Search Results
Found 1 results
510(k) Data Aggregation
(149 days)
The DeltaScan Monitor provides the binary DeltaScan Output based on a technical index of polymorphic delta (PMD) waveshape detections made in the EEG from the bipolar Fp2 and Pz channel on adult patients (over 60 years of age) to aid in the diagnosis of acute encephalopathy.
DeltaScan should only be used by a healthcare provider as a component of a complete clinical evaluation or as support for the clinician's decision to pursue further testing. The device is NOT to be used as a stand-alone method in the evaluation or diagnosis of acute encephalopathy.
The intended patient is a hospitalized, awake adult, who is at risk of acute encephalopathy and delirium as decided by the responsible licensed healthcare physician or a medical professional working under the responsibility of a licensed healthcare physician.
The use environment is in hospitals:
· non-sterile environments;
· ICUs, wards, and other patient evaluation locations;
The DeltaScan Monitor is intended to be used in combination with the DeltaScan Patch (K222671) through a proprietary connector design.
The DeltaScan Monitor provides EEG signal acquisition and analysis technology intended for use as an adjunct to clinical judgment. The DeltaScan Monitor provides support in clinical decision-making by providing an assessment for a patient having acute encephalopathy or not, based on a measure of the detected polymorphic delta (PMD) waves in the EEG.
The DeltaScan Monitor consists of a Monitor and a Patch connector. The Patch connector contains the EEG amplifier hardware. The Monitor contains electronics for galvanic isolation to the EEG cable with Patch connector, storage of EEG recording and log files (eMMC memory chip), processing capacity to run software (DeltaScan Monitor Application, or DMA), user interface elements (e.g., screen, keys, recording button), battery (FEY PA-IEC-LNB162Q.R001), and the charging circuitry. EEG data is collected by the DeltaScan Monitor using a DeltaScan Patch. Collected EEG signals are amplified, digitized, and then processed by the software algorithms to provide the user with the DeltaScan Output. The DeltaScan Monitor Application is stand-alone software running on an Embedded Linux OS.
This document describes the acceptance criteria and the study that proves the device (DeltaScan Monitor R2) meets these criteria.
1. Table of Acceptance Criteria and Reported Device Performance
| Acceptance Criteria (Pre-specified Criteria) | Reported Device Performance (Study Result) | Pass/Fail | Discussion and Sensitivity Analysis | Rationale for Safety and Effectiveness |
|---|---|---|---|---|
| ICU | ||||
| Null hypothesis: NPV < 0.80 (lower bound of CI ≥ 0.8) | NPV = 0.85 CI = [0.77, 0.92] | Fail | The reasons why the end-point for NPV was missed is clear: 1) due to COVID-19 restriction, the number of patients on ICUs is less than aimed for, resulting in larger confidence intervals, and 2) the prevalence estimates were too low in the power calculation, resulting in too high pre-specified criteria. A sensitivity analysis that adjusts the prevalence to the study protocol estimate for prevalence (35%) shows that: NPV = 0.89 [0.85, 0.94]. Under these conditions, the end-points would have been met. | We conclude that reasons for the missed end-point on NPV are clear and understood. Overall, NPV values are reasonably high, while NPV+PPV values exceed the pre-specified criterium. Both NPV and PPV results are robust for some variation in study assumptions (sensitivity analysis). When considering both NPV and PPV, the performance shows safety and effectiveness. |
| NPV + PPV ≥ 1 | NPV + PPV = 1.62 CI = [1.50, 1.72] | Pass | ||
| Ward | ||||
| Null hypothesis: NPV < 0.85 (lower bound of CI ≥ 0.85) | NPV = 0.83 CI = [0.76, 0.89] | Fail | The reason why the end-point for NPV was missed is clear: the prevalence estimates were too low in the power calculation, resulting in too high pre-specified criteria. A sensitivity analysis that adjusts the prevalence to the study protocol estimate for prevalence (25%) shows that: NPV = 0.90 [0.87, 0.93]. Under these conditions, the end-points would have been met. | We conclude that reason for the missed end-point on NPV are clear and understood. Overall, NPV values are reasonably high, while NPV+PPV values exceed the pre-specified criterium. Both NPV and PPV results are robust for some variation in study assumptions (sensitivity analysis). When considering both NPV and PPV, the performance shows safety and effectiveness. |
