Search Results
Found 1 results
510(k) Data Aggregation
(283 days)
AnX Robotica, Inc.
The NaviCam Stomach Capsule is intended for visualization of the stomach of adults (≥22 years old) with a BMI less than 38. The system can be used in clinics and hospitals, including ER settings.
The NaviCam Capsule Endoscopy System with NaviCam Stomach Capsule is an endoscopic capsule imaging system intended to obtain images of the stomach and duodenum. In contrast to passive capsule endoscopy systems, it uses external magnetic fields to allow the position of the capsule to be controlled by an operator. The NaviCam Capsule Endoscopy System with NaviCam Stomach Capsule consists of an ingestible capsule, a data recorder, a locator, and a controller.
This response focuses on the provided text to describe the acceptance criteria and the studies performed for the NaviCam Capsule Endoscope System.
Acceptance Criteria and Device Performance for NaviCam Capsule Endoscope System
The NaviCam Capsule Endoscope System with NaviCam Stomach Capsule underwent various non-clinical (bench) and clinical studies to demonstrate its safety and effectiveness.
1. Table of Acceptance Criteria and Reported Device Performance
Non-Clinical Acceptance Criteria and Performance:
| Test | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Bite Force | The capsule should withstand applied force up to (b) (4) | Passed |
| Temperature Safety | △T should not be more than (b) (4) | Passed |
| Magnetic Force Measurement | The acceptance criteria for this test are: (b) (4) | Passed |
| Magnetic Field Test | 1) The maximum value of magnetic flux density on the surface of the NaviCam stomach capsule must be less than or equal to (b) (4) 2) The magnetic field of the capsule at (b) (4) must be less than (b) (4) | Passed |
| Battery Life | The acceptable result of the battery test is at least (b) (4) pictures (equal to (b) (4) hours battery life time, (b) (4) captured by the capsule. | Passed |
| pH Test | 1) After soaking, there is no change in the capsule weight. Considering measurement error, the variation in weights obtained by the balance shall not exceed (b) (4) 2) After soaking, there is no change on the surface of the capsule front and rear shells. 3) Resolution and color reproduction are not affected. | Passed |
| Color Performance | For a capsule endoscope, because of its size, Field of View (FOV), small field size and the way of imaging based on the illumination of its built-in LEDs, for improved image color reproduction the total color difference of a sample (ΔE) should be kept as no more than (b) (4) | Passed |
| Image Resolution | For (b) (4) working distance, module transfer function (MTF) (b) (4) is not less than (b) (4) | Passed |
| Field of View | (b) (4) is accepted, that is (b) (4) (b) (4) degree. | Passed |
| Geometric Distortion | Distortion value not larger than (b) (4) is accepted. | Passed |
| Depth of View (DOV) | For reflectance USAF1951 angular resolution test, DOV is (b) (4) mm to (b) (4) in air or under water. | Passed |
| Peak Illuminance | (b) (4) is accepted, that is (b) (4) to (b) (4) lux. | Passed |
| Image Intensity Uniformity | (1) The two-dimensional distribution of the IIU space is basically a spatially symmetric distribution. (2) The four-dimensional distribution of the IIU at (b) (4) is basically symmetrical from the center. (3) The calculated minimum IIU value in the four directions of (b) (4) should not be less than (b) (4) | Passed |
| Photobiological Safety | The device must meet light hazard exposure limits per IEC 62471:2006. | Passed |
Clinical Acceptance Criteria and Performance (Comparative Studies):
While specific numeric acceptance criteria for sensitivity, specificity, PPV, and NPV are not explicitly stated as acceptance criteria in the clinical sections, the consistently high reported performance against the "gold standard" of gastroscopy is used to demonstrate effectiveness. The FDA's conclusion on benefit/risk references these results.
| Metric | Comparative Study (350 patients) | CFDA Study (99 subjects) |
|---|---|---|
| Sensitivity (for gastric focal lesions) | 90.4% (95% CI: N/A, see table for per-lesion sensitivity) | 90.9% |
| Specificity (for gastric focal lesions) | 94.7% (95% CI: 91.9% - 97.5%) | 94.8% |
| Diagnostic Accuracy (for gastric focal lesions) | 93.4% (95% CI: 90.8% - 96.0%) | 93.9% |
| Positive Predictive Value (PPV) | 87.9% (95% CI: 81.7% - 94.0%) | 83.3% |
| Negative Predictive Value (NPV) | 95.9% (95% CI: 93.4% - 98.4%) | 97.3% |
| Sensitivity for lesions ≥5mm | ≥87.5% (77.3-97.8%) | Not explicitly stated, grouped with overall sensitivity |
| No missed clinically significant lesions | MCE did not miss any lesions of significance (tumors, large ulcers) | Not explicitly stated, but high NPV implies this |
| Effective Visualization Rate (CFDA Study) | N/A (Comparative study focuses on lesion detection) | Dentate line: 92%, Cardia: 97%, Fundus: 94%, Gastric body: 99%, Gastric antrum: 99%, Pylorus: 99%, Duodenal bulb: 96% |
| Adverse Events | No adverse events or capsule retention | No adverse events or capsule retention |
2. Sample Sizes and Data Provenance
- Comparative Study:
- Test Set Sample Size: 350 patients.
