(168 days)
The CORA System is intended for in vitro diagnostic use to provide semi-quantitative indications of the hemostasis state of a venous blood sample. The CORA System records the kinetic changes in a venous sample of 3.2% cirrated whole blood as the sample clots, and retracts in real time. The system output consists of a table of numerical values for parameters R, K, Angle, MA, and FLEV.
The CORA System provides specific blood modifiers, in the form of reagents dried-in-place within CORA Cartridges.
Results from the CORA analysis should not be the sole basis for a patient diagnosis, but should be evaluated together with the patient's medical history, the clinical picture and, if necessary, further hemostasis tests.
The indication for CORA System use is with adult patients where an evaluation of their blood hemostasis properties is desired. Hemostasis evaluations are commonly used to assess clinical conditions in cardiology procedures to assess hemorrhage or thrombosis conditions before, during and following the procedure.
The CORA System consists of a four-channel diagnostic analyzer with integrated computer module, system reagents (CK, CRT, CKH, and CFF), and Abnormal Quality Control material and microfluidic test cartridges.
To perform a test, a disposable CORA Cartridge is inserted into the analyzer. Blood or WQC material is added to an entry port on the cartridge and drawn into the cartridge under analyzer control. The amount of the sample drawn into the cartridge is automatically determined by the volume of the blood chambers in the cartridge. Once in the disposable, the sample is metered into as many as four separate analysis channels, depending upon the assays being performed. Reconstitution of reagents dried within the cartridge is accomplished by moving the sample back and forth through reagent chambers, under the control of microfluidic valves and bellows (pumps) within the cartridge. After each sample has been mixed with reagent, it is delivered to a test cell where it is monitored for visco-elastic changes due to coagulation. Excess sample material is moved under microfluidic control into an enclosed waste chamber within the cartridge.
The CORA technology is based on a disposable containing up to four independent measurement cells. Each cell consists of a short vertically-oriented injection molded tube (ring) with a diameter of 2.5mm and a length of 4.5mm. Detection of clotting in the CORA System is performed optically. Under control of the analyzer, approximately 20ul of prepared sample is delivered to the tube, where a meniscus naturally forms at each end of the tube. The tube is positioned so that the lower meniscus partially blocks light traveling from a collimated source toward a photodiode.
During testing, a piezoelectric actuator drives the measurement cell(s) through a motion profile composed of summed sinusoids at different frequencies. The profile has a maximum amplitude of under 10um and contains frequencies from 10-500Hz. Some, but not all, of the measurement cell motion will induce motion in the sample meniscus, which will be detected by the photodiode. The resulting motion of the meniscus is monitored optically and recorded by the analyzer to calculate the resonant frequency and modulus of elasticity (stiffness) of the sample. By performing a Fast Fourier Transform (FFT) on meniscus motion data, it is possible to determine the frequencies of input motion that caused the greatest deflection of the sample (these are called the resonant frequencies).
Here's an analysis of the acceptance criteria and study details for the CORA® (Coagulation Resonance Analysis) System, based on the provided text:
Important Note: The provided text is a 510(k) summary, which focuses on demonstrating substantial equivalence to a predicate device. It typically does not contain explicit "acceptance criteria" for performance in the same way a clinical trial protocol would. Instead, it demonstrates performance that aligns with a predicate device and adheres to established analytical performance guidelines (like CLSI standards). The "acceptance criteria" as presented below are inferred from the demonstrated performance and the context of regulatory submission for a device of this type.
