(224 days)
Ethos Treatment Management is indicated for use in managing and monitoring radiation therapy treatment plans and sessions.
Ethos Treatment Planning is indicated for use in generating and modifying radiation therapy treatment plans.
Ethos Treatment Management is a software product designed to help radiation therapy medical professionals manage treatments for patients with malignant or benign diseases for whom radiation therapy is indicated. It allows the physician to create and communicate radiation treatment intent (RT intent) to the treatment planner, review and approve candidate plans, and monitor treatment progress. It is intended to be used with a treatment planning system to treat or alleviate disease in humans by streamlining the treatment management and monitoring processes.
Ethos Treatment Planning is a standalone software device designed to generate and modify radiation therapy treatment plans and manage treatment sessions. The device supports the traditional and adapted treatments, in which the scheduled plan is adapted to the patient's anatomy at the time of treatment.
Here's a breakdown of the acceptance criteria and the study details for the AI segmentation models within the Ethos Treatment Management 3.0 and Ethos Treatment Planning 2.0 devices, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance (AI Model Validation for Contouring)
| Validation Characteristic | Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|---|
| Contour Quality | Minor or no adjustments needed in at least 80% of test cases. | Consistently produced contours that needed minor or no adjustments in at least 80% of the test cases. |
| Quantitative Metric (DICE coefficient) | Comparison benchmark against references published in the literature. | Used as a comparison benchmark, especially when introducing a model for a new organ. |
| Model Type | Not explicitly stated as acceptance criteria, but a characteristic of the model. | Convolutional neural networks with static weights; do not continuously learn. |
| Image Resolution Handling | Not explicitly stated as acceptance criteria, but a characteristic of the model's operation. | Operates on suitable image resolutions, patient images are resampled before inference, and label maps are sampled back onto the patient image grid. |
2. Sample Size Used for the Test Set and Data Provenance
- Test Set Sample Size: 1045 scans from various body sites.
- Full body: (part of 179 total patients)
- Head and Neck: (part of 1173 total patients)
- Thorax: (part of 600 total patients)
- Abdomen: (part of 527 total patients)
- Bowel: (part of 507 total patients)
- Pelvis: (part of 1192 total patients)
- Data Provenance: The largest number and percentage proportionally of scans originated from patients in the United States. Other country origins are not specified but implied to be varied due to "various healthcare facilities worldwide" mentioned for expert evaluation. The data appears to be retrospective, collected from patients with existing treatment indications for various cancers.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
- Number of Experts: Not explicitly stated, but referred to as "Experts" (plural).
- Qualifications of Experts: Radiation oncologists, dosimetrists, and physicists from various healthcare facilities worldwide, with "significant clinical experience in segmentation of CT imaging for the different disease sites covered by the AI models."
4. Adjudication Method for the Test Set
- Adjudication Method: Not explicitly stated as a formal adjudication method (like 2+1 or 3+1). The text mentions that "Experts... evaluated the quality of the contours across test sets to assess the need and the type of contour adjustments." This suggests a consensus-based or individual expert assessment of the AI-generated contours against their clinical judgment, but not a specific multi-reader adjudication protocol for the initial ground truth creation for the test set. For the model validation of contour quality, experts assessed the AI output.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done
- A formal MRMC comparative effectiveness study comparing human readers with AI vs. without AI assistance was not explicitly described in the provided text. The validation process involved experts evaluating the AI-generated contours to assess "the time saved on contouring tasks," which hints at an indirect measure of assistive benefit, but not a direct MRMC study.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
- Yes, a standalone performance assessment was done. The "Contouring performance undergoes rigorous evaluation through verification and validation processes." This included quantitative metrics like the DICE similarity coefficient. The validation also focused on the AI models producing contours that required "minor or no adjustments in at least 80% of the test cases," which is a metric of the AI's standalone output quality observed by experts.
7. The Type of Ground Truth Used
- Ground Truth Type: Expert consensus based on "human anatomy experts" following "RTOG and DAHANCA clinical guidelines." Pathology or outcomes data were not mentioned as ground truth for segmentation.
8. The Sample Size for the Training Set
- Training Set Sample Size: 4769 scans from various body sites.
- Full body: (part of 179 total patients)
- Head and Neck: (part of 1173 total patients)
- Thorax: (part of 600 total patients)
- Abdomen: (part of 527 total patients)
- Bowel: (part of 507 total patients)
- Pelvis: (part of 1192 total patients)
9. How the Ground Truth for the Training Set Was Established
- Ground Truth Establishment for Training Set: "Ground truth annotations were established by human anatomy experts as part of the algorithm development following RTOG and DAHANCA clinical guidelines. A single set of contours was produced for each training image. These clinical experts have significant clinical experience in segmentation of CT imaging for the different disease sites covered by the AI models. To ensure accuracy, contour definitions available in contouring guidelines are established prior to contouring tasks."
§ 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.