(214 days)
The BC-3200 auto hematology analyzer is a quantitative, automated hematology analyzer and leukocyte differential counter to be used in clinical laboratories for In Vitro Diagnostic purpose.
The intended use of BC-3200 Auto Hematology Analyzer is to identify the normal patient, with all normal system-generated parameters, and to flag or identify patient results that require additional studies.
The BC-3200 Auto Hematology Analyzer is a quantitative, automated hematology analyzer and leukocyte differential counter for In Vitro Diagnostic Use in clinical laboratories. It is only to be used by trained medical professionals to identify the normal patient. with all normal system-generated parameters, and to flag or identify patient results that require additional studies. The analyzer provides analysis results of 16 parameters (listed below) of human blood and three histograms.
| Parameter | Abbreviation |
|---|---|
| White Blood Cell or leukocyte | WBC |
| Lymphocyte | Lymph# |
| Mid-sized cell | Mid# |
| Granulocyte | Gran# |
| Lymphocyte percentage | Lymph% |
| Mid-sized cell percentage | Mid% |
| Granulocyte percentage | Gran% |
| Red Blood Cell or erythrocyte | RBC |
| Hemoglobin Concentration | HGB |
| Mean Corpuscular (erythrocyte) Volume | MCV |
| Mean Cell (erythrocyte) Hemoglobin | MCH |
| Mean Cell (erythrocyte) Hemoglobin Concentration | MCHC |
| Red Blood Cell (erythrocyte) Distribution Width | RDW |
| Hematocrit | HCT |
| Platelet | PLT |
| Mean Platelet Volume | MPV |
| White Blood Cell Histogram | WBC Histogram |
|---|---|
| Red Blood Cell Histogram | RBC Histogram |
| Platelet Histogram | PLT Histogram |
The BC-3200 Auto Hematology Analyzer system consists of the analyzer, reagents (M-30D DILUENT, M-30R RINSE, M-30CFL LYSE, M-30E E-Z CLEANSER and M-30P PROBE CLEANSER), controls (BC-3D Hematology Control), calibrator (SC-CAL PLUS Hematology Calibrator) and accessories.
The two independent measurement methods used in this analyzer are: the Coulter method for determining the WBC, RBC, and PLT data and the colorimetric method for determining the HGB.
Acceptance Criteria and Device Performance Study for BC-3200 Auto Hematology Analyzer
This document summarizes the acceptance criteria and study findings for the BC-3200 Auto Hematology Analyzer, as described in the provided 510(k) summary.
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria detailed in the document are primarily derived from comparisons to a legally marketed predicate device (COULTER® A.T diff 2TM Analyzer) and general performance expectations for hematology analyzers. The reported device performance is presented in various tables throughout the summary.
Performance Parameters and Criteria (Derived from Predicate Device Comparison & Industry Standards):
| Performance Characteristic | Acceptance Criteria (from Predicate/Industry Std) | BC-3200 Reported Performance | Section in Document |
|---|---|---|---|
| Reproducibility | |||
| WBC (%CV) | ≤3.0% (for 6.0-15.0 x 10^3 /μL) | 0.93 - 1.85% | Table 1, 2, 3 |
| RBC (%CV) | ≤3.0% (for 3.00-6.00 x 10^6 /μL) | 0.60 - 1.76% | Table 1, 2, 3 |
| HGB (%CV) | ≤2.0% (for 12.0-18.0 g/dL) | 0.4 - 1.1% | Table 1, 2, 3 |
| MCV (%CV) | ≤3.0% (for 80.0-100.0 fL) | 0.25 - 0.62% | Table 1, 2, 3 |
| PLT (%CV) | ≤7.0% (for 200-500 x 10^3 /μL) | 1.58 - 8.39% | Table 1, 2, 3 |
