Background	Choose Top of pageABSTRACTBackground <<Thoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination VettingProceduresModality ConsiderationsConclusionReferences

A cluster of pneumonia cases associated with a novel coronavirus was first identified in the city of Wuhan, China, in December 2019 [1]. In February 2020, the disease caused by this virus was named COVID-19 by the World Health Organization [2], and the coronavirus itself was designated severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by the Coronavirus Study Group of the International Committee on Taxonomy of Viruses [3].

Transmission is thought to occur primarily via respiratory droplets. Respiratory secretions released when a SARS-CoV-2–positive patient coughs, sneezes, or talks may reach the mucous membranes of another person either directly or indirectly through touching of a contaminated surface and subsequent touching of the mucous membranes. SARS-CoV-2 can remain viable on surfaces for many hours [4]. This has an impact on imaging room turnover between patients. In addition, it has been estimated that up to 86% of patients are infected by individuals with undocumented cases (i.e., individuals with mild, limited, or no symptoms), on the basis of models developed using information from reported infections within China [5].

Patients with COVID-19 most commonly present with respiratory illness, including fever (98–99% of patients), dry cough (59–76%), and myalgias or fatigue (44–70%) [6, 7]. Less common symptoms are headache (8% of patients), diarrhea (3%) [6, 7], and anosmia [8]. However, many patients with COVID-19 have no symptoms.

The complications most commonly resulting in ICU admission are acute respiratory distress syndrome (61%) and arrhythmia (44%) [8]. Laboratory findings that have also been associated with worse outcomes include elevated liver enzyme levels, an elevated lactate dehydrogenase level, an elevated d-dimer level, and acute kidney injury [9].

The scientific and medical communities are actively working to develop and validate effective treatments and a vaccine; however, at the time of the writing of this article, treatment is primarily supportive care.

Thoracic Imaging	Choose Top of pageABSTRACTBackgroundThoracic Imaging <<Abdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination VettingProceduresModality ConsiderationsConclusionReferences

Chest radiography is less sensitive in detecting subtle lung parenchymal changes, which can be better appreciated on CT. Nonetheless, chest radiography remains the first imaging test for any patient presenting with symptoms of respiratory infection, and chest radiographic findings could be normal or could show focal or multifocal ill-defined airspace opacities in patients who are infected with SARS-CoV-2 (Fig. 1).

View larger version (272K)

Fig. 1A —53-year-old man with coronavirus disease (COVID-19) pneumonia and medical history of hypertension and type 2 diabetes mellitus. Patient presented to emergency department with chief complaint of worsening progressive shortness of breath of 2 days' duration. One week earlier, patient had tested positive for virus that causes COVID-19 when he presented with flulike symptoms with fever, body aches, headache, and sore throat. A, Portable chest radiograph obtained in emergency department shows bilateral, ill-defined, somewhat hazy airspace opacities in both lungs, with greater number of opacities seen in right lung compared with left lung.

View larger version (281K)

Fig. 1B —53-year-old man with coronavirus disease (COVID-19) pneumonia and medical history of hypertension and type 2 diabetes mellitus. Patient presented to emergency department with chief complaint of worsening progressive shortness of breath of 2 days' duration. One week earlier, patient had tested positive for virus that causes COVID-19 when he presented with flulike symptoms with fever, body aches, headache, and sore throat. B, Follow-up portable chest radiograph obtained 10 hours after radiograph in A for evaluation of proper positioning of central catheter and nasogastric tube shows progression of disease in both lungs.

