(62 days)
EndoSize is a software solution that is intended to provide Physicians and Clinical Specialists with additional information to assist them in reading and interpreting DICOM CT scan images of structures of the heart and vessels.
EndoSize enables the user to visualize and measure (diameters, lengths, volumes, angles) structures of the heart and vessels.
Indications for Use:
EndoSize enables visualization and measurement of the heart and vessels for preoperational planning and sizing for cardiovascular interventions and surgery, and for postoperative evaluation.
General functionalities are provided such as:
- Segmentation of cardiovascular structures
- Automatic and manual centerline detection
- Visualization of CT scan images in every planes, 2D review, 3D reconstruction, Volume Rendering, MPR, Stretched CMPR
- Measurement and annotation tools
- Reporting tools
EndoSize is a stand-alone software application that runs on any standard Windows or Mac OSX based computer. It enables Physicians and Clinical Specialists to select patient CT scan studies from various data sources, view them, and process the images thanks to a comprehensive set of tools. EndoSize is intended to provide a clinical decision support system during the preoperative planning of endovascular surgery.
EndoSize contains five modules dedicated to different types of endovascular interventions, EndoSize EVAR, EndoSize FEVAR, EndoSize TEVAR, EndoSize TAVI and EndoSize Peripheral. These modules can be marketed in combination or as separate solutions. It is also possible to market custom versions of EndoSize to Stent manufacturers, based on the modules listed above. The differences between EndoSize and a custom version of EndoSize (user interface, manufacturer logo, manufacturer stent catalogue in the software, optional features of a generic module), do not modify neither the functioning nor the safety of the software.
One custom version of EndoSize is marketed under the trademark "Intelix". This version includes the modules Intelix AFX and/or Intelix AFX2 and/or Intelix NELLIX which are customized versions of EndoSize EVAR module for specific endografts.
EndoSize enables assessment and measurement of different vascular structures such as vessels, valves, aneurysms, and other anomalies. It provides simple to assess the feasibility of endovascular procedures. EndoSize can combine 2D scan slices into comprehensive 3D models of the patient, and can display supporting DICOM CT scan data. The software accurately represents different types of tissue, making it easier to diagnose anomalies in scans. It works with DICOM CT scan images and can access multiple DICOM data files and PACS server.
The provided document does not detail specific acceptance criteria or a comprehensive study demonstrating that the device meets these criteria in a quantitative manner. Instead, it is a 510(k) summary for a software update (EndoSize version 3.1, K160376), asserting substantial equivalence to a previously cleared device (EndoSize version 3.0, K141475).
The document focuses on:
- Product description and intended use: EndoSize is software for visualizing and measuring heart and vessel structures from CT scans for pre-operational planning, sizing, and post-operative evaluation in cardiovascular interventions and surgery.
- Changes from predicate device: Primarily updates to catalogs, minor UI/UX improvements, and new tools like calcium estimation, C-arm angle recording, and NASCET value calculation, which are based on existing functionalities.
- Performance Data (Conformance): The device is stated to conform to DICOM standards, ISO 14971 (risk management), and IEC 62304 (software life-cycle).
- Bench Testing: It states that every specification is validated by bench tests, including importation, patient management, display, processing, module functioning, measurement, and report creation/exportation. Any modifications undergo the same bench testing and regression testing.
Therefore, many of the requested details about acceptance criteria, specific study design, sample sizes, expert involvement, and ground truth establishment are not present in this regulatory submission document. This type of 510(k) relies on the argument of substantial equivalence, where the new features are described as based on existing, cleared technology, and the changes do not raise new questions of safety or effectiveness. The "performance data" section focuses on software development life cycle processes and internal validation via bench testing, rather than a formal clinical performance study with defined acceptance criteria and human readers.
Here's a breakdown of what can be inferred or is explicitly stated from the document regarding your questions:
1. A table of acceptance criteria and the reported device performance
Not explicitly provided. The document states:
- "Every specification of the EndoSize software is validated by a bench test before release."
- "Bench testing includes: Tests of Importation of DICOM images, Patient Manager tests, Tests of image display and processing, Functioning tests of the different modules..., Measurement tests, Reports creation and exportation tests."
- "Every modification to the EndoSize software is validated by the same bench testing as described above."
This implies that acceptance criteria are defined for these bench tests, but the specific numerical or qualitative criteria and the results (e.g., "99% of images imported successfully," "measurements were within X% of ground truth") are not publicly disclosed in this summary. The "reported device performance" is essentially that it "successfully undergone bench testing" and "performs as well or better than the predicate device."
2. Sample sizes used for the test set and the data provenance
Not specified. The document mentions "bench testing" which usually refers to internal laboratory testing on a set of pre-defined test cases, not necessarily a large patient image dataset for clinical validation. The provenance of any data used for these internal bench tests (e.g., country of origin, retrospective/prospective) is not disclosed.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
Not specified. Given the nature of a software update (special 510(k)) and the focus on "bench testing" of software functionalities, it's highly probable that ground truth for internal validation was established by software engineers and potentially clinical experts employed by the manufacturer, but no details on their number or qualifications are provided.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set
Not specified. This level of detail on ground truth establishment is not present.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
No, an MRMC study was not detailed or performed for this submission. The device is not presented as an AI-assistant for human interpretation in the sense of a diagnostic aid that changes reader performance. It's an image processing and measurement tool. The primary purpose of this 510(k) (a 'Special 510(k)') is to demonstrate substantial equivalence of updated software, not to prove clinical utility with a new type of performance study.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Not explicitly detailed as a formal study for this submission. The document describes the software's capabilities (segmentation, centerline detection, measurements, etc.) and states these were validated via "bench testing." This implies internal, automated, or semi-automated tests of the algorithms' outputs, which aligns with "standalone" performance, but not in the context of a rigorous, independently assessed clinical study.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
Not specified. For internal bench tests on image processing and measurement software, ground truth likely involves:
- Known input data: Using CT scans with pre-defined anatomical structures or simulated data where "true" measurements are known.
- Manual measurements: Highly precise, manually performed measurements by trained personnel (possibly clinical experts) on images to serve as a reference for comparison with software's automated measurements.
8. The sample size for the training set
Not specified. This document focuses on the software update and its validation (bench testing), not the development of a machine learning model, so a "training set" in that context is not discussed. If any machine learning components were present, the original K141475 submission might have briefly touched upon it, but this document contains no information.
9. How the ground truth for the training set was established
Not applicable/Not specified. As there's no mention of a machine learning training set for this specific submission, this information is not provided.
§ 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).