(237 days)
The Tosoh Automated Glycohemoglobin Analyzer HLC-723G8 is intended for in vitro diagnostic use for the quantitative measurement of % hemoglobin A Ic (HbA 1c) (DCCT/NGSP) and mmol/mo hemoglobin A1c (IFCC) in whole blood specimens. This test is to be used as an aid in diagnosis of diabetes and as an aid in identifying patients who may be at risk for developing diabetes.
The Tosoh Automated Glycohemoglobin Analyzer HLC-723G8 is an automated High Performance Liquid Chromatography (HPLC) system that separates and reports stable A 1c (sA1c) percentage in whole blood. The operational portion of the G8 is composed of a sampling unit, liquid pump, degasser, column, detector, microprocessors, sample loader, floppy disk drive unit, operation panel and a printer. The Tosoh Automated Glycohemoglobin Analyzer HLC-723G8 uses non-porous ion- exchange high performance liquid chromatography (HPLC) for rapid, accurate and precise separation of the stable form of HbA Ic from other hemoglobin fractions. The G8 uses a cation exchange column and separates the usual hemoglobin components in the blood into six fractions, A 1a, A 1b, F, LA 1c, sA Ic, and A0. The separation is done by eluting the hemoglobins from the column with a stepwise elution of three elution buffers containing different salt concentrations. The result report includes a sample ID, date, percentage and retention time of each fraction, sA1c percentage and total A1 percentage (A1a + A1b + sA1c), along with a chromatogram of the elution pattern of the hemoglobin fractions. If a sample contains a hemoglobin variant, the column elutes the material depending upon its charge.
Here's a breakdown of the acceptance criteria and the study details for the Tosoh Automated Glycohemoglobin Analyzer HLC-723G8, based on the provided 510(k) summary:
Acceptance Criteria and Device Performance
| Criterion | Acceptance Criteria (Stated or Implied) | Reported Device Performance |
|---|---|---|
| Linearity / Reportable Range | Previously established for 4.0% - 16.9% HbA1c (from predicate device comparison table) | 4.0% - 16.9% HbA1c |
| Interference | Variance greater than assigned value x 1.00 ± 5% for potential interferents. No interference from HbA2, HbS, HbC, HbD. Flag for HbE. | No Significant Interference Observed at Specified Concentrations: |
| Acetylated Hb: 50 mg/dL | ||
| Albumin: 5000 mg/dL | ||
| Aldehyde Hb: 25 mg/dL | ||
| Ascorbic Acid: 25 mg/dL | ||
| Carbamylated Hb: 25 mg/dL | ||
| Bilirubin C: 21 mg/dL | ||
| Labile Hb: 1000 mg/dL | ||
| Lipemia: 1000 mg/dL | ||
| Rheumatoid Factor: 550 IU/mL | ||
| Bilirubin F: 18 mg/dL | ||
| Hemoglobin Variant Interference: | ||
| HbS: No interference (at ~6% and ≥8% HbA1c levels) | ||
| HbC: No interference (at ~6% and ≥8% HbA1c levels) | ||
| HbE: Yes, interference. A flag is displayed, and %HbA1c result is not reported. | ||
| HbD: No interference (at ~6% and ≥8% HbA1c levels) | ||
| HbF: No interference (at ~6% and ≥8% HbA1c levels) | ||
| HbA2: No interference (at ~6% and ≥8% HbA1c levels) | ||
| Precision | Not explicitly stated as a numerical criterion for this device, but predicate device states "equal to or less than 1.5% for concentrations in the range of 5.3% to 12.1% HbA1c" for between-analyzer and between-lot. For this device, "0.7 - 1.8% (total precision) and 0.3 - 0.9% (within run, precision)" is mentioned in the comparison table as the performance of the new device. | Within-run Precision (CV%): Ranges from 0.18% (12% D1) to 1.00% (5% A2). |
| Total Precision (CV%): Ranges from 0.68% (6.5% B1) to 1.83% (5% A1). | ||
| Correlation (Method Comparison) | r ≥ 0.991; Y ≈ 1.0x + 0.0y (implied from predicate, which showed Y = 1.015x + 0.015) | Compared to Primus ultra2: |
| Slope (Deming): 0.996 (0.971 to 1.021) | ||
| Intercept (Deming): 0.073 (-0.110 to 0.255) | ||
| Correlation Coefficient (R): 0.991 | ||
| Bias: 0.047 (0.657%) | ||
| Total Error (TE) | %TE of ≤6% is acceptable. Goal at 6% for (TE)95 (e.g., 0.30 at 5.0% Decision Level, 0.72 at 12.0% Decision Level) | %TE (at 5%, 6.5%, 8%, 12%): 5.8%, 2.8%, 3.0%, 3.1% respectively (for the first method of calculation) |
| (TE)95 (at 5%, 6.5%, 8%, 12%): 0.23, 0.13, 0.20, 0.34 respectively (for the second method of calculation) |
Study Details for the HLC-723G8
The provided document describes the performance characteristics and studies conducted for the Tosoh Automated Glycohemoglobin Analyzer HLC-723G8.
-
Sample size used for the test set and the data provenance:
- Interference Studies:
- Various numbers of samples for each potential interferent, tested at two HbA1c levels (e.g., Acetylated Hb, Albumin, etc., were tested at certain concentrations, some up to 1000 mg/dL to 5000 mg/dL).
- Hemoglobin Variant Interference: 10 samples for each variant (HbS, HbC, HbE, HbD, HbF, HbA2), tested at two HbA1c levels (approximately 6% and ≥8%).
