Artificial Intelligence in Dental Imaging Practice among Oral Radiologists from the City of Chennai: A Qualitative Analysis

Abstract

Background

Artificial intelligence (AI) has the potential to significantly transform the role of the doctor and revolutionize the practice of medicine and radiology. Studies have documented the clinical potentials of AI in medical imaging practice to improving patient care. There are no studies done in India that assess the attitude and perspectives of oral radiologists toward AI in dental radiology.

Aim

The present study aimed to identify barriers and enablers for the implementation of AI in dental radiology.

Materials & Methods

The study employs a qualitative design using an open-ended instrument. The sample size was 35 oral radiologists who were working within the district of Chennai and convenience/snowball sampling was done. Data obtained were analyzed using qualitative content analysis. Six themes of concerns were generated: expectant tool; career insecurity; cost of new technology; equipment preservation and data insecurity; service delivery quality; need for expanding AI awareness. A questionnaire developed along the Theoretical Domains Framework and the Capabilities, Opportunities, and Motivations influencing Behaviors model was used to guide interviews. Mayring's content analysis was employed to point out barriers and enablers.

Results

In the current study, nine themes were identified as enablers, four as conflicting themes, and seven as barriers for the acceptance of AI-integrated radiology. The factors that influence the radiology field and their attitude toward the introduction of AI in dental radiology were assessed thoroughly. Both stakeholders emphasized chances and hopes for AI along with marginal fear of job insecurity. A range of enablers for implementing AI in dental diagnostics was identified (e.g., the chance for higher diagnostic accuracy, a reduced workload, more comprehensive reporting, and better patient–provider communication).

Conclusion

AI has an important role to play in the global health care setting. The current study reported the attitudes and apprehensions of oral radiologists toward AI integration in radiology. Decision-makers and radiology industry may want to consider the barriers and conflicting themes that appear as challenges to foster implementation of AI in dentistry.

Keywords

artificial intelligence - oral radiologists - qualitative analysis - medical imaging

Introduction

Artificial intelligence (AI) has evolved as a revolution in all fields of science including health and medicine. There has been a high rate of research and studies conducted on AI-assisted medical diagnostics, especially radiographic image analysis. Many studies have shown that AI-based algorithms have proved to produce high accuracy and sensitivity in the identification and characterization of radiographic findings.[1] Apart from having an accuracy similar to or higher than those of experienced specialists, the AI-based systems can assist in patient triage, decision-making, image postprocessing, quality control, and automated, comprehensive reporting. Many algorithms have already been developed for the detection of dental caries, periapical lesions, bone loss, tooth fractures, and other dental anomalies from dental radiographs.[2] Lewis and colleagues summarized the impacts of AI in radiology where they stated that AI can lead to workflow improvement, production of high-quality images, improvement in image interpretation, radiomics analysis, dose optimization, automated scheduling, and reduced scan time. They also stated that AI can significantly bring down the workload of radiographers, radiologists, and other medical imaging staff.[3] However, before the integration of any novel and complex technology into conventional health care setups, the health care providers have to be sufficiently trained and given awareness about the technology. Radiologists may have numerous queries, expectations, doubts, and fears about new concepts of AI like machine learning, patient data protection, ethical guidelines, accountability, levels of predictability and risk of bias, etc. All these factors will collectively determine the utility of AI among radiologists and there is a need to understand the factors that act as barriers and enablers to facilitate implementation of AI in dental radiology practice. A qualitative analysis can yield comprehensive insight into the mindset of the included subjects. A theoretical framework can ensure deep levels of understanding of the beliefs and fears behind a new technology thus fostering appropriate action for improving the implementation of AI.

The current study aimed to qualitatively analyze the perspectives of oral and maxillofacial radiologists on the use of AI in dental radiology and assess the in-depth interviews using the Theoretical Domains Framework (TDF) and Behavior change wheel (BCW) associated Capability, Opportunity, Motivation and Behavior (COM-B) model.

