(206 days)
The Philips MammoDiagnost DR is indicated for generating mammographic images that can be used for screening and diagnosis of breast cancer. The MammoDiagnost DR is intended to be used in the same clinical applications as traditional film/screen systems.
The Philips MammoDiagnost DR digital mammography system is designed to perform screening and diagnostic mammography procedures on standing, seated or recumbent patients. The system consists of a gantry with an operator console including the Eleva Workspot (EWS), consisting of an acquisition workstation (AWS) and a 19" touch screen monitor. The system's rotating X-ray tube (0.3 / 0.15 Focal Spot, Anode: Tungsten/Molybdenum) offers 4 anode tracks. The gantry provides motorized elevation, rapid rotation and compression via two foot switches. An operator radiation shield and a removable recessed patient face shield are included in the system as well as the following paddles: Compression paddles (7" x 9,5" x 12"), Axilla and Spot compression. The system is equipped with a solid state amorphous selenium detector, providing a matrix size of 3584 x 2816.
The acquisition workstation provides a high performance central processing unit with a minimum of 500GB hard disk, and a 4GB memory. A 19" flat panel touch screen color monitor on pivot mount allows the users a selectable viewing position. Acquired images are provided in DICOM MG format which allows its integration into a digital mammography environment including diagnostic review workstations, CAD Systems and PACS.
This document describes the Philips MammoDiagnost DR Full Field Digital Mammography System, detailing its performance in non-clinical and clinical evaluations.
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria for this device are not explicitly defined as numerical thresholds in the provided text, but rather through demonstration of substantial equivalence to predicate devices and acceptable image quality. The reported device performance is based on various technical measurements and a clinical image attribute review.
| Performance Metric | Reported Device Performance | Acceptance Criteria (Implied) |
|---|---|---|
| Sensitometric Response | Linear relationship between median LSB and detector entrance dose (y = 3.5877x, R-squared = 1). | Demonstrated linear and consistent response to varying X-ray exposures, comparable to predicate devices. (No explicit threshold, but perfect R-squared indicates strong performance). |
| Spatial Resolution (MTF) | Measured according to IEC 62220-1-2. Binned MTF vs. Frequency plot provided (specific values not tabulated but plotted for Binned Vertical, Binned Horizontal, and Binned Average). Note: The image descriptions show MTF values from 0.0 to 1.2 on Y-axis for frequency up to 6 LP/mm. | Performance comparable to predicate devices (Siemens Mammomat Novation and Inspiration). The MTF and DQE are stated to be "very similar" to predicate devices. |
| Noise Power Spectrum (NPS) | Measured at 8 dose levels according to IEC 62220-1-2. Horizontal and vertical NPS plots provided (specific values not tabulated but plotted). | Acceptable noise characteristics across various dose levels, comparable to predicate devices. |
| Signal-to-Noise Ratio (SNR) / DQE | Horizontal and vertical DQE measured according to IEC 62220-1-2, with plots showing DQE from 0% to 80% across various XSpec/YSpec values (uGy) for frequencies up to 6 LP/mm. | High efficiency of SNR transfer, comparable to predicate devices. The DQE is stated to be "very similar" to predicate devices. |
| Dynamic Range | DQE over dose plotted at low and mid frequencies (e.g., DQE @ 3lp/mm between 0% and 70% for Entrance Dose up to 900 microGrays, with xSpec and ySpec increasing with dose). | Sufficient signal response over a range of X-ray exposures, comparable to predicate devices. |
| Repeated Exposure Test (Lag Behavior) | Ghost image factor over time plotted (mean, average, and a limitation line at 0.30). No specific numerical result beyond the plot presented. | Acceptable lag behavior (minimal ghosting) within specified limits (implies ghost image factor below 0.30). |
| Phantom Image Tests (ACR, CDMAM) | ACR Phantom Scores: Normal Dose (Fibers 5.5, Speck Groups 4.0, Masses 4.0), High Dose (Fibers 6.0, Speck Groups 4.0, Masses 4.0) for both softcopy and hardcopy. Artifacts scored 0 in all cases. CDMAM results following EUREF guidelines and published literature. "Excellent detection capabilities" claimed. | Image quality demonstrates visibility of various features in ACR and CDMAM phantoms, comparable to or exceeding predicate device performance and meeting accreditation standards. |
| Automatic Exposure Control (AEC) | CNR in both normal and high dose contact modes changed by less than 10% per centimeter of PMMA thickness. | Maintains optimal image quality while minimizing patient dose across varying breast equivalent thicknesses, with consistent performance (less than 10% CNR change per cm PMMA). |