- Data Provenance: Not explicitly stated, but clinical samples are implied as "known concentrations of HbA1c" were spiked. The samples are likely clinical specimens from an unspecified region, retrospectively used for these studies.
- Precision Study: Four concentrations of unaltered EDTA whole blood specimens. Tested with 3 lots and 3 instruments.
- Each specimen/reagent combination had 2 replicates in a single run, 2 times a day for 20 nonconsecutive days, totaling 40 runs and 80 determinants per data set. This implies a large number of measurements, but the exact number of unique patient blood specimens for the "four concentrations" is not specified.
- Data Provenance: Unaltered EDTA whole blood specimens. Likely clinical samples, but country of origin/retrospective/prospective is not specified.
- Correlation (Method Comparison) Study:
- Sample Size: 120 unaltered EDTA whole blood specimens.
- Data Provenance: "Frozen specimens were utilized for this study." Implies retrospective samples, but country of origin is not specified.
- Sample Distribution: The 120 samples covered a range of HbA1c levels:
- ≤ 5%: 5 samples (4.2%)
- 5 – 6%: 15 samples (12.5%)
- 6 – 6.5%: 30 samples (25.0%)
- 6.5 – 7%: 30 samples (25.0%)
- 7 – 8%: 20 samples (16.7%)
- 8 – 9%: 10 samples (8.3%)
-
9%: 10 samples (8.3%)
- Interference Studies:
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- For this in vitro diagnostic device, "ground truth" is established by a reference method/device, not human experts.
- Interference Study: Ground truth was based on "known concentrations of HbA1c" (likely measured by a standardized reference method) and evaluated against a "variance greater than the assigned value x 1.00 ± 5%".
- Hemoglobin Variant Interference Study: Ground truth was established by "reference values" for %HbA1c and IFCC mmol/mol, likely from a reference method.
- Correlation Study: The comparative method was the Primus ultra2, analyzed by a "secondary NGSP reference laboratory." The Primus ultra2 is an established HPLC system for HbA1c. The qualification of individuals operating this reference laboratory is not specified, but its designation as an "NGSP reference laboratory" implies adherence to high standards for HbA1c measurement.
-
Adjudication method (e.g. 2+1, 3+1, none) for the test set:
- Not applicable in this context. Adjudication by multiple human experts is typically used for subjective assessments (e.g., image interpretation). For quantitative diagnostic devices, performance is typically assessed against a 'gold standard' reference method.
-
If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:
- Not applicable. This is an automated in vitro diagnostic device, not an AI-assisted human reader system.
-
If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
- Yes, this is a standalone device. The entire submission describes the algorithm-only performance of the Automated Glycohemoglobin Analyzer HLC-723G8. It is designed to quantitatively measure HbA1c without human intervention in the measurement process itself, beyond sample loading and general operation.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc):
- The ground truth for performance evaluation was based on reference methods and certified standards.
- For correlation/comparison, the Primus ultra2 (analyzed by an NGSP reference laboratory) served as the comparative method.
- The device's assigned HbA1c values are certified via the National Glycohemoglobin Standardization Program (NGSP) and are traceable to both the International Federation of Clinical Chemistry (IFCC) and the Diabetes Control and Complications Trial (DCCT). This indicates that the ground truth is based on established and internationally recognized standardization programs and reference measurement procedures for HbA1c.
- The ground truth for performance evaluation was based on reference methods and certified standards.
-
The sample size for the training set:
- The document does not explicitly mention a "training set" in the context of machine learning or AI. This is a traditional HPLC system. Its development likely involved internal optimization and validation studies, but these are not described as distinct "training" and "test" sets in the way an AI/ML algorithm would. The linearity of the assay was "previously established" under a prior 510(k) (K071132).
-
How the ground truth for the training set was established:
- As a traditional HPLC system, there isn't a "training set" in the common understanding of AI/ML. The device's calibration and method development would rely on known standards and reference materials, traceable to NGSP/IFCC, to ensure accurate and reproducible measurements. The "ground truth" for developing such a system would involve calibrated reference materials with known HbA1c concentrations.
§ 862.1373 Hemoglobin A1c test system.
(a)
Identification. A hemoglobin A1c test system is a device used to measure the percentage concentration of hemoglobin A1c in blood. Measurement of hemoglobin A1c is used as an aid in the diagnosis of diabetes mellitus and as an aid in the identification of patients at risk for developing diabetes mellitus.(b)
Classification. Class II (special controls). The special controls for this device are:(1) The device must have initial and annual standardization verification by a certifying glycohemoglobin standardization organization deemed acceptable by FDA.
(2) The premarket notification submission must include performance testing to evaluate precision, accuracy, linearity, and interference, including the following:
(i) Performance testing of device precision must, at a minimum, use blood samples with concentrations near 5.0 percent, 6.5 percent, 8.0 percent, and 12 percent hemoglobin A1c. This testing must evaluate precision over a minimum of 20 days using at least three lots of the device and three instruments, as applicable.
(ii) Performance testing of device accuracy must include a minimum of 120 blood samples that span the measuring interval of the device and compare results of the new device to results of a standardized test method. Results must demonstrate little or no bias versus the standardized method.
(iii) Total error of the new device must be evaluated using single measurements by the new device compared to results of the standardized test method, and this evaluation must demonstrate a total error less than or equal to 6 percent.
(iv) Performance testing must demonstrate that there is little to no interference from common hemoglobin variants, including Hemoglobin C, Hemoglobin D, Hemoglobin E, Hemoglobin A2, and Hemoglobin S.
(3) When assay interference from Hemoglobin F or interference with other hemoglobin variants with low frequency in the population is observed, a warning statement must be placed in a black box and must appear in all labeling material for these devices describing the interference and any affected populations.