Materials and Methods

Study design and Selection of Participants

One author conducted the interviews. The author is an oral and maxillofacial radiologist who is working as a clinician for the past 7 years. She has experience in the organization, conduction, and analysis of qualitative studies. Data saturation is the most common guiding principle for assessing the adequacy of purposive samples in qualitative research. The concept of saturation was developed by Glaser and Strauss (1967)[4] as “theoretical saturation” and was part of their influential grounded theory approach to qualitative research. An essential criterion for determining saturation of data is that it should be confirmed only after no new information was found in two or three consecutive interviews or focus groups.[5] [6] [7]

In the current study, sample size was formulated based on the theoretically rooted approach which sets an initial minimum sample size of n = 40 participants, with the potential of recruiting more participants until no new information emerges at the point of data saturation.[8] This is a standard sampling technique implemented by numerous qualitative studies done in homogenous populations. According to Hennink et al in the systematic review on sample sizes for data saturation in qualitative research, the authors found that qualitative studies can reach saturation at relatively small sample sizes (an average of 9–17 interviews) in homogenous populations.[9] Similarly, we achieved data saturation on reaching 35 interviews, which consisted a total of 35 dental professionals who held different job roles like academic and clinical roles in the private and public sectors in the city of Chennai, Tamil Nadu. All participants were interviewed in January and February 2024 using convenient sampling. Ethical clearance was obtained from our institutional review board (SDC/2024/013). The participants were explained about the study and they were asked to sign the informed consent forms. All methods were performed according to the guidelines and regulations of the Helsinki Declaration of 1975, as revised in 2000.

Theoretical Domains Framework

The present qualitative study was built on 35 semi-structured individual interviews with oral radiologists. The interview guides were developed based on TDF. TDF facilitated a systematic and comprehensive assessment of attitudes, apprehensions, and expectations of dentists concerning dental diagnostics and barriers and enablers for AI-based radiographic diagnosis. After a thematic content analysis, the BCW and its associated COM-B model were applied. The themes were finally classified as barriers, facilitators, or conflicting themes. Both TDF and BCW have been employed by several qualitative studies in dental research. When they are combined, they help in developing a theoretical framework that links behavior determinants and interventions to enable behavior change. In the current study, the necessary behavior change is the adoption of AI-supported dental radiology by the oral and maxillofacial radiologists in their routine clinical practice. The interview guides were based on themes that highlighted the attitudes, apprehensions, and expectations of dentists concerning dental diagnostics and barriers and enablers for AI-based radiographic diagnosis. The questionnaire was designed through the consolidation of previous studies.[10] [11] It consisted of a mixed set of open- and closed-ended questions based on 20 themes. Interview guides covering the 14 domains of the TDF were developed by two authors: knowledge, skills, social/professional role and identity, beliefs about capabilities, optimism, beliefs about consequences, reinforcement, intentions, goals, memory, attention and decision processes, environmental context and resources, social influences, and emotions and behavioral regulation. Each domain contained one or more questions ([Tables 1] and [2]).

Table 1 Interview guide

• Interviewee's details: Institution/designation, current position, years of experience
1.	Environmental context and resources • Advantages of digital radiographs:  Can you list out a few advantages of digital imaging? • Time management:  Do you feel you can save time if we use AI-integrated radiology?  If yes, can you explain how time can be managed using AI integration in hospital imaging systems? • Quick and comprehensive reporting:  Do you believe AI-based imaging software can provide a quicker and more detailed report when compared with manual reporting?
2.	Social influences • AI considered as an added professional skill:  Do you consider AI-based imaging as an added skill? If yes, are you interested to learn and upskill yourself?
3.	Reinforcement • Reinforcement in hospital setups:  Hospitals are making AI-based imaging mandatory due to various reasons. Do you agree with their move?  Do you think this move will motivate doctors to adapt to AI?
4.	Beliefs about consequences • High accuracy, quick OP disposal:  How can AI-integrated radiology help in quick disposal of OP in hospital setups?  Can AI provide higher accuracy than manual reporting?
5.	Social/professional role and identity • Optimism to learn new skill, effective tool for research and teaching:  Are you enthusiastic to learn AI imaging?  How do you think AI will help you in enhancing education and research among your students?
6.	Skills • Lack of training and courses for learning AI-integrated radiology:  What are the factors that stop you from learning AI tools?  Have you heard of courses that train dentists in AI?
7.	Environmental context and resources • Maintenance and quality management of AI  Do you think maintenance of AI software is laborious and expensive?  If yes, explain?
8.	Beliefs about consequences • Job insecurity  What is the likelihood of radiologists being replaced by AI software? Do you fear losing your job?
9.	Emotions • Fear of bias, fear of software corruption and loss of data  Patient data security  What is your take on patient data security in AI-based imaging software?
10.	Knowledge: • Lack of exposure to AI in dental curriculum  Is there adequate exposure to AI in the dental undergraduate and postgraduate curriculum?  Can you explain briefly?
11.	Environmental context and resources • Impact of AI  How do you think AI can revolutionize patient care in Indian health care setups?
12.	Social professional role/identities • Reliance on AI  Do you feel you will start excessively depending on AI software? If yes, how do you think this scenario can be handled?

