Search Results

The MiroCam® Capsule Endoscope System is intended for visualization of the small bowel mucosa.

• It may be used in the visualization and monitoring of lesions that may indicate Crohn's disease not detected by upper and lower endoscopy.
• It may be used in the visualization and monitoring of lesions that may be a source of obscure bleeding (either overt or occult) not detected by upper and lower endoscopy.
• It may be used in the visualization and monitoring of lesions that may be potential causes of iron deficiency anemia (IDA) not detected by upper and lower endoscopy.
  It may be used as a tool in the detection of abnormalities of the small bowel and this device is indicated for adults and children from two years of age.
  The Suspected GI Bleeding Indicator (SGIB) is intended to mark frames of the video suspected of containing blood or red areas.

The MiroCam® Capsule Endoscope System is comprised of the following core components:

• MiroCam® Capsule: MC1600-B
• MiroCam® Receiver: MR2000
• Data Cable / Data Belt (Optional): MR1000-D / MR1000-D(S) or MR1000-D(M) or MR1000-D(L)
• MiroView™ Software: MiroView™ U 4.0
  The general usage workflow of the MiroCam® system is as follows, the MiroCam® capsule captures images of the GI tract, which are sent via Human Body Communication to sensor pads which are affixed to the patient body. The sensor pads are connected to the receiver by the data cables. The image data is stored on the receiver for the duration of the patient procedure. After removing the receiver set from the patient body, the receiver is connected via USB to the Commercial PC and the image data is uploaded. Following upload, the physician (Gastroenterologist) reviews the patient image data for suspected abnormalities of the small bowel.

The provided text describes the MiroCam® Capsule Endoscope System and its substantial equivalence to predicate devices, including performance testing results for the MiroView™ Express 3.0 software. While it refers to "acceptance criteria" through comparison with predicate devices and mentions performance testing, explicit, quantifiable acceptance criteria are not clearly laid out in a dedicated table format with corresponding performance metrics for the AI component (Suspected GI Bleeding Indicator - SGIB).

However, the closest information to acceptance criteria for the software component's performance (specifically the MiroView™ Express 3.0 with its "informatics algorithm" for detecting significant images) can be inferred from the clinical study summary.

Here's an attempt to extract and present the information based on the provided text:

Table of Acceptance Criteria (Inferred) and Reported Device Performance

The document does not provide a formal table of acceptance criteria with "pass/fail" thresholds. Instead, it describes a clinical study to evaluate the performance of the "express view reading mode" (which includes an "informatics algorithm") for the MiroCam® Capsule Endoscope System, particularly concerning the detection of "significant lesions" and "significant images." The implicit acceptance criterion is that the new system's performance, particularly in terms of sensitivity and reading time, is comparable or improved compared to standard reading, while also demonstrating clinical utility.

Study Details Proving Device Meets Acceptance Criteria:

1. Sample Size and Data Provenance:

   • Test Set Sample Size: 83 patients with obscure gastrointestinal bleeding.
   • Data Provenance: Retrospective and prospective (films were read, then a second reading was performed). The study was multicenter, conducted across 10 centers in France.

2. Number of Experts and Qualifications for Ground Truth:

   • Number of Experts: Three experts.
   • Qualifications of Experts: Not explicitly stated beyond "three experts." It can be inferred they are gastroenterologists or specialists in capsule endoscopy given the context of the study.

3. Adjudication Method for Ground Truth:

   • Method: A "consensus reading and review" was done by the three experts. This implies a 3-expert consensus model, likely where all three agreed, or a majority vote determined the ground truth.

4. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

   • Was it done? Yes, a comparative study was performed comparing "standard mode" reading and "express view mode" reading (which includes the AI algorithm).
   • Effect Size of Human Reader Improvement with AI (Express View) vs. Without AI Assistance (Standard View):
     • Reading Time: The mean reading time without AI assistance (Standard Reading) was 39.7 minutes. The mean reading time with AI assistance (Express View Mode) was 19.7 minutes. This represents a significant reduction of approximately 50.4% in reading time (39.7 - 19.7 = 20 minutes reduction) for human readers when assisted by the express view mode and its algorithm.
     • Sensitivity (Clinical Lesions): Standard reading: 93.3% sensitivity. Express view reading: 82.2% sensitivity. The document states this difference was "NS" (not statistically significant). While not an "improvement" in sensitivity for the human reader due to AI, the implication is that the efficiency gain (reduced reading time) was achieved without a statistically significant compromise in sensitivity for identifying significant lesions by the human reader.
     • Detection of Significant Images by Algorithm: The informatics algorithm detected 94.3% of significant images (66/70), which was higher than both standard reading (82.9%) and express view reading (78.6%) by human readers. This suggests direct algorithmic assistance in highlighting potentially crucial frames.

5. Standalone (Algorithm Only) Performance:

   • Was it done? Yes, in a way. The "informatics algorithm" within the MiroView™ Express 3.0 was evaluated for its ability to detect "significant images" independently of the human review process. It is stated that "The informatics algorithm detected 66/70 images (94.3%) thus missing four small-bowel angiodysplasia." This represents a standalone performance metric for the algorithm's image detection capability compared to the ground truth.

6. Type of Ground Truth Used:

   • Ground Truth: Expert consensus. Specifically, "A consensus reading and review were done by three experts, and considered to be the gold standard." This ground truth was established for identifying "significant lesions" and "significant images."

7. Training Set Sample Size:

   • The document does not specify the sample size used for the training set of the "informatics algorithm" or the MiroView™ Express 3.0 software. The clinical study described appears to be a validation study for an already developed algorithm.

8. How Ground Truth for Training Set Was Established:

   • The document does not provide information on how the ground truth for the training set was established. The clinical study details describe the ground truth establishment for the test set used in this specific performance evaluation.

Ask a specific question about this device