Imaging Publications in the COVID-19 Pandemic: Applying New Research Results to Clinical Practice
2020; Radiological Society of North America; Volume: 297; Issue: 1 Linguagem: Inglês
10.1148/radiol.2020201724
ISSN1527-1315
Autores Tópico(s)COVID-19 and healthcare impacts
ResumoHomeRadiologyVol. 297, No. 1 PreviousNext Original ResearchFree AccessSpecial ReportImaging Publications in the COVID-19 Pandemic: Applying New Research Results to Clinical PracticeJohn Eng , David A. BluemkeJohn Eng , David A. BluemkeAuthor AffiliationsFrom the Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, 1800 Orleans St, Baltimore, MD 21287 (J.E.); and Department of Radiology, University of Wisconsin School of Medicine and Public Health, Madison, Wis (D.A.B.).Address correspondence to J.E. (e-mail: [email protected]).John Eng David A. BluemkePublished Online:Apr 23 2020https://doi.org/10.1148/radiol.2020201724MoreSectionsPDF ToolsAdd to favoritesCiteTrack Citations ShareShare onFacebookXLinked In Abstract Online supplemental material is available for this article.Summary Applying the growing body of coronavirus disease 2019 imaging evidence to daily clinical practice requires thoughtful use of established clinical epidemiology principles.Introduction Scientific data and their dissemination are essential elements of an effective response to the current pandemic. As the global outbreak of coronavirus disease 2019 (COVID-19) has unfolded, biomedical journals including Radiology have been working to publish results of clinical experience and research in a timely manner (1). New imaging research aims to help radiologists around the world to be prepared for the arrival of patients with COVID-19 in their practices. In the past months, clinicians have seen hundreds of research headlines on all aspects of diagnosis, monitoring, and treatment of COVID-19. The plethora of new research challenges our ability to interpret and apply new research results. Inevitably, we attempt to intuitively extrapolate new research results to my patient and my hospital, perhaps without having time to discern critical features in the research study population and study design. In this article, we review principles of clinical study design and clinical epidemiology as applied to examples from the imaging literature concerning CT diagnosis of COVID-19. Our goal is to assist the clinician in their critical appraisal of the rapidly growing COVID-19 imaging literature and to illustrate how these new research results may apply to daily practice.Move Beyond Sensitivity and Specificity to Predictive Value Most articles examining the diagnostic performance of CT focus on its sensitivity for diagnosing COVID-19 pneumonia against the reference standard, reverse-transcriptase polymerase chain reaction (RT-PCR) for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (2). A few articles also consider the specificity of CT. Although a logical starting point, sensitivity and specificity, by definition, indicate the diagnostic performance of CT when the true COVID-19 status is already known. Because a patient's COVID-19 status is typically unknown at presentation, sensitivity and specificity are not the most clinically useful indicators of diagnostic test performance. Instead, we need to know the converse: How likely is the patient to have COVID-19 given positive or negative CT findings? This last question defines the concept of predictive value. In the current pandemic, positive predictive value (PPV) is the probability of having COVID-19 when a radiologist determines that the CT scan is positive and has findings that are consistent with COVID-19. The negative predictive value (NPV) is the probability of not having the disease when the CT scan is negative (or normal). PPV and NPV can be calculated with well-known formulas (Fig 1) (3,4), but the calculations are not easy to perform in your head.Figure 1: Calculation of positive and negative predictive values.Figure 1:Download as PowerPoint To obtain a better intuitive understanding of the PPV and NPV formulas, we consider an example of diagnosing COVID-19 pneumonia with CT. In this example, suppose CT has a sensitivity of 97% (5) and an (optimistic) specificity of 80% for COVID-19. In addition, assume a COVID-19 prevalence of 15% (the approximate percentage of positive results found in RT-PCR tests performed at Johns Hopkins Medicine hospitals in April 2020). Inserting these assumptions into the formulas in Figure 1, the calculated PPV is only 46% but the NPV is 99% (Fig 2).Figure 2: Plot of positive and negative predictive values under two different