Search Results

CT-MASS has been developed for the objective and reproducible analysis of multi-slice, multi-phase left and right ventricular function from cardiac CT data sets. The CT-MASS software package can be used to semi-automatically calculate and display various parameters such as: EDV, BSV, stroke volume, ejection factor, peak ejection and filling rates, myocardial mass, regional wall thickness, as well as well as wall thickening/thinning, and regional wall motion. This is including the axial to short axis reformat.

When interpreted by a trained physician these parameters may be useful in supporting the determination of a diagnosis.

CT-MASS is a professional state-of-the-art analytical software tool designed for UNIX, Linux as well as Windows platforms. CT-MASS facilitates the import and visualization of multi-slice, multi-phase CT data sets encompassing the cardiac chambers via CD-Rom and digital network. This CT-MASS functionality is independent of the CT equipment vendor. CT-MASS provides objective and reproducible global and regional two-, three- and four-dimensional clinically relevant parameters describing left and right ventricular heart function, such as ventricular volumes, regional wall thickness and wall thickening/thinning CT-MASS is intended to support all clinicians, i.e. cardiologists, radiologists, and referring physicians involved in the noninvasive assessment of heart function.

This 510(k) submission for CT-MASS does not contain a specific section detailing the acceptance criteria and a study proving the device meets those criteria with a table of performance metrics. The submission focuses on substantial equivalence to predicate devices and provides general information about verification and validation tests and evaluations by hospitals and literature, but it lacks the granular detail requested.

However, based on the provided text, we can infer some information due to the device's nature as an analytical software tool for cardiac function. The core of its safety and effectiveness relies on its ability to objectively and reproducibly analyze cardiac CT data.

Here’s an attempt to construct the response based on the available information and typical expectations for such a device, while clearly identifying what is stated and what is inferred or missing:

Acceptance Criteria and Device Performance Study for CT-MASS

This 510(k) submission (K033774) for CT-MASS does not explicitly define a table of acceptance criteria or present a detailed study with specific performance metrics against those criteria. The submission primarily focuses on demonstrating substantial equivalence to predicate devices (K013422 "CardIQ Function" and K020796 "CardIQ Analysis III") and asserts that "potential hazards are controlled by a risk management plan for the software development process... including hazard analysis, verification and validation tests." It also mentions "Evaluations by hospitals and literature" support the safety and effectiveness.

Given the intended use of "objective and reproducible analysis of multi-slice, multi-phase left and right ventricular function," the acceptance criteria would implicitly relate to the accuracy and reproducibility of the calculated cardiac parameters.

1. Table of Acceptance Criteria and Reported Device Performance

As specific acceptance criteria and detailed performance metrics are not provided in the submitted document, the table below represents inferred or expected criteria for this type of device, with no reported performance values available in this submission.

2. Sample Size Used for the Test Set and Data Provenance

The submission does not specify a sample size for a test set or provide details on data provenance (e.g., country of origin, retrospective/prospective). It generally refers to "verification and validation tests" and "Evaluations by hospitals."

3. Number of Experts Used to Establish Ground Truth and Qualifications

The submission does not specify the number of experts used to establish ground truth or their qualifications. It mentions the software is "intended to support all clinicians, i.e. cardiologists, radiologists, and referring physicians."

4. Adjudication Method for the Test Set

The submission does not describe any adjudication method (e.g., 2+1, 3+1, none) for a test set. It states that "The ventricular contours and regions-of-interest will be interpreted by the operator, who can choose to accept or reject the outlines," implying a human-in-the-loop scenario where the operator ultimately validates or corrects the automated outlines.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

A Multi-Reader Multi-Case (MRMC) comparative effectiveness study is not mentioned in this submission. The effectiveness is primarily asserted through substantial equivalence to predicate devices and general claims of supporting clinical diagnosis and decision-making. No effect size for human reader improvement with AI assistance is provided.

6. If a Standalone Performance Study Was Done

A standalone performance study (algorithm only without human-in-the-loop performance) is not explicitly detailed in this submission. The description emphasizes the software's ability to "semi-automatically calculate" parameters and that the operator "can choose to accept or reject the outlines," indicating a human-in-the-loop design. However, the "verification and validation tests" could conceptually include aspects of standalone algorithm testing.

7. The Type of Ground Truth Used

The type of ground truth used is not specified. For cardiac function analysis, ground truth typically involves:

• Expert Consensus: Manual expert tracings/measurements from cardiac imaging (e.g., CT, MRI, echocardiography) often used as a reference.
• Pathology: Less common for structural/functional parameters, but might be relevant for mass if biopsy correlations were involved.
• Outcomes Data: Unlikely to be the direct ground truth for individual parameter accuracy, but could be used in larger clinical utility studies.
  Given the device's function, expert consensus on cardiac contours and measurements from high-quality imaging would be the most probable, albeit unstated, ground truth.

8. The Sample Size for the Training Set

The submission does not specify the sample size for a training set. As a 510(k) from 2004, the concept of "training set" for AI/ML validation was less formalized in regulatory submissions compared to today. The device is described as an "analytical software tool" and not explicitly termed an AI/ML product, although algorithms are inherently involved.

9. How the Ground Truth for the Training Set Was Established

The submission does not describe how the ground truth for a training set (if one was used) was established.

Ask a specific question about this device