(130 days)
WinSCAN is an automated imaging system that scans peripheral blood smears for nucleated blood cells. WinSCAN software provides for operator interaction and decision making to identify nucleated red blood cells (nRBC), and enumerate them as #nRBC/100 white blood cells (WBC).
The WinSCAN System is an automated imaging system that scans peripheral blood smears for nucleated blood cells. WinSCAN software provides for operator interaction and decision making to identify nucleated red blood cells (nRBC), and enumerate them as #nRBC/100 white blood cells (WBC). The operator selects the recommended parameters for nucleated cells, starts the scan and can walk away. The system locates and stores color images of the nucleated cells for each slide, the operator reviews the data on the WinSCAN monitor, and classifies the nucleated cells as nRBC or WBC based on stain and nuclear characteristics. Identification parameters include light adsorption, size and shape.
Automatic relocation, capture and archiving of the cell images are performed by the instrument based upon operator selection. The instrument also can be used in the manual mode to systematically scan a slide and examine each field.
The WinSCAN System consists of the following components:
- Intel-based PC with Windows operating system .
- Monitor, kevboard, mouse .
- Switchstick control unit t
- Color printer for images and texts .
- Microscope with brightfield capability and 10x, 20x, 40x and 50x . objectives
- Set of transmission and excitation filters .
- Camera and image acquisition board t
- Motorized devices: stage, filter wheels, focus drive 0
- Motor controller unit and manual controls .
The Intel-based PC operates the instrument through the software program that coordinates control of the automated microscope. The base operating system of the PC consists of MicroSoft Windows 3.1 and DOS version 6.0. When the WinSCAN icon is selected from the Program Manager, WinSCAN software is downloaded and operation of the system can begin. The WinSCAN program includes all facilities to scan, relocate, and acquire images from a microscope slide. The WinSCAN software also performs image acquisition and scan functions, as well as archiving and relocation of images and facilitates operator review of cells. Classification of cells is performed by the operator.
The Monitor. Kevboard, Mouse, and switchstick control unit provide user interfaces to the system. The monitor displays all information to the user, including menus for navigating through the WinSCAN operation, image acquisition results and allows for operator visual review for operator classification of the nucleated cells found on a slide. The keyboard allows for user entry of scan parameters, slide numbers, user identification, comments, and annotation of results. The mouse/keyboard is used to interact with the instrument and software, including selecting menus and functions of the WinSCAN software and instrument.
A color printer is included with the system that allows the user to print results of scanning or acquired images of cells.
Here's a breakdown of the acceptance criteria and study information for the WinSCAN Automated Imaging System, based on the provided text:
Acceptance Criteria and Device Performance
| Acceptance Criteria | Reported Device Performance |
|---|---|
| Correlation coefficient with manual method | >/= 0.98 |
| Precision (Coefficient of Variation) | 24% - 26% CV (compared to 23% CV for manual microscopy) |
Study Details
2. Sample sizes for the test set and data provenance:
The document does not explicitly state the sample size used for the test set. It mentions "30 samples" for precision testing but doesn't specify if this is the full test set for correlation.
Data Provenance: Not specified.
3. Number of experts used to establish the ground truth for the test set and their qualifications:
The document does not provide information on the number of experts or their qualifications used to establish ground truth for the test set. The device itself relies on "operator interaction and decision making to identify nucleated red blood cells" and "operator review," implying human classification is integral, but it doesn't specify how the ground truth for testing was established.
4. Adjudication method for the test set:
The document does not specify an adjudication method. Given that cell classification is performed by an operator even in the device's intended use, it's possible that the "manual microscopy method" against which the device was compared served as a form of ground truth or reference, but no formal adjudication process is described.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, and its effect size:
The document mentions a comparison to the "standard manual microscope method" and states that "Results of studies demonstrate that WinSCAN is an effective tool that aids the operator in locating and classifying nucleated cells." It does not explicitly describe an MRMC study designed to measure the improvement in human readers' performance with AI assistance. The study focuses on the correlation of the device's output (when used by an operator) with manual methods, and the precision of the system.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
No, a standalone study was not performed. The WinSCAN system is explicitly designed for "operator interaction and decision making" and "operator review" for cell classification. The device "aids the operator in locating" cells, but the final classification is human-driven.
7. The type of ground truth used:
The ground truth appears to be based on the "standard manual microscope method." This implies human expert classification from traditional microscopy.
8. The sample size for the training set:
The document does not specify a sample size for a training set. Given the date (1998) and the description of the device as an "automated imaging system" that aids an operator in classification rather than an autonomous decision-making AI, it's likely that the "AI" component is more geared towards image processing, cell location, and presentation, rather than complex machine learning that requires a distinct training phase as understood today.
9. How the ground truth for the training set was established:
As no training set is mentioned (or implied in a modern AI sense), the method for establishing its ground truth is not applicable/not provided. The device's "identification parameters include light adsorption, size and shape," suggesting rule-based or conventional image processing techniques, rather than supervised machine learning requiring labeled training data.
§ 864.5260 Automated cell-locating device.
(a)
Identification. An automated cell-locating device is a device used to locate blood cells on a peripheral blood smear, allowing the operator to identify and classify each cell according to type. (Peripheral blood is blood circulating in one of the body's extremities, such as the arm.)(b)
Classification. Class II (performance standards).