(148 days)
The Pill Sense System is a prescription only device consisting of a reusable receiver and singleuse ingestible capsule, intended to be used for the detection of blood in the upper gastrointestinal tract in hemodynamically stable adults suspected of having upper gastrointestinal bleeding (UGIB).
Pill Sense is not a standalone diagnostic device, but an adjunct for clinical decision making. A negative or normal result obtained by Pill Sense System does not exclude presence of pathology, if symptoms persist, further evaluation should be performed.
The Pill Sense system is designed to detect blood in the upper gastrointestinal tract. The overall system consists of a Pill Sense Capsule and Pill Sense Receiver. The capsule and receiver components are packaged and sold separately or as a system.
The capsule is a single-patient use, battery-powered capsule that features sensors that detect blood and wirelessly transmit data to an external receiver. The capsule makes its way through the gastrointestinal tract and is then passed naturally from the body. The reusable receiver collects and displays real-time information gathered by the capsule. The receiver interprets the data and displays a sensor output value. It plots these values on a chart using data acquisition and when completed displays a "blood detected" or "no blood detected" message. The software supports entering of the patient information, pairing of the receiver and capsule and data interpretation and display.
Within the ingestible capsule, four lights with distinct wavelengths and a photodetector are located within a sensing well in the center of the capsule. Each light is sequentially pulsed to emit light into the capsule sensing well and based on the liquid in the sensing gap, the emitted light is partially or fully absorbed. The remaining light is received by the photodetector, which converts an electrical signal proportional to the amount of light.
Once a "blood detected" message is displayed, it remains on the screen and does not change during the rest of the monitoring. If the monitoring is stopped by the user before 5 minutes has elapsed, an "Inconclusive results" message will be displayed. Only one final readout is given at the end of the 5minute monitoring period, and there is no quantification of blood detected. The overall time from capsule activation, ingestion and display of a reading takes less than 10 minutes. By design, the user is required to monitor for a minimum of 5 minutes and the monitoring automatically stops after 10 minutes. As soon as a "Blood Detected" message is received, the user can stop monitoring at any time. If deemed necessary, the user can continue monitoring for up to 40 minutes.
Acceptance Criteria and Device Performance for Pill Sense System
1. Table of Acceptance Criteria and Reported Device Performance
Clinical Performance Testing:
| Parameter | Acceptance Criteria | Reported Device Performance (95% CI) |
|---|---|---|
| Sensitivity | > 75% | 92.9% (76.5%, 99.1%) |
| Specificity | > 60% | 90.6% (82.9%, 95.6%) |
Non-Clinical/Bench Testing:
| Test | Acceptance Criteria | Device Performance (Summary) |
|---|---|---|
| Visual Inspection & Dimensional Verification | External surface free from extraneous matter or surface defects impacting performance/safety. | Met (Implied by supporting biocompatibility and overall performance) |
| Bite Resistance | No cracks, sharp edges that could cause harm. | Verified (No damage to capsule) |
| pH Solution Test | No damage to capsule. | Verified (No damage to capsule) |
| Capsule Leak Resistance Test | Absence of dye solution within the capsule enclosure. | Verified (Absence of dye solution) |
| Blood Detection Test | Capsules in blood solution produce "blood detected." Capsules in gastric fluid only produce "No blood detected." Capsules in non-blood red liquids/foods produce "no blood detected." Capsules in blood with red liquids/food produce "blood detected." | Met (Implied by successful clinical performance) |
| Battery Life (Capsule) | Operate for a minimum of 6 hours. | Met (Implied by successful monitoring sessions) |
| Wireless (Capsule & Receiver) | Communicate wirelessly. Transmit enough data to successfully complete monitoring session at 0.5 meters range in simulated model. | Met (Implied by successful data transmission) |
| Battery Life (Receiver) | Operate for a minimum of 1 hour. | Met (Implied by successful receiver operation) |
2. Sample Size and Data Provenance
-
Clinical Study (DETECT-1):
- Test Set Sample Size: 124 subjects for the modified Intent-to-Treat (mITT) cohort.
- 28 subjects with upper GI bleeding (confirmed by EGD).
- 96 subjects without upper GI bleeding (confirmed by EGD).
- Data Provenance: Prospective, single-center, single-arm comparative clinical investigation.
- Country of Origin: Not explicitly stated, but typically US-based for FDA de novo submissions without specific mention otherwise.
- Test Set Sample Size: 124 subjects for the modified Intent-to-Treat (mITT) cohort.
-
Non-Clinical/Bench Studies: Sample sizes for these tests are not explicitly stated in the provided text but are generally smaller and conducted in a controlled laboratory environment.
3. Number of Experts and Qualifications for Ground Truth
- Clinical Study (DETECT-1):
- Number of Experts: The ground truth for the presence or absence of upper GI bleeding was established by performing Esophagogastroduodenoscopy (EGD). EGDs are typically performed by one expert (a gastroenterologist or similarly qualified physician).
- Qualifications of Experts: Gastroenterologists or other qualified physicians trained in performing and interpreting EGDs. The text states "The investigators performing the EGD were blinded to the results provided by the Pill Sense System," implying qualified medical professionals. Specific years of experience are not mentioned.
4. Adjudication Method
- Clinical Study (DETECT-1): For the primary endpoint (comparison of Pill Sense results with EGD findings), the adjudication method involved direct comparison. The "investigators performing the EGD were blinded to the results provided by the Pill Sense System." This implies EGD served as the independent reference standard, and no specific adjudication process between multiple EGD readers or between Pill Sense and EGD was described beyond direct comparison. Essentially, the EGD result was considered the definitive ground truth against which the device's performance was measured.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- No, an MRMC comparative effectiveness study was not conducted or described. The study focused on the standalone performance of the Pill Sense System against a clinical reference standard (EGD), without evaluating the effect size of human readers improving with or without AI assistance. The Pill Sense System is described as providing a "blood detected" or "no blood detected" message, and "is not an algorithm-only without human-in-the-loop performance."
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance)
- Yes, a standalone performance was conducted. The Pill Sense System itself (the capsule and receiver) produced the "blood detected" or "no blood detected" message, which was then compared directly to the EGD findings. The device is not presented as an AI-assisted interpretation tool for human readers but as a direct diagnostic aid providing its own result. The text states "The receiver interprets the data and displays a sensor output value. It plots these values on a chart using data acquisition and when completed displays a 'blood detected' or 'no blood detected' message." This indicates an automated, standalone output from the device.
7. Type of Ground Truth Used
- Clinical Study (DETECT-1): The primary ground truth used was endoscopic findings (Esophagogastroduodenoscopy - EGD). The EGD was performed within 4 hours of the Pill Sense system reading and determined the presence ("Blood") or absence ("No Blood") of upper GI bleeding in subjects.
8. Sample Size for the Training Set
- Not explicitly stated for the clinical study. The provided document focuses on the "DETECT-1" study as a clinical investigation for performance evaluation (test set). Information regarding a distinct training set for the Pill Sense System's algorithm development is not included in the provided text. It's common for such devices to be developed and optimized using internal data before a pivotal clinical trial for regulatory submission.
9. How Ground Truth for Training Set Was Established
- Not explicitly stated in the provided documentation. As the document focuses on the DETECT-1 study as the primary clinical evidence, details on how the ground truth for any potential earlier training set was established are not provided.
N/A