assumptions of sensitivity and specificity. For the "optimistic" assumption, sensitivity is 97% and specificity is 80%. For the "pooled" assumption, sensitivity is 94% and specificity is 37%, which are the pooled estimates from the meta-analysis by Kim et al (2). COVID-19 = coronavirus disease 2019, NPV = negative predictive value, PPV = positive predictive value.Figure 2:Download as PowerPoint Both predictive values, especially the PPV, are numerically different from the assumed sensitivity and specificity. Why? First, the PPV and NPV formulas include an additional component—the prevalence of COVID-19—and the PPV and NPV are heavily influenced by the prevalence of illness in a given population (6). Second, intuition may naturally associate PPV with sensitivity and NPV with specificity. However, this intuition is not correct: Unless the disease prevalence is much greater than 50%, the PPV calculation is mathematically dominated by specificity rather than sensitivity (7) (Fig 3). Because CT is not a specific test for COVID-19 pneumonia, there will be many cases (100% − 46% = 54%, or about one of every two cases) falsely diagnosed as COVID-19 by radiologists who note lung opacities as "diagnostic" of COVID-19 pneumonia. Therefore, in our first example, there is only about a 50-50 chance that the positive CT findings are actually due to COVID-19 pneumonia.Figure 3: Determinants of the predictive value of a diagnostic test.Figure 3:Download as PowerPoint On the other hand, our example with NPV of 99% mentioned earlier suggests that a negative CT scan could rule out COVID-19 infection. This is because the NPV calculation is mathematically dominated by sensitivity, which is high, rather than specificity (Fig 3). If CT is very sensitive in the detection of COVID-19 pneumonia, very few cases will be missed; thus, a patient with a negative CT scan will be unlikely to be a missed case (7). Although the preceding example was meant to be illustrative, the calculation demonstrates similar results—relatively low PPV and high NPV for CT—across plausible ranges of sensitivity, specificity, and prevalence (Fig 2). Note further that the example has no mention of "quality" of the research publication or how many patients were studied. Rather, it focuses on the effect of the clinical environment on predictive value. A community hospital in the midwestern portion of the United States may have a community prevalence of COVID-19 of 1%–2%; a single nursing home in the same community may have a prevalence of 40%, approaching that reported in Wuhan, China (2). CT at these two sites in the same community would have vastly different predictive values for COVID-19 pneumonia. Appendix E1 (online) includes a spreadsheet with which the reader can try their own calculations. For example, if specificity is lowered to a less optimistic value of 50%, which is the upper limit of the confidence interval reported by Kim et al (2), the calculated PPV in our example would be a lowly 26%.Pay Close Attention to Characteristics of the Study Population Sensitivity, specificity, PPV, and NPV are strictly valid only for the population from which they are obtained. Study populations in COVID-19 publications may be vastly different not only between each other but also when compared with your hospital's clinical population. In this section, we discuss the impact of study populations on the interpretation and clinical application of research results (Fig 4).Figure 4: Key study population characteristics to assess when evaluating COVID-19 publications. COVID-19 = coronavirus disease 2019, RT-PCR = reverse-transcriptase polymerase chain reaction, SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.Figure 4:Download as PowerPoint We start by comparing the distinctly different study populations reported by Ai et al (5) and Bai et al (8) in Radiology. Both are studies of CT to identify COVID-19 pneumonia but with apparently opposite results concerning CT performance. Nevertheless, both study results are valid in the context of their individual study populations. Ai et al (5) reported their experience with 1014 patients suspected of having SARS-CoV-2 infection, finding CT to have a sensitivity of 97% and specificity of 25%. But the prevalence of COVID-19 was 59%, which is far higher than is being reported in the United States at the present time. In theory, sensitivity and specificity are inherent properties of a diagnostic test that are independent of disease prevalence (4). However, in practice, a difference in disease prevalence is often a surrogate marker for underlying population differences