Abbreviations: AI, artificial intelligence; OP, outpatient.

Table 2 Themes and sub-themes with examples of participant responses

S. No.	Themes	Examples of responses from participants
1.	Environmental context and resources Advantages of digital radiographs Time management Quick and comprehensive reporting	“I think digital radiographs save time as they can be easily retrieved whenever needed. It can also be shared with other consultants through cloud storage links.” “Artificial intelligence can definitely help in quick imaging especially in government hospitals when the crowd is very high and becomes unmanageable due to lack of staff…”
2.	Social Influences AI considered as an added professional skill	“Many of my colleagues have already started using AI imaging software and related tools… They are holding short-term skill-based courses on AI for doctors and dentists and I feel I should join them so that I stay updated…'
3.	Reinforcement Reinforcement in hospital setups	“Some hospitals have made it mandatory to work with AI-based radiology softwares…such reinforcement may work positively for some and for others it may induce unnecessary stress and raise a lack of confidence…”
4.	Beliefs about consequences High accuracy Quick OP disposal	“In our OP, on a daily basis we take 80 to 100 extraoral radiographs in a span of 2–3 hours…Manual errors can happen though rarely…I feel that AI can definitely produce accurate reports…”
5.	Social/professional role and identity Optimism to learn new skill Effective tool for research and teaching	“I always like to learn something new and update myself from time to time” “We have started doing small research projects with AI models and ability in detecting different lesions on CBCT, OPG images. The results are reliable and such research projects are interesting for our PG and UG students….”
6.	Skills Lack of training and courses for learning AI-integrated radiology	“Though we have YouTube videos on AI-based software usage…there is no comprehensive training available exclusively for head and neck imaging using AI…. So, we need to depend on the software vendors for support frequently…”
7.	Environmental context and resources Maintenance and quality management of AI	“AI-integrated software systems need high maintenance…Frequent support from the vendor is needed for quality assurance.”
8.	Beliefs about consequences Job insecurity	“Definitely we have a fear that we may be replaced and too many of us may have to compete for a few positions…” “AI can provide comprehensive diagnosis and reporting in seconds and so the hospitals/institutions may lay off radiologists….”
9.	Emotions Fear of bias Fear of software corruption and loss of data Patient data security	“There is always a fear of loss of data due to software malfunction or other malwares” “Patient data security is also doubtful and many softwares lack robust features that enhance data security…”
10.	Knowledge: Lack of exposure to AI in dental curriculum	“In dental undergraduate and even postgraduate curriculum, there is insufficient introduction to AI…” “Practical classes on uploading images in AI and segmenting imaging should be taught…”
11.	Environmental context and resources Impact of AI	“Globally AI has spread rapidly and soon we will have to learn to work with AI-integrated radiology…”
12.	Social professional role/identities Reliance on AI	“I hope I don't start depending on AI completely…. overdependence is again a problem!”

Abbreviations: AI, artificial intelligence; CBCT, cone-beamed computed tomography; OPG, orthopantomogram.

Data Collection

A pilot test was done with the developed interview guide within the research group which comprised a set of five oral radiologists. The objective of the pilot study was to test the interview guide and assess its clarity and relevance to the current topic among the oral radiologists. The interview questions were easy to understand and the oral radiologists did not report any difficulties. The original questionnaire guide was used for the main study without any further refinement. The interviews lasted 30 to 40 minutes each and were recorded. The interview was structured in two sections. In the first part, the participants explained how they perceived radiographic diagnostics in dentistry. In the second part, they were asked questions about the factors they perceived as barriers to and enablers in AI assisted radiology. The participants expressed their apprehensions, expectations, and thoughts in a dialogue manner, with the interviewer refocusing on barriers and facilitators when the conversation deviated from the main issues of discussion.