| Inter-Laboratory Precision | Not explicitly stated as a numerical criterion compared to predicate. Implied to be acceptable if variability is low. | CV% ranges shown for WBC (1.20-2.35%), Gran (0.49-5.20%), Lymph (3.60-6.02%), Mid (2.86-8.84%), RBC (1.34-1.81%), HGB (1.11-1.78%), MCV (1.43-1.99%), PLT (1.89-5.56%). | Table 4 |
| Linearity | |||
| WBC | ±0.3 or ±5% (0-99.9 x 10^3 /μL) | All reported error percentages are within ±5% where absolute error is not applicable, or less than 3.94 for samples in the middle range. | Table 5 |
| RBC | ±0.05 or ±5% (0-7.0 x 10^6 /μL) | All errors are within ±5%. | Table 6 |
| HGB | ±0.2 or ±3% (0-25.0 g/dL) | All errors are within ±3%. | Table 7 |
| PLT | ±10 or ±10% (0-999 x 10^3 /μL) | All errors are within ±10%. | Table 8 |
| Carryover | |||
| WBC | ≤2.0% | 0% | Table 9, 10 |
| RBC | ≤2.0% | 0.46% (whole blood), 0% (control) | Table 9, 10 |
| HGB | ≤2.0% | 0.46% (whole blood), 0% (control) | Table 9, 10 |
| PLT | ≤2.0% | 0% | Table 9, 10 |
| Correlation to Predicate Device | Correlation coefficient (r) close to 1, slope close to 1, intercept close to 0, mean difference acceptable. Specific criteria are not explicitly stated but implied by comparison. | WBC: r=0.9994, Slope=1.0097; RBC: r=0.9971, Slope=0.9916; HGB: r=0.9982, Slope=0.9951; PLT: r=0.9961, Slope=0.8882. Other parameters also show strong correlations (mostly >0.97). | Table 11 |
| Correlation to Manual Differential | Correlation coefficient (r) close to 1. Specific criteria not explicitly stated. | Lymph%: r=0.95; Mid%: r=0.57; Gran%: r=0.94 | Table 12 |
| Ability to Flag Abnormal WBC Histograms | Not explicitly stated, but high agreement and low false negative rates are desirable. The predicate device's flagging ability would be the implicit benchmark. | Agreement (%): 82.5; False Positive Ratio (%): 10.6; False Negative Ratio (%): 45 | Table 13 |
2. Sample Sizes and Data Provenance
- Test Set Sample Sizes:
- Reproducibility (Imprecision): 11 replicate tests for each of nine samples (three low, three normal, three high concentrations).
- Inter-Laboratory Precision: Three samples (low, normal, high concentrations), each run twice on two BC-3200 devices in two different laboratories.
- Linearity: Diluted samples tested at multiple concentrations (from 0% to 100%), each concentration run twice. The exact number of initial samples before dilution is not specified.
- Carryover: High concentration sample run three times (i1, i2, i3), then low concentration sample run three times (j1, j2, j3). Repeated for high-level control.
- Correlation to Predicate Device: Ranges from 98 to 103 samples depending on the parameter (e.g., 103 for WBC, RBC, HGB, HCT, MCV, MCH, MCHC, RDW; 98 for Lymph#, Mid#, Gran#, Lymph%, Mid%, Gran%; 102 for MPV).
- Correlation to Manual Differential: 196 samples.
- Abnormal WBC Histograms Flagging: 200 samples.
- Reference Ranges: 121 donors.
- Data Provenance: The document does not explicitly state the country of origin for the data or whether the studies were retrospective or prospective. Given the submitter's location (Shenzhen, P. R. China) and the context of a 510(k) submission, it is likely that parts of the studies were conducted in China and/or possibly in collaboration with clinical sites. The studies appear to be prospective for performance characterization (e.g., reproducibility, linearity) and likely involved prospective collection of samples for correlation and flagging studies.
