A Thesis submitted to The University of Arizona College of Medicine - Phoenix in partial fulfillment of the requirements for the Degree of Doctor of Medicine. / The purpose of this survey was to expand the limited knowledge regarding non-radiologist physician preferences in radiology report styles and content.

Revisión por pares

No description available.

O câncer de mama é uma triste realidade no contexto mundial, sendo o segundo tipo mais frequente entre as mulheres, representando um grave problema de saúde pública. Para diagnosticá-lo, realiza-se a mamografia, interpretada por médicos radiologistas devidamente treinados. Porém, diferentes fatores dificultam a interpretação da imagem mamográfica, como a fadiga visual, stress e distração do profissional, podendo ocasionar falhas ou atraso no diagnóstico. É possível minimizar a taxa de erros se a análise for realizada por meio da dupla leitura, realizada por dois radiologistas ou auxiliada por esquemas computacionais (CAD), no intuito de funcionar como uma segunda opinião. O objetivo do presente estudo foi avaliar a influência do CAD na interpretação de imagens mamográficas de radiologistas experientes, tanto como ferramenta auxiliar na detecção de nódulos e de agrupamentos de calcificações (CADe) quanto como classificador de nódulos (CADx). Os resultados obtidos mostraram que o CAD influenciou positivamente na precisão diagnóstica quando utilizado como classificador, aumentando a sensibilidade e a especificidade dos radiologistas, com a vantagem de não aumentar significativamente a taxa de falsos positivos. Tal fato demonstra que o CAD auxiliou no diagnóstico, sendo ferramenta plausível de ser utilizada na análise mamográfica. Mais estudos devem ser realizados seguindo esta metodologia a fim de realizar novos testes com maior número das amostras e de analisadores. Futuros estudos poderiam explorar especificamente a atuação dos esquemas CADx quando aplicados exclusivamente na análise de mamas densas, em que os sinais radiológicos não são tão evidentes, o que é um grande entrave para diagnósticos corretos. / Breast cancer is a sad reality all over the world, the second most frequent type among women, therefore a serious public health concern. For it diagnosis, mammography is performed and interpreted by qualified medical radiologists. However, several factors have limited the correct interpretation of mammographic images, as visual fatigue, stress and professional distraction, which may result in failed or delayed diagnosis. Its possible to minimize the error rate performing the analysis by means of double reading or by the support of computational schemes (CAD) in order to act as a second opinion. The aim of this study was evaluate the impact of CAD on the interpretation of mammographic images performed by experienced radiologists, aiming to detect nodules and clusters of calcifications (CADe) and as classifier nodules tool (CADx). The results showed that CAD has positively influenced the diagnostic accuracy when used as a classifier, increasing the sensitivity and specificity of radiologists with the advantage of not significantly increasing the rate of false positives. This shows that the CAD really improved the diagnosis, being a useful tool in the analysis of mammographic images.. Additional tests including more samples and analyzers should be performed according to this approach. Future studies could specifically explore the performance of CADx schemes when applied exclusively to the analysis of dense breasts, where the radiological signs are not so obvious, which is a serious difficulty for the medical examination.

