(60 days)
aPROMISE is intended to be used by healthcare professionals and researchers for acceptance, transfer, storage, image display, manipulation, quantification and reporting of digital medical images. The system is intended to be used with images acquired using nuclear medicine (NM) imaging using PSMA PET/CT. The device provides general Picture Archiving and Communications System (PACS) tools as well as a clinical application for oncology including marking of regions of interest and quantitative analysis.
aPROMISE (automated PROstate specific Membrane Antigen Imaging SEgmentation) consists of a cloud-based software platform with a web interface where users can upload body scans of PSMA PET/CT image data in the form of DICOM files, review patient studies and share study assessments within a team. The software complies with the Digital Imaging and Communications in Medicine (DICOM) 3 standard.
Multiple scans can be uploaded for each patient and the system provides a separate review for each study. The review page display studies in a 4-panel view showing PET, CT, PET/CT fusion and maximum intensity projection (MIP) simultaneously and includes the option to display each view separately. The device is used to review entire patient studies, using image visualization and analysis tools for users to identify and mark regions of interest (ROIs). While reviewing image data, users can mark ROIs by selecting from pre-defined hotspots that are highlighted when hovering with the mouse pointer over the segmented region, or by manual drawing, i.e selecting individual voxels in the image slices to include as hotspots. Selected or drawn hotspots are subject to automatic quantitative analysis. The user can review the results of this quantitative analysis and determine which hotspots should be reported as suspicious lesions.
To create a report the signing user is required to confirm quality control, and electronically sign the report preview. Signed reports are saved in the device and can be exported as a JPG or DICOM file.
Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:
Important Note: The provided text is an FDA 510(k) summary, which focuses on demonstrating substantial equivalence to a predicate device rather than presenting a standalone clinical study report with detailed acceptance criteria and performance data in the typical scientific paper format. Therefore, some information, especially very specific numeric acceptance criteria and detailed study results, is inferenced or described broadly as "met predetermined acceptance criteria" because the exact values are not explicitly stated in this regulatory document.
1. Table of Acceptance Criteria and Reported Device Performance
The FDA 510(k) summary does not present specific numeric acceptance criteria (e.g., minimum accuracy percentage, specific sensitivity/specificity thresholds) or detailed reported device performance in a precise tabular format. Instead, it describes functional and analytical performance goals and states that these were met.
For the purpose of this request, I've constructed a table based on the descriptions provided in the "Bench Testing" section:
| Acceptance Criteria Category/Description | Reported Device Performance |
|---|---|
| Digital Phantom Validation Study | |
| Accuracy, Linearity, and Limit of Detection of SUV (Standardized Uptake Value) and Volume quantification against known values of a digital reference object (NEMA phantom). | All SUV and volume quantification tests of aPROMISE met their predetermined acceptance criteria. (Specific numeric thresholds for accuracy, linearity, and LOD are not provided in the document.) |
| Comparison to Predicate (KSWVWR - K160334) | |
| Equivalent performance for standard functions in marking and quantitative assessments of user-defined regions of interest in PSMA PET/CT. | Demonstrated equivalent performance when compared to the predicate KSWVWR (K160334). (Specific metrics of equivalence are not provided in the document.) |
| Analytical Performance in Clinical Study (Reproducibility & Consistency) | |
| Enables automated quantification of tracer uptake in reference organs that are more reproducible and consistent than those obtained manually by clinicians. | Demonstrated that aPROMISE enables the automated quantification of tracer uptake in reference organs that are more reproducible and consistent than those obtained manually. (Specific metrics for reproducibility and consistency are not provided in the document.) |
| Analytical Performance in Clinical Study (Sensitivity for Pre-selection) | |
| High sensitivity in pre-selection of regions of interest determined to be suspicious for metastatic disease. | Demonstrated that aPROMISE has high sensitivity in pre-selection of regions of interest that are determined to be suspicious for metastatic disease. (Specific sensitivity metric is not provided; also note the subject device states "All detected high intensity ROIs can be shown for review by the user without pre-selection," which seems to contradict the predicate's ANN pre-selection method description, but the document clarifies "showing all detected high intensity regions (no preselection as of the predicate)" as the subject device's approach while still achieving high sensitivity in detecting these regions for user review). |
2. Sample Size Used for the Test Set and Data Provenance
The document refers to a "Digital Phantom Validation Study" and an "Analytical Performance in Clinical Study."