Although CT is more sensitive than radiography in detecting subtle lung parenchymal changes, normal chest CT findings do not imply that a person does not have COVID-19, and an abnormal CT finding is not specific for a COVID-19 diagnosis. The radiographic findings of COVID-19 vary according to the stage of the disease. In the early stages of the disease, there may be no imaging findings or very subtle nodular ground-glass opacities (GGO), which enlarge over time to resemble a more rounded configuration and which later progress to denser consolidations [10] (Fig. 2). The halo sign, the crazy paving pattern caused by interlobular septal thickening and GGO and the reverse halo sign that mimics organizing pneumonia have all been described [11] (Fig. 3). The distribution of these findings is almost always peripheral without subpleural sparing and is bilateral with lower lobe dominance. Rounded lung opacities were present on up to one-third of chest CT scans of patients with COVID-19 who had symptoms [12]. Li et al. [13] reported seeing a spectrum of findings in the very early stages of COVID-19, ranging from normal to subtle centrilobular nodules to more classic multifocal GGO. The Society of Thoracic Radiology, the American College of Radiology (ACR), and the Radiological Society of North America recently produced a consensus statement on reporting chest CT findings related to COVID-19; this statement includes a clear description of the typical, indeterminate, and atypical appearances of the disease on chest CT [14].

View larger version (316K)

Fig. 2 —46-year-old man suspected of having virus that causes coronavirus disease (COVID-19) with increasing shortness of breath worrisome for pulmonary embolism. Contrast-enhanced chest CT scan shows characteristic bilateral, peripheral rounded ground-glass opacities.

View larger version (268K)

Fig. 3A —Two patients with coronavirus disease (COVID-19) pneumonia. A, 76-year-old who presented with memory loss. Image from neck CT angiography shows incidentally noted ground-glass opacities with crazy paving pattern (arrow).

View larger version (314K)

Fig. 3B —Two patients with coronavirus disease (COVID-19) pneumonia. B, 72-year-old man with esophageal stent who presented with hemoptysis. Chest CT performed on this patient showed ground-glass opacities and some focal nodular areas of consolidation. Patient later tested positive for virus that causes coronavirus disease (COVID-19). Contrast-enhanced chest CT scan shows ground-glass opacity in left lower lobe with focal nodular areas of consolidative change (solid arrow). Subtle ground-glass opacity (dashed arrow) is also seen in right lower lobe.

Pleural effusions, cavitation, and enlarged lymph nodes are rarely encountered in the early stages of COVID-19 but may become apparent in later stages. For patients who undergo intubation in the ICU, radiographic findings often show progression, with increasing multifocal lung consolidation ultimately leading to the appearance of acute respiratory distress syndrome. In patients with underlying lung conditions, such as chronic obstructive pulmonary disease or interstitial lung disease, findings may have an atypical pattern, which can pose difficulties in making a diagnosis.

We have seen patients initially admitted for trauma who showed incidental GGO not related to trauma or who even had these changes develop after admission to the hospital, suggesting either coexistence of COVID-19 in individuals without symptoms who were involved in motor vehicle collisions or spread of infection among hospitalized patients.

In a longitudinal study, Wang et al. [10] described the temporal changes in CT findings for 90 patients with COVID-19 pneumonia. In their cohort, the extent of CT abnormalities progressed rapidly after symptom onset and peaked during days 6–11 of illness [10]. GGO were the predominant pattern of abnormalities after symptom onset. The percentage of patients with a mixed pattern of abnormalities (GGO and dense consolidation) peaked during days 12–17 of illness and became the second most prevalent pattern thereafter. Sixty-six of the 70 patients (94%) who were discharged from the hospital had residual disease on final CT scans, with GGO the most commonly seen pattern.

At UAB, the following guidelines were adopted for chest imaging of any patients suspected of being infected with SARS-CoV-2 (Table 1) and were approved by UAB's COVID-19 task force. To better prepare the imaging technologists for every patient encounter, ordering providers were asked to specifically state whether any patient undergoing an imaging study was suspected of having COVID-19. These institutional guidelines were developed on the basis of recommendations from the American College of Radiology that were updated on March 22, 2020 [15]. According to the recommendations, CT should not be used as a screening examination or a first-line test for patients suspected to have COVID-19. Rather, it should be used sparingly and be limited to hospitalized patients showing symptoms of COVID-19 who also have specific clinical indications for CT. When CT is performed for patients with COVID-19, the scanner should be appropriately cleaned before subsequent patients are scanned. The ACR further recommends use of portable radiography units in ambulatory care facilities for instances when chest radiography is clinically warranted. The advantages of using these units are that they can be easily cleaned and patients need not be brought to the radiology department.

