Search Results

The COOK Hemospray device is intended for hemostasis of non-variceal bleeds in the GI (gastrointestinal) tract. It is for prescription use only.

Hemospray is an inert, bentonite powder developed for endoscopic hemostasis. The powder is delivered by use of a carbon dioxide powered delivery system and through a catheter inserted through the working channel of an endoscope which provides access to the bleed. Each device contains approximately 20g of powder.

This document is a 510(k) Premarket Notification for a medical device called the Hemospray Endoscopic Hemostat. The submission primarily addresses a minor design change to an existing predicate device to correct a failure mode related to the activation knob. Because the changes are focused on a revised manufacturing process and an improved safety feature (the activation knob's tensile strength) rather than a change in the therapeutic mechanism or intended use, many of the typical acceptance criteria for AI/ML-based devices, such as those related to diagnostic performance or multi-reader studies, are not applicable here.

Therefore, I cannot populate the table and sections as requested for an AI/ML device's acceptance criteria and study proving performance. This document concerns a physical medical device and primarily focuses on mechanical performance and safety, not on the performance of a software algorithm.

However, I can extract information related to the device's functional acceptance criteria as demonstrated through bench testing for the design modification.

Here's how I can interpret and present the available information in the context of your request, focusing on the specific modification made to the physical device:

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

Note: The acceptance criteria and performance data provided relate to the mechanical integrity of the modified activation knob, not to an AI/ML algorithm's diagnostic or therapeutic performance.

2. Sample size used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective)

• Sample Size: Not explicitly stated as a number of units, but "All tested samples" implies a complete test of the sample set prepared.
• Data Provenance: The testing was "bench testing" conducted by the manufacturer, Wilson-Cook Medical, Inc., located in Winston-Salem, North Carolina. This is prospective testing performed for the purpose of this 510(k) submission.

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

• Not applicable. This study is not evaluating a diagnostic or AI/ML device requiring clinical expert ground truth for image interpretation or diagnosis. It focuses on the mechanical performance of a device component. The "ground truth" for the new design specification was established based on a human factors study evaluating the maximum force a user could apply.

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

• Not applicable. No adjudication method is described, as the evaluation is based on objective physical measurements (tensile strength) against a defined design specification, rather than subjective interpretations.

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

• Not applicable. This study is not an MRMC study and does not involve AI assistance. It evaluates a mechanical design change to a physical hemostatic device.

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

• Not applicable. This study does not involve an algorithm.

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

• The "ground truth" for the new tensile strength design specification was established based on a human factors study that evaluated the maximum force a user could apply to the red activation knob during use. The success of the device was then determined by objective mechanical testing to confirm that the component met this established design specification.

8. The sample size for the training set

• Not applicable. This is not an AI/ML device; therefore, there is no "training set." The human factors study provided data to establish the design specification.

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

• Not applicable. As above, no training set for an AI/ML algorithm. The design specification (which could be considered a form of "ground truth" for the mechanical test) was established through a "human factors study that evaluated the maximum force a user could apply to the red activation knob during use."

Ask a specific question about this device

The COOK Hemospray® Endoscopic Hemostat is used for hemostasis of non-variceal gastrointestinal bleeding.

The Hemospray® Endoscopic Hemostat is a prescription use device consisting of a hemostatic agent and a delivery system. The hemostatic agent is (b) (4) | bentonite powder, a naturallysourced aluminum phyllosilicate clay. The delivery system is an endoscopic accessory used for spraying the powder onto the bleeding surface. The delivery device consists of a 220cm polyethylene application catheter, a handle with a pressurized CO2 cartridge, and a powder chamber containing approximately 20g of the Hemospray® material is propelled through the application catheter by release of CO2 from the cartridge located in the device handle. A trigger valve with an activation button allows the physician to control the amount of powder delivered to the affected area. Each actuation of Hemospray® delivers of CO2 and (b) gram of (b) (4) bentonite. Use of the complete 20-gram canister can produce a volume increase within the lumen of the bowel of 3 liters. In the presence of blood, the (b) (4) bentonite may swell 10-15% in volume.

The delivery system is offered in two configurations (i.e., two different outer diameters of the powder delivery catheter). The HEMO-7 version is used with a 2.8mm endoscope accessory channel, and the HEMO-10 version is used in a 3.7mm endoscope accessory channel.

The Hemospray® Endoscopic Hemostat is designed for hemostasis of non-variceal gastrointestinal bleeding. The acceptance criteria for the device are derived from a combination of non-clinical, animal, and clinical studies, as well as literature reviews and post-market data. The studies aim to demonstrate the device's biocompatibility, sterility, shelf life, and its ability to achieve acute hemostasis while minimizing risks such as re-bleeding, perforation, and other adverse events.

Here's a breakdown of the acceptance criteria and the studies proving the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Sizes and Data Provenance

• Test Set (Clinical Data):

  • Pilot Study: 20 patients. Hong Kong, China. Prospective.
  • SEAL Registry: 89 evaluable patients (N=92 enrolled, 3 excluded). Canada, Denmark, England, France, Germany, Italy, Holland. Post-market, observational survey.
  • HALT Study: 64 enrolled patients (interim report). EU and Canada. Post-market, single-arm study.
  • APPROACH Study: 50 enrolled patients. Canada. Post-market, single-arm, prospective observational cohort study.
  • Literature Review: 30 relevant studies, comprising 522 patients. Geographically diverse (Europe, Asia, Canada). Retrospective and prospective studies, case series, case reports, and registry-derived studies.
  • Emergency Use: 5 critically ill patients. United States. Pre-market.
  • Total Clinical Data (overall summary): 750 patients.

