Search Results

The REMbrandt software is intended for Polysomnography studies and allows recording, displaying, printing and storage of physiological signals to assist in the diagnosis of various sleep disorders and sleep related respiratory disorders. The REMbrandt software allows: Automated analysis of physiological signals that is intended for use only in adults; An optional audio/visual alert for user defined threshold on calibrated DC input. These alerts are not intended for use as life support such as vital signs monitoring or continuous medical surveillance in intensive care units. Sleep report templates which summarize recorded and scored sleep data using simple measures including count, average, maximum and minimum values as well as data ranges for trended values; The REMbrandt software does not provide any diagnostic conclusion about the patient's condition and is intended to be used only by qualified and trained medical practitioners, in research and clinical environments.

The REMbrandt software consists of three applications, DataLab, Analysis Manager and REMbrandt Manager, which run on a desktop or laptop computer and require no specialized hardware. They are Windows based applications used by trained medical professionals to investigate sleep disorders. The REMbrandt software collects and digitizes the electrical voltages of patient physiological signals. After collecting and saving the signals, it provides tools and detectors to analyze the signals, which aid in the interpretation of a sleep study. The software consists of four main functional areas: A. Data Acquisition & Display (REMbrandt DataLab), B. Scoring/Review & Analysis (REMbrandt Analysis Manager), C. Report Generation (REMbrandt Analysis Manager), D. Archiving & Data Management (REMbrandt Manager). The REMbrandt software contains eight (8) computer-assisted scoring analyzers. All automatic detection tools are provided as time saving aids to assist trained medical practitioners in the review and analysis of vast amounts of data. Each computer-assisted scoring analyzer runs a specific type of event marking or numeric value processing in the study and each can be enabled individually as needed at the discretion of the user. The scoring rule parameters used in the computer-assisted scoring analyzers depend on available input signals in the study as well as user defined settings. All output from computer assisted scoring analyzers require medical professional review and acceptance.

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:

Acceptance Criteria and Device Performance Study for REMbrandt

The REMbrandt software includes eight computer-assisted scoring analyzers. The provided study specifically evaluates the performance of the Respiratory, Limb Movement, and Snore Event Assisted-scoring Detectors.

1. Table of Acceptance Criteria and Reported Device Performance

The study evaluates the "Positive Percent Agreement (PPA)" and "False Detection Rate per Hour (FD/h)" for the assisted-scoring detectors. While explicit acceptance criteria (e.g., "PPA must be >X%") are not directly stated as numerical thresholds in the provided text, the objective is to establish that the REMbrandt performance is "equivalent to the performance of the predicate device" and that its performance is "comparable to the manual markings of expert reviewers" and "clinically equivalent to the Reference standard." We can infer that the reported values demonstrate this equivalence or comparability.

Table: REMbrandt Event Detection Assisted Scoring Detectors Performance

The conclusion states that "Compared to the Reference standard, REMbrandt assisted-scoring detectors showed performance levels comparable to the manual markings of expert reviewers. The device performance is clinically equivalent to the Reference standard (majority rule) as constructed for this study, similar to results reported in the literature and to performance reported for other commercially available devices."

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

• Sample Size: 50 diagnostic PSG sleep studies (one study per subject) for each event type evaluated.
• Data Provenance: The text states, "All subjects involved in this study were adult (>18 years old) subjects with a clinical indication for a sleep study. The subject data were de-identified and applied as subject data to this study." The country of origin is not specified, but the submission is to the US FDA, implying a potentially diverse or US-centric origin, though not explicitly stated. The study is retrospective, as existing de-identified PSG studies were collected and then analyzed by experts and the device.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

• Number of Experts: Three (3) experienced and certified PSG specialists.
• Qualifications of Experts: This group included "one board certified sleep specialist." The other two are described as "experienced and certified PSG specialists."

4. Adjudication Method for the Test Set

The adjudication method used to establish the "Reference standard" (ground truth) was a majority rule: "at least two out of three expert scorings (medical professionals certified on PSG recording and analysis) agree on the presence of an event within an epoch."

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No, a multi-reader multi-case (MRMC) comparative effectiveness study was not conducted to assess how much human readers improve with AI vs. without AI assistance. This study focused on the standalone performance of the REMbrandt assisted-scoring detectors compared to a human-established ground truth. The software's role is described as "time saving aids to assist trained medical practitioners," implying a human-in-the-loop workflow, but a comparative effectiveness study with and without AI assistance for human readers was not part of this submission's performance testing.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance)

Yes, a form of standalone performance was implicitly done. The REMbrandt assisted-scoring detectors were run independently ("Separate from the expert review, all PSG studies were also analyzed by REMbrandt assisted-scoring detectors at default values"). The output of these detectors (algorithm only) was then compared against the "Reference standard" (expert consensus). The device is described as providing "computer-assisted event marking analyzers" and that "All output from computer assisted scoring analyzers require medical professional review and acceptance," indicating it's an aid, not a diagnostic tool unto itself. However, the performance metrics (PPA and FD/h) were derived from the algorithm's output before human review.

7. Type of Ground Truth Used

The type of ground truth used was expert consensus, specifically defined by a "majority rule" (at least two out of three expert scorings agree) on the presence of an event within an epoch. The criteria for event scoring (Apnea, Hypopnea, Limb Movement, Snore, Arousals) were clearly defined and applied by the experts.

8. Sample Size for the Training Set

The document does not specify the sample size for the training set. The clinical study summary focuses exclusively on the validation/test dataset.

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

The document does not provide information on how the ground truth for the training set (if any was used for algorithm development) was established. The "Clinical Study Summary" describes the ground truth process only for the test set.

Ask a specific question about this device

Ask a specific question about this device