Search Results

QOCA® image Smart RT Contouring System is a post-processing software intended to automatically contour DICOM CT imaging data using deep-learning-based algorithms.

Contours that are generated by QOCA® image Smart RT Contouring System may be used as input for clinical workflows including external beam radiation therapy treatment planning. QOCA® image Smart RT Contouring System must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by OOCA® image Smart RT Contouring System. The output of QOCA® image Smart RT Contouring System in the format of RTSTRUCT objects are intended to be used by radiation oncology department.

QOCA® image Smart RT Contouring System does not provide a user interface for data visualization. System settings, user settings, progress status, and other functionalities are managed via a web-based interface.

The software is not intended to automatically detect or contour lesions. Only DICOM images of adult patients are considered to be valid input.

QOCA® image Smart RT Contouring System is a post-processing software used to automatically contour DICOM CT imaging data using deep-learning-based algorithms. OOCA® image Smart RT Contouring System contouring workflow supports CT inout data and produces RTSTRUCT outputs. Contours that are generated by QOCA® image Smart RT Contouring System may be used as input for clinical workflows including external beam radiation therapy treatment planning.

The output of QOCA® image Smart RT Contouring System, in the form of RTSTRUCT objects, are intended to be used by radiation oncology department. The output of QOCA® image Smart RT Contouring System must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by QOCA® image Smart RT Contouring System.

QOCA® image Smart RT Contouring System includes the following functionality:

• Automated contouring of organs at risk (OAR) workflow
  • Input - DICOM CT
  • Output - DICOM CT (Original), DICOM RTSTRUCT
• Web-based interface of system settings, user settings, and checking progress status

QOCA® image Smart RT Contouring System is intended to be used on adults undergoing treatment that requires the identification of anatomical structures in the body considered to be OAR. QOCA® image Smart RT Contouring System is intended to be used in the head, neck, and pelvis regions.

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

Acceptance Criteria and Device Performance

Note: The reported device performance (Model Performance) shows the Dice Similarity Coefficient (DSC) as the primary metric for acceptance criteria. Hausdorff Distance 95 (HD95) is also provided as a secondary metric for model performance, but specific acceptance criteria for HD95 are not explicitly stated in the table. The text states "The subject device achieved a median DSC > 0.80," indicating an overarching criterion as well.

Study Details

1. Sample Size used for the test set and the data provenance:

   • Sample Size (Test Set): 220 cases (110 head and neck CT images and 110 pelvis CT images).
   • Data Provenance:
     • 50 cases from Taiwan (for each anatomical site, totaling 100 cases).
     • 60 cases from United States public datasets (TCIA - The Cancer Imaging Archive) (for each anatomical site, totaling 120 cases).
   • Type of Study: Retrospective performance study.
   • Independence: This test set is explicitly stated to be independent of the data used for nonclinical tests (which included training, validation, and a smaller test set).

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • The document does not explicitly state the number of experts used to establish the ground truth for the test set, nor their specific qualifications.
   • It only mentions that "Ground truth annotations were established following CT-based delineation of organs at risk in the head and neck region: DAHANCA, EORTC, GORTEC, HKNPCSG, NCIC CTG, NCRI, NRG Oncology and TROG consensus guidelines and Pelvic Normal Tissue Contouring Guidelines for Radiation Therapy: A Radiation Therapy Oncology Group Consensus Panel Atlas." This implies that the ground truth was created by human experts adhering to well-established clinical guidelines for radiation therapy contouring.

3. Adjudication method for the test set:

   • The document does not specify an adjudication method (e.g., 2+1, 3+1). It only mentions that ground truth was "established following" various consensus guidelines.

4. If a multi-reader multi-case (MRMC) comparative effectiveness study was done:

   • No, a multi-reader multi-case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance was not done. The study described is a standalone performance validation of the AI algorithm.

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

   • Yes, a standalone performance study was done. The section title is "Segmentation Performance Test" and it states, "The standalone performance of the subject device has been validated in a retrospective performance study on CT data previously acquired for RT treatment planning."

6. The type of ground truth used:

   • Expert Consensus/Clinical Guidelines: The ground truth annotations for the test set were established by human experts "following CT-based delineation of organs at risk" based on several recognized consensus guidelines (DAHANCA, EORTC, GORTEC, HKNPCSG, NCIC CTG, NCRI, NRG Oncology and TROG for Head and Neck; RTOG Consensus Panel Atlas for Pelvis).

7. The sample size for the training set:

   • Total Initial Data: 317 cases of head and neck images and 351 cases of pelvic images (total 668 cases).
   • Training Set Size: These initial cases were distributed in an "8:1:1 ratio into Training datasets, Validation datasets, and Test datasets."
     • Therefore, the training set would be approximately 8/10ths of the total initial data:
       • Head and Neck training: 0.8 * 317 ≈ 254 cases
       • Pelvic training: 0.8 * 351 ≈ 281 cases
       • Total Training Set: Approximately 535 cases (254 + 281).
   • Note: This training set data is distinct from the 220 cases used for the final standalone performance test.

8. How the ground truth for the training set was established:

   • The document states that the initial data (used for training, validation, and an internal test set) was "retrospectively collected from 2000 to 2021 from two hospitals in Taiwan".
   • It doesn't explicitly detail the process of ground truth establishment for the training set, but given the context of medical imaging for radiation therapy, it's highly implied that these contours were also created by clinical experts (e.g., radiation oncologists or dosimetrists) at those hospitals, likely following standard clinical practices. The subsequent "Segmentation Performance Test" details how ground truth for the final evaluation set was established ("following CT-based delineation... consensus guidelines"), suggesting a similar rigorous approach for the data used in training.

Ask a specific question about this device