Search Results

The PhysCade System is intended for the analysis, display, and storage of cardiac electrophysiological data and maps for analysis by a physician.

The clinical significance of utilizing the PhysCade System to help identify areas with intra-cardiac atrial electrograms exhibiting local early activated potentials and other features of interest during atrial arrhythmias has not been established by clinical investigations.

The PhysCade™ System (PhysCade) is an artificial intelligence (AI) enabled device intended to assist clinicians in their management of patients with heart rhythm disorders (arrhythmias). PhysCade is a medical decision support system which post-processes electrograms (EGMs) collected inside the heart during electrophysiology (EP) mapping procedures using compatible diagnostic EP catheters. The PhysCade software has advanced algorithms that analyze the collected EGMs to provide information on regions of interest in the heart that may be useful to the clinician to support clinical decisions together with other available patient-related information.

PhysCade provides specialized analyses of data from a compatible multipolar catheter. The primary output (coPILOT) indicates the predominant earliest site of activation relative to the catheter electrode array. Supporting outputs include determining activation times of successive beats in the EGM signal on each electrode (coMAP), voltage at each electrode, signal quality, and sequences of propagation over multiple beats of the arrhythmia on the catheter.

The PhysCade System consists of a computer workstation, display, and custom software and is not connected to other devices or medical equipment.

The provided FDA 510(k) Clearance Letter for the PhysCade System gives some information about the device's development and testing, particularly regarding its AI algorithms. However, it does not explicitly detail the acceptance criteria or the specific results of a study proving the device meets those criteria, nor does it provide the requested levels of detail for the ground truth establishment, expert qualifications, or MRMC study results.

Based on the information provided, here's what can be extracted and inferred for the requested points. Where information is not present, it is explicitly stated.

Acceptance Criteria and Device Performance Study

The document states, "Design Validation confirmed that the AI/ML system is accurate for its intended use." This indicates that performance testing was conducted. However, the specific quantitative acceptance criteria (e.g., a specific F1 score, accuracy, sensitivity, or precision threshold) and the resulting performance metrics the device achieved are not explicitly stated in this document.

Table of Acceptance Criteria and Reported Device Performance:

Study Details:

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

   • Test Set Sample Size:
     • Number of electrograms: ~5 Million
     • Number of patients: 109
   • Data Provenance: The document does not explicitly state the country of origin.
   • Retrospective or Prospective: The document does not explicitly state whether the data was collected retrospectively or prospectively. It references "datasets with the following characteristics" suggesting pre-collected (likely retrospective) data.

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

   • Number of Experts: NOT PROVIDED.
   • Qualifications of Experts: NOT PROVIDED. The document only mentions that the device is for "analysis by a physician" and is "intended to be operated by nurses, technicians, physicians, or other personnel who are trained and approved by each treating facility." It does not specify who established the ground truth or their qualifications.

3. Adjudication method for the test set:

   • Adjudication Method: NOT PROVIDED. The document does not describe how disagreements, if any, among experts establishing ground truth were resolved.

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:

   • MRMC Study: The document does not mention a multi-reader multi-case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance. The performance testing described focuses on the device's accuracy ("Design Validation confirmed that the AI/ML system is accurate for its intended use"), rather than human performance improvement.
   • Effect Size: NOT APPLICABLE as no MRMC study is mentioned.

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

   • Standalone Performance: While not explicitly stated as "standalone performance," the phrase "Design Validation confirmed that the AI/ML system is accurate for its intended use" strongly implies a standalone performance evaluation of the algorithm's output against established ground truth. The device "provides information on regions of interest in the heart that may be useful to the clinician to support clinical decisions," suggesting the AI's output is evaluated directly.

6. The type of ground truth used:

   • Type of Ground Truth: The document states the device "analyzes the collected EGMs to provide information on regions of interest in the heart that may be useful to the clinician." It further notes the "primary output (coPILOT) indicates the predominant earliest site of activation relative to the catheter electrode array." The ground truth would therefore pertain to the identification of these "regions of interest" or the "earliest site of activation" based on expert interpretation of electrophysiological data. The exact method of establishing this ground truth (e.g., expert consensus on EGM analysis, correlation with other diagnostic modalities, or clinical outcomes) is NOT EXPLICITLY STATED.

7. The sample size for the training set:

   • Training Set Sample Size:
     • Number of electrograms: ~15 Million
     • Number of patients: 174

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

   • Ground Truth Establishment for Training Set: The document does not explicitly state how the ground truth for the training set was established. It only describes the characteristics of the training, tuning, and test cohorts. It is commonly understood that ground truth for training data is established by similar means to test data (e.g., expert annotation), but this is not detailed here.

Ask a specific question about this device