Search Results
Found 1 results
510(k) Data Aggregation
(575 days)
The NMR LipoProfile® -2 test, used with the NMR Profiler, an automated NMR spectrometer, measures lipoprotein particles to quantify LDL particle number (LDL-P), HDL cholesterol (HDL-C), and triglycerides in serum and plasma using nuclear magnetic resonance (NMR) spectroscopy. LDL-P and these NMR-derived concentrations of triglycerides and HDL-C are used in conjunction with other lipid measurements and clinical evaluation to aid in the management of lipoprotein disorders associated with cardiovascular disease. This test is performed and provided as a service by LipoScience Laboratory.
The NMR LipoProfile Test involves measurement of the 400 MHz proton NMR spectrum of a plasma or serum sample, deconvolution of the composite signal at ~0.8 ppm to produce the signal amplitudes of the lipoprotein subclasses that contribute to the composite plasma signal, and conversion of these subclass signal amplitudes to lipoprotein subclass concentrations. The 0.8 ppm plasma NMR signal arises from the methyl group protons of the lipids carried in the VLDL, and HDL subclasses of varying diameter. The NMR signals from the various lipoprotein subclasses have unique and distinctive frequencies and lineshapes, each of which are accounted for in the deconvolution analysis model. Each subclass signal amplitude is proportional to the number of subclass particles emitting the signal, which enables subclass particle concentrations to be calculated from the subclass signal amplitudes derived from the spectral deconvolution analysis. LDL subclass particle concentrations, in units of nanomoles of particles per liter (nmol/L), are summed to give the reported total LDL particle concentration (LDL-P). By employing conversion factors that assume that the various lipoprotein subclass particles have cholesterol and triglyceride contents characteristic of normolipidemic individuals, HDL cholesterol and triglyceride concentrations are also derived.
AI Device Acceptance Criteria and Study Summary
Here's an analysis of the provided text regarding the NMR LipoProfile-2 Assay and NMR Profiler Instrument Test System:
1. Acceptance Criteria and Reported Device Performance
The document describes several analytical and clinical performance aspects. While explicit "acceptance criteria" for all metrics are not always stated with numerical thresholds, they are implied through the performance data and comparisons to predicate devices. For analytical sensitivity and linearity, specific ranges are provided.
Table 1: Acceptance Criteria (Implied/Stated) and Reported Device Performance
| Metric | Acceptance Criteria (Implied/Stated) | Reported Device Performance |
|---|---|---|
| Analytical Sensitivity: | Acceptable precision and accuracy with total error ≤20%. | |
| - LDL-P (LOQ) | - | 300 nmol/L |
| - HDL-C (LOQ) | - | 10 mg/dL |
| - Triglycerides (LOQ) | - | 25 mg/dL |
| Assay Precision: | Acceptable intra-assay and inter-assay variability. (No explicit numeric criteria stated, but low CVs are implied as acceptable) | |
| - LDL-P (Intra-assay) | Pool 1: Mean 2222 nmol/L, SD 49.1, %CV 2.2; Pool 2: Mean 1042 nmol/L, SD 47.7, %CV 4.6 | |
| - LDL-P (Inter-assay) | Pool 1: Mean 1925 nmol/L, SD 66.7, %CV 3.5; Pool 2: Mean 1053 nmol/L, SD 68.4, %CV 6.5 | |
| - HDL-C (Intra-assay) | Pool 1: Mean 41 mg/dL, SD 0.54, %CV 1.3; Pool 2: Mean 57 mg/dL, SD 0.42, %CV 0.7 | |
| - HDL-C (Inter-assay) | Pool 1: Mean 42 mg/dL, SD 1.17, %CV 2.8; Pool 2: Mean 56 mg/dL, SD 0.83, %CV 1.5 | |
