(160 days)
Segasist P-AC contouring software is a standalone software application for Windows platforms that assists clinicians in generating estimates of the anatomy boundary contours of the prostate gland in Computed Tomography (CT) scans, Magnetic Resonance (MR) images and ultrasound (sonography) scans to aid in patient diagnosis, treatment planning and post-treatment monitoring. The software is intended to be used to provide clinicians with tools to efficiently contour/delineate the prostate gland in volume data and save the results in DICOM and BMP format. The clinician has the ability to use the saved contours directly or import them in other software tools to perform the task at hand.
The clinician retains the ultimate responsibility for making the pertinent diagnosis and patient management decisions based on their standard practices and visual comparison of the individual images. The Segasist P-AC software tool is a compliment to manual contouring techniques.
Segasist P-AC (Segmentation Assistant for Prostate – Auto-Contouring) is a standalone atlas-based segmentation software tool for auto-contouring of the prostate gland from different input image modalities (Computed Tomography (CT) scans, Magnetic Resonance (MR) images, ultrasound scans). The software can read, write and display DICOM images from/to local directories, and offers the possibility of defining regions of interest (ROls) around the prostate gland in order to delineate the prostate for contouring, visual assessment, and size and volume calculation purposes, either manually, or via semi-automated or automated processes.
The Segasist P-AC software is a tool that has been designed and developed to assist clinicians (radiologists, oncologists, medical physicists etc.) in performing contouring/delineation of the prostate gland in images in multiple modalities more efficiently. The software is capable of segmenting the prostate gland in individual slices, in a choice of different modalities, and for any given view (axial, sagittal, or coronal). This is done by requiring some user input (clicks or drawing ROIs).
For volume prostate data. Segasist P-AC calculates the prostate volume in cubic centimeters and displays the contours on each slice. The results (contours) can be saved as DICOM or binary images (BMP), which can be edited/modified at any time, completely dismissed or accepted and saved by the end user.
The efficiency of contouring performed by the Segasist P-AC software may be improved by generating/using an advanced atlas using gold standard images created by the experienced clinician(s). This requires the software to be trained (atlas creation) before being used. The software can be delivered pre-trained with the comprehensive atlas or the end user can generate their own atlas; a well-established practice for atlas-based segmentation software products.
Segasist P-AC also offers a built-in editor, enabling the user to edit, modify, or change the extracted prostate boundaries to their desired configuration based on their medical and clinical knowledge and experience. The results provided by the Segasist P-AC software needs to be approved by the experienced clinician and can always be modified or corrected by him/her. In addition, the end user can delineate the prostate gland manually using the P-AC software, if necessary or desired. As a result, when Segasist P-AC generates a result, the expert user always has the final decision to override the software result, if deemed appropriate in his/her clinical judgment. It is up to the expert user to accept the result without any change, reject it completely and delineate manually, or modify the Segasist P-AC result and then save it. The software does not provide any auto-detection or auto-saving functionalities. Regardless of the accuracy of Segasist P-AC result. it is always the experienced clinician that remains the decision maker regarding the acceptability of the computed segmentation. Therefore, the final decision on diagnosis, treatment and overall management of the patient is not based on the software result.
Segasist P-AC software does not alter the original input images of the prostate gland. nor does it change the final results obtained once approved by the clinical expert.
Segasist P-AC offers several features and functionalities such as, but not limited to:
- . Import/Export DICOM images
- . Saving Contours to DICOM or BMB format
- . Semi-automated Segmentation
- . Auto-Segmentation (fast slice-to-slice auto-segmentation with minimal user interactions)
- . Volume Segmentation and measurement
- . Edge Enhancement (contour enhancement by user controlled edge snapping).
- . Standard Functionalities for Image Visualization (windowing, contrast, brightness, zoom, panning etc)
- . Advanced Functionalities for Contour Editing For Manual Segmentation (drawing, inflating, deflating, shifting, cut & add etc)
- . User access to modify the resulting contours at any time
Here's an analysis of the Segasist Prostate Auto-Contouring Software based on the provided document:
1. Table of Acceptance Criteria and Reported Device Performance
The provided 510(k) summary does not explicitly state quantitative acceptance criteria or specific reported device performance metrics (e.g., Dice score, Hausdorff distance, sensitivity, specificity) for the Segasist P-AC software. The document focuses on demonstrating substantial equivalence to predicate devices and describes the software's functionalities and validation approach.
