Please use this identifier to cite or link to this item:
https://www.um.edu.mt/library/oar/handle/123456789/107733
Title: | Exploring radiologists’ perception and use of artificial intelligence software in mammography |
Authors: | Vella, Cheryl (2022) |
Keywords: | Radiologists -- Malta Artificial intelligence Breast -- Radiography -- Malta Neural networks (Computer science) -- Malta Deep learning (Machine learning) -- Malta |
Issue Date: | 2022 |
Citation: | Vella, C. (2022). Exploring radiologists’ perception and use of artificial intelligence software in mammography (Master’s dissertation). |
Abstract: | Purpose: To explore and establish radiologists’ perception on the use of artificial intelligence (AI) software in mammography and to determine whether their perceptions differ after using an AI software in clinical practice. Methodology: The first phase involved a cross- sectional survey amongst local and European radiologists. The questionnaire used a Likert scale to evaluate the current use, perceived benefits and limitations of using AI mammography software in clinical practice. In the second phase, the local association was asked to send another invite to the local radiologists so as to invite those interested to an information and demonstration session about AI software provided by a local vendor. Interested radiologists willing to participate in the second phase had the opportunity to use the AI simulator remotely, following which they were once again asked to complete the identical questionnaire sent out in the first phase. Results: A total of 25 radiologists interpreting mammograms in Europe completed the questionnaire, of whom 7 worked in Malta. Forty percent (n=10) of the radiologists had utilised AI to interpret mammograms for at least 25% of the mammograms in their practices, while none of the local radiologists used AI. According to the questionnaire the radiologist found AI mostly useful in interpreting screening mammograms (mean rating score of 5.68), reducing the reporting time (5.12) and improving the accuracy of breast cancer detection, while liability issues (3.58) were identified as the main barrier. After trying out an AI software the mean value of the local radiologists on using AI as a triage tool in mammography improved from 4.43 to 5.33 however this change was not significant as the p- value was 0.383. Conclusion: Overall, both Maltese and European radiologists perceived the integration of AI in a mammography screening as beneficial. However, training and more exposure to AI software is necessary to integrate this tool in clinical practice locally. |
Description: | M.Sc. Radiography(Melit.) |
URI: | https://www.um.edu.mt/library/oar/handle/123456789/107733 |
Appears in Collections: | Dissertations - FacHSc - 2022 Dissertations - FacHScRad - 2022 |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
2218HSCRAD501005011444_1.PDF Restricted Access | 3.18 MB | Adobe PDF | View/Open Request a copy |
Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.