View Larger Version

TABLE 1: Suggested Chest Imaging Pathways for Patients With Confirmed and Suspected SARS-CoV-2 Infection as Implemented at the University of Alabama in Birmingham, Alabama

For patients admitted to the hospital or seen in the emergency department, portable radiography is frequently ordered for those who present with cough or fever. To limit exposure to other patients in the radiology department and limit patient transfers, radiography can be performed using a portable unit. When examining such patients, the technologist has to don personal protective equipment (PPE), and once the radiographic examination is complete, the technologist needs to do? the PPE gear, preferably with a trained observer monitoring removal of the PPE. Another method used at one institution was to perform chest radiography either through the glass of an isolation room door or at a greater-than-usual distance (i.e., 10–15 feet [3.0–4.6 meters]) across a semi-isolation antechamber adjacent to an isolation room [16].

Abdominopelvic Imaging	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic Imaging <<NeuroradiologyMusculoskeletal ImagingExamination VettingProceduresModality ConsiderationsConclusionReferences

The abnormal laboratory test findings associated with more severe cases of COVID-19, including an elevated d-dimer level, an elevated serum creatinine level resulting from acute renal failure, and elevated liver function test levels, may prompt requests for imaging. In addition, imaging may be requested to evaluate other concomitant medical conditions, such as the transplant viability in patients who have previously undergone solid organ transplants. At Emory University and at UAB, radiologists and sonographers worked together to create abbreviated ultrasound protocols designed to answer the clinical question at hand while minimizing sonographer exposure times.

For example, at Emory University, before the COVID-19 pandemic, the standard unilateral lower extremity ultrasound protocol used to evaluate for deep vein thrombus was built onto ultrasound equipment using vendor software (Scan Assistant [GE Healthcare] and SmartExam [Philips Healthcare]) and consisted of a minimum of 24 saved images, including gray-scale images with and without compression, color Doppler ultrasound images, and spectral images covering the area from the groin through the infrapopliteal veins. The protocol was reformulated for SARS-CoV-2–positive patients and PUI so that when examination results are negative, sonographers now must save only gray-scale images documenting compression of the common femoral vein, femoral vein, deep femoral vein, popliteal vein, and infrapopliteal veins with a color Doppler and spectral image saved at a single level. If there is a deep vein thrombus, sonographers are to add color Doppler imaging to determine whether the thrombus is occlusive or nonocclusive.

In addition, a point-of-care ultrasound protocol developed by the critical care team serves as a first-pass screening examination before initiation of a request for a diagnostic study that is performed by registered sonographers and interpreted by diagnostic radiologists [17]. Also at Emory University, renal ultrasound is not performed for SARS-CoV-2–positive patients unless it is approved by a nephrologist, because an elevated serum creatinine level is known to be part of the disease process and because performing renal ultrasound examinations for all SARS-CoV-2–positive patients in this scenario is thought to provide a low yield and would unnecessarily expose sonographers and expend PPE.

Although diarrhea has been described as an uncommon symptom of COVID-19, currently available information does not indicate that COVID-19 itself results in an acute abdomen. However, SARS-CoV-2–positive patients may have other comorbidities or other concomitant conditions that warrant abdominal imaging with CT or MRI. Also, characteristic GGO in the lung may be visible in the lung bases of SARS-CoV-2–positive patients undergoing CT of the abdomen and pelvis (Fig. 4).

View larger version (231K)

Fig. 4A —42-year-old man with coronavirus disease (COVID-19) pneumonia. A, Axial (A) and coronal (B) images from contrast-enhanced CT of abdomen and pelvis obtained to evaluate abdominal pain and diarrhea show rounded, peripheral ground-glass pulmonary opacities (arrows).