Data Analysis, Findings, Reporting

Interviews were transcribed and analyzed using MaxQDA software. A coding tree was built based on the 14 TDF domains. To increase inter- and intra-coder reliability, the authors performed transcription and analysis independently and reviewed the results. Mayring's principles of inductive and deductive content analysis were executed and the identified themes were classified as barriers, facilitators, and conflicting themes. Themes related to the COM-B model were highlighted. The qualitative analysis followed the consolidated criteria for reporting qualitative studies (COREQ) checklist.

Results

The characteristics of the oral and maxillofacial radiologists who participated in the study are given in [Table 3]. Statements were collected from the interview transcripts and coded according to the TDF. The identified themes were organized according to the COM-B model. Relevant themes were found in nine domains of the TDF. [Table 4] sums up the identified themes, which were further classified as enablers, conflicting themes, and barriers, associated with Capability, Opportunity, and Motivation and along the TDF domains. We identified nine themes as enablers, four as conflicting themes, and seven as barriers (total: 20 themes).

Table 3 Characteristics of oral radiologists who were included in the study

S. No.	Characteristics of participants	Total number of participants: N = 35
1.	Sex
	Male	13
	Female	22
2.	Age
	31–40 years	18
	41–50 years	6
	51–60 years	4
	≥61 years	7
3.	Job position
	Academician	19
	Clinician	7
	Specialty radiology practice	6
	Public hospital	3
4.	Work experience
	1–5 years	10
	6–10 years	8
	More than 10 years	17

Table 4 Barriers, enablers, and conflicting themes identified in the study

Enablers:		Frequency distribution: (%)
A. Capability and opportunity	Environmental context and resources • Advantages of digital radiographs • Time management • Quick and comprehensive reporting	32 (91.4%)
	Social influences • AI considered as an added professional skill	31 (89%)
B. Motivation	Reinforcement • Reinforcement in hospital setups	30 (86%)
	Beliefs about consequences • High accuracy • Quick OP disposal	32 (91%)
	Social/professional role and identity • Optimism to learn new skill • Effective tool for research and teaching	33 (94%)
Barriers:
A. Capability	Skills Lack of training and courses for learning AI integrated radiology	34 (97%)
B. Opportunity	Environmental context and resources Maintenance and quality management of AI	33 (94%)
C. Motivation	Beliefs about consequences Job insecurity	33 (94%)
	Emotions • Fear of bias • Fear of software corruption and loss of data Patient data security	32 (91%)
Conflicting themes:
A. Capability	Knowledge: Lack of exposure to AI in dental curriculum	31 (89%)
B. Opportunity	Environmental context and resources Impact of AI	30 (86%)
C. Motivation	Social professional role/identities Reliance on AI	34 (97%)

Abbreviation: AI, artificial intelligence; OPG, orthopantomogram.

Enablers for AI-integrated Oral Radiology

Under enablers, a majority of dentists considered the advantages of digital radiographs and digital databases and effective time management as an opportunity. They also said that AI may help with quick and comprehensive reporting as an opportunity for integrating AI in oral radiology. AI was also believed to be an added professional skill as AI is already used in most of the disciplines of medicine. The sub-themes under motivation were reinforcement, beliefs about consequences, and social role and identity. Dentists said that there could be reinforcement in hospital setups that could lead to an increased use of AI in dental setups. High accuracy and quick outpatient disposal were considered advantages that motivate the use of AI in dentistry. Increased optimism among dentists of all age groups was prevalent. The use of AI as an effective tool for research and teaching was considered by the dentists as a motivating factor.

Barriers to AI-Integrated Oral Radiology

A total of six sub-themes were classified as barriers. Many dentists considered the lack of training and the unavailability of training courses on AI as a major barrier. Maintenance and quality management of AI was another barrier. Lack of resources in many public and private dental setups in India was a major setback. Under motivation, job insecurity and various fears like fear of bias, fear of software malfunction, and patient data security were considered barriers to the application of AI in oral radiology.

Conflicting Themes for AI-Integrated Oral Radiology

A total of five sub-themes were identified as conflicting in nature. Lack of exposure to the applications of AI in dental curricula and reliance of oral radiologists on AI were highlighted by many dentists. A possibility of over-reliance on AI can lead to loss of social and professional identity. Job insecurities and doubts about data protection were considered conflicting sub-themes as demotivating factors.

