Search Results

The RefleXion Medical Radiotherapy System (RMRS) is indicated for treatment planning and precise delivery of image-guided radiation therapy, stereotactic radiotherapy, or stereotactic radiosurgery for tumors or other targeted tissues anywhere in the body when radiation treatment is indicated. while minimizing the delivery of radiation to vital healthy tissue. The megavoltage X-ray radiation is delivered in a rotational, modulated, image-guided format in accordance with the physician approved plan.

The RefleXion Medical Radiotherapy System is also indicated for FDG-guided treatment which includes modeling, planning and precise delivery of FDG-guided radiation therapy, a type of Biology-guided Radiotherapy (BgRT), in five or fewer fractions for adults. It is indicated for tumor volumes in lung and bone subject to potential motion and positional uncertainty that have each been assessed with on-board PET/CT prior to delivery for adequate localization, sufficient FDG metabolic activity, local contrast and consistent biodistribution to meet the RMRS requirements, while minimizing the delivery of radiation to vital healthy tissue. BgRT involves the detection of signals from F18 during active beam delivery as a guide to deliver megavoltage X-ray radiotherapy in a rotational, modulated format in accordance with a physician approved treatment plan.

For complete fludeoxyglucose F18 prescribing information, refer both to the current medical imaging agent labeling and to this device labeling under "FDG Medical Imaging Agent Information".

The RMRS is a linear accelerator capable of delivering intensity-modulated radiation therapy, stereotactic body radiation therapy, stereotactic radiotherapy, stereotactic radiosurgery, and biology-guided radiotherapy (BgRT).

The RMRS is comprised of six major subsystems as shown in Figure 1.

The major subsystems of the RefleXion Medical Radiotherapy System are: 1) Compact linear accelerator, 2) primary collimation, 4) PET scanner, 5) kVCT, and 6) MV x-ray detectors.

The linear accelerator portion of the device was previously cleared under a 510(k) premarket notification (K190978). The subject device in K190978 included the PET hardware, but the PET hardware and associated software was inactive and outside the scope of K190978. The subject of the current submission is the addition of the BgRT functionality and the activation of the PET hardware.

Biology-guided Radiation Therapy

BgRT is a type of volumetric modulated arc therapy that allows radiation dose delivery in a slice-by-slice fashion using a multi-pass couch motion based on the collection and processing of PET data from the radiotracer fludeoxyglucose F18 (FDG). BgRT also allows the dose delivery to be adjusted in real-time to account for target motion in the lung and bone.

The PET system is comprised of two PET detector arcs mounted on the ring gantry that rotate with the system to generate complete tomographic samples of radiotracer distribution required for the BgRT algorithm.

Rapidly acquired "limited time sampled" (LTS) PET imaging data are used to form accumulated lines of response (projection data). These data are then used to guide the beam using the mapping calculated during BgRT planning, maintaining the ability to track the target in real-time as it moves within a pre-defined volume called the biology-tracking zone (BTZ).

Here's a breakdown of the acceptance criteria and study proving the device meets them, based on the provided text, structured as requested:

Acceptance Criteria and Device Performance

1. Table of Acceptance Criteria and Reported Device Performance

The provided text details several performance tests and their acceptance criteria, primarily within the "SUMMARY OF NONCLINICAL/BENCH STUDIES" and "SUMMARY OF CLINICAL INFORMATION" sections. Due to the diverse nature of these tests, a single comprehensive table is challenging. Instead, I've organized them by the type of test (Bench Studies and Clinical Studies) and extracted the relevant acceptance criteria and results.

Bench Studies: Dosimetric Accuracy and Motion Tracking

Clinical Study: Cohort II (BgRT Delivery Consistency and Workflow)

2. Sample Size Used for the Test Set and Data Provenance

• Bench Studies: The sample sizes for the bench studies are not explicitly stated in terms of a numerical count of "test sets" or cases. Instead, they describe the methodology (e.g., "large anthropomorphic phantom," "ArcCHECK Model") and the types of configurations tested (e.g., respiratory/non-respiratory motion, single step shifts, various target-to-background contrasts, edge cases, multi-target). The nature of these tests suggests a comprehensive exploration of the device's performance under simulated conditions.

  • Provenance: In silico (simulated) and phantom studies. No specific country of origin is mentioned for the bench testing.

• Clinical Studies:

  • Cohort I: 8 patients were enrolled, with 6 having evaluable images for the secondary endpoints.
  • Cohort II: 9 patients were enrolled (4 with bone tumors, 5 with lung tumors).
  • Provenance: Prospective clinical trial conducted at two sites in the U.S.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

• Bench Studies: Ground truth for bench studies was established through the controlled and measurable nature of the phantom setups and physical dosimetric measurements (film, ion chambers). There's no mention of human experts establishing ground truth in terms of image review or clinical assessment for the bench tests.