O câncer de mama é uma triste realidade no contexto mundial, sendo o segundo tipo mais frequente entre as mulheres, representando um grave problema de saúde pública. Para diagnosticá-lo, realiza-se a mamografia, interpretada por médicos radiologistas devidamente treinados. Porém, diferentes fatores dificultam a interpretação da imagem mamográfica, como a fadiga visual, stress e distração do profissional, podendo ocasionar falhas ou atraso no diagnóstico. É possível minimizar a taxa de erros se a análise for realizada por meio da dupla leitura, realizada por dois radiologistas ou auxiliada por esquemas computacionais (CAD), no intuito de funcionar como uma segunda opinião. O objetivo do presente estudo foi avaliar a influência do CAD na interpretação de imagens mamográficas de radiologistas experientes, tanto como ferramenta auxiliar na detecção de nódulos e de agrupamentos de calcificações (CADe) quanto como classificador de nódulos (CADx). Os resultados obtidos mostraram que o CAD influenciou positivamente na precisão diagnóstica quando utilizado como classificador, aumentando a sensibilidade e a especificidade dos radiologistas, com a vantagem de não aumentar significativamente a taxa de falsos positivos. Tal fato demonstra que o CAD auxiliou no diagnóstico, sendo ferramenta plausível de ser utilizada na análise mamográfica. Mais estudos devem ser realizados seguindo esta metodologia a fim de realizar novos testes com maior número das amostras e de analisadores. Futuros estudos poderiam explorar especificamente a atuação dos esquemas CADx quando aplicados exclusivamente na análise de mamas densas, em que os sinais radiológicos não são tão evidentes, o que é um grande entrave para diagnósticos corretos. / Breast cancer is a sad reality all over the world, the second most frequent type among women, therefore a serious public health concern. For it diagnosis, mammography is performed and interpreted by qualified medical radiologists. However, several factors have limited the correct interpretation of mammographic images, as visual fatigue, stress and professional distraction, which may result in failed or delayed diagnosis. Its possible to minimize the error rate performing the analysis by means of double reading or by the support of computational schemes (CAD) in order to act as a second opinion. The aim of this study was evaluate the impact of CAD on the interpretation of mammographic images performed by experienced radiologists, aiming to detect nodules and clusters of calcifications (CADe) and as classifier nodules tool (CADx). The results showed that CAD has positively influenced the diagnostic accuracy when used as a classifier, increasing the sensitivity and specificity of radiologists with the advantage of not significantly increasing the rate of false positives. This shows that the CAD really improved the diagnosis, being a useful tool in the analysis of mammographic images.. Additional tests including more samples and analyzers should be performed according to this approach. Future studies could specifically explore the performance of CADx schemes when applied exclusively to the analysis of dense breasts, where the radiological signs are not so obvious, which is a serious difficulty for the medical examination.

Objetivo: Verificar graus de concordância diagnóstica pela Angio-TC entre médicos da emergência, residentes de radiologia e radiologistas torácicos em casos suspeitos de tromboembolismo pulmonar (TEP). Material e Métodos: Foram retrospectivamente estudados 123 pacientes consecutivos com suspeita de TEP, de maço/2012 a fevereiro/2013, os quais realizaram Angio-TC pulmonar, obtida com colimação por multidetector 64x0,5. As imagens foram inicialmente interpretadas isoladamente por dois médicos da Emergência e por dois médicos residentes da Radiologia, e subsequentemente por dois Radiologistas Torácicos em consenso, verificando-se o grau de concordância interobservador entre eles quanto à presença de TEP. Resultados: O grau de concordância entre os Residentes da radiologia e os Radiologistas torácicos foi muito bom (Índice Kappa de 0,82 e 0,81); entre os Médicos da Emergência e os Radiologistas torácicos foi baixo ou moderado (Kappa de 0,37 e 0,42.), com um índice de 40,0% de relatos falso-positivos, aproximadamente. Aproximadamente 40,0% das interpretações falso-positivas correlacionaram-se principalmente com casos de TEP lobar e segmentar, enquanto que as falso-negativas ocorreram mais vezes com casos de TEP segmentar e subsegmentar. O grau de concordância geral entre todos os observadores mostrou-sei também moderado (Kappa de 0,50). Conclusão: Usando a Angio-TC no diagnóstico do TEP, o grau de concordância entre os radiologistas torácicos e os residentes da radiologia foi muito bom; entretanto, a concordância entre os médicos da emergência com os radiologistas torácicos foi baixa, com tendência a superestimar o diagnóstico da condição. / Purpose: To assess interobserver agreement rates between the Radiology resident, the thoracic radiologist and the Emergency physician for diagnosing pulmonary embolism (PE) in Computed Tomography Pulmonary Angiography (CTPA) examinations. Methods: Two Emergency physicians and two Radiology residents retrospectively evaluated 123 CTPA scans at our institution, and reported the presence of PE or not, individually. Two thoracic radiologists then reviewed the images, and a consensus was reached. Statistical analysis was performed, in order to provide the interobserver agreement. Results were expressed in kappa values for subsequent comparison. Results: Very good agreement in CTPA reading was observed between Radiology residents and thoracic radiologists (kappa index of 0.82 and 0.81). Fair and moderate agreement (kappa index of 0.39 and 0.42) was demonstrated between Emergency physicians and thoracic radiologists, with a rate of 40% false-positive reports, approximately, mainly in cases of both lobar and segmental involvement, whereas false-negative occurred more times in cases of segmental and subsegmental PE. The overall agreement was also moderate (kappa index of 0.50). Conclusion: The inter-observer agreement in CTPA examinations between radiology residents and thoracic radiologists in PE diagnoses was good, but it was low between the emergency physicians and the radiologists, with a tendency of overestimating that diagnoses.