Discussion

AI has been gaining momentum in every aspect of health delivery and medical diagnostics is no exception. AI in health care has emerged as a promising domain that can revolutionize patient care, especially in a densely populated country like India. The acute shortage of specialists in India coupled with inadequate health care infrastructure is likely to increase the burden of diseases. AI-assisted radiological tools have an extensive scope to detect and diagnose subtle patterns and lesions at unprecedented speeds that can enhance patient outcomes and reduce the burden on the health care system. Studies that have assessed medical images like radiographs, clinical photos, and histological sections using AI-powered radiology demonstrated remarkable performance with accuracies similar to and sometimes higher than expert physicians.[12] [13] [14] The applications of AI have also entered dental practice and dentists are closely adapting to this imminent change in their patient care system. Our study aimed to assess the attitude of dentists toward AI for dental diagnostics and to identify barriers and enablers for implementing AI in dentistry. In total, 9 themes as enablers, 4 as conflicting themes and 7 as barriers (a total of 20 themes) were identified covering 9 of the domains of the TDF, and touching all three relevant behavior determinants, capability, opportunity, and motivation.

India stood at the forefront of adapting robotic technology in health care in the 1990s. Currently, the surgical robotics market in India is predicted to expand fivefold by 2025. Similarly, AI-driven radiology is actively welcomed and gaining acceptance among oral radiologists.[15] [16] In our analysis, several factors were identified as enablers for the implementation of AI in oral radiology. First, the participants said that the option to generate a fully automated diagnostic report in a few seconds using AI algorithms is a huge time saver, especially in public and private setups that handle large volumes of patients daily. The participants felt that AI can aid in effective time management with quick OP (outpatient) disposal. The automated comprehensive reporting was welcomed by all of the participants who said that such an AI integration can help in reducing labor-intensive workload due to exhaustive manual reporting. Participants highlighted the fact that in many Indian Institutions, there is a shortage of radiographers and radiology specialists. AI can help in reducing fatigue and burnout among oral radiologists who have to carry out every task all by him/herself, beginning with the consultation to the radiology reporting and referral. Even a simple AI algorithm that can differentiate “normal” from “abnormal” can help streamline operations, especially in low-resource settings. Similar enablers with a positive outlook on AI were recorded in an earlier qualitative analysis conducted in 2021 in Germany.[10] In our study, dentists considered AI as an added professional skill that can improve job prospects and boost their career trajectory. Under motivation, the dentists gave an optimistic set of themes that enable the integration of AI in dental imaging setups. The participants said that most of the hospitals in India and across the globe are reinforcing the application of AI on all fronts of diagnostic radiology. This is a major motivator that allows dentists to upgrade themselves and learn the practical aspects of AI. Factors like high accuracy and quick OP disposal were also stated as motivating factors that directly help in providing quick diagnostic decisions with reduced medical errors. The themes that contributed to professional role and identity were optimism to learn a new skill and AI utilization for research and academic purposes. These enablers were highly valid and similar to a study that evaluated the use of AI tools in health care crisis. The authors said that AI helped in improving technical accuracy in medical diagnosis, and treatment outcomes and helps cater to a larger set of patients a lot more easily. AI-powered radiology holds promise in analyzing massive data with superhuman speed and can detect and diagnose the most common ailment to the most complex ones.[17]

The dentists considered the lack of training and the unavailability of AI-based courses as barriers. In India, AI is a relatively a new subject and it has not been introduced formally in the medical and dental curriculum. This demands the need to advocate deep learning methodologies through additional courses. However, AI-centric imaging courses are very minimal thus leading to a gap in skill development. Second, the dentists considered maintenance of AI-based equipment and economic challenges as major barriers due to reduced resources in many public and private hospitals. The dentists expressed their uncertainties about the constant availability of maintenance personnel and their services for managing the AI-based software and systems. Dentists also stated that fear of errors, software corruption, data loss, and data insecurity were potential barriers to the acceptance of AI in oral radiology. Machine learning technologies can develop biases that may make it difficult to establish accountability. These algorithms can predict a higher prevalence of disease or lesions. The software can malfunction and lead to loss of patient data. There is also a gray area where there is no clarity on how accountability and liability will be determined if a wrong diagnosis is given by the oral radiologist due to a glitch in the AI software. Clear and strict laws have to be designed with defined boundaries of the health care system such that neither party solely faces any legal issues. Such potential ethical, occupational, and medical issues have to be avoided through the implementation of strict policies that meet the professional requirements of AI in health care so that ethical and safe health delivery is possible across the nation.[18] The Government of India has established various initiatives like the Niti Aayog's “National Strategy for AI” and the “International Centre for Transformative AI (ICTAI), 2020,” etc. These initiatives focus on research projects on health care and in framing certification systems for AI developers and vendors. Further investment by the government in AI policies can help define guidelines and foster the health care ecosystem organically.