• Clinical Studies:

  • Ground Truth for Planning and Treatment decisions:

    • Site Investigator (SI): Involved in the "localization decision" and "investigator assessment" for acceptable BgRT plans. Their specific qualifications are not detailed beyond being a "site investigator."
    • Agreement Standard (AS): Used in conjunction with the SI for localization decisions. The nature or composition of the AS is not explicitly defined (e.g., whether it's a consensus of experts, a reference standard from a central review).
    • Radiation Oncologist: Approved the detailed treatment plan and decided whether radiotherapy delivery with PET guidance was warranted based on metabolic characteristics (AC and NTS).
    • Medical Physicist or Dosimetrist: Developed the detailed treatment plan and conducted patient-specific quality assurance.
    • Clinical Expertise: Decisions on BgRT plan acceptability and patient positioning were made by clinical staff.

  • "Cleared, third-party diagnostic PET/CT": This served as a comparative ground truth for evaluating PET imaging performance, implying that its established diagnostic accuracy serves as a reference.

4. Adjudication Method for the Test Set

• Bench Studies: No specific "adjudication method" involving human experts is applicable, as the ground truth was based on physical measurements and controlled simulations.

• Clinical Studies:

  • For endpoint 1 (Cohort I) and 2 (Cohort II) (localization decisions), agreement was assessed between a "Site Investigator (SI)" and an "Agreement Standard (AS)." The method for establishing the "Agreement Standard" (e.g., consensus, expert panel) is not specified. It appears to be a direct comparison against a predefined reference, rather than a multi-reader, multi-case adjudication process.
  • For endpoint 2 (Cohort I) and 3 (Cohort II) (concordance with diagnostic PET/CT), the comparison was against a "cleared, third-party diagnostic PET/CT," which serves as the established reference without explicit "adjudication" in the sense of multiple human readers resolving disagreements.
  • No formal 2+1 or 3+1 adjudication method involving multiple independent expert readers is described for the clinical study data in the provided text. Decisions and assessments appear to be made by the "Investigator" or clinical staff.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done, and what was the effect size of how much human readers improve with AI versus without AI assistance

• No MRMC comparative effectiveness study involving human readers assisting with or being assisted by the RMRS (which incorporates AI/algorithmic guidance) is described in the provided text.
• The clinical studies focus on the RMRS's performance against predefined criteria and existing diagnostic modalities, not on evaluating human reader improvement with AI assistance.

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

• The "SUMMARY OF NONCLINICAL/BENCH STUDIES" section, particularly the detailed performance metrics for motion tracking, dosimetric coverage, and gamma index evaluations, represent the standalone algorithm performance in a controlled, simulated environment. These tests involved the device delivering radiation based on its algorithms and PET data, without human real-time intervention in the delivery process, and assessing the physical outcome.
• In Cohort II, the "emulated BgRT dose distribution in silico" and the concordance between "physical and digital phantoms of emulated BgRT delivery" also reflect the standalone performance of the algorithm in generating and confirming deliverable fluence based on human patient PET emissions. While human input is involved in initial planning and approval, the core evaluation of dose delivery consistency and accuracy is algorithmic.

7. The Type of Ground Truth Used

• Bench Studies: Ground truth was established through physical measurements (e.g., film dosimetry, ion chamber measurements) in highly controlled phantom environments, along with the precise knowledge of programmed motion and simulated targets. This constitutes a manufactured or precisely controlled ground truth.
• Clinical Studies:
  • Comparative Reference: For PET imaging performance, a "cleared, third-party diagnostic PET/CT" served as a comparative ground truth for localization decisions and "plan proceed" decisions.
  • Clinical Assessment/Investigator Opinion: For whether RefleXion PET data could be used to generate an "acceptable BgRT plan," the ground truth was based on Investigator assessment.
  • Dosimetric QA Standards: For the achievement of intended dose distribution and plan quality assurance, the ground truth was defined by established radiotherapeutic quality assurance standards (e.g., gamma index meeting 3 mm/3% deviation standard, DVH metrics).
  • In Silico Emulation: For the primary endpoint in Cohort II, the consistency of the emulated dose distribution was compared against the "approved BgRT plan." This is a form of design-specific ground truth derived from the planning system's output.

8. The Sample Size for the Training Set

• The provided text describes performance testing and clinical studies for validation. It does not mention a separate training set or its sample size for the development of any AI/algorithm components within the RMRS. The information focuses solely on the validation data. It's possible the device's algorithms were developed internally or on proprietary datasets not detailed in this submission summary.