1. Table of Acceptance Criteria and Reported Device Performance
Given the nature of the submission (510(k) summary), the "acceptance criteria" are implied by the precision limits set and met, and the acceptable agreement with the predicate device.
| Test Parameter | Acceptance Criteria (Implied / Stated) | Reported Device Performance (CV% or Correlation with Predicate) |
|---|---|---|
| Analytical Precision (All Reagents) | ||
| R parameter | CV ≤ 15% | All reported CVs for R ≤ 14.8% (Hypo, Normal, Hyper, Anticoagulant Patients) |
| K parameter | CV ≤ 25% | All reported CVs for K ≤ 23% (Hypo, Normal, Hyper, Anticoagulant Patients) |
| Alpha (Angle) | CV ≤ 10% | Most reported CVs for Angle ≤ 8.7%; a few outliers noted (up to 27.4%) with justification that they are not clinically significant. |
| MA parameter | CV ≤ 10% | Most reported CVs for MA ≤ 8.4%; one CFF MA Hypo outcome (11.6%) with justification that the SD is small and not clinically significant. |
| FLEV parameter | (No explicit CV% provided, but expected to be within reasonable limits for clinical use) | Most reported CVs for FLEV ≤ 5% (Hypo, Normal, Hyper, Anticoagulant Patients) |
| Method Comparison (Correlation with TEG 5000) | ||
| All parameters (R, K, Alpha, MA, FLEV) | Excellent agreement expected, with R-values generally above | R-values for all parameters (CK, CKH, CRT, CFF) ranged from 0.680 to 0.938, demonstrating strong correlation. |
| Reader Study | ||
| Agreement between CORA and TEG 5000 readings | Excellent agreement expected | Demonstrated "Excellent agreement" between the two devices based on 30 test outcomes per reagent category, per reader. |
2. Sample Size Used for the Test Set and Data Provenance
-
Analytical Precision:
- Normal Donors (original study): Blood draws from 3 donors (1 Hypo, 1 Normal, 1 Hyper) across 5 non-consecutive days, by 2 operators, using 3 reagent lots and 12 analyzers. Each test had 2 replicates. This results in 120 measurements for each parameter per level (Hypo, Normal, Hyper).
- Spiked Samples (supplemental study): 3 donors for each reagent-parameter and spiking type (hypocoagulable, hypercoagulable). For each donor-reagent-parameter-hypo/hyper type, there were 12 outcomes (combinations of 3 operators, 3 reagent lots, 12 instruments, 2 replicates per instrument).
- Anticoagulant Patients (supplemental study): 4 Dabigatran patients and 2 Warfarin patients for each reagent-parameter. 12 outcomes for each patient-reagent-parameter.
- Data Provenance: Not explicitly stated, but performed in Coramed's laboratory. Given it's a US company, it's likely US-derived. This is a retrospective analysis of laboratory test samples.
-
Reference Ranges:
- Sample Size: Up to 55 normal volunteer subjects from each of three sites, totaling 157 samples. At least 151 valid results for every parameter for all reagents.
- Data Provenance: Three clinical sites. Subjects chosen to represent demographic populations of these three areas (age, race, gender). This is a prospective study from normal volunteers.
-
Method Comparison:
- Sample Size: Not explicitly stated as a single number but conducted at three clinical sites on patient samples. Up to 10% contrived samples were added to broaden the comparison range.
- Data Provenance: Three clinical sites: Mayo Clinic, University of Pittsburgh Medical Center, and Sinai Hospital, Baltimore. Subjects were patients undergoing cardiovascular surgery or cardiology procedures, with blood samples drawn pre- and post-surgery and in the ICU. This is a prospective study from clinical patients, with some contrived samples.
-
Reader Study:
- Sample Size: 30 test outcomes compared for each reagent test category (e.g., CK, CKH, CRT, CFF).
- Data Provenance: Three sites.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications
-
Analytical Precision, Linearity, Sensitivity & Specificity, Heparin Neutralization: The ground truth for these analytical studies is based on the known concentrations/conditions of the samples (e.g., spiked samples, known interferents) and laboratory measurements. No human "experts" established ground truth in the interpretative sense.
-
Reference Ranges: Ground truth is established by obtaining samples from "normal volunteer subjects" and excluding outliers based on standard criteria. There is no mention of "experts" establishing a ground truth for individual samples.
-
Method Comparison: Ground truth is the results obtained from the predicate TEG 5000 device.
-
Reader Study:
- Number of Experts: Nine readers (three experienced TEG 5000 doctors at each of the three sites).