In this analysis, the participants expressed a few themes as conflicting that reduced their trust and willingness to adopt AI in routine clinical work. Although the majority of participants accepted AI-integrated oral radiology, few participants stated job insecurity as a concern that prevents them from adapting to AI-based imaging techniques. In a period where technology is releasing innovations at an express speed, it is ironic that dentists fear their replacement by AI algorithms. These barriers were similar to a study by Antwi et al among radiographers. The authors said that the radiographers feared the negative impact of AI on their workforce. They were anxious that it could lead to the loss of jobs across Africa and even make radiology courses/education programs irrelevant. Similar barriers such as job insecurities were reported in other studies too.[19] However, experts say that till date AI has not replaced any job in the health care sector. AI will not replace anyone's job but will help in advancing health care. While automation of jobs is a possibility, several other factors like the cost of developing AI, costs of software, and regulatory and social acceptance especially in health care can curb the replacement of health care workers at least for the next two decades.[11] [20] [21] Radiologists help in diagnosing diseases, designing treatment plans, and are in constant patient interaction. They are needed in multi-specialty setups for image-guided intervention procedures. Indeed, radiologists perform much more than interpretating images. On the other hand, AI algorithms are trained to perform single tasks only. Deep learning techniques are trained for specific image recognition and classification. However, a high volume of such narrow detection tasks is needed to train the algorithm to identify all the findings on a medical image. All of these aspects enable only a limited incursion of AI into the health care industry with only a minor risk of job loss.[22]

Apart from job insecurities, the participants stated that a lack of exposure to AI in the dental curriculum is a major setback that makes it difficult to adapt AI. Introducing faculty development programs and skill development programs on AI in dental imaging can be of benefit to the fraternity. A few of them expressed their concern on the possibility of over-reliance on AI that could lead to possible misdiagnosis or a tendency to overlook minute details which would have been noted in a manual image interpretation. There is also another possibility of radiologists losing their attention and perception skills due to over-reliance on AI technologies. The American College of Radiology Data Science Institute has developed a comprehensive catalogue of AI use case documents which will guide radiologists in the applications of AI in routine clinical environments.[23] [24] [25] It is needless to say that AI is to be used only as a supplement and not as a replacement for an oral radiologist. A radiologist should always develop his independent judgement and diagnosis before considering the output from the AI algorithm. Oral radiologists should use assistive technologies only in conjunction with their critical observation and interpretative skills to prevent any diagnostic errors.

Conclusion

AI has an important role to play in the global health care setting. With the high levels of accuracy and precision, it can assist in quick and easy diagnosis and patient care. The current study reported the attitudes and apprehensions of oral radiologists toward AI integration in radiology. The findings of the study suggest that a total of nine themes are enablers, four are conflicting themes, and seven are barriers for the acceptance of AI-integrated radiology. Enablers for implementing AI in dental diagnostics were higher diagnostic accuracy, comprehensive reporting, and better patient–provider communication. Though there are a total of seven barriers and four conflicting themes that appear as challenges, streamlined vendors who develop algorithms according to proper regulations will enhance their acceptance among radiologists. Development of AI algorithms tailored to the Indian context is possible through the collaboration of major stakeholders, researchers, and industry personnel. The current study will serve as a tool for understanding the perception of oral radiologists in the integration of AI in clinical practice. It can help to develop a conducive environment with public and private partnerships for the seamless integration of AI with oral radiology.

Conflict of Interest

None declared.

Data Availability Statement

The data will be made available upon request to the author.

Ethical Approval

The ethical approval was obtained from the IEC of SRM dental college with approval number (SDC/2024/013).

