Journal:The current state of knowledge on imaging informatics: A survey among Spanish radiologists
| Full article title | The current state of knowledge on imaging informatics: A survey among Spanish radiologists |
|---|---|
| Journal | Insights into Imaging |
| Author(s) | Eiroa, Daniel; Antolín, Andreu; Ascanio, Mónica F.d.C.; Ortiz, Violeta P.; Escobar, Manuel; Roson, Nuria |
| Author affiliation(s) | Hospital Universitario Valle de Hebrón, Hospital Universitario Nuestra Señora de Candelaria |
| Primary contact | Email: danieldomingo dot eiroa dot idi at gencat dot cat |
| Year published | 2022 |
| Volume and issue | 13 |
| Page(s) | 34 |
| DOI | 10.1186/s13244-022-01164-0 |
| ISSN | 1869-4101 |
| Distribution license | Creative Commons Attribution 4.0 International |
| Website | https://insightsimaging.springeropen.com/articles/10.1186/s13244-022-01164-0 |
| Download | https://insightsimaging.springeropen.com/track/pdf/10.1186/s13244-022-01164-0.pdf (PDF) |
 |
This article should be considered a work in progress and incomplete. Consider this article incomplete until this notice is removed. |
Abstract
Background: There is growing concern about the impact of artificial intelligence (AI) on radiology and the future of the profession. The aim of this study is to evaluate general knowledge and concerns about trends on imaging informatics among radiologists working in Spain (residents and attending physicians). For this purpose, an online survey among radiologists working in Spain was conducted with questions related to knowledge about terminology and technologies, need for a regulated academic training on AI, and concerns about the implications of the use of these technologies.
Results: A total of 223 radiologists answered the survey, of whom 76.7% were attending physicians and 23.3% residents. General terms such as "AI" and "algorithm" had been heard of or read in at least 75.8% and 57.4% of the cases, respectively, while more specific terms were scarcely known. All the respondents considered that they should pursue academic training in medical informatics and new technologies, and 92.9% of them reckoned this preparation should be incorporated in the training program of the specialty. Patient safety was found to be the main concern for 54.2% of the respondents. Job loss was not seen as a peril by 45.7% of the participants.
Conclusions: Although there is a lack of knowledge about AI among Spanish radiologists, there is a will to explore such topics and a general belief that radiologists should be trained in these matters. Based on the results, a consensus is needed to change the current training curriculum to better prepare future radiologists.
Key points: Spanish radiologists desire to delve deeper into imaging informatics. Patient safety and adaptation to new technologies are the main concerns. A change on radiology education is needed to include artificial intelligence.
Keywords: artificial intelligence, medical informatics, medical education, surveys and questionnaires, radiology
Introduction
There is no doubt that the upsurge of machine learning (ML) and deep learning (DL) algorithms, paired with the high amount of digital data generated in radiology, is changing this medical specialty. ML is already used in different imaging modalities such as CAD (computer-aided design) systems for breast cancer screening on mammography [1] or nodule detection on thoracic CT or radiography. [2] DL algorithms, in particular convolutional networks, are a promising technique for processing medical imaging data not only in tasks like image classification, object detection, segmentation, or registration [3], but also on dose optimization, creation and maintenance of biobanks, and structured reporting among others. [4]
More than 50,000 articles are returned when the search “Radiology” AND “Artificial Intelligence” OR “Deep Learning” OR “Machine Learning” is performed in the PubMed medical research engine, with a “quasi-exponential” slope for the last 10 years. Such is the concern that both the European Society of Radiology (ESR) and the Radiological Society of North America (RSNA) have their own specialized internet portals dedicated to artificial intelligence (AI) [5, 6], and the latter has even published a peer-reviewed journal fully dedicated to it. [7]
For this reason, there is growing concern among radiologists about the future of the profession. Some believe that radiologists will become obsolete in a few years, and others, such as the aforementioned societies [4, 8], have a more conservative stance in which AI will enhance the role of the radiologist and turn the job from volume-based to value-based. [9] Regardless of particular opinions, the irruption of AI in the radiological field, as well as its progressive integration into clinical practice, will bring a radical change in radiology as we currently know it.
The aim of this study is to evaluate general knowledge and concerns about trends on imaging informatics among radiologists currently working in Spain (both residents and attending physicians). All those respondents who had completed residency at the time of the survey are referred to as "attending physicians" throughout the text.
Methods
An online survey using Google Forms was designed by the authors, composed of 20 questions related to the level of knowledge about trending terminology and technologies according to the most recent and relevant literature [4, 8, 10], the need for a regulated academic training, and concerns about the implications of the widespread use of these technologies in the clinical setting, both ethics- and workforce-related. A summary of the survey is displayed in Table 1.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A link to the survey was distributed among radiologists working in Spain, who were asked to share and publicize it, as widely as possible, among colleagues throughout the country after requesting their permission. It was also shared by some of the regional subsidiaries of the Sociedad Española de Radiología Médica (SERAM). It remained open for 62 days, between July 30 and September 30, 2019. Radiologists not working in Spain were excluded.
Responses were stored in a spreadsheet (Google Forms) that was later transformed into a comma-separated value file that was loaded into a Jupyter Notebook using the Python (v 3.4) pandas (v 1.1.4) library for data exploration and statistical analysis. To facilitate analysis and drawing of conclusions, the answers in the Concerns section were grouped into three categories: not concerned (options 1 and 2), indifferent (option 3), and concerned (options 4 and 5). In the instances where group comparison was made between attending physicians and residents, the Chi-square was used (scikit-learn v 0.24.0). Yates’ correction for continuity was applied where necessary. Statistical significance was accepted at p < 0.05.
Confidence intervals are not provided since the survey was purely descriptive. The results are expressed in percentages of the total answers throughout the manuscript.
Results
In the span of two months, a total of 223 radiologists answered the survey, of whom 171 (76.7%) were attending physicians and 52 (23.3%) resident physicians. When comparing to the current distribution of members in SERAM (836 residents and 5139 nonresidents) [11], we found that we had a greater proportion of residents than expected (p < 0.05).
Regarding attending radiologists, 50.9% worked exclusively in the public setting, while 5.8% worked only in the private sector and 38.6% combines public and private dedication. The same proportion (39.2%) had either fewer than 10 years or more than 20 years of working experience.
As per the residents, 44.2% were in their second year of specialty. Upon finishing, 63.5% desire to work in the public setting, mostly with some private dedication (55.8%). 32.7% had not yet decided on their preferred work setting. A summary of the results is shown in Table 2.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
With respect to the terminology and technologies, most of the underlying technologies used in deep learning remained unknown to the survey participants, including Python (71.7%), R (81.2%), PyTorch (96.4%) and TensorFlow (81.2%). Conversely, general terms like "artificial intelligence" and "algorithm" had been heard of or read in at least 75.8% and 57.4% of the cases, respectively (Fig. 1). Statistical significance was found for "artificial intelligence," "algorithm," "backpropagation," "blackbox," and "TensorFlow" (Table 3). According to Pearson's residuals, this significance is mainly due to a higher proportion of residents showing occasional use or knowledge only.
|