that can affect the results of testing and treatment. For example, one possible explanation for the unexpectedly low specificity reported by Ai et al is that radiologists were employing a low interpretation threshold for diagnosing COVID-19 with CT—perhaps including pneumonia patterns that are seen with COVID-19 but are less typical. A low threshold involves a tradeoff between sensitivity and specificity that favors sensitivity. Such a tradeoff might be justified in an epidemic area with an extremely high COVID-19 prevalence. Other possible explanations for an unexpectedly low specificity—not necessarily applicable to the study by Ai et al—include an imperfect reference standard (eg, RT-PCR tests with variable performance due to varying stages of viral shedding) or high local prevalence of infections that mimic COVID-19 pneumonia. Bai et al (8) reported a sensitivity range of 73%–93% and a specificity range of 93%–100% among the American radiologists in their study. But their study was a reader study and, by design, not population-based. In their study, the prevalence of COVID-19 approximated 50%, which is customary for well-designed reader studies. More important, the study "population" was contrived, again by design. All CT cases, both positive and negative, were abnormal. Furthermore, all the COVID-19 "negative" cases were patients with confirmed viral pneumonia of another type. Obviously, this study "population" is nothing like clinical practice. In real practice, the expected COVID-19–negative cases would include patients with thoracic conditions other than viral pneumonia or even patients with no thoracic disease at all. Because COVID-19 is a new disease, no data from prospective, randomly sampled populations can exist at the beginning of the outbreak. Initial studies must rely on retrospective data collected from patients who happened to have both CT and RT-PCR, the reference test. These studies are subject to substantial selection bias, a common problem with retrospective studies. These biases are well-recognized and discussed elsewhere (9,10). Given all of the preceding caveats, one might wonder whether such studies are of any value to clinical practice. They do indeed have value, if appropriate attention is paid to the characteristics of each study population. Even if it involved a population with high COVID-19 prevalence, the study from Ai et al (5) provided a preliminary estimate of the sensitivity and specificity of CT early in the pandemic—at a time when the diagnostic performance of CT was unknown in any type of population. The results from Bai et al (8) suggest that non–COVID-19 viral pneumonias do not resemble COVID-19, because their study's high specificity meant few non–COVID-19 viral pneumonias were mistakenly classified as COVID-19. In addition, COVID-19 pneumonia can resemble other viral pneumonias, because their study's moderate sensitivity meant some COVID-19 pneumonias were mistakenly classified as another type.Where Do We Go from Here? To see how results reported in the COVID-19 literature fit into your clinical practice, it is crucial to understand the relationship between each study population and that encountered in your own clinical environment. Marked differences are likely to exist between your practice versus research study populations. As illustrated by the preceding examples, population differences can result in apparently conflicting research results, which can lead to the conclusion that the "true" sensitivity and specificity are yet unknown (11). In reality, there is probably no single "true" sensitivity, specificity, or predictive value that would apply to all clinical practice settings. Meta-analyses are "studies of studies" that attempt to provide a statistical summary of clinical parameters (eg, sensitivity and specificity) across all published research on the topic. A traditional meta-analysis of a diagnostic test begins by assuming the existence of underlying "true" values of sensitivity and specificity. A meta-analysis then attempts to estimate these values by pooling the results statistically from multiple studies. For CT diagnosis of COVID-19, the assumption of a single set of true values is, unfortunately, not a good assumption. A high degree of "heterogeneity" for CT has been reported (2). For example, Kim et al (2) provide statistical confidence intervals for sensitivity and specificity instead of a single number (a so-called "point estimate"). It is left to the reader to determine the most applicable location within the confidence interval. For CT diagnosis of COVID-19, statistical