Search Results

The Günther Tulip Filter implant is intended for the prevent pulmonary embolism (PE) via placement in the vena cava in the following situations:

• · Pulmonary thromboembolism when anticoagulant therapy is contraindicated;
• · Failure of anticoagulant therapy in thromboembolic diseases;
• · Emergency treatment following massive PE where anticipated benefits of conventional therapy are reduced; and
• · Chronic, recurrent PE where anticoagulant therapy has failed or is contraindicated.

The Günther Tulip Filter implant may be retrieved if clinically indicated; please refer to the "Optional Filter Retrieval" section of the Instructions for Use for more information.

The product is intended for percutaneous placement via a femoral or jugular vein for filtration of inferior vena cava (IVC) blood to prevent PE.

The Günther Tulip Filter Set consists of a filter composed of a paramagnetic cobalt chromium alloy (50 mm long when compressed to a diameter of 30 mm), preloaded on a femoral filter introducer; a 7 French coaxial introducer system (compatible with a .035 inch wire guide); and a 10 French predilator with hydrophilic coating for vessel access. The introducer dilator has eight sideports and two radiopaque markers 30 mm apart (end-to-end). The Günther Tulip Filter implant is designed to act as a permanent filter or retrievable filter.

The provided text is related to a 510(k) premarket notification for a medical device called the "Günther Tulip® Vena Cava Filter Set." This document focuses on demonstrating that the updated device is substantially equivalent to a previously cleared predicate device. It explicitly states that no performance testing was warranted for this submission because no changes were made to the device's design, manufacturing, sterilization, or principles of operation.

Therefore, the document does not contain the information requested regarding acceptance criteria, device performance results, sample sizes for test sets, expert involvement, adjudication methods, MRMC studies, standalone performance, or ground truth establishment for a new device's performance evaluation. The submission relies on the performance data from the predicate device (K172557).

To answer your request, if this were a submission requiring new performance testing for an AI/ML medical device, the following information would be necessary:

Hypothetical Study Information (Illustrative, NOT based on the provided document)

Since the provided document states "No performance testing was warranted," I cannot extract the requested information. However, if this were an AI/ML device requiring performance evaluation, the following is an example of what would be expected for a submission.

Acceptance Criteria and Device Performance (Hypothetical Example for an AI/ML Device)

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Size and Data Provenance

• Test Set Sample Size: 500 cases (e.g., medical images or patient records).
• Data Provenance: Retrospective, multi-center data collected from hospitals in the United States (40%), Europe (30%), and Asia (30%). Care was taken to include diverse patient demographics and disease prevalence.

3. Number and Qualifications of Experts for Ground Truth

• Number of Experts: 3 independent expert readers.
• Qualifications: All experts were board-certified radiologists with at least 10 years of experience in the relevant modality (e.g., diagnostic imaging for venous thromboembolism) and had specific expertise in the condition being assessed.

4. Adjudication Method for the Test Set

• Adjudication Method: 2+1 Adjudication. Initial independent reads by two experts. If their assessments disagreed, a third, senior expert independently reviewed the case, and their decision served as the tie-breaker and final ground truth.

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

• MRMC Study Conducted: Yes.
• Effect Size of Human Reader Improvement:
  • Average Sensitivity Improvement with AI Assistance: +8.5% (from 80% without AI to 88.5% with AI).
  • Average Specificity Improvement with AI Assistance: +4.2% (from 85% without AI to 89.2% with AI).
  • Average Reading Time Reduction with AI Assistance: 15% faster per case.
  • Overall Diagnostic Accuracy Improvement (AUROC): Relative improvement of 5% in reader AUROC on challenging cases.

6. Standalone Performance (Algorithm Only)

• Standalone Performance Done: Yes. The "Reported Device Performance" in the table above reflects the standalone (algorithm-only) performance results on the independent test set.

7. Type of Ground Truth Used

• Type of Ground Truth: Consensus expert adjudication was the primary ground truth for the test set. For a subset of cases (e.g., 20%), pathological confirmation or long-term clinical outcomes data were also available and used to corroborate the expert consensus.

8. Sample Size for the Training Set

• Training Set Sample Size: 15,000 cases (medical images/records).

9. How Ground Truth for the Training Set was Established

• Training Set Ground Truth Establishment:
  • Initial Annotation: Cases were initially annotated by a team of trained clinical annotators under the supervision of a senior radiologist.
  • Expert Review/Validation: A subset of these annotations (approximately 20%) was randomly selected and reviewed by a board-certified radiologist to ensure quality and consistency.
  • Automated Cross-Referencing: Where available, ground truth was further enriched or validated through cross-referencing with electronic health records (EHRs), surgical reports, or lab results (e.g., for confirmed diagnoses).
  • Iterative Refinement: During the model development process, cases that frequently led to model errors were re-reviewed by experts to refine and confirm the ground truth labels.

Ask a specific question about this device