Micrima MARIA Data Collection for Machine Learning Study
- Conditions
- Breast Cancer
- Interventions
- Device: MARIA
- Registration Number
- NCT04894955
- Lead Sponsor
- Micrima, Ltd.
- Brief Summary
The MARIA® breast imaging system is a CE-marked radiofrequency (RF) medical imaging device. The system employs an electromagnetic imaging technique that exploits the dielectric contrast between normal and cancerous tissues. The use of artificial intelligence will provide additional, novel functionality to the MARIA® system but requires participant data in order to develop and validate the machine learning algorithms that aim to increase the accuracy and overall clinical utility of the device. This study aims to collect data from sites for this purpose.
- Detailed Description
Previous and existing trials have been focused on and designed to assess sensitivity of different versions of MARIA® when used on symptomatic subjects. This study will consider those measures, but the primary endpoint will be the collection of the data required for developing and validating machine learning-based enhancements of the MARIA® system. Data that are diverse in terms of subject characteristics, including disease state, presentation of disease, and variation of normal anatomy are requisite for the creation of accurate and well-characterized machine intelligence. As such, the recruitment criteria is broad in order to encompass all kinds of subjects that are representative as possible of the general population.
At the current stage in development of the MARIA® system a need has been identified to acquire additional data for validation of a lesion classification algorithm specific to women with larger breasts. This need has guided the study size calculation; however, additional data from all shell sizes continues to be important for further lesion classifier development and the development of additional functionality, including automated breast density assessment. Therefore, as part of this study subjects across the entire intended use population will be recruited.
The data points that are collected via this study have been selected as those that are required for the purposes of machine learning first and foremost, with any secondary endpoints utilising only existing available information. In addition to this data, a questionnaire will be provided to each participant to obtain valuable feedback on the scan experience.
All subjects will undertake the same scan procedure, requiring them to have both breasts scanned using the MARIA® system. The MARIA® scan will be in addition to any standard diagnostic procedures, including imaging, that comprise the standard of care and will have no effect on the participant's diagnosis or treatment. While there is no direct benefit nor detrimental effect from this study to the participants, the study has potentially significant implications for future learning and for the breast imaging landscape.
Recruitment & Eligibility
- Status
- UNKNOWN
- Sex
- Female
- Target Recruitment
- 400
Not provided
Not provided
Study & Design
- Study Type
- INTERVENTIONAL
- Study Design
- SINGLE_GROUP
- Arm && Interventions
Group Intervention Description Main study group MARIA -
- Primary Outcome Measures
Name Time Method Validation of MARIA breast classifier for malignant/benign tissue using clinical information 12 months The ground truth diagnosis for each participant will be obtained from clinical diagnosis information pertaining to histology and imaging. This will ultimately lead to a binary classification for that patient of either a benign or malignant diagnosis. This will then be compared to the classification that the AI has determined for that area of the breast volume.
- Secondary Outcome Measures
Name Time Method - Participant tolerance using data obtained via a questionnaire 12 months Participants asked for their opinion on the MARIA scan experience, likes and dislikes, and also comparison to other modalities.
- Proportion of participants in whom breast cancer is correctly detected using MARIA® compared with mammography, in the overall participants, and stratified by mammographic breast density (BIRADS A-D) and by histological tumour type (DCIS, IDC, ILC) 12 months Further analysis on participants identified with breast cancer, but comparing directly to mammography, and stratifying with breast density and histology.
- Proportion of participants with invasive breast cancer in whom breast cancer is correctly identified by MARIA® 12 months This will compare the number of participants that have been formally diagnosed, via biopsy, with breast cancer to those identified with breast cancer by MARIA
Trial Locations
- Locations (1)
Leeds Teaching Hospitals NHS Foundation Trust
🇬🇧Leeds, United Kingdom