View larger version (395K)

Fig. 4B —42-year-old man with coronavirus disease (COVID-19) pneumonia. B, Axial (A) and coronal (B) images from contrast-enhanced CT of abdomen and pelvis obtained to evaluate abdominal pain and diarrhea show rounded, peripheral ground-glass pulmonary opacities (arrows).

Neuroradiology	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiology <<Musculoskeletal ImagingExamination VettingProceduresModality ConsiderationsConclusionReferences

Although the most common clinical presentation of COVID-19 in patients is acute respiratory illness, there is growing awareness of the neurotropic potential of coronaviruses. In particular, this group of viruses has an affinity for the angiotensin-converting enzyme 2 receptor, which is used for intracellular entry and is expressed in various tissues, including the CNS [18]. Neurologic symptoms, such as headache, dizziness, altered mental status, ataxia, anosmia, and ageusia, have been reported in SARS-CoV-2–positive patients [8, 19]. Furthermore, CNS invasion has been reported for most of the coronaviruses, with brainstem infection seen in patients infected with SARS-associated coronavirus and experimental animals, and this may contribute to respiratory failure in SARS-CoV-2–positive patients [20].

Brain involvement by coronaviruses has also been documented on imaging. Diffuse, bilateral nonenhancing lesions on brain MRI have been reported in patients with Middle East respiratory syndrome–associated coronavirus [21]. A recent case report of a single patient with COVID-19 showed findings similar to those of acute hemorrhagic necrotizing encephalopathy on head CT and brain MRI [22]. It is unclear whether these imaging findings reflect direct viral CNS invasion or whether they result from other processes, such as anoxia or an intracranial cytokine storm. Regardless, brain imaging of SARS-CoV-2–positive patients may be indicated because, in addition to COVID-19–associated direct CNS involvement, these patients may have other comorbidities requiring neuroimaging or may have CNS disorders such as stroke develop.

Musculoskeletal Imaging	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal Imaging <<Examination VettingProceduresModality ConsiderationsConclusionReferences

To our knowledge, no musculoskeletal imaging findings are known to be associated with COVID-19. During this pandemic, musculoskeletal imaging has been restricted to reduce patient, provider, and technologist exposure. Examples of allowable imaging are primarily related to tumor, infection, and trauma evaluation. Trauma evaluation not only applies to imaging of fractures but also allows the diagnosis of ligament and tendon injuries, which require urgent repair to reduce long-term morbidity. When possible, the SARS-CoV-2 infection status of these patients is determined before imaging is performed. At UAB, all advanced musculoskeletal imaging is approved by a radiologist or radiology resident.

Examination Vetting	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination Vetting <<ProceduresModality ConsiderationsConclusionReferences

From the beginning, one of the biggest challenges faced by radiologists and clinicians was delineating between imaging that was to be considered essential and that which was considered nonessential. At UAB, once general guidance was set by the U.S. Centers for Disease Control and Prevention and the hospital, more specific delineation between essential and nonessential imaging was made at a department and section level or modality level with input from our referring clinics and clinicians. At UAB, imaging was limited to emergent or essential imaging only, with essential imaging loosely categorized as imaging centered around oncology, infection, acute ischemic events, and acute neurologic changes. Each radiology section was regularly updated with a list of scheduled examinations. The individual sections would then evaluate their upcoming examinations to determine whether they fit into one of the aforementioned categories or whether they clearly were nonessential. For examinations that did not fall neatly into one of these categories, radiologists within the sections were strongly encouraged to reach out to the ordering clinician for clarification, verification, or both regarding whether the imaging was deemed essential. For some modalities, such as ultrasound, it was relatively easy to create a list of examinations that fit or did not fit into one of the aforementioned categories. For other modalities, however, there were often gray areas. The department stressed the need to engage referring clinicians and involve them in the decision-making process whenever there was any doubt as to whether the examination was deemed to be essential or emergent. As a result, the referring clinicians were critical partners in helping the hospital and our department limit exposure to patients, radiologists, and staff.