- Qualifications: "Experienced TEG 5000 doctors." Specific years of experience or board certification are not detailed.
4. Adjudication Method for the Test Set
-
Analytical Precision, Linearity, Sensitivity & Specificity, Heparin Neutralization, Reference Ranges, Method Comparison: No adjudication method is described as these are quantitative analytical studies or comparisons to a predicate device. Ground truth is either inherent in the sample preparation or derived from the predicate device's measurements.
-
Reader Study: No explicit adjudication method (e.g., 2+1, 3+1) is described. The study assessed "agreement between the two devices" based on the doctors' readings of both device outputs. It implies that the TEG 5000 reading was considered the reference for comparison, but not necessarily an adjudicated ground truth in the sense of multiple experts independently labeling cases and resolving discrepancies.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
-
Was an MRMC study done? Yes, a "Reader Study" was conducted, which has elements of an MRMC study in that it involved multiple readers and multiple cases.
-
Effect Size of Human Readers Improve with AI vs. Without AI Assistance: This information is not applicable and not provided in the document. The CORA System is a diagnostic instrument that provides "semi-quantitative indications of the hemostasis state," similar to the predicate TEG 5000. It does not incorporate AI nor does it aim to "assist" human readers in interpreting complex images or deriving diagnoses. It measures parameters directly; the "Reader Study" was about comparing the readings of the outputs from the CORA System to the predicate TEG 5000, not about human performance with/without AI assistance.
6. Standalone Performance (Algorithm only without human-in-the-loop performance)
- Was a standalone performance study done? Yes, the entire set of analytical performance studies (Analytical Precision, Interference, Linearity, Sensitivity and Specificity, Heparin Neutralization) and the Method Comparison study effectively represent standalone (algorithm only) performance. The CORA System generates numerical outputs (R, K, Angle, MA, FLEV) directly. The human "reader" in the "Reader Study" is interpreting the output of the device, not interacting with an algorithm to refine a diagnosis. Therefore, the core function of the device is standalone.
7. Type of Ground Truth Used
- Analytical Precision, Interference, Linearity, Sensitivity and Specificity, Heparin Neutralization: Ground truth was established by known sample properties (e.g., spiked concentrations of substances, known degrees of hemodilution) or through controlled laboratory measurements.
- Reference Ranges: Ground truth was established by samples from "normal volunteer subjects", with outliers removed using standard statistical criteria.
- Method Comparison: Ground truth was the predicate device (TEG 5000) measurements. This is a common approach in 510(k) submissions to demonstrate equivalence.
- Reader Study: Ground truth was readings from the predicate device (TEG 5000) outcomes by experienced clinicians, to which the CORA System's outcomes were compared. It's essentially a comparison of the device outputs' interpretability.
8. Sample Size for the Training Set
The document does not explicitly describe a "training set" for an AI/algorithm in the conventional sense. The CORA System is a measurement device that senses physical changes and calculates parameters. The development of its algorithms (how it analyzes frequency data to provide parameters) would have involved extensive engineering and potentially iterative testing, but this is distinct from the "training set" concept often associated with machine learning.
The analytical studies (precision, linearity, etc.) and method comparison effectively validate the final algorithms/calculations after their development.
9. How the Ground Truth for the Training Set Was Established
As noted above, the concept of a "training set" and its "ground truth" in the context of an AI/ML algorithm isn't directly applicable here. The device's parameters are derived from physical measurements and established formulas related to viscoelastic properties of blood. The underlying physics and physiology are the "ground truth" upon which the measurement technique is built. Any internal calibration or algorithm development would have relied on known samples or precisely measured physical responses, but these are not explicitly detailed as a "training set" in the submission.
§ 864.5425 Multipurpose system for in vitro coagulation studies.
(a)
Identification. A multipurpose system for in vitro coagulation studies is a device consisting of one automated or semiautomated instrument and its associated reagents and controls. The system is used to perform a series of coagulation studies and coagulation factor assays.(b)
Classification. Class II (special controls). A control intended for use with a multipurpose system for in vitro coagulation studies is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 864.9.