Search Results

The Advanced Radiation Therapy Systems, Inc., "Advanced Treatment Planning System" is used to plan patient treatments for radiation therapy with external photons and electrons.

The Advanced Treatment Planning System (ATPS) from Advanced Radiation Therapy Systems, Inc. (ARTS) is a software product that runs on a Silicon Graphics, Inc. UNIX Workstation in conjunction with specified accessory hardware. The ATPS provides the user the tools to easily perform patient treatment planning for the application of electron and photon radiation therapy utilizing input of patient anatomy from Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) devices. Incorporation of 3D visualization software from Advanced Visualization Systems (AVS), combined with X windows and Motif graphics, results in a graphical user interfacc that is both flexible and casy to use. See the User's Manual in Tab 5 for a more detailed description and examples of graphics. The ATPS provides both two dimensional (2D) and three dimensional (3D) dose calculation algorithms for photons and electrons.

The provided text describes a Premarket Notification [510(k)] Summary for a device called "Advanced Treatment Planning System" (ATPS) by Advanced Radiation Therapy Systems, Inc. (ARTS). This document is a regulatory submission to the FDA, asserting substantial equivalence to predicate devices, and as such, it does not contain a detailed study proving the device meets specific acceptance criteria in the way a clinical trial detailed report would.

However, based on the information provided, we can infer some aspects related to acceptance criteria and how a "study" (in this context, the 510(k) submission itself and the comparison with predicates) demonstrates compliance.

Here's an attempt to answer your questions based on the provided text, with the understanding that typical detailed study data is not present in this type of summary document:

1. A table of acceptance criteria and the reported device performance

The document does not explicitly state "acceptance criteria" with numerical targets and then explicitly report "device performance" against those targets in a table format. Instead, the core of a 510(k) submission for a treatment planning system relies on substantial equivalence to predicate devices. This implies that the device's performance, particularly its dose calculation algorithms and user interface, must be comparable or not raise new questions of safety and effectiveness compared to the predicate devices.

The "acceptance criteria" are implicitly the functional and performance benchmarks set by the predicate devices. The "reported device performance" is demonstrated through the "Predicate Comparison Table" (mentioned but not included in the provided text for "Technological Characteristics"). This table would detail how the ATPS's features and performance parameters align with or are similar to the predicate devices.

Implicit Acceptance Criteria (Inferred from 510(k) process):

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

The document does not provide details on a specific "test set" sample size or data provenance. In a 510(k) for a treatment planning system, the "test" often involves demonstrating that the dose calculation algorithms produce results consistent with established physics principles and industry standards, and that the software functions as intended across various clinical scenarios. This might involve:

• Benchmarking against publicly available phantom data: Often used to validate dose calculation accuracy.
• Comparison to predicate device outputs: Running the same patient cases on both the new device and the predicate to show comparable treatment plans.
• Internal validation cases: Cases designed by the manufacturer to test specific features and edge cases.

However, the specific numbers of such cases, their origin, or whether they were retrospective or prospective are not mentioned in this summary.

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 document does not specify the number or qualifications of experts used to establish ground truth. For a device like a treatment planning system, "ground truth" for dose calculations is typically established through:

• Physics principles and accepted dosimetry measurements: E.g., measurements in phantom tanks with ion chambers or film.
• Comparison to other validated algorithms: Using highly accurate research algorithms or established commercial systems as a reference.
• Clinical expert review: Radiation oncologists and medical physicists review generated treatment plans for clinical acceptability and reasonableness.

Given this is a 510(k) summary, these details are usually found in the full submission, not the summary.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

The document does not mention any adjudication method. This is typically relevant for studies involving human interpretation (e.g., image reading) where disagreement among experts needs to be resolved. For a treatment planning system, the "adjudication" is more about adherence to physical laws and clinical standards, and consistency with predicate device performance.

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

The document does not indicate that a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was performed. MRMC studies are primarily for devices that assist human readers in tasks like diagnosis (e.g., CAD systems for mammography). A Radiation Therapy Treatment Planning System is a tool for physicists and oncologists to create treatment plans, not primarily to "read" or interpret images in a diagnostic sense where AI offers an assistive reading benefit in an MRMC context. Therefore, this type of study would not typically be applicable or described for this device.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

The ATPS is inherently a "human-in-the-loop" device. It provides "the user the tools to easily perform patient treatment planning." While the dose calculation algorithms themselves operate "stand-alone" on computed data, the overall "performance" of the system is the combination of the algorithms and the user's ability to create an optimal plan.
Therefore, "standalone" algorithm performance would have been assessed internally (e.g., dose calculation accuracy validation), but the device itself is designed for human interaction. The 510(k) likely focused on validating the accuracy of the underlying algorithms and the usability of the interface, not a "standalone vs. human-in-the-loop" comparison in the same way as AI diagnostic tools.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

The document does not explicitly state the type of ground truth used. For a treatment planning system, the "ground truth" for validation would primarily involve:

• Physics dosimetry measurements: Using phantoms and physical detectors to verify the accuracy of calculated dose distributions.
• Reference calculations: Comparing the ATPS's dose calculations to those from well-validated and accepted algorithms or benchmark software.
• Clinical acceptability by physicists and oncologists: Reviewing treatment plans to ensure they meet clinical standards and safely deliver the prescribed dose to the target while sparing healthy tissue.

The most concrete "ground truth" for the dose calculation algorithms would be based on physical measurements and established physics models, often vetted by expert medical physicists.

8. The sample size for the training set

The document does not mention a "training set" or its sample size. The ATPS is described as a software product providing dose calculation algorithms. Such algorithms, especially those developed in the late 1990s, would typically be based on analytical models of radiation transport (e.g., Clarkson, convolution/superposition for photons; various electron algorithms) derived from fundamental physics, rather than machine learning models that require labeled training data in the modern sense. Therefore, a "training set" as understood in AI/ML contexts would not be applicable here. The algorithms are "trained" by incorporating physics principles and parameters, not by being fed a dataset.

9. How the ground truth for the training set was established

As there is no mention of a "training set" in the context of machine learning, this question is not applicable based on the provided text. The "ground truth" for the underlying physics models would be derived from fundamental scientific principles, experimental measurements in physics labs, and Monte Carlo simulations, all validated by expert physicists over decades.

Ask a specific question about this device