• Training Set: Not explicitly stated for specific clinical studies, as many were observational or post-market with concurrent data collection. For the algorithm (if applicable, which is not the primary focus for a physical device in this context), this information would be separate. For the manufacturing/design aspects, non-clinical and bench testing would conceptually be "training data" for product development cycles.

  • Biocompatibility: GLP porcine gastric bleeding animal study (subchronic toxicity and implantation).
  • Performance - Animal: 6 Hemospray® treated animals, 3 sham control (total N=9). Porcine gastric bleeding model.

3. Number of Experts to Establish Ground Truth for Test Set & Qualifications

• For the clinical studies (Pilot, HALT, APPROACH), the ground truth for outcomes like hemostasis, re-bleeding, and adverse events was established by treating physicians (endoscopists/gastroenterologists) at the respective clinical sites. The text doesn't specify the exact number or years of experience for individual experts, but it implies qualified medical professionals for each study. For example, the HALT study had 10 enrollment sites, meaning at least 10 primary investigators or teams. The APPROACH study had 4 clinical sites.
• The SEAL Registry explicitly states that "all physician participants were trained on the use of the Hemospray® kit by Wilson-Cook Medical, Inc. personnel." This implies a level of qualification and standardized training for those assessing outcomes.
• The overall literature review compiled data reported by clinician-authors in their respective publications.

4. Adjudication Method for the Test Set

• The document does not explicitly detail a centralized adjudication method (e.g., 2+1, 3+1 structure with an independent committee) for the clinical studies. For single-center or multi-center, physician-reported outcomes, the assessment of hemostasis, re-bleeding, and adverse events would typically be performed by the local site investigators guided by the study protocol definitions.
• For the SEAL Registry, "No queries or audits to identify and correct missing data were performed," indicating a lack of formal adjudication beyond initial reporting.
• The HALT study mentions endoscopically-confirmed recurrent bleeds, suggesting a visual assessment by the clinician performing the endoscopy.

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

• No MRMC comparative effectiveness study was done. The studies presented are primarily single-arm studies evaluating the performance of Hemospray® alone or in conjunction with other standard therapies.
• The document does compare Hemospray® results to "Standard of Care per ASGE" in Table 20 (APPROACH study results) and generally discusses how Hemospray® "rivaled standard of care (SOC) with 63% initial hemostasis and 38% re-bleed rate" for patients on antithrombotic therapy. However, these are comparisons to historical or published SOC rates, not from a direct head-to-head MRMC study.
• No effect size of human readers improving with AI vs. without AI assistance is provided, as this is a physical medical device, not an AI diagnostic tool.

6. Standalone (Algorithm Only) Performance Study

• Not applicable as the Hemospray® Endoscopic Hemostat is a physical medical device. Performance is inherently tied to its application by a human operator (physician) in a clinical setting.

7. Type of Ground Truth Used

• Clinical Studies (Pilot, HALT, APPROACH, Emergency Use): Clinician assessment and observation (endoscopic visualization) of bleeding cessation, re-bleeding, and adverse events, as defined by study protocols. Outcomes data (e.g., mortality).
• Animal Study: Gross and histologic assessment of regional and systemic effects by veterinary pathologists/researchers.
• Bench Studies: Objective measurements against predefined engineering specifications (e.g., pressure, density, functional checks).
• Biocompatibility: Laboratory testing results based on ISO and ASTM standards.
• Literature Review: Aggregated clinical outcomes reported in peer-reviewed publications.

8. Sample Size for the Training Set

• Animal Study: 6 Hemospray® treated animals, 3 sham control (total N=9). While not a "training set" in the machine learning sense, this GLP porcine study served as crucial in vivo data for validating effectiveness and safety prior to widespread human use.
• Biocompatibility: Various numbers for different tests (e.g., guinea pigs for sensitization, rabbits for pyrogenicity, cell lines for cytotoxicity). The "subchronic toxicity and implantation" was evaluated in the porcine animal study.
• Non-clinical/Bench Studies: Performed on device prototypes and finished devices; specific sample sizes for each bench test are not always detailed but are implicit in the "Passed" results.
• First-in-Human Pilot Study (20 patients): Could be considered early "training" or "feasibility" data for refining clinical protocols and understanding human safety.

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

• Animal Study (Porcine Model): Ground truth for hemostasis was established by direct observation during the procedure and confirmed by follow-up. Safety endpoints (gross and histologic assessment of regional and systemic effects) were established by detailed pathological examination of tissues collected at termination by trained veterinary pathologists.
• Biocompatibility Testing: Ground truth established according to standardized international (ISO, ASTM) and national (USP) testing protocols, which define specific biological responses (e.g., non-cytotoxic, non-irritant) under controlled laboratory conditions.
• Bench Testing: Ground truth established through engineering specifications and predefined pass/fail criteria for physical and functional parameters of the device and its components, measured using calibrated equipment.

Ask a specific question about this device