Search Results

The Model 9100 PFT/DICO is a pulmonary function testing device which uses Morgan Scientific's ComPAS2 software to measure subject respiratory parameters including FVC, SVC, MVV, CPF, RMS, SNIP, DLCO, MBN2 and SBN2.

The device is PC-based and designed for lung function testing on adults and pediatrics, 6 years and older, in a variety of professional healthcare environments e.g., primary care, hospitals, pharmaceutical research centers and physicians' offices.

The Model 9100 PFT/DICO is intended for the assessment of respiratory function through the measurement of dynamic lung volumes i.e., spirometry and other lung functions i.e., diffusing capacity.

The Model 9100 PFT/DICO is composed of various sensors and valves with associated low level firmware. The firmware interfaces with the Morgan Scientific's ComPAS2 software (K213872) that resides on an on-board computer. The Model 9100 also provides for user input and present resulting data on an integral display.

The ComPAS2 software controls valves and reads unprocessed data from the sensors in the Model 9100then determines respiratory parameters including FVC, SVC, MVV, CPF, RMS (MIP and MEP), SNIP, DLCO, MBN2 and SBN2. The Model 9100 PFT/DICO firmware does not determine any respiratory parameters.

The ComPAS2 software uses flow and volume from the Vitalograph pneumotachograph spirometer to display the flow and volume information measured directly from patient effort. ComPAS2 also utilizes gas analyzer readings from the Model 9100 patient test benchmark to display dilution lung volume data and single / multi breath diffusion data measured directly from patient effort. This information is then provided in a report format.

The provided text describes the regulatory clearance of the Vitalograph Model 9100 PFT/DICO, a pulmonary function testing device, and its substantial equivalence to a predicate device. However, it does not contain information about a study proving the device meets acceptance criteria related to a machine learning or AI model's performance.

The document outlines performance testing conducted for the device's electrical, mechanical, and biocompatibility aspects, as well as software verification and validation. It explicitly states that the device uses "Morgan Scientific's ComPAS2 software to measure subject respiratory parameters," but there's no indication that this software includes an AI or machine learning component that would require a study with human-in-the-loop performance, expert ground truthing, or MRMC studies typically associated with AI/ML medical devices.

Therefore, many of the requested details about acceptance criteria for an AI model's performance and associated study specifics (sample size for test/training, number of experts, adjudication, MRMC, standalone performance, ground truth type) cannot be extracted from this document.

Instead, the document focuses on demonstrating substantial equivalence to a predicate device based on similar indications for use, technological characteristics, and principles of operation, supported by standard bench testing and software validation.

Here's an attempt to answer the request based only on the provided text, highlighting the absence of AI/ML-specific details:

Device: Vitalograph Model 9100 PFT/DICO

Study Type: This document describes a 510(k) premarket notification for substantial equivalence, supported by bench testing, software verification/validation, and compliance with various standards. It is not an AI/ML performance study. The "study that proves the device meets the acceptance criteria" refers to the entire body of evidence submitted for 510(k) clearance, rather than a specific AI model's performance study.

1. A table of acceptance criteria and the reported device performance

The document defines performance specifications and states that testing supported the safety and performance, implying these specifications were met. The specific "acceptance criteria" for the overall device's performance are embedded in the compliance with standards and the "similar" comparisons to predicate/reference devices.

Note on "Acceptance Criteria" for AI: The document does not describe acceptance criteria for an AI or machine learning model. The stated accuracies (e.g., flow, volume, gas sensors) are for the physical measurement components of the device, not a predictive algorithm based on complex data interpretation.

2. Sample size used for the test set and the data provenance

• Sample Size for Test Set: Not specified for any performance testing, other than the implication that tests were sufficient to meet specific standards (e.g., ATS/ERS waveforms, drop tests, cleaning validations). There is no test set in the context of an AI/ML model's performance.
• Data Provenance: Not applicable in the context of typical AI/ML data provenance (e.g., country of origin, retrospective/prospective clinical data). The performance tests are largely bench-based or simulated.

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

• Not applicable. There is no mention of human experts establishing ground truth for a test set, as would be done for an AI/ML interpretation task. Ground truth for the device's measurements would be established by reference standards or highly accurate laboratory equipment.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

• Not applicable. No expert adjudication process is described.

5. 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, an MRMC comparative effectiveness study was not done. The device measures physiological parameters; it does not "assist" human readers in interpreting complex medical images or data.

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

• Not applicable in the context of an AI/ML algorithm. The device itself is the "standalone" entity that performs measurements. The software (ComPAS2) controls the device and processes the raw sensor data, but there's no indication of it being a standalone AI/ML interpreter.

7. The type of ground truth used (expert concensus, pathology, outcomes data, etc)

• The ground truth for the device's performance relies on calibration standards, reference instruments, and established engineering/medical device testing protocols (e.g., ATS/ERS guidelines for spirometry, ISO standards for gas analysis accuracy, and various electrical/mechanical standards). There is no "expert consensus," "pathology," or "outcomes data" used for performance validation in the AI/ML sense.

8. The sample size for the training set

• Not applicable. There is no mention of an AI/ML model that would require a training set. The ComPAS2 software and device firmware are likely developed using traditional software engineering and embedded system development methods, not machine learning model training.

9. How the ground truth for the training set was established

• Not applicable, as there is no AI/ML training set.

Ask a specific question about this device