averaging of prevalence and predictive value are not meaningful because of marked heterogeneity of the study populations in the existing literature. When judging study populations, COVID-19 prevalence is a major indicator of similarity of those populations. However, readers must also consider other important characteristics such as age, other demographic factors, comorbid diseases, COVID-19 stage at time of imaging, and spectrum of disease severity, including the proportion of critically ill patients and those with no symptoms at time of diagnosis (Fig 4). Research studies usually assume equal access to medical care, but logistical, political, educational, and socioeconomic factors have affected access to COVID-19 testing. Current estimates of COVID-19 prevalence are based on the prevalence among those who have been tested, which is distinct from the prevalence in the entire population, so apparent COVID-19 prevalence may change as testing becomes more widespread. Clinical practice factors are also important to consider, such as how patients suspected of having SARS-CoV-2 infection are selected for CT or RT-PCR testing and the CT interpretation criteria employed. The proportion of infected patients who are asymptomatic is especially important because these patients are associated with a different spectrum of CT findings (12) and may be selected differently for RT-PCR testing. All of these differences can affect the apparent performance and clinical role of CT in diagnosis and management of COVID-19 both in research populations and clinical practice. Differences in patient populations are acknowledged in the COVID-19 consensus statement from the Fleischner Society, whose guidelines are stratified according to disease risk and severity (13). COVID-19 reporting guidelines (11) represent an initial effort to standardize CT interpretation thresholds, which would otherwise be a source of hidden variability in the clinical population. Rapid dissemination of clinical evidence will continue to be an important component in responding to the COVID-19 pandemic. As the disease's newness fades, opportunities may arise to conduct prospective studies on randomly sampled populations. Even with these more rigorously designed studies, attention will still be necessary to compare study populations with those encountered in clinical practice. Attention will also be required if COVID-19 prevalence or severity shifts in clinical populations. Thoughtful application of clinical epidemiology principles is necessary to place new COVID-19 evidence in a proper clinical context that continues to evolve.Disclosures of Conflicts of Interest: J.E. disclosed no relevant relationships. D.A.B. disclosed no relevant relationships.Author Contributions Author contributions: Guarantors of integrity of entire study, J.E., D.A.B.; study concepts/study design or data acquisition or data analysis/interpretation, J.E., D.A.B.; manuscript drafting or manuscript revision for important intellectual content, J.E., D.A.B.; approval of final version of submitted manuscript, J.E., D.A.B.; agrees to ensure any questions related to the work are appropriately resolved, J.E., D.A.B.; literature research, J.E., D.A.B.; and manuscript editing, J.E., D.A.B.References1. Lei J, Li J, Li X, Qi X. CT imaging of the 2019 novel coronavirus (2019-nCoV) pneumonia. Radiology 2020;295(1):18. Google Scholar2. Kim H, Hong H, Yoon SH. Diagnostic performance of CT and reverse transcriptase-polymerase chain reaction for coronavirus disease 2019: a meta-analysis. Radiology 2020 Apr 17:201343 [Epub ahead of print]. Google Scholar3. Zhou XH, Obuchowski NA, McClish DK. Statistical methods in diagnostic medicine. New York, NY: Wiley, 2002; 43–48. Google Scholar4. Weinstein MC, Fineberg HV. Clinical decision analysis. Philadelphia, Pa: Saunders, 1980; 75–94. Google Scholar5. Ai T, Yang Z, Hou H, et al. Correlation of chest CT and RT-PCR testing for coronavirus disease 2019 (COVID-19) in China: a report of 1014 cases. Radiology 2020;296(2):E32–E40. Google Scholar6. Gordis L. Epidemiology. 3rd ed. Philadelphia, Pa: Elsevier, 2004; 82–87. Google Scholar7. Guyatt GH, Sackett DL, Haynes RB. Evaluating diagnostic tests. In: Haynes RB, Sackett DL, Guyatt GH, Tugwell P, eds. Clinical epidemiology. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins, 2006; 277–282. Google Scholar8. Bai HX, Hsieh B, Xiong Z, et al. Performance of radiologists in differentiating COVID-19 from non-COVID-19 viral pneumonia at chest CT. Radiology 2020;296(2):E46–E54. Google Scholar9. Begg CB, McNeil BJ. Assessment of radiologic tests: control of bias and other design considerations. Radiology 1988;167(2):565–569. Google Scholar10. Eng J, Siegelman SS. cImproving radiology research methods: what is being asked and who is being studied? Radiology 1997;205(3):651–655. Google Scholar11. Simpson S, Kay FU, Abbara S, et al. Radiological Society of North America expert consensus statement on reporting chest CT findings related to COVID-19. Endorsed by the Society of Thoracic Radiology, the American College of Radiology, and RSNA. Radiol Cardiothorac Imaging 2020;2(2):e200152. Google Scholar12. Inui S, Fujikawa A, Jitsu M, et al. Chest CT Findings in cases from the cruise ship "Diamond Princess" with coronavirus disease 2019 (COVID-19). Radiol Cardiothorac Imaging 2020;2(2):e200110. Google Scholar13. Rubin GD, Ryerson CJ, Haramati LB, et al. The role of chest imaging in patient management during the COVID-19 pandemic: a multinational consensus statement from the Fleischner Society. Radiology 2020;296(1):172–180. Google ScholarArticle HistoryReceived: Apr 20 2020Revision requested: Apr 21 2020Accepted: Apr 21 2020Published online: Apr 23 2020Published in print: Oct 2020 FiguresReferencesRelatedDetailsCited ByComputed tomography severity score as a predictor of disease severity and mortality in COVID-19 patients: A systematic review and meta-analysisJayPrakash, NaveenKumar, KhushbooSaran, Arun KumarYadav, AmitKumar, Pradip KumarBhattacharya, AnupaPrasad2023Jun1 | Journal of Medical Imaging and Radiation Sciences, Vol. 54, No. 2Infectious DiseasesBahadırErtürk, ZamirKemal Ertürk2023 | , Vol. 18Heterogeneity Analysis of Chest CT Predict Individual Prognosis of COVID-19 PatientsBoYang, BeiZhang, LichenGao, JianZhang, HuaimingQiu, WencaiHuang2022 | Current Medical Imaging Formerly Current Medical Imaging Reviews, Vol. 18, No. 3COVID-19 Imaging: What We Know Now and What Remains UnknownJeffrey P. Kanne, Harrison Bai, Adam Bernheim, Michael Chung, Linda B. Haramati, David F. Kallmes, Brent P. Little, Geoffrey Rubin, Nicola Sverzellati, 9 February 2021 | Radiology, Vol. 299, No. 3Editor's Note: 2020—A Year Like No Other for RadiologyDavid A. Bluemke, , 22 December 2020 | Radiology, Vol. 298, No. 2Passive Microwave Radiometry for the Diagnosis of Coronavirus Disease 2019 Lung Complications in KyrgyzstanBatyrOsmonov, LevOvchinnikov, ChristopherGalazis, BerikEmilov, MustafaKaraibragimov, MederSeitov, SergeyVesnin, AlexanderLosev, VladislavLevshinskii, IllarionPopov, ChingizMustafin, TuratKasymbekov, IgorGoryanin7 February 2021 | Diagnostics, Vol. 11, No. 2Diagnostic Performance of Chest CT for SARS-CoV-2 Infection in Individuals with or without COVID-19 SymptomsKristof De Smet, Dieter De Smet, Thomas Ryckaert, Emanuel Laridon, Birgit Heremans, Ruben Vandenbulcke, Ingel Demedts, Bernard Bouckaert, Stefaan Gryspeerdt, Geert A. Martens, 10 August 2020 | Radiology, Vol. 298, No. 1Diagnostic accuracy and interobserver variability of CO-RADS in patients with suspected coronavirus disease-2019: a multireader validation studyDavideBellini, NicolaPanvini, MarcoRengo, SimoneVicini, MiriamLichtner, TizianaTieghi, DeaIppoliti, FedericaGiulio, ElenaOrlando, MarioIozzino, Maria GraziaCiolfi, SarahMontechiarello, Ugod'Ambrosio, Emanueled'Adamo, ChiaraGambaretto, StefanoPanno, VanessaCaldon, CesareAmbrogi, IacopoCarbone2021 | European Radiology, Vol. 31, No. 4COVID-19 pneumonia: high diagnostic accuracy of chest CT in patients with intermediate clinical probabilityAnne LaureBrun, AlexiaGence-Breney, JulieTrichereau, Marie ChristineBallester, MarcVasse, Marie LaureChabi, FrançoisMellot, Philippe A.Grenier2021 | European Radiology, Vol. 31, No. 4Structured reporting in portable chest radiographs: An essential tool in the diagnosis of COVID-19AndrewYates, Philip J.Dempsey, SebastianVencken, Peter J.MacMahon, Barry D.Hutchinson2021 | European Journal of Radiology, Vol. 134Practical clinical and radiological models to diagnose COVID-19 based on a multicentric teleradiological emergency chest CT cohortPaulSchuster, AmandineCrombé, HubertNivet, AliceBerger, LaurentPourriol, NicolasFavard, AlbanChazot, FlorianAlonzo-Lacroix, EmileYoussof, Alexandre BenCheikh, JulienBalique, BasilePorta, FrançoisPetitpierre, GrégoireBouquet, CharlesMastier, FlavieBratan, Jean-FrançoisBergerot, VivienThomson, NathanBanaste, GuillaumeGorincour2021 | Scientific Reports, Vol. 11, No. 1Bounding infection prevalence by bounding selectivity and accuracy of tests: with application to early COVID-19JörgStoye2021 | The Econometrics Journal, Vol. 25, No. 1Radiological Society of North America (RSNA) Expert Consensus Statement Related to Chest CT Findings