| - Trig (Intra-assay) | Pool 1: Mean 189 mg/dL, SD 2.0, %CV 1.1; Pool 2: Mean 75 mg/dL, SD 1.2, %CV 1.5 | |
| - Trig (Inter-assay) | Pool 1: Mean 219 mg/dL, SD 2.9, %CV 1.3; Pool 2: Mean 80 mg/dL, SD 1.7, %CV 2.1 | |
| Linearity: | Wide varying target concentrations with acceptable percent bias. | |
| - LDL-P (Linear Range) | - | 300-6000 nmol/L |
| - HDL-C (Linear Range) | - | 7-160 mg/dL |
| - Triglycerides (Linear Range) | - | 5-2700 mg/dL |
| Reportable Range: | - | |
| - LDL-P | - | 300 - 3500 nmol/L |
| - HDL-C | - | 7 - 140 mg/dL |
| - Triglycerides | - | 5 - 1100 mg/dL |
| Interfering Substances: | No appreciable interference at clinically relevant concentrations. | No appreciable interference by Endogenous substances (Bilirubin, Creatinine, Hemoglobin, Urea, Uric acid) or Exogenous substances (Acetaminophen, Aspirin, Clopidogrel, Enalapril, Fenofibrate, Furosemide, Glipizide, Hydralazine, Hydrochlorothiazide, Ibuprofen, Isosorbide dinitrate, Metformin, Metoprolol, Naproxen, Niacin, Nifedipine, Piroxicam, Simvastatin, Thiazolidinedione, Triamterene) at tested concentrations. The concentrations tested were representative of the highest blood concentrations expected for the highest therapeutic doses of these compounds. |
| Method Comparison (HDL-C): | Demonstrate substantial equivalence to a predicate chemistry analyzer system. (Implicit: strong correlation and similar mean values.) | 5,362 plasma samples tested. R-squared = 0.897. NMR LipoProfile mean: 50.7 mg/dL; Predicate mean: 51.3 mg/dL. |
| Method Comparison (Triglycerides): | Demonstrate substantial equivalence to a predicate chemistry analyzer system. (Implicit: strong correlation and similar mean values.) | 5,362 plasma samples tested. R-squared = 0.929. NMR LipoProfile mean: 128.7 mg/dL; Predicate mean: 123.9 mg/dL. |
| Clinical Performance (LDL-P): | Statistically significant relationship to CVD risk, aiding in the management of lipoprotein disorders. (Implicit: demonstrating predictive value for cardiovascular events). | VA-HIT Study (Baseline): Odds Ratio 1.31 (95% CI, 1.09-1.57), p=0.004 for a new CHD event with 1-SD increment of LDL-P in placebo group. EPIC-Norfolk Study (Multivariable): Statistically significant association of LDL-P quartiles with incident CAD events (p=0.02, highest quartile HR 1.37 (1.04-1.83)). Women's Health Study: Statistically significant association of LDL-P quintiles with incident CVD events (p<0.001, highest quintile HR 2.51 (1.91-3.30)). |
2. Sample Sizes and Data Provenance
Test Set (for method comparison):
- Sample Size: 5,362 plasma samples (for HDL-C and Triglycerides method comparison).
- Data Provenance: From individuals who were part of the Multi-Ethnic Study of Atherosclerosis (MESA). The specific country of origin is not explicitly stated, but MESA is a US-based study. This data is retrospective.
Test Set (for clinical performance of LDL-P measurement):
- VA-HIT Study: 1,061 plasma samples (364 cases, 697 controls). Data is retrospective from a previous randomized placebo-controlled clinical trial. The trial was conducted in the US (Veterans Affairs).
- EPIC-Norfolk Study: 2,888 serum samples (1,003 cases, 1,885 controls). Data is retrospective from the prospective EPIC-Norfolk study conducted in Norfolk, United Kingdom.
- Women's Health Study: 27,673 plasma samples (1,015 CVD events). Data is retrospective from a randomized, double-blind, placebo-controlled trial of US-based female healthcare professionals.
