(53 days)
The Prelude Planning Software for the electron beam IORT treatment can be used for any malignant and benign tumor. For Prelude no limitation is given to the patient population. Local/Regional recommendations or guidelines may indicate patient who will benefit from IORT more than from other treatment modalities.
In general, since Prelude is tailored for the planning with the Mobetron®, it can be used for IORT treatment planning, if a patient is prescribed to be treated with the Mobetron®.
The Prelude software supports the IORT treatment workflow. Prelude Dosimetric measurement data of the radiation device can be displayed by selecting the machine parameters. Upon that information the user can easily plan the treatment and the software calculate the required parameters for the IORT devices.
For quality assurance the machines parameters can be recorded and visualized.
For the calculation of the output factors or the monitor units either the IAEA or AAPM protocol is followed.
The software is intended to be used by medical professionals in the area of radiation therapy.
The main purpose is to plan the technical parameters required to perform an electron beam IORT to treat both malignant and benign tumors
This document describes the MedCom GmbH Prelude planning software, K180308, for electron beam Intraoperative Radiation Therapy (IORT) treatment planning.
1. A table of acceptance criteria and the reported device performance
The provided text does not contain a specific table of acceptance criteria with corresponding device performance metrics in the format typically used for AI/ML device submissions (e.g., sensitivity, specificity, AUC, FROC analysis). Instead, the document focuses on general software safety, effectiveness, and usability assessments, emphasizing that the software meets its intended use and is safe and effective.
The performance details are described qualitatively rather than quantitatively against specific acceptance criteria. Key performance aspects reported include:
| Acceptance Criteria (Inferred) | Reported Device Performance |
|---|---|
| Software Functionality and Intended Use | Confirmed that the Prelude System meets its intended use. Provides expected parameters for simulated treatment data. |
| Usability | User could successfully create a treatment plan, defining all necessary treatment parameters (beam energy, applicator diameter, prescribed dose, etc.). Clearly understandable where to enter parameters and their impact on dose distribution. Workflow requires crucial parameters before calculation and approval. |
| Treatment Plan Approval and Storage | Complete treatment plan successfully approved by user (with sufficient rights) and saved into the database. User able to review and confirm all treatment parameters by verifying the report and accessing the plan from the database. |
| Radiation Dose Distribution Visualization | Fast dose distribution visualization with energy mixing. Based on measured beam data. Note: does not account for tissue inhomogeneities. |
| Quality Assurance (QA) Management | Integration of patient and treatment data into one platform/database facilitates data analysis and reporting. QA management features allow streamlining workflow and tracking equipment performance. |
| Safety and Risk Management | Tested software does not create any new risk. Safe and usable in clinical environment. All identified risks reduced to acceptable level. Overall residual risk acceptable. Probability of serious injury evaluated as "improbable." No issues detected that would prevent clinical use. Considered risks from similar devices. |
| Effectiveness (Comparative to Existing Procedures/Tools) | The software improves efficiency of the IORT procedure by integrating various treatment planning and QA tools. Clinical evaluation shows the system is effective and comparable to existing procedures. State of the art like other tools on the market. |
2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
- Sample Size for Test Set: The document states that "Clinical patient data was simulated." It does not specify a numerical sample size for this simulated data.
- Data Provenance: The data used for testing was "dosimetric measurement data from a Mobetron® device" and "simulated" clinical patient data. The country of origin for the simulated data or the Mobetron® device's data is not explicitly mentioned, but MedCom GmbH is located in Germany. The study appears to be a retrospective analysis of simulated data and device performance.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)
The concept of "ground truth" as typically defined for AI/ML diagnostic devices (e.g., truth established by pathology or expert consensus on a test set) is not directly applicable here in the same way. The device is a planning software, not a diagnostic one.
Instead, the "truth" or correctness of the outputs was assessed through:
- Comparison to "expected parameters" for the simulated input data.
- Usability testing with "users" and "Mobetron users and other experts in that field" who provided feedback.
- Risk assessment team included "application specialists and a medical expert besides the development team and quality managers with risk management experience." Specific numbers and detailed qualifications of these individuals are not provided, beyond stating they were "medical expert" and "experts in that field."
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
No formal adjudication method (like 2+1 or 3+1 for resolving disagreements among multiple readers) is described for a test set in the context of diagnostic decision-making. The testing involved verification that the software produced "expected parameters" for simulated data and that users could successfully create plans and approve them, and that the software assisted in optimizing the treatment delivery.
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 multi-reader multi-case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance is mentioned. The software is a planning tool, not an AI diagnostic assistant. Its purpose is to support medical professionals in radiation therapy by assisting in treatment planning and QA, not to be a diagnostic aid that would typically involve an MRMC study.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
The evaluation described is intrinsically "human-in-the-loop" as Prelude is a planning software intended to be used by medical professionals. The software assists in calculations and visualizations, but the user defines parameters, approves plans, and interprets outputs. The testing included assessing user interaction and ability to create and approve plans.
7. The type of ground truth used (expert concensus, pathology, outcomes data, etc)
The "ground truth" for evaluating this planning software was based on:
- Expected parameters: For the simulated clinical patient data, the software's calculated output parameters were compared against "expected parameters" for a treatment.
- Usability Feedback/Expert Opinion: Evaluation of whether users could successfully create treatment plans, define parameters, and whether the workflow was understandable. This suggests expert review of the software's functionality and output.
- Established Radiation Therapy Protocols: The software follows IAEA or AAPM protocols for calculation methods, implying adherence to recognized standards.
8. The sample size for the training set
The document does not describe the use of machine learning models requiring a distinct "training set." Prelude is a treatment planning software that performs calculations based on measured dosimetric data and established physical principles (IAEA/AAPM protocols), rather than a system trained on a large dataset of patient images or outcomes. Therefore, the concept of a "training set" in the context of AI/ML is not applicable here.
9. How the ground truth for the training set was established
As there is no mention of a training set for machine learning, the question of how its ground truth was established is not applicable. The software's calculations leverage "dosimetric measurement data of the radiation device," which serves as input to its algorithms based on physics principles, not as training data for a learning model.
§ 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.