| Patient Radiation Dose | Mean Glandular Dose (mGy): 0.44 (2cm), 0.87 (4cm), 1.91 (6cm) for 50/50 fibroglandular/adipose tissue. | Patient dose kept as low as reasonably achievable, within acceptable limits for mammography, similar to predicate devices. |
| Clinical Image Attribute Review | All 6 mammography cases reviewed by two expert radiologists were deemed of "sufficiently acceptable quality for clinical mammographic usage" and passed the review. | Images have acceptable quality for mammographic screening and diagnostic usage, and are substantially equivalent to a predicate device as determined by an MQSA review, with a score of 1 (worse) - 5 (best) and an overall pass/fail. |
2. Sample Size Used for the Test Set and Data Provenance
- Non-Clinical Tests (Sensitometric, MTF, NPS, DQE, Dynamic Range, Repeated Exposure, AEC, Patient Dose): These tests were conducted under controlled laboratory conditions using phantoms (e.g., American College of Radiology (ACR) accreditation phantom, CDMAM contrast-detail mammography phantom, PMMA blocks for AEC and dose). No human subject data was involved in these specific tests.
- Clinical Image Attribute Review Test Set: 6 mammographic image cases.
- Data Provenance: The images were acquired from the Philips FFDM system, MammoDiagnost DR. The origin (country) of the patients is not specified, but the study implies a retrospective review of existing cases acquired on the device. Patients were "non-pregnant women over the age of 40," and "women with breast implants were not included."
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
- Clinical Image Attribute Review: 2 expert Radiologists.
- Qualifications: They met "the radiologists' qualifications specified in the Class II Special Controls Guidance Document: Full-Field Digital Mammography System." This typically implies board-certified radiologists with specific mammography training and experience, meeting MQSA requirements.
4. Adjudication Method for the Test Set
- Clinical Image Attribute Review: The review was performed by two expert radiologists. Each case received a score of 1 (worse) - 5 (best) for image attributes and an overall pass/fail. The document states, "The two expert breast imaging subspecialist radiologists agreed that the six image sets...are of sufficiently acceptable quality." This indicates a consensus or agreement model, rather than a strict 2+1 or 3+1, as only two radiologists were used and their agreement on "sufficiently acceptable quality" was the outcome. There is no mention of a third adjudicator for disagreements.
5. If a Multi Reader Multi Case (MRMC) Comparative Effectiveness Study was done
- No, a formal Multi Reader Multi Case (MRMC) comparative effectiveness study comparing human readers with AI assistance versus without AI assistance was not explicitly conducted or described. The clinical evaluation focused solely on the quality of images produced by the device when reviewed by human experts, not on the impact of AI on human reader performance.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
- This device, the Philips MammoDiagnost DR, is a full-field digital mammography system for image acquisition, not an AI algorithm for image analysis. Therefore, a standalone performance evaluation of an algorithm without human-in-the-loop is not applicable to this submission. The "performance" described relates to the imaging system's ability to produce high-quality images.
7. The Type of Ground Truth Used
- Non-Clinical Tests: Engineered physical phantoms (ACR, CDMAM, PMMA blocks) and direct physical measurements based on established standards (IEC 62220-1-2) served as the "ground truth" for objective technical performance metrics (e.g., MTF, DQE, dose).
- Clinical Image Attribute Review: Expert consensus of two MQSA-qualified breast imaging subspecialist radiologists served as the "ground truth" for assessing visual image quality attributes and overall clinical acceptability.
8. The Sample Size for the Training Set
- This submission describes a digital mammography acquisition system, not an AI-powered diagnostic algorithm. Therefore, there is no "training set" in the context of machine learning. The system's performance is based on its hardware and software design, calibrated and validated through engineering tests and clinical image review.
9. How the Ground Truth for the Training Set Was Established
- As stated above, there is no "training set" for an AI algorithm in this submission. The "ground truth" for the device's technical specifications and image quality was established through objective physical measurements using phantoms and expert human review of clinical images.
§ 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.