Test Set (for analytical performance - sensitivity, precision, linearity, interfering substances):
- Analytical Sensitivity: Serum specimens with low initial concentrations were serially diluted. 20 replicates of each dilution were analyzed. Exact number of initial specimens not stated.
- Assay Precision: 2 patient serum pools tested with 20 replicates each (intra-assay) and over 20 different runs (inter-assay).
- Linearity: 2 serum pools prepared and diluted to produce 12 different samples. 6 replicates analyzed for each.
- Interfering Substances: 6 plasma pools spiked with potential interferents.
3. Number of Experts and Qualifications for Ground Truth
The document does not mention the use of experts to establish ground truth for the test sets in the context of radiologists or similar clinical diagnosticians. The ground truth for the method comparison studies was established by commercially available chemistry analyzer systems which are considered standard clinical methods.
For the clinical outcome studies (VA-HIT, EPIC-Norfolk, Women's Health Study), the "ground truth" for disease events (e.g., CHD event, CAD, CVD event) would have been established by clinical adjudication committees or established diagnostic criteria within those long-term epidemiological studies. The qualifications of these individuals are not specified in this 510(k) summary, but they would typically involve cardiologists and other medical specialists.
4. Adjudication Method for the Test Set
Since the ground truth for clinical outcomes was based on events recorded in large epidemiological studies, there was no "adjudication method" in the sense of comparing human reads against an AI output. The lipid measurements from the NMR LipoProfile device were used as a predictor for these retrospectively determined clinical events. For the method comparison studies, the comparison was directly against existing commercial chemistry analyzer systems.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No Multi-Reader Multi-Case (MRMC) comparative effectiveness study was mentioned. The device is an in vitro diagnostic device, not an imaging AI designed to assist human readers. Its performance is measured directly against established lab methods and its predictive power for clinical outcomes.
6. Standalone Performance Study
Yes, the studies described are primarily standalone (algorithm only without human-in-the-loop performance).
- The "Method Comparison" studies (HDL-C, Triglycerides) directly compare the device's output to that of predicate devices.
- The "Clinical Performance of NMR LipoProfile test for LDL-P Measurement" studies assess the association of the device's LDL-P output with future cardiovascular events directly, not in conjunction with human interpretation.
7. Type of Ground Truth Used
- Analytical Performance (Sensitivity, Precision, Linearity, Interfering Substances): Ground truth was established internally using laboratory standards, dilutions, and spiked samples, often by comparing observed values to target values or calculated expected values.
- Method Comparison (HDL-C, Triglycerides): Ground truth was established by predicate commercial chemistry analyzer systems.
- Clinical Performance (LDL-P): Ground truth for the outcomes (CHD events, CAD, CVD events) was established by outcomes data (clinical events recorded during follow-up in the respective large-scale epidemiological studies: VA-HIT, EPIC-Norfolk, Women's Health Study).
8. Sample Size for the Training Set
The document does not specify a distinct "training set" size. The NMR LipoProfile Test involves "deconvolution of the composite signal at ~0.8 ppm" and relies on an "algorithm resident on the LipoProfile Analysis Server [which] is the foundation of the LipoProfile assay." This suggests that the algorithm was developed and likely trained using a set of data, but details about this training dataset (size, ground truth, etc.) are not provided in this 510(k) summary, as is common for in vitro diagnostic devices where the focus is often on analytical and clinical validation of the final product.
9. How the Ground Truth for the Training Set Was Established
As no specific training set is detailed, the method for establishing its ground truth is not explicitly mentioned. However, based on the technology description (deconvolution of NMR spectra), the "ground truth" for developing the algorithm would likely involve characterized lipoprotein standards and/or meticulously analyzed human samples where lipoprotein subclass concentrations were determined by established gold-standard methods or extensive biochemical profiling. These are typically developed through iterative research and development processes by the manufacturer.
Ask a specific question about this device
Page 1 of 1