• References

• 1 Pesapane F, Codari M, Sardanelli F. Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine. Eur Radiol Exp 2018; 2 (01) 35
• 2 Hinton G. Deep learning-a technology with the potential to transform health care. JAMA 2018; 320 (11) 1101-1102
• 3 Lewis SJ, Gandomkar Z, Brennan PC. Artificial intelligence in medical imaging practice: looking to the future. J Med Radiat Sci 2019; 66 (04) 292-295
• 4 Glaser B, Strauss A. The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine; Chicago: 1967
• 5 Coenen M, Stamm TA, Stucki G, Cieza A. Individual interviews and focus groups in patients with rheumatoid arthritis: a comparison of two qualitative methods. Qual Life Res 2012; 21 (02) 359-370
• 6 Francis JJ, Johnston M, Robertson C. et al. What is an adequate sample size? Operationalising data saturation for theory-based interview studies. Psychol Health 2010; 25 (10) 1229-1245
• 7 Morse WC, Lowery DR, Steury T. Exploring saturation of themes and spatial locations in qualitative public participation geographic information systems research. Soc Nat Resour 2014; 27 (05) 557-571
• 8 Krueger RA. Focus Groups: A Practical Guide for Research. Newbury Park, CA: Sage Publications; 1988
• 9 Hennink M, Kaiser BN. Sample sizes for saturation in qualitative research: A systematic review of empirical tests. Soc Sci Med 2022; 292: 114523
• 10 Müller A, Mertens SM, Göstemeyer G, Krois J, Schwendicke F. Barriers and enablers for artificial intelligence in dental diagnostics: a qualitative study. J Clin Med 2021; 10 (08) 1612
• 11 Antwi WK, Akudjedu TN, Botwe BO. Artificial intelligence in medical imaging practice in Africa: a qualitative content analysis study of radiographers' perspectives. Insights Imaging 2021; 12 (01) 80
• 12 Ciompi F, Chung K, van Riel SJ. et al. Towards automatic pulmonary nodule management in lung cancer screening with deep learning. Sci Rep 2017; 7: 46479
• 13 Nibali A, He Z, Wollersheim D. Pulmonary nodule classification with deep residual networks. Int J Comput Assist Radiol Surg 2017; 12 (10) 1799-1808
• 14 Herweh C, Ringleb PA, Rauch G. et al. Performance of e-ASPECTS software in comparison to that of stroke physicians on assessing CT scans of acute ischemic stroke patients. Int J Stroke 2016; 11 (04) 438-445
• 15 Hinton G. Deep learning-a technology with the potential to transform health care. JAMA 2018; 320 (11) 1101-1102
• 16 Bal M, Amasyali MF, Sever H, Kose G, Demirhan A. Performance evaluation of the machine learning algorithms used in inference mechanism of a medical decision support system. ScientificWorldJournal 2014; 2014: 137896
• 17 Meskó B, Hetényi G, Győrffy Z. Will artificial intelligence solve the human resource crisis in healthcare?. BMC Health Serv Res 2018; 18 (01) 545
• 18 Chan S, Siegel EL. Will machine learning end the viability of radiology as a thriving medical specialty?. Br J Radiol 2019; 92 (1094) 20180416
• 19 Amisha MP, Malik P, Pathania M, Rathaur VK. Overview of artificial intelligence in medicine. J Family Med Prim Care 2019; 8 (07) 2328-2331
• 20 Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer 2018; 18 (08) 500-510
• 21 Botwe BO, Akudjedu TN, Antwi WK. et al. The integration of artificial intelligence in medical imaging practice: perspectives of African radiographers. Radiography (Lond) 2021; 27 (03) 861-866
• 22 Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J 2019; 6 (02) 94-98
• 23 Low LL, Lee KH, Hock Ong ME. et al. Predicting 30-day readmissions: performance of the LACE index compared with a regression model among general medicine patients in Singapore. BioMed Res Int 2015; 2015: 169870
• 24 Nait Aicha A, Englebienne G, van Schooten KS, Pijnappels M, Kröse B. Deep learning to predict falls in older adults based on daily-Life trunk accelerometry. Sensors (Basel) 2018; 18 (05) 1654
• 25 Char DS, Shah NH, Magnus D. Implementing machine learning in health care – addressing ethical challenges. N Engl J Med 2018; 378 (11) 981-983

Publikationsverlauf

Artikel online veröffentlicht:
11. Juni 2025