in COVID-19 Versus CO-RADS: Comparison of Reporting System Performance Among Chest Radiologists and End-User PreferenceSiobhan B.O' Neill, DanielleByrne, Nestor L.Müller, SabeenaJalal, WilliamParker, SavvasNicolaou, Ana-MariaBilawich2021 | Canadian Association of Radiologists Journal, Vol. 72, No. 4Role of Imaging in Diagnosis and Management of COVID-19: A Multiorgan Multimodality Imaging ReviewVinithraVaradarajan, MahsimaShabani, BharathAmbale Venkatesh, Joao A. C.Lima2021 | Frontiers in Medicine, Vol. 8Diagnostic accuracy and false-positive rate of chest CT as compared to RT-PCR in coronavirus disease 2019 (COVID-19) pneumonia: A prospective cohort of 612 cases from India and review of literatureReddyRavikanth2021 | Indian Journal of Radiology and Imaging, Vol. 31, No. S 01COVID-19'DA KARDİYOTORASİK RADYOLOJİK GÖRÜNTÜLEME VE YAPAY ZEKANIN ROLÜVeysel AtillaAYYILDIZ2021 | SDÜ Tıp Fakültesi Dergisi, Vol. 28, No. COVİD-19 ÖZEL SAYISuboptimal Quality and High Risk of Bias in Diagnostic Test Accuracy Studies at Chest Radiography and CT in the Acute Setting of the COVID-19 Pandemic: A Systematic ReviewDominika Suchá, Robbert W. van Hamersvelt, Andor F. van den Hoven, Pim A. de Jong, Helena M. Verkooijen, 30 July 2020 | Radiology: Cardiothoracic Imaging, Vol. 2, No. 4Comparison of admission chest computed tomography and lung ultrasound performance for diagnosis of COVID-19 pneumonia in populations with different disease prevalenceDavideColombi, MarcelloPetrini, GabrieleMaffi, Gabriele D.Villani, Flavio C.Bodini, NicolaMorelli, GianlucaMilanese, MarioSilva, NicolaSverzellati, EmanueleMichieletti2020 | European Journal of Radiology, Vol. 133Artificial intelligence for the detection of COVID-19 pneumonia on chest CT using multinational datasetsStephanie A.Harmon, Thomas H.Sanford, ShengXu, Evrim B.Turkbey, HolgerRoth, ZiyueXu, DongYang, AndriyMyronenko, VictoriaAnderson, AmelAmalou, MaximeBlain, MichaelKassin, DilaraLong, NicoleVarble, Stephanie M.Walker, UlasBagci, Anna MariaIerardi, ElviraStellato, Guido GiovanniPlensich, GiuseppeFranceschelli, CristianoGirlando, GiovanniIrmici, DominicLabella, DimaHammoud, AshkanMalayeri, ElizabethJones, Ronald M.Summers, Peter L.Choyke, DaguangXu, MonaFlores, KakuTamura, HirofumiObinata, HitoshiMori, FrancescaPatella, MaurizioCariati, GianpaoloCarrafiello, PengAn, Bradford J.Wood, BarisTurkbey2020 | Nature Communications, Vol. 11, No. 1Accompanying This ArticleCOVID-19 Update #7 (5/05/20)May 5 2020Default Digital Object SeriesRecommended Articles Chest CT Findings 1 Year after COVID-19: Another Piece of the Post-Pandemic PuzzleRadiology2023Volume: 308Issue: 1Radiological Society of North America Expert Consensus Document on Reporting Chest CT Findings Related to COVID-19: Endorsed by the Society of Thoracic Radiology, the American College of Radiology, and RSNARadiology: Cardiothoracic Imaging2020Volume: 2Issue: 2Natural Language Processing and Machine Learning for Detection of Respiratory Illness by Chest CT Imaging and Tracking of COVID-19 Pandemic in the United StatesRadiology: Cardiothoracic Imaging2021Volume: 3Issue: 1COVID-19 in the Radiology Literature: A Look BackRadiology: Cardiothoracic Imaging2022Volume: 4Issue: 4Diagnostic Accuracy of North America Expert Consensus Statement on Reporting CT Findings in Patients Suspected of Having COVID-19 Infection: An Italian Single-Center ExperienceRadiology: Cardiothoracic Imaging2020Volume: 2Issue: 4See More RSNA Education Exhibits A Comprehensive Imaging Review of COVID-19 Pneumonia with Focus on RSNA Expert Consensus, Fleischner Society Statement and ACR Recommendations: Challenges, Opportunities and Future Direction Digital Posters2020Ultra-low-dose Lung CT For COVID-19 Viral Pneumonia Using Deep Learning Reconstruction.Digital Posters2021Reviewing Idiopathic Interstitial Pneumonia's Radiographic Features According to the Latest American Thoracic Society / European Respiratory Society UpdatesDigital Posters2020 RSNA Case Collection Severe Coronavirus disease 2019 (COVID-19)RSNA Case Collection2020COVID-19 presenting with fever, headache, and chest tightnessRSNA Case Collection2020Spontaneous pneumomediastinum in COVID-19 pneumoniaRSNA Case Collection2020 Vol. 297, No. 1 PodcastSupplemental MaterialAbbreviations Abbreviations: COVID-19 coronavirus disease 2019 NPV negative predictive value PPV positive predictive value RT-PCR reverse-transcriptase polymerase chain reaction SARS-CoV-2 severe acute respiratory syndrome coronavirus 2 Metrics Altmetric Score Open AccessThis article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.PDF download
Referência(s)