equivalence to predicate devices.

Table 5-1 explicitly shows "Yes" for MDxNet across all listed common features (2D Image Review, Multiplanar Reformatting, 3D Volume Rendering, Maximum Intensity Projection, Image Archiving, Image Filming, Image Transfer/Network Connectivity, Examination of 2D image as 3D volume, Comparison of Multiple Scans, General medical imaging indications, Computer Platform, Image Display, Operating System). |
| Accuracy and Performance to Specification (general system performance) | "bench testing was conducted to establish MDxNet's accuracy, performance to specification, as well as testing to accepted industry standards..." (Section 5.3)

"Test results to date included elsewhere support the conclusion that the actual device, software and hardware satisfies the design intent. Actual device performance as tested internally and by third parties (including controlled clinical trials) conforms to the system performance standards and has met expectations." (Section 5.6) |
| Adherence to Quality System Regulations (e.g., FDA 21 CFR 820) | "Continual nonclinical testing is being conducted through internal company procedures according the FDA 21CFR820 Quality System and Current Good Manufacturing Practices." (Section 5.6) |
| Hardware Testing | "All hardware testing will be conducted and meet the specified acceptance criteria before the device is marketed." (Section 4, Hardware & Software Information) |
| Software Level of Concern | "The level of concern relative to the software has been determined as minor using the decision tree provided in the revised FDA Software Guidance, May 11, 2005." (Section 4, Hardware & Software Information) |

Study Details (Focusing on Functional & Safety Verification)

As noted, this is not an AI performance study. The "clinical studies" mentioned are for system validation, not diagnostic accuracy of an AI algorithm.

1. Sample size used for the test set and the data provenance:

   • Test Set Sample Size: The document states that "Two MDxNet Systems were developed to acquire potential end user input and to allow validation of the software, hardware, server and integration under Investigational Review Board (IRB) approval in a medical clinic setting." It also mentions "controlled clinical trials." However, no specific sample size (number of patients or images) for any particular "test set" is provided.
   • Data Provenance: "Medical clinic setting" and "third parties (including controlled clinical trials)." No specific countries are mentioned, but the company is US-based (Providence, RI). The studies appear to be prospective as they were conducted for validation.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • N/A. The document does not describe a clinical diagnostic "ground truth" establishment process for imaging analysis that would require experts. The device is a tool, not an AI for diagnosis. The validation focuses on whether the system functions as intended and safely integrates other FDA-approved devices (Spirometry, ECG).

3. Adjudication method for the test set:

   • N/A. No expert adjudication is described as the device is not performing diagnostic interpretations that would require it.

4. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:

   • No. This is not an AI-assisted device. Therefore, no MRMC study comparing human readers with and without AI assistance was conducted or is applicable.

5. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:

   • N/A. The MDxNet is a system that allows users to process and view images. It is not a standalone diagnostic algorithm.

6. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

   • N/A. For the image processing and display features, "ground truth" would relate to the accuracy of image reconstruction, archiving, and retrieval, which are engineering and system performance validations rather than clinical diagnostic ground truth. For the integrated ECG and Spirometry, the "ground truth" would be the direct output of those separate FDA-approved devices. The main focus of the study was about the integration and display of these clinical results, not their diagnostic determination by MDxNet.

7. The sample size for the training set:

   • N/A. The document does not mention any "training set" in the context of machine learning or AI.

8. How the ground truth for the training set was established:

   • N/A. No training set for AI is mentioned.

Summary of Device Context:

The CyberMDx MDxNet, as described in this 2006 510(k) submission, is a general-purpose image processing server. Its primary function is to:

• Acquire, digitize, archive, and retrieve 2D ultrasound images.
• Perform 3D reconstruction of ultrasound images.
• Provide tools for image review, multiplanar reformatting, volume rendering, and maximum intensity projection.
• Integrate data from external FDA-approved devices (like spirometers and 12-lead ECGs) for capture, review, comparison, reporting, and archiving.

The submission demonstrates substantial equivalence to predicate devices by virtue of its functional features as an imaging workstation/server, not by demonstrating AI-driven diagnostic performance. The "studies" mentioned are for systematic validation of the device's ability to perform these functions safely and effectively, and to integrate data from other medical devices, within a clinical setting under IRB approval.

Ask a specific question about this device