| NPV + PPV ≥ 1 | NPV + PPV = 1.66 CI = [1.55, 1.75] | Pass | ||
| ICU + Ward (pooled) | ||||
| Null hypothesis: NPV < 0.80 (lower bound of CI ≥ 0.8) | NPV = 0.84 CI = [0.79, 0.88] | Fail | The reason why the end-point for NPV was missed is clear: the prevalence estimates were too low in the power calculation, resulting in too high pre-specified criteria. A sensitivity analysis that adjusts the prevalence to the study protocol estimate for prevalence on the ICU (35%) shows that: NPV = 0.87 [0.84, 0.90]. Under these conditions, the end-points would have been met. | We conclude that reason for the missed end-point on NPV are clear and understood. Overall, NPV values are reasonably high, while NPV+PPV values exceed the pre-specified criterium. Both NPV and PPV results are robust for some variation in study assumptions (sensitivity analysis). When considering both NPV and PPV, the performance shows safety and effectiveness. |
| NPV + PPV ≥ 1 | NPV + PPV = 1.63 CI = [1.55, 1.71] | Pass |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size for Test Set: 434 patients (195 on ICUs and 239 on wards) fulfilled inclusion, but not exclusion, criteria.
- Data Provenance:
- Country of Origin: The Netherlands (geographically distinct clinics: 6 ICUs and 15 wards).
- Retrospective or Prospective: Prospective. The study "DeltaStudy" was designed to evaluate diagnostic performance and repeatability, involving the collection of EEGs with DeltaScan and clinical data on ICUs and wards.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications
- Acute Encephalopathy (EEG reference standard): 3 separate EEG experts. Qualifications are not explicitly detailed beyond being "EEG experts." They visually assessed 4-minutes of EEG data for the presence of polymorphic delta activity.
- Delirium (clinical reference standard): 3 clinical delirium experts. Qualifications are not explicitly detailed beyond being "clinical delirium experts." They assessed clinical data including researcher's interview based on DSM-5 criteria A-C, Electronic Health Record data, and description of patient behavior.
4. Adjudication Method for the Test Set
The ground truth for both acute encephalopathy and delirium was established using consensus (majority vote) among the experts.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No MRMC study comparing human readers with and without AI assistance was reported. The study focused on the standalone diagnostic performance of the DeltaScan Monitor in comparison to expert consensus (ground truth).
6. Standalone (Algorithm Only) Performance
Yes, a standalone performance study was conducted. The "DeltaScan Output" (binary positive/negative for acute encephalopathy) was determined by the DeltaScan Monitor from EEG recordings, and these outputs were compared against the expert committee's estimated diagnoses for acute encephalopathy and delirium.
7. Type of Ground Truth Used
- Acute Encephalopathy: Expert consensus from 3 EEG experts visually assessing EEG data for polymorphic delta activity.
- Delirium: Expert consensus from 3 clinical delirium experts assessing clinical data and DSM-5 criteria.
8. Sample Size for the Training Set
The document mentions that the DeltaScan was calibrated based on a "previous clinical calibration dataset" from Numan et al., 2019, BJA. This dataset contained 321 EEG recordings. This is the sample size for the calibration dataset, which effectively serves as a training or development set for the algorithm's scoring and thresholding.
9. How the Ground Truth for the Training Set Was Established
For the calibration dataset (Numan et al., 2019, BJA), the ground truth for acute encephalopathy and delirium was established through expert labels. The document states that the calibration dataset contained "321 EEG recordings with expert labels for acute encephalopathy and delirium." This suggests a similar expert review process to the test set, where experts provided their diagnoses to create the ground truth.
Ask a specific question about this device
Page 1 of 1