Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology
Artificial intelligence (AI) uses specific algorithms to come to conclusions and process human intelligence possibly better than humans can. Thus far, AI has had relatively little impact in the field of obstetrics and gynecology, although in the field providers rely on ultrasound images to make diag...
| Hoofdauteurs: | , , |
|---|---|
| Formaat: | Journal article |
| Taal: | English |
| Gepubliceerd in: |
Wolters Kluwer Health
2021
|
| _version_ | 1826298814123737088 |
|---|---|
| author | Drukker, L Noble, JA Papageorghiou, AT |
| author_facet | Drukker, L Noble, JA Papageorghiou, AT |
| author_sort | Drukker, L |
| collection | OXFORD |
| description | Artificial intelligence (AI) uses specific algorithms to come to conclusions and process human intelligence possibly better than humans can. Thus far, AI has had relatively little impact in the field of obstetrics and gynecology, although in the field providers rely on ultrasound images to make diagnoses. This article suggests that AI could be useful in the healthcare setting by providing new insights into screening, diagnoses, prediction, and management.<br>
The AI term “deep learning” can be used to describe advances in pattern recognition. This tool can be particularly useful in healthcare settings that rely heavily on imaging reports, such as radiology or ultrasound units. Artificial intelligence technologies that utilize deep learning have been approved for use to increase productivity through automated screening, aiding in diagnosis management, or prioritizing certain imaging reports.<br>
Although there are many benefits to the introduction of AI technology into the healthcare setting, there are also concerns and limitations. One of the concerns is the potential career impact. However, this does not automatically imply that this area will face unemployment by the introduction of AI. It could, instead, prompt a new way in which people work in these fields. One of the biggest advantages is that sonographers and maternal-fetal medicine experts would not have to spend as much time completing routine tasks and would have more time to perform tasks that add value and positively impact patient care. Another advantage that AI has is that the data provided are consistently reproducible. In other words, the data provided are consistent over time, whereas the data provided by a clinician may vary by level of training, fatigue, and distractions. Further, AI technology can read thousands of scans per day, whereas a sonologist can read approximately only 50 to 100 scans per day.<br>
There are many ethical dilemmas to consider when exploring the possibility of introducing AI into the healthcare setting. For example, should AI technology be able to choose which patient receives the last intensive care unit bed? There can also be privacy issues that arise. Artificial intelligence technology relies on a large amount of data in order to come to a diagnosis or treatment plan for each specific patient, but usually this sensitive patient information is handled only by a healthcare professional. If AI were to be used, this information may also have to go through a third party. Additionally, sometimes AI can be wrong. This then begs the question of who is at fault when AI makes an incorrect diagnosis or prompts incorrect management.<br>
The authors present a high-level view of the benefits and limitations of AI use in the healthcare setting. Deep learning can be useful for image pattern recognition and can provide quality assurance but can also lead to privacy concerns. The authors conclude that healthcare professionals should begin documenting their experience with AI so that more data can be collected on its practical use in the healthcare setting. |
| first_indexed | 2024-03-07T04:52:33Z |
| format | Journal article |
| id | oxford-uuid:d57765b8-efd8-4e67-a949-dad7c2cf453d |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T04:52:33Z |
| publishDate | 2021 |
| publisher | Wolters Kluwer Health |
| record_format | dspace |
| spelling | oxford-uuid:d57765b8-efd8-4e67-a949-dad7c2cf453d2022-03-27T08:26:09ZIntroduction to artificial intelligence in ultrasound imaging in obstetrics and gynecologyJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:d57765b8-efd8-4e67-a949-dad7c2cf453dEnglishSymplectic ElementsWolters Kluwer Health2021Drukker, LNoble, JAPapageorghiou, ATArtificial intelligence (AI) uses specific algorithms to come to conclusions and process human intelligence possibly better than humans can. Thus far, AI has had relatively little impact in the field of obstetrics and gynecology, although in the field providers rely on ultrasound images to make diagnoses. This article suggests that AI could be useful in the healthcare setting by providing new insights into screening, diagnoses, prediction, and management.<br> The AI term “deep learning” can be used to describe advances in pattern recognition. This tool can be particularly useful in healthcare settings that rely heavily on imaging reports, such as radiology or ultrasound units. Artificial intelligence technologies that utilize deep learning have been approved for use to increase productivity through automated screening, aiding in diagnosis management, or prioritizing certain imaging reports.<br> Although there are many benefits to the introduction of AI technology into the healthcare setting, there are also concerns and limitations. One of the concerns is the potential career impact. However, this does not automatically imply that this area will face unemployment by the introduction of AI. It could, instead, prompt a new way in which people work in these fields. One of the biggest advantages is that sonographers and maternal-fetal medicine experts would not have to spend as much time completing routine tasks and would have more time to perform tasks that add value and positively impact patient care. Another advantage that AI has is that the data provided are consistently reproducible. In other words, the data provided are consistent over time, whereas the data provided by a clinician may vary by level of training, fatigue, and distractions. Further, AI technology can read thousands of scans per day, whereas a sonologist can read approximately only 50 to 100 scans per day.<br> There are many ethical dilemmas to consider when exploring the possibility of introducing AI into the healthcare setting. For example, should AI technology be able to choose which patient receives the last intensive care unit bed? There can also be privacy issues that arise. Artificial intelligence technology relies on a large amount of data in order to come to a diagnosis or treatment plan for each specific patient, but usually this sensitive patient information is handled only by a healthcare professional. If AI were to be used, this information may also have to go through a third party. Additionally, sometimes AI can be wrong. This then begs the question of who is at fault when AI makes an incorrect diagnosis or prompts incorrect management.<br> The authors present a high-level view of the benefits and limitations of AI use in the healthcare setting. Deep learning can be useful for image pattern recognition and can provide quality assurance but can also lead to privacy concerns. The authors conclude that healthcare professionals should begin documenting their experience with AI so that more data can be collected on its practical use in the healthcare setting. |
| spellingShingle | Drukker, L Noble, JA Papageorghiou, AT Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title | Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title_full | Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title_fullStr | Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title_full_unstemmed | Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title_short | Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| title_sort | introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology |
| work_keys_str_mv | AT drukkerl introductiontoartificialintelligenceinultrasoundimaginginobstetricsandgynecology AT nobleja introductiontoartificialintelligenceinultrasoundimaginginobstetricsandgynecology AT papageorghiouat introductiontoartificialintelligenceinultrasoundimaginginobstetricsandgynecology |