- Data Provenance: Retrospective, conducted at a tertiary center in China.
- Chinese Food and Drug Administration (CFDA) Study:
- Test Set Sample Size: 99 subjects.
- Data Provenance: Not explicitly stated as retrospective or prospective within the provided text for this specific study, but it's a "self-controlled comparative trial" suggesting a prospective design where subjects undergo both procedures. Country of origin is China (implied by CFDA).
- Literature Reviews:
- Screening study for gastric cancer: 3182 asymptomatic Chinese individuals.
- Superficial gastric neoplasia study: 10 subjects.
- Locator detection study: Not specified per subject, but performed on patients where the capsule was detected by X-ray.
3. Number of Experts and Qualifications for Ground Truth
- Comparative Study: The gastroscopy procedures were used as the gold standard. While the number and specific qualifications of the endoscopists performing the gastroscopy are not detailed in the provided text, it can be assumed that they were qualified medical professionals performing conventional gastroscopy. The interpretation of MCE and gastroscopy results to calculate performance metrics would have been done by clinicians.
- CFDA Study: Similar to the comparative study, gastroscopy was the ground truth. The determination of "complete," "incomplete," or "unable to be explored" for visualization rates was "based on the determination of the physician," indicating expert involvement. No specific number or qualifications are provided.
- Literature (Screening for gastric cancer): For the 7 suspected malignancies identified by MCE, the confirmation was "pathologically," indicating histopathologists provided the definitive ground truth. No specific number or qualifications are given for these pathologists.
- Literature (Superficial gastric neoplasia): Diagnostic agreement of MCE, ESD (endoscopic submucosal dissection), and pathology were compared. Pathology would be the ultimate ground truth here.
4. Adjudication Method for Test Set
The adjudication method is not explicitly described as a formal consensus process (e.g., 2+1, 3+1). The studies directly compared MCE findings to gastroscopy results, which served as the "gold standard" or control. Discrepancies would likely have been noted, but a formal adjudication process for discordant cases between MCE and gastroscopy is not elaborated upon in the text.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No specific MRMC comparative effectiveness study, designed to measure the effect size of human readers improving with AI vs. without AI assistance, is detailed in the provided text. The studies compared the device's performance (MCE) directly against conventional gastroscopy (human readers performing a different procedure), not human readers reviewing MCE images with and without AI assistance.
6. Standalone (Algorithm Only) Performance
The "NaviCam Capsule Endoscope System with NaviCam Stomach Capsule" is described as an endoscopic system where an operator controls the capsule. The images captured by the capsule are then presumably reviewed by a human clinician. Therefore, the reported performance metrics (sensitivity, specificity, etc.) represent the performance of the system in use by a clinician, not a standalone algorithm that automatically detects lesions without human intervention. The text does not describe an "algorithm only" performance evaluation.
7. Type of Ground Truth Used
- Clinical Studies (Comparative and CFDA): The primary ground truth for lesion detection was conventional gastroscopy. For a subset of suspected malignancies, pathology was used for confirmation. For visualization rates in the CFDA study, it was physician determination.
- Literature (Screening for gastric cancer): Confirmed malignancies were referenced against pathology.
- Literature (Superficial gastric neoplasia): Comparisons were made between MCE, ESD, and pathology.
8. Sample Size for the Training Set
The provided text details clinical validation studies and refers to leveraging clinical data from two studies and 11 literature articles. It does not specify a dedicated training set size for the NaviCam's internal algorithms (e.g., image processing, magnetic control). The clinical studies are described as comparative evaluations against a gold standard, not as studies used to train the device's analytical aspects. The device itself (the capsule and its control system) is the subject of the performance and safety studies, not a developing AI algorithm that requires a distinct training set.
9. How the Ground Truth for the Training Set Was Established
As no distinct training set for an AI algorithm (in the context of automated lesion detection or diagnosis) is described, the method for establishing its ground truth is not applicable based on the provided text. The device's "training" in a broader sense would be through iterative engineering and testing during its development, informed by medical knowledge and clinical requirements, rather than a specific supervised learning approach on a labeled dataset. The clinical studies serve to validate the overall system's performance, using gastroscopy and pathology as the truth.
Ask a specific question about this device
Page 1 of 1