- Sample Size for Test Set:
- Digital Phantom Validation Study: "digital reference object (NEMA phantom)" - This implies a single, well-defined digital phantom, not a large sample size of patient data.
- Analytical Performance in Clinical Study: The specific sample size (number of patients or scans) for this "clinical study" is not specified in the provided document.
- Data Provenance:
- Digital Phantom Validation Study: A digital NEMA phantom is a synthetic dataset. The document does not specify its origin.
- Analytical Performance in Clinical Study: The document refers to it as a "Clinical Study" but does not specify the country of origin of the data, nor whether it was retrospective or prospective. Given the context of a 510(k) submission and the lack of detailed clinical trial information, it is likely that this was a retrospective analysis of existing clinical data, but this is not explicitly stated.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications
- Number of Experts: The document states that the "Analytical Performance in Clinical Study" compared the performance of aPROMISE "to that of clinicians." However, the exact number of clinicians (experts) involved in establishing the ground truth or serving as comparators is not specified.
- Qualifications of Experts: The document refers to them as "clinicians." No further specific qualifications (e.g., number of years of experience, subspecialty training like nuclear medicine physician or radiologist) are provided.
4. Adjudication Method for the Test Set
The document does not describe any specific adjudication method (e.g., 2+1, 3+1, none) used for establishing ground truth from the "clinicians." It simply states that the device's performance was compared to "that of clinicians" for manual quantification and for identifying suspicious regions. It would imply that the clinicians' outputs themselves served as the comparative "ground truth" or reference for assessing the device's analytical performance.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- Was an MRMC study done? Information provided does not indicate that a formal MRMC comparative effectiveness study was performed where human readers' performance with and without AI assistance was measured. The "Analytical Performance in Clinical Study" primarily assessed the device's standalone analytical capabilities (reproducibility/consistency of quantification and sensitivity of pre-selection) compared to manual clinician performance, not an AI-assisted human workflow versus unassisted human workflow.
- Effect size of improvement: Since a formal MRMC study as described (human readers with and without AI) was not detailed, there is no information provided on the effect size of how much human readers improve with AI vs. without AI assistance.
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study
Yes, aspects of standalone performance were analyzed.
- The Digital Phantom Validation Study explicitly assessed the algorithm's accuracy, linearity, and limit of detection for SUV and volume quantification against known phantom values, which is a standalone assessment.
- The Analytical Performance in Clinical Study also evaluated the device's automated quantification of tracer uptake and its pre-selection of regions of interest, which are functions performed by the algorithm before user interaction. It specifically states "aPROMISE enables the automated quantification..." and "aPROMISE has high sensitivity in pre-selection of regions of interest..." This indicates standalone algorithmic functions were evaluated.
7. Type of Ground Truth Used
- For Digital Phantom Validation: The ground truth was known values from a digital reference object (NEMA phantom). This is a synthetic, precisely defined ground truth.
- For Analytical Performance in Clinical Study: The ground truth or reference standard for comparison appears to be the manual quantification and determination of suspicious regions by "clinicians." This implies "expert consensus" or "expert readings" served as the benchmark, though the exact process (e.g., single expert, multiple experts with consensus, adjudication) is not detailed. There is no mention of pathology or long-term outcomes data used as ground truth.
8. Sample Size for the Training Set
The document does not provide any information regarding the sample size of the training set used for developing the aPROMISE algorithms (e.g., for organ segmentation or hotspot detection). The 510(k) summary focuses on the validation of the finished device.
9. How the Ground Truth for the Training Set Was Established
Since the document does not specify the training set used, it does not describe how the ground truth for any potential training set was established.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).