Objetivo: Verificar graus de concordância diagnóstica pela Angio-TC entre médicos da emergência, residentes de radiologia e radiologistas torácicos em casos suspeitos de tromboembolismo pulmonar (TEP). Material e Métodos: Foram retrospectivamente estudados 123 pacientes consecutivos com suspeita de TEP, de maço/2012 a fevereiro/2013, os quais realizaram Angio-TC pulmonar, obtida com colimação por multidetector 64x0,5. As imagens foram inicialmente interpretadas isoladamente por dois médicos da Emergência e por dois médicos residentes da Radiologia, e subsequentemente por dois Radiologistas Torácicos em consenso, verificando-se o grau de concordância interobservador entre eles quanto à presença de TEP. Resultados: O grau de concordância entre os Residentes da radiologia e os Radiologistas torácicos foi muito bom (Índice Kappa de 0,82 e 0,81); entre os Médicos da Emergência e os Radiologistas torácicos foi baixo ou moderado (Kappa de 0,37 e 0,42.), com um índice de 40,0% de relatos falso-positivos, aproximadamente. Aproximadamente 40,0% das interpretações falso-positivas correlacionaram-se principalmente com casos de TEP lobar e segmentar, enquanto que as falso-negativas ocorreram mais vezes com casos de TEP segmentar e subsegmentar. O grau de concordância geral entre todos os observadores mostrou-sei também moderado (Kappa de 0,50). Conclusão: Usando a Angio-TC no diagnóstico do TEP, o grau de concordância entre os radiologistas torácicos e os residentes da radiologia foi muito bom; entretanto, a concordância entre os médicos da emergência com os radiologistas torácicos foi baixa, com tendência a superestimar o diagnóstico da condição. / Purpose: To assess interobserver agreement rates between the Radiology resident, the thoracic radiologist and the Emergency physician for diagnosing pulmonary embolism (PE) in Computed Tomography Pulmonary Angiography (CTPA) examinations. Methods: Two Emergency physicians and two Radiology residents retrospectively evaluated 123 CTPA scans at our institution, and reported the presence of PE or not, individually. Two thoracic radiologists then reviewed the images, and a consensus was reached. Statistical analysis was performed, in order to provide the interobserver agreement. Results were expressed in kappa values for subsequent comparison. Results: Very good agreement in CTPA reading was observed between Radiology residents and thoracic radiologists (kappa index of 0.82 and 0.81). Fair and moderate agreement (kappa index of 0.39 and 0.42) was demonstrated between Emergency physicians and thoracic radiologists, with a rate of 40% false-positive reports, approximately, mainly in cases of both lobar and segmental involvement, whereas false-negative occurred more times in cases of segmental and subsegmental PE. The overall agreement was also moderate (kappa index of 0.50). Conclusion: The inter-observer agreement in CTPA examinations between radiology residents and thoracic radiologists in PE diagnoses was good, but it was low between the emergency physicians and the radiologists, with a tendency of overestimating that diagnoses.