9. How the Ground Truth for the Training Set Was Established

• Since no training set and its use in algorithm development are explicitly described, there is no information provided on how its ground truth was established.

Ask a specific question about this device

The RefleXion Medical Radiotherapy System is indicated for treatment planning and precise delivery of image-guided radiation therapy, stereotactic radiotherapy, or stereotactic radiosurgery for tumors or other targeted tissues anywhere in the body when radiation treatment is indicated, while minimizing the delivery of radiation to vital healthy tissue. The megavoltage X-ray radiation is delivered in a rotational, modulated, image-guided format in accordance with the physician approved plan.

The RefleXion Medical Radiotherapy System (RMRS), a hybrid imaging-therapy system, is capable of delivering intensity-modulated radiation therapy (IMRT), stereotactic body radiation therapy (SBRT), stereotactic radiotherapy (SRT) and stereotactic radiosurgery (SRS) utilizing an on-board kilovoltage CT (kVCT) system for patient localization and field sizes based on jaws which are 40 cm wide, with slice width choices of 1.0 cm and 2.0 cm to optimize patient treatment. It is also equipped with a binary multileaf collimator (MLC) used to create intensity modulated radiation fields at a source-axis distance of 85 cm. It can achieve a nominal dose rate of 850 cGy/min.

The RMRS consists of the following clinical subsystems: 6 MV photon radiotherapy delivery, Kilovoltage (KV) X-ray CT imaging, MV X-ray detection, and treatment planning. The LINAC along with the fixed primary collimator, adjustable collimators, and multileaf collimator (MLC) mount in the therapy plane. The kVCT imaging plane is separate from the therapy plane and provides images used to localize the patient prior to treatment. Additionally, the system is designed architecturally to include PET imaging components for future product enhancements. Although the PET subsystem hardware is present in the system, its functionality is currently disabled.

Here's an analysis of the acceptance criteria and study information provided in the document:

1. Acceptance Criteria and Reported Device Performance

Notes: The document states that "Results of bench testing showed conformance to applicable requirements and specifications" for all listed performance tests. While specific numerical results are provided for high and low contrast resolution, they are presented as comparisons to the predicate device rather than explicit acceptance criteria and direct performance measurements against those criteria. However, the implicit indication is that these improved or comparable values meet expectations.

2. Sample Size Used for the Test Set and Data Provenance

The document states, "No animal studies or clinical tests have been included with this pre-market submission." This indicates that the testing was primarily bench testing and in-vitro verification and validation. Therefore, the concept of a "test set" with "data provenance" in the clinical sense (e.g., patient data from a specific country, retrospective/prospective) is not applicable here.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

Not applicable. As per point 2, no clinical test set was used that would require expert-established ground truth. The "ground truth" for the performance tests would be established through a combination of:

• Physical phantoms and measurement devices: For tests like dosimetric accuracy, isocenter coincidence, and geometric accuracy.
• Standards and specifications: The acceptance criteria themselves are drawn from industry standards and internal design specifications.

4. Adjudication Method for the Test Set

Not applicable, as no clinical test set requiring expert adjudication was used.

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

Not applicable. The RefleXion Medical Radiotherapy System is a radiation treatment and imaging system, not an AI-assisted diagnostic or interpretation tool for human readers. There is no mention of such a human reader study.

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

The performance data presented in the table are for the standalone device performance (e.g., mechanical accuracy, dosimetric accuracy, imaging performance). The device's treatment planning software is part of the system, and its accuracy is tested (e.g., "Treatment Planning System Dose accuracy tests"), indicating a standalone assessment of the algorithm's output. These results are implied to be "conforming to applicable requirements and specifications."

7. The Type of Ground Truth Used

The ground truth for the verification and validation tests was established through:

• Physical measurements and industry standards: For mechanical, dosimetric, and imaging performance tests (e.g., phantoms, dosimeters, rulers, calibrated test equipment against known values).
• Design specifications and requirements: The device's performance was measured against its predefined technical specifications and relevant international standards (e.g., IEC standards).

8. The Sample Size for the Training Set

Not applicable. The document describes a medical device (a linear accelerator and its associated systems) and its performance, not a machine learning or AI model that requires a "training set" in the context of data-driven algorithm development. While the system undoubtedly has complex software, the document focuses on its overall and component performance validation rather than the training of a predictive model.

9. How the Ground Truth for the Training Set was Established

Not applicable, as there is no "training set" discussed in this submission.

Ask a specific question about this device