For patients that are PUI or are known to be infected with SARS-CoV-2, there needs to be a coordinated approach that includes radiologists, radiology technologists, and referring clinicians, to avoid the spread of SARS-CoV-2 to other patients or health care employees. At all institutions, radiologists are actively involved in vetting examinations ordered for PUI and SARS-CoV-2–positive patients. For example, radiologists are reviewing prior imaging to determine whether the indication provided for a requested ultrasound examination has already been addressed with prior imaging, such as a recently obtained previous CT scan that revealed no hydronephrosis. For ultrasound examinations that are determined to be indicated, sonographers and radiologists, often in consultation with the ordering provider, are working together so that the ultrasound examination can be focused on the clinical question. For example, a complete abdominal ultrasound examination typically includes multiple images of the liver, gallbladder, bile duct, pancreas, both kidneys, spleen, aorta, and inferior vena cava. Sonographers and radiologists are now reviewing the electronic medical record to determine the true reason for the examination and, when possible, are changing the examination to a limited abdominal or retroperitoneal ultrasound focusing on the area of interest (e.g., the liver or kidneys). Although some examinations (e.g., complete obstetric ultrasound examination for fetal anatomy assessment) cannot be easily modified, follow-up obstetric ultrasound examinations for evaluation of growth can be abbreviated to the essential images, such as biometry (three images for head, abdomen, and femur measurement) and four quadrants for amniotic fluid assessment. Cine clips can be obtained to prevent the need for additional imaging to allow for troubleshooting at the time of interpretation by the radiologist. Where possible, remote viewing capabilities can be developed so the resident or attending radiologists can view the images and provide guidance in real time to the sonographer.

Neuroimaging is unique in its reliance on MRI, a resource that is more restricted in its availability and includes unique challenges because of the magnetic field. At Emory University, an algorithm was developed for imaging PUI and SARS-CoV-2–positive patients with neurologic symptoms. Initially, and most importantly, the indication is reviewed to determine the appropriateness, the preferred modality, and the acuity. The modality choice may be overlooked and may lead to unnecessary exposure of health care employees and further downtime for equipment during cleaning. For example, head CT is often ordered as the initial imaging examination of a patient with neurologic symptoms because of its ready availability. However, many patients require brain MRI to obtain a more definitive diagnosis. Therefore, for most patients, we think it is more appropriate to bypass CT and move directly to MRI.

The acuity of the requested study is another important factor that needs to be considered before neuroimaging or neuroradiology procedures are performed, particularly for PUI. Given the increasing availability of SARS-CoV-2 testing and the development of newer, more rapid testing methods, delaying none-mergent studies for patients who are PUI may obviate the issue if they have negative SARS-CoV-2 test results. With the development of in-house testing at these academic institutions, many nonemergent studies can be delayed until SARS-CoV-2 test results are available.

Procedures	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination VettingProcedures <<Modality ConsiderationsConclusionReferences

Special considerations are centered around procedures. In particular, early in the pandemic, many hospitals had shortages of PPE, significantly limiting the procedural capabilities of most departments for PUI, SARS-CoV-2–positive, and noninfected patients. In addition, SARS-CoV-2 testing was not readily available in the early stages, placing patients and staff at risk. Several fluoroscopic gastrointestinal procedures are considered aerosolizing procedures, such as modified barium swallows and upper gastrointestinal examinations. Other procedures, such as thoracentesis, have the potential to elicit cough, whereas other examinations, such as thyroid node fine-needle aspiration biopsies, head and neck biopsies, and vascular and interventional radiology procedures that require access from the neck and upper extremities, placed clinicians at higher risk. For such reasons, these examinations required a higher level of triage to ensure the safety of staff and clinicians as well as conservation of PPE. At UAB, the current policy for procedures requiring moderate or deep sedation is that patients be tested 48 hours before their procedure is performed to confirm their SARS-CoV-2 infection status; testing of patients is also required for all aerosolizing and potential aerosolizing procedures, such as feeding tube placement or upper gastrointestinal examinations during which patients may be asked to cough to elicit reflux. All inpatients and patients undergoing an invasive procedure (including interventional procedures and biopsies) are now required to be tested at the time of admission at the University of Washington.