Objetivo: Verificar graus de concordância diagnóstica pela Angio-TC entre médicos da emergência, residentes de radiologia e radiologistas torácicos em casos suspeitos de tromboembolismo pulmonar (TEP). Material e Métodos: Foram retrospectivamente estudados 123 pacientes consecutivos com suspeita de TEP, de maço/2012 a fevereiro/2013, os quais realizaram Angio-TC pulmonar, obtida com colimação por multidetector 64x0,5. As imagens foram inicialmente interpretadas isoladamente por dois médicos da Emergência e por dois médicos residentes da Radiologia, e subsequentemente por dois Radiologistas Torácicos em consenso, verificando-se o grau de concordância interobservador entre eles quanto à presença de TEP. Resultados: O grau de concordância entre os Residentes da radiologia e os Radiologistas torácicos foi muito bom (Índice Kappa de 0,82 e 0,81); entre os Médicos da Emergência e os Radiologistas torácicos foi baixo ou moderado (Kappa de 0,37 e 0,42.), com um índice de 40,0% de relatos falso-positivos, aproximadamente. Aproximadamente 40,0% das interpretações falso-positivas correlacionaram-se principalmente com casos de TEP lobar e segmentar, enquanto que as falso-negativas ocorreram mais vezes com casos de TEP segmentar e subsegmentar. O grau de concordância geral entre todos os observadores mostrou-sei também moderado (Kappa de 0,50). Conclusão: Usando a Angio-TC no diagnóstico do TEP, o grau de concordância entre os radiologistas torácicos e os residentes da radiologia foi muito bom; entretanto, a concordância entre os médicos da emergência com os radiologistas torácicos foi baixa, com tendência a superestimar o diagnóstico da condição. / Purpose: To assess interobserver agreement rates between the Radiology resident, the thoracic radiologist and the Emergency physician for diagnosing pulmonary embolism (PE) in Computed Tomography Pulmonary Angiography (CTPA) examinations. Methods: Two Emergency physicians and two Radiology residents retrospectively evaluated 123 CTPA scans at our institution, and reported the presence of PE or not, individually. Two thoracic radiologists then reviewed the images, and a consensus was reached. Statistical analysis was performed, in order to provide the interobserver agreement. Results were expressed in kappa values for subsequent comparison. Results: Very good agreement in CTPA reading was observed between Radiology residents and thoracic radiologists (kappa index of 0.82 and 0.81). Fair and moderate agreement (kappa index of 0.39 and 0.42) was demonstrated between Emergency physicians and thoracic radiologists, with a rate of 40% false-positive reports, approximately, mainly in cases of both lobar and segmental involvement, whereas false-negative occurred more times in cases of segmental and subsegmental PE. The overall agreement was also moderate (kappa index of 0.50). Conclusion: The inter-observer agreement in CTPA examinations between radiology residents and thoracic radiologists in PE diagnoses was good, but it was low between the emergency physicians and the radiologists, with a tendency of overestimating that diagnoses.

Thesis (MSc (Radiography))--Cape Peninsula University of Technology, 2019 / Introduction: After-hours diagnostic imaging is essential in the majority of large public hospitals, as it plays a vital role in the treatment and management of patients. Radiologists are not always available after hours to provide reports on radiographic images since, nationally and globally, there is a shortage of these professionals. Radiographic images are frequently interpreted by emergency physicians after hours. Thus, while diagnostic imaging represents an essential component in patient care (including management and treatment), according to the literature, a significant cause of clinical error occurs through the misinterpretation of radiographic images by emergency physicians. The aim of this study was to determine emergency physicians’ views on whether there was a need for an after-hours diagnostic radiology reporting service in emergency departments at some public hospitals, in the Durban Metropole. It is important to note that in addition to the above, this study calculated the number of radiographic examinations performed after-hours, and the number that was reported by the radiologist during office hours, since there was no radiology cover after hours. Methods: A descriptive cross-sectional quantitative survey design was employed using a self-administered questionnaire as a data collection instrument completed by emergency physicians at four public hospitals. In addition, additional data was collected to determine the number of radiographic examinations that had been performed after hours, at the selected four public hospitals over a period of three months, as well as the number of radiographic examinations that was reported on. This enabled the authors to determine the number of radiographic examinations that went unreported during this study period. Results: A total of 39 emergency physicians participated in the survey, with a mean and median age of 39.46 and 38 years, respectively (SD = 9.11 years). The results of this study showed that between 0.1% and 0.6% of radiographic examinations performed after hours were reported on by radiologists during office hours, for this study period. This implies that less than 1% of all examinations produced after hours at the four public hospitals, received a radiology report. Emergency physicians felt that the interpretation of images took up valuable time. The survey found that there was near total consensus amongst respondents on whether they prefer after-hours reporting to be performed by a radiologist as 46.2% (n = 18) of the respondents strongly agreed and 41.0% agreed (n = 16). Furthermore, a total of 35.9% (n=14) of respondents agreed and 43.6% (n=17) strongly agreed, that having a reporting radiographer reporting on radiographic images after-hours, would benefit patient flow. The survey also found that 92% of the sampled emergency physicians agreed (59.0% strongly agreed and 33.3% agreed, respectively) that there was a need for further training in the interpretation of radiographic images. Discussion: From the above results, it is evident that since the majority of radiographic examinations went unreported after hours, the task to interpret the radiographic images is left to the emergency physicians as part of their patient management. Conceivably, this added image interpretation results in a further increase in the workload of emergency physicians. It is therefore not surprising that emergency physicians preferred that after-hours reporting of radiographic images be done by radiologists. According to the literature, reporting radiographers also play a role in alleviating the workload of emergency physicians and improving patient flow, by providing a report for the radiographic images during after-hours. Thus, reporting radiographers afford emergency physicians additional time to concentrate on patient treatment, resulting in faster patient throughput. Reporting on radiographic images is not yet included in the scope of the South African radiographer. The findings of this study, though, suggested that there was a need for emergency physicians to undergo training in the interpretation of radiographic images. Conclusion: The study recommends that an after-hours reporting service be considered for the four public hospitals concerned. It is recommended that the heads of the emergency and radiology departments further consider offering courses on radiographic image interpretation for emergency physicians.