3. Number of Experts and Qualifications for Ground Truth
- Correlation to Manual Differential: The document implies that manual differential counts were performed by experts to establish ground truth for the 196 samples. The number of experts and their specific qualifications (e.g., years of experience as a clinical pathologist or medical technologist performing manual differentials) are not specified.
- Ability to Flag Abnormal WBC Histograms: Similarly, manual differential was used as the ground truth for flagging abnormal WBC histograms for 200 samples. The number of experts and their qualifications are not specified.
4. Adjudication Method
The document does not describe any specific adjudication method (e.g., 2+1, 3+1) used for establishing ground truth in any of the studies (e.g., manual differential counts). It is implied that a single expert or standard laboratory practice was used for manual differentials.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No Multi-Reader Multi-Case (MRMC) comparative effectiveness study was mentioned or performed. The studies focus on comparing the device's performance to a predicate device and manual methods, not on how human readers' performance improves with or without AI assistance. The device in question (BC-3200 Auto Hematology Analyzer) is an automated hematology analyzer, not an AI-assisted diagnostic tool for human readers.
6. Standalone Performance Study
Yes, standalone performance studies were extensively conducted. The entire "Performance characteristics" section (starting from [2]) describes the algorithm-only performance of the BC-3200 Auto Hematology Analyzer. This includes:
- Reproducibility: Testing the consistency of the device's measurements.
- Inter-Laboratory Precision: Assessing variability across different devices/laboratories.
- Linearity: Evaluating the device's accuracy across different concentration ranges.
- Carryover: Measuring the residual effect from a previous high-concentration sample.
- Correlation: Comparing the device's results to a predicate device and manual methods.
- Ability to flag abnormal WBC histograms: Assessing the device's automated flagging capability.
- Reference Ranges: Establishing normal ranges for the device.
All these studies demonstrate the standalone performance of the BC-3200 without human-in-the-loop performance modifications.
7. Type of Ground Truth Used
The types of ground truth used include:
- Reference values from highly controlled measurements/calibrators: For reproducibility and linearity studies, where the true value is often defined by precise dilutions or calibrated controls.
- Predicate device measurements: For correlation studies against the COULTER® A.T diff 2TM Analyzer (Table 11).
- Expert Consensus / Expert Manual Differential: For correlation of differential counts (Lymph%, Mid%, Gran%) and for evaluating the ability to flag abnormal WBC histograms (Table 12 and 13). This is typically established by trained laboratory personnel performing microscopic examination and enumeration.
- Donor Samples: For establishing reference ranges (Table 14).
8. Sample Size for the Training Set
The document does not explicitly mention a "training set" or "test set" in the context of machine learning model development. This is an automated hematology analyzer, not an AI/ML-based diagnostic system that would typically undergo distinct training and testing phases with labeled datasets in the machine learning sense. The described studies are performance validation studies which effectively act as a test set for the device's overall design and algorithm, but there isn't a separate, large, labeled dataset used specifically for "training" an AI model in the conventional sense.
9. How the Ground Truth for the Training Set was Established
As mentioned above, the device is an automated hematology analyzer utilizing the Coulter method and colorimetric method, not an AI/ML system that requires a training set in the typical sense. Therefore, the concept of establishing ground truth for a training set does not directly apply to the information provided in this 510(k) summary. The ground truth for the performance validation studies was established through the methods described in point 7.
§ 864.5220 Automated differential cell counter.
(a)
Identification. An automated differential cell counter is a device used to identify one or more of the formed elements of the blood. The device may also have the capability to flag, count, or classify immature or abnormal hematopoietic cells of the blood, bone marrow, or other body fluids. These devices may combine an electronic particle counting method, optical method, or a flow cytometric method utilizing monoclonal CD (cluster designation) markers. The device includes accessory CD markers.(b)
Classification. Class II (special controls). The special control for this device is the FDA document entitled “Class II Special Controls Guidance Document: Premarket Notifications for Automated Differential Cell Counters for Immature or Abnormal Blood Cells; Final Guidance for Industry and FDA.”