(93 days)
The Senographe Essential FFDM systems generate digital mammographic images that can be used for screening and in the diagnosis of breast cancer. The Senographe Essential FFDM systems are intended to be used in the same clinical applications as traditional film-based mammographic systems.
Premium View is an image-processing algorithm, which increases the visibility of breast structures. The main advantage is to provide a single breast image, where the contrast in the fatty tissues is similar to that obtained by setting WW (window width) and WL (window level) for optimum visualization of fatty tissues, and the contrast in the fibro-glandular tissue is similar to that obtained by setting WW and WL for optimal visualization of fibro-glandular tissues.
PVi is an option that can simplify the presentation of mammographic images, improve workflow, and streamline the review process of images with very dark or bright areas by presenting the image with the WW and WL optimized for review with minimal need for the user to make adjustments for the various tissue areas. This could be especially useful with patients who have very thin breasts, very dense breasts, or implants that currently require multiple adjustments of WW and WL to review.
eContrast accommodates different customer display preferences in screening and diagnosis, with 5 different levels, while keeping the PV i option for implants. eContrast postprocessing is also available for stereo images.
The subject of this submission is a software-only option to Senographe Essential Full Field Digital Mammography (FFDM) system called eContrast. eContrast is an image post-processing algorithm that will introduce a modification to the previously approved Premium View (PV) / Premium View i (PVi) (K110798). eContrast processing will offer 6 levels of contrast strength for image viewing, where the desired combination of image sharpness, image smoothness, level of tissue penetration, and level of contrast may be selected by the radiologist. The final image appearance varies according to the selected level. eContrast is a Software only option.
The provided text describes the GE Healthcare eContrast software option, an image post-processing algorithm for Full Field Digital Mammography (FFDM) systems. The study detailed is a clinical evaluation to determine the clinical acceptance of images processed with this algorithm.
Here's a breakdown of the requested information based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state quantitative acceptance criteria or corresponding reported device performance metrics in a readily quantifiable format (e.g., specific AUC values, sensitivity, specificity thresholds, or reader improvement percentages).
Instead, the clinical evaluation's objective was to "determine if the images retrospectively processed with the new processing still provided acceptable clinical image quality." This indicates a qualitative assessment of clinical acceptance rather than a quantitative performance metric.
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size for Test Set: The evaluation was based on patients with:
- Two fatty breasts
- Two dense breasts
- Two images that were "in between" (referring to breast density).
- The total number of patients or images is not explicitly stated, but it seems to be a small, representative set (presumably 6 images or 6 sets of images representing these categories).
- Data Provenance: The images were acquired from patients classified as BIRADS 1 or 2 (which typically means normal or benign findings), ensuring a range of breast densities. The origin country is not specified, but GE Healthcare is a multinational company with a French submitter address. The study was retrospective, as it involved "retrospectively processed" images.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
The document does not specify:
- The exact number of experts (it mentions "radiologist" in the device description, implying radiologists are the end-users and implicitly the experts for this type of image review).
- Their specific qualifications or years of experience.
The evaluation's objective was about "clinical acceptance," suggesting radiologists likely made these judgments.
4. Adjudication Method for the Test Set
The adjudication method is not explicitly stated. Given the focus on "clinical acceptance" by radiologists, it's implied that radiologists reviewed the images, but how their opinions were combined or adjudicated (e.g., consensus, majority vote) is not detailed.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done
No. The study described is a "clinical image evaluation" to assess the "clinical acceptance" of images processed with eContrast. It does not mention a comparative effectiveness study involving human readers with and without AI assistance, nor does it provide an effect size for human reader improvement.
6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done
The study's objective was to determine if retrospectively processed images "still provided acceptable clinical image quality," which implies a human reviewer interacting with the processed images. Therefore, a purely standalone (algorithm-only) performance evaluation in terms of clinical decision-making was not presented for regulatory purposes. The eContrast is an image post-processing algorithm, and its output is intended for human interpretation.
7. The Type of Ground Truth Used
The ground truth used was based on clinical acceptance of image quality by (presumably) radiologists. The study did not rely on pathology, outcomes data, or a pre-established consensus on disease presence/absence, but rather on the subjective assessment of whether the processed image quality was acceptable for clinical review. The selection criteria of BIRADS 1 or 2 patients suggests that the "ground truth" for the cases themselves was their benign status, but the evaluation was on the acceptability of the image processing.
8. The Sample Size for the Training Set
The document does not specify the sample size used for training the eContrast algorithm. eContrast is described as an "image post-processing algorithm" and its "Technology" section describes specific image manipulation steps (Collimator Detection, Pseudo-log Transformation, Thickness Equalization, Dynamic Range Management, Auto-contrast), which might be rule-based or involve learned parameters. If machine learning was used for any of these steps, the training data size is not provided.
9. How the Ground Truth for the Training Set Was Established
As the document does not specify a training set or a machine learning-based approach requiring ground truth in the traditional sense (e.g., for classification tasks), it does not describe how ground truth for a training set was established. It appears to be an image processing algorithm designed to enhance image features based on predefined criteria for contrast and sharpness, rather than a diagnostic algorithm that predicts disease.
§ 892.1715 Full-field digital mammography system.
(a)
Identification. A full-field digital mammography system is a device intended to produce planar digital x-ray images of the entire breast. This generic type of device may include digital mammography acquisition software, full-field digital image receptor, acquisition workstation, automatic exposure control, image processing and reconstruction programs, patient and equipment supports, component parts, and accessories.(b)
Classification. Class II (special controls). The special control for the device is FDA's guidance document entitled “Class II Special Controls Guidance Document: Full-Field Digital Mammography System.”See § 892.1(e) for the availability of this guidance document.