However, based on the text, the implicit acceptance criteria are that the device "works as intended" and "was acceptable for clinical use, and did not introduce any new concerns of safety or effectiveness compared to predicate products or manual contouring of the prostate gland." The performance is reported as meeting these general criteria.
| Acceptance Criteria (Implicit) | Reported Device Performance |
|---|---|
| Works as intended (i.e., accurately contours the prostate gland with user input across various modalities). | "This bench testing deemed that the Segasist P-AC software works as intended..." The software successfully segments the prostate gland in individual slices, calculates volume, and allows for editing. |
| Acceptable for clinical use. | "...was acceptable for clinical use..." The software is designed to assist clinicians in generating anatomy boundary contours for diagnosis, treatment planning, and monitoring, with the clinician retaining ultimate responsibility. |
| Does not introduce new concerns of safety or effectiveness compared to predicate products or manual contouring. | "...and did not introduce any new concerns of safety or effectiveness compared to predicate products or manual contouring of the prostate gland." Substantial equivalence to predicate devices is claimed. |
| Output can be edited/modified/overridden by the end user (clinician). | "The results provided by the Segasist P-AC software needs to be approved by the experienced clinician and can always be modified or corrected by him/her." "the expert user always has the final decision to override the software result." |
| Compatible with DICOM and BMP formats for import/export. | "The software can read, write and display DICOM images... The results (contours) can be saved as DICOM or binary images (BMP)." |
| Functions across CT, MR, and ultrasound modalities. | "standalone atlas-based segmentation software tool for auto-contouring of the prostate gland from different input image modalities (Computed Tomography (CT) scans, Magnetic Resonance (MR) images, ultrasound scans)." |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size for Test Set: The document states that "imported prostate image datasets from the various imaging modalities were used as input for testing of the software functionalities in accordance with the software validation/verification plans." It also mentions "sufficient numbers to support the intended use of the device." However, a specific number for the test set sample size is not provided.
- Data Provenance: The document does not explicitly state the country of origin for the image data used in testing. It only mentions "imported prostate image datasets." It implies these were retrospective clinical test cases, as it refers to "bench testing using imported images from the various imaging modalities" and "clinical test cases."
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
- Number of Experts: The document details the software development process involved "the input and collaboration of experienced and trained professionals, such as radiologists, oncologists or other highly qualified medical clinicians." For the independent testing, it states "independently by experienced and trained medical professionals who are representative of the commercial end users." However, a specific number of experts used to establish ground truth for the test set is not provided.
- Qualifications of Experts: The experts involved in development and testing are described as "radiologists, oncologists or other highly qualified medical clinicians that are proficient in reading, evaluating and interpreting images of the prostate produced by MR, CT or ultrasound devices." Specific years of experience are not mentioned, but the description emphasizes their proficiency and experience.
4. Adjudication Method for the Test Set
The document does not explicitly describe an adjudication method (like 2+1 or 3+1). It states that the "input was captured in a written and approved Software Requirement Specifications Document" and testing was conducted "both internally at Segasist Technologies and independently by experienced and trained medical professionals." The final contours generated by the software are subject to review and modification by the clinician, indicating a human-in-the-loop approach where the expert has the final say. However, for the initial ground truth used to evaluate the algorithm itself, the method is not detailed.
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, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done as described in this document. The submission focuses on demonstrating substantial equivalence and the software's functional validation, rather than an explicit comparative effectiveness study showing improvement with AI assistance for human readers using quantitative metrics (e.g., sensitivity, specificity, reading time reduction). The device is positioned as a "compliment to manual contouring techniques," implying assistance, but without a formal study to quantify the improvement.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
Yes, a standalone performance assessment was implicitly done. The document describes "bench testing" using "imported images" to determine if the "Segasist P-AC software works as intended." While the intended use involves human oversight and modification, the initial evaluation of the software's ability to generate contours (before human intervention) constitutes a form of standalone performance assessment. The "Segasist P-AC calculates the prostate volume in cubic centimeters and displays the contours on each slice," representing its standalone output.
7. The Type of Ground Truth Used (Expert Consensus, Pathology, Outcomes Data, etc.)
The ground truth for the test set appears to be expert consensus or expert-defined contours. The document states that "input and collaboration of experienced and trained professionals" guided development, and "independent... medical professionals" tested the software using "clinical test cases." The creation of "an advanced atlas using gold standard images created by the experienced clinician(s)" further supports that expert-defined contours were used as the reference. There is no mention of pathology or outcomes data being used as ground truth for contouring accuracy directly, though the software's use is indicated for diagnosis and treatment planning, where such data would eventually be relevant.
8. The Sample Size for the Training Set
The document mentions that the software "can be delivered pre-trained with the comprehensive atlas or the end user can generate their own atlas." Training involves "atlas creation." However, the specific sample size used for the pre-trained comprehensive atlas (the training set) is not provided.
9. How the Ground Truth for the Training Set Was Established
The ground truth for the training set (the "comprehensive atlas" or user-generated atlases) is established by experienced clinician(s). The document states, "The efficiency of contouring performed by the Segasist P-AC software may be improved by generating/using an advanced atlas using gold standard images created by the experienced clinician(s)." This indicates that human experts manually contoured the images that form the basis of the atlas used for training.
§ 892.5050 Medical charged-particle radiation therapy system.
(a)
Identification. A medical charged-particle radiation therapy system is a device that produces by acceleration high energy charged particles (e.g., electrons and protons) intended for use in radiation therapy. This generic type of device may include signal analysis and display equipment, patient and equipment supports, treatment planning computer programs, component parts, and accessories.(b)
Classification. Class II. When intended for use as a quality control system, the film dosimetry system (film scanning system) included as an accessory to the device described in paragraph (a) of this section, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 892.9.