Bakgrund: Artificiell intelligens förväntas lösa en rad olika problemområden inom hälso-sjukvården, däribland röntgen. Både radiologer och röntgensjuksköterskor förväntas ta till sig ny information och forskning och utveckla området framåt. Trots detta finns det relativt lite information gällande personalens uppfattningar om artificiell intelligens. Syfte: Syftet med studien var att belysa röntgensjuksköterskor och radiologers uppfattningar av artificiell intelligens Metod: En kvalitativ litteraturstudie. Sökningarna genomfördes i databaserna Pubmed, Cinahl, Web Of Science och Google Scholar. Åtta artiklar analyserades och kvalitetsgranskades. Resultat: Resultatet delades in i kategorierna: Det finns oklarheter om AI:s fulla betydelse med subkategorierna: En önskan om mer utbildning, Osäkerheter kring AI:s värde inom röntgen samt förbättrad vård trots farhågor. Följt av den sista kategorin: Yrket står inför en utveckling med subkategorierna, En delad syn gällande AI:s påverkan på arbetsrollen, En främst positiv skiftning i jobbmöjlighet samt långsiktiga förhoppningar för professionerna. Resultatet diskuteras sedan mot kompetensbeskrivningen som teori. Konklusion: Det fanns en positiv inställning till AI, trots farhågor och ovissheter. Dock krävs mer forskning för att fastslå AI:s påverkan på verksamheten och professionen. Något som är tydligt är att personalen kräver ytterligare kunskaper men också mer tid till att läsa på om forskning gällande AI för att kunna arbeta evidensbaserat. / Background: Artificial intelligence is expected to solve a number of different problem areas in healthcare, including the radiography department. Both radiologists and radiographers are expected to absorb new information and research and bring the field forward. Despite this, there is relatively little information regarding the staff's perceptions of artificial intelligence. Aim: The aim of the study was to illustrate radiographers and radiologists’ perceptions of artificial intelligence Methods: A qualitative literature study. Searches were performed in Pubmed, Cinahl, Web of Science and Google Scholar. Eight articles were analyzed and quality reviewed. Results: The results where divided into the categories: There are uncertainties about AI:s full meaning with the subcategories: A desire for more training, Uncertainties about the value of AI in radiography and improved care despite the concerns. Followed by the last category: The professions are facing a development with the subcategories, A divided view regarding the impact of AI on the work role, A mainly positive shift in job opportunities and long-term hopes for the professions. The result is then discussed against the competence description as theory. Conclusion: There was a positive attitude towards AI, despite the concerns and uncertainties. However, more research is required to decide AI's full impact on the profession. Something that is clear is that personnel requires additional knowledge but also more time to study research regarding AI in order to be able to work based on evidence.