Modality Considerations	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination VettingProceduresModality Considerations <<ConclusionReferences

Portable ultrasound and radiography examinations are being performed for SARS-CoV-2–positive patients and PUI so that examination rooms do not have to undergo a terminal cleaning that would render them temporarily offline for use by other patients. At institutions where ultrasound equipment is available in ICUs, such as the University of Washington, point-of-care ultrasound, and procedures such as paracentesis or thoracentesis, are being performed at the bedside by ICU providers when possible. At Emory University, smaller, mobile ultrasound machines from outpatient facilities have been redeployed to facilitate greater access to point-of-care ultrasound in critical care units and the emergency department. These outpatient facilities were closed initially in response to restrictions on the performance of elective care, and they remain closed because of shelter-in-place orders and social distancing recommendations intended to flatten the curve of new infections. Technologists follow institutional guidelines for PPE use and machine cleaning when performing portable examinations.

To our knowledge, to date none of the institutions have shortened CT or MRI acquisition protocols for SARS-CoV-2–positive patients and PUI. After imaging, the CT or MRI examination room then undergoes a terminal cleaning according to local infection control recommendations. This means that the machine is temporarily offline for use by other patients and that adherence to institutional guidelines (e.g., patient masking) is maintained. At Emory University, each modality-specific examination room has been evaluated by environmental facilities engineers and infection prevention specialists to determine downtime, as customized by the rate of air exchange in the room. In addition, mobile high-efficiency particulate air filters have been placed in rooms that have been designated as locations of first choice for the imaging of PUI and SARS-CoV-2–positive patients. High-efficiency particulate air filters were not placed in MRI rooms because of concerns related to the impact on the magnetic field. Imaging PUI and SARS-CoV-2–positive patients with MRI poses challenges related to the static magnetic field. At UAB, for nonintubated PUI and SARS-CoV-2–positive patients, full PPE, including an N95 mask for the patient and staff, is required. Some of the masks contain ferromagnetic metallic strips (e.g., with the strips placed across the nose or fasteners located at the edges of the masks) that may experience translational and torque forces, which could be strong enough to unseal the mask. This may allow the dispersal of infectious respiratory droplets from patients or expose staff to these droplets. Therefore, it is important to be aware of mask composition before a mask is used in MRI. Imaging of intubated PUI and SARS-CoV-2–positive patients further raises the issues of MRI-safe ventilators and patient transfers. The MRI-safe ventilators in use at Emory University have a closed loop with a high-efficiency particulate air filter in place. However, if patients are transferred to the MRI-safe ventilator within the MRI suite, terminal cleaning is required. Emory University has instituted a process in which the MRI-safe ventilator is transported to the patient's room and the ventilator exchange occurs at the bedside. The patient is then transported to the MRI examination room using the closed-loop, MRI-safe ventilator, which decreases exposure to other patients and health care workers. The process is reversed once the imaging study is completed. As long as the closed-loop system remains intact, terminal cleaning is not necessary and a less-thorough, routine cleaning may be performed. If the closed loop is broken, if the patient were to remove their mask during an examination, or if both of these situations should occur, terminal cleaning is required, resulting in examination room downtime of approximately 3 hours, significantly decreasing MRI patient throughput.

Conclusion	Choose Top of pageABSTRACTBackgroundThoracic ImagingAbdominopelvic ImagingNeuroradiologyMusculoskeletal ImagingExamination VettingProceduresModality ConsiderationsConclusion <<References

The COVID-19 pandemic has prompted radiology departments across the United States to modify their workflows to allow optimized inpatient imaging of PUI and SARS-CoV-2–positive patients when necessary, while also preserving the health of other patients and health care workers. Imaging findings of COVID-19 in the chest have been extensively described in recent publications. Imaging workflows will likely evolve further as the radiology community continues to learn more about COVID-19, including transmission of SARS-CoV-2, its natural history, and imaging manifestations in other anatomic areas.