Search Results

HERMES Medical Imaging suite that provides software applications used to process, display, analyze and manage nuclear medicine and other medical imaging data transferred from other workstation or acquisition stations.

The base product design of Hermes Medical Imaging Suite v5.7 is the same as for the Hermes Medical Imaging Suite v5.6 (K153056). The following modules have been added in this applicatio to Hybrid Viewer NM Processing: Colonic Transit, Remnant Liver, Parathyroid, Dosimetry, Classic DMSA, Oesophageal Transit/Reflux, HIDA, Salivary Gland, Bone3Phase Analysis and Uniformity.

Here's an analysis of the provided text regarding the acceptance criteria and study details for the Hermes Medical Imaging Suite v5.7:

Understanding the Device:

The device, HERMES Medical Imaging Suite v5.7, is a software application suite for processing, displaying, analyzing, and managing nuclear medicine and other medical imaging data. It has added several new modules compared to its predecessor (v5.6).

Challenges in Extracting Information:

The provided text is a 510(k) summary, which is a regulatory document focused on demonstrating substantial equivalence to a predicate device. It doesn't present a traditional "acceptance criteria" table or a single, comprehensive study report with detailed methodologies. Instead, it describes a series of comparative tests against predicate devices and manual calculations. Therefore, the "acceptance criteria" are implied by the comparisons and the statement that "the testing results supports that all the software specifications have met the acceptance criteria." The reported performance is the outcome of these comparisons.

1. Table of Acceptance Criteria and Reported Device Performance

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

• Hybrid Recon (Myocardial SPECT):

  • Sample Size: 10 subjects.
  • Data Provenance: Not explicitly stated, but implicitly retrospective as they were "compared" to a previous version and involved "stress&rest myocardial SPECT scan with Tc99m tetrofosmin." No country of origin is mentioned.

• Hybrid Recon (Phantom Studies):

  • Sample Size: Jaszczak phantom and IEC Phantom (physical phantoms, not patient data).
  • Data Provenance: Not applicable in the same way as patient data. These are controlled experimental setups.

• Classic DMSA, Colonic Transit, Oesophageal Transit/Reflux, 3Phase Bone Analysis, HIDA, Salivary Gland Analysis:

  • Sample Size: Not explicitly stated for these modules. The phrasing "compared to manual calculations" or "compared to Xeleris 3.1" does not provide the number of cases used in these comparisons.
  • Data Provenance: Not explicitly stated. Assumed to be retrospective if patient data was used, or simulated/phantom data for manual calculations.

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

• Number of Experts: Not explicitly stated.
• Qualifications of Experts: Not explicitly stated.
  • For the modules compared against "manual calculations (from first principles)," the "expert" is implied to be a human capable of performing these calculations, but specific qualifications are not provided.
  • For comparisons against predicate devices, the predicate devices themselves are the "truth" or reference, not human experts establishing ground truth for the test set.

4. Adjudication Method for the Test Set

• The document describes direct comparisons of results generated by the device (v5.7) against predicate devices (v5.6, v5.3, GE Xeleris 3.1) or against "manual calculations."
• There is no mention of an adjudication method (like 2+1 or 3+1 consensus) involving multiple human experts for determining a ground truth label for the test cases used. The reference is either an older software version or predefined manual calculation methods.

5. If a Multi Reader Multi Case (MRMC) Comparative Effectiveness Study was done

• No, an MRMC comparative effectiveness study was not done. The studies described are comparisons of one software version against another, or against manual calculations, focusing on the output consistency and accuracy of the software itself, not on how human readers perform with or without AI assistance.

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

• Yes, the described studies are standalone algorithm-only performance tests. The comparisons are between algorithm versions or between an algorithm and manually calculated results. There is no mention of a human interacting with the AI outputs as part of the performance evaluation.

7. The Type of Ground Truth Used

The ground truth used varies by the type of comparison:

• Hybrid Recon (Myocardial SPECT): The outputs from the predicate device (Hermes Medical Imaging Suite v5.6) were used as the reference for comparison.
• Hybrid Recon (Phantom Studies): The known true activity concentrations within the Jaszczak and IEC phantoms were used as the ground truth.
• Classic DMSA: The outputs from the predicate device (Classic DMSA application in Medical Imaging Suite v5.3) were used as the reference.
• Colonic Transit: The outputs from the predicate device (GE Xeleris 3.1 review and workstation) were used as the reference.
• Oesophageal Transit/Reflux, 3Phase Bone Analysis, HIDA, Salivary Gland Analysis: Manual calculations (from first principles) were used as the ground truth. This implies an established, understood method of calculation is the reference.

8. The Sample Size for the Training Set

• Not provided. The document focuses entirely on verification and validation testing, and the 510(k) summary typically does not include details about the training set for the software, especially for traditional medical imaging software (not AI/ML in the modern sense) where "training" might refer to software development and bug fixing rather than data-driven model training.

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

• Not applicable / Not provided. As no information on a training set (in the context of machine learning) is given, how its ground truth was established is also not mentioned. The software modules appear to be based on established algorithms and calculations for nuclear medicine image processing, rather than being "trained" on a dataset with external ground truth labels.

Ask a specific question about this device