Introduction
In December 2019 a clustering of pneumonia was reported for the first time in Wuhan,
Hubei, China [1]. A coronavirus was identified as the pathogen, which has since been given the name
Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). The disease caused by
SARS-CoV-2 is called COVID-19 which has since spread worldwide. In Germany the numbers
continue to increase, and the level of COVID-19 lethality in Europe cannot yet be
accurately estimated. However, lethality appears to be particularly higher in older
patients compared to seasonal influenza [2]
[3]. Based on experience in China, computed tomography (CT) is very helpful in COVID-19
in that it can produce an image that is suggestive for diagnosis, and the sensitivity
of CT is in some cases higher than that of the current gold standard, real-time polymerase
chain reaction (RT-PCR) [4]
[5].
The aim of this article is to present a case with COVID-19 pneumonia and to provide
an overview of the existing radiological literature on COVID-19.
Case Report
The inpatient admission of a 59-year-old male was carried out with outpatient-confirmed
SARS-CoV-2 infection with positive RT-PCR from a throat swab and relevant clinical
symptoms. The patient had previously visited South Tirol. Six days prior to hospital
admission, the patient exhibited initial symptoms in the form of chills and fever.
He had no pre-existing conditions.
At the time of admission there was a high fever of 40.0 °C and conspicuous tachypnea
with a respiratory rate of 28/min with sufficient oxygenation (ambient air saturation
95 %). Clinically, there was also a severe general feeling of malaise with obvious
fatigue. Laboratory results showed leukopenia with 3.8 × 109/l, lymphopenia with 0.96 × 109/l (25 %), an only slightly increased CRP with 14.7 mg/l and a low procalcitonin of
0.11 ng/ml. D-dimer as risk factor for instability was not increased at 370 µg. Interleukin-6
as inflammatory parameter was slightly elevated at 36.4 pg/ml, LDH was 330 U/l and
creatinine kinase was 412 U/l. Albumin was initially within the normal range at 4.0 g/dl.
The native thin-slice volume CT of the thorax performed on the day of admission ([Fig. 1]) showed multifocal, bilateral ground glass opacities, partly with overlapping reticulations
appearing as crazy paving. Consolidation was present, but much less pronounced than
the ground glass opacities, which were somewhat round and geographically arranged.
It was noticeable that they were sometimes sharply delineated from the neighboring
healthy lung parenchyma. A positive air bronchogram was evident within the changes.
All changes were in the periphery of the lungs, in the craniocaudal plane the base
of the lungs, and here the left lower lobe, were most affected.
Fig. 1 Non-contrast enhanced thin slice Volume-CT at day of admission: MPRs in axial a–c, coronal d and sagittal e orientation. Bilaterally multiple patchy ground glass opacities and crazy paving
are seen. Some of the opacities are round and some geopgraphic shaped. Partially,
the lesions are sharply demarcated against the surrounding healthy lung. The left
lower lobe and the periphery of the lungs are predominantly involved. Images courtesy
PD Dr. med W. Kurre.
Due to the presence of a risk constellation with clinically significant compromise,
high fever, respiratory impairment and CT-morphologically pronounced parenchymal alterations,
a decision was quickly made in favor of a therapy including antivirals (Lopinavir-Ritonavir)
and a broadly-calculated antibiotic therapy. With persistent fever up to 39.6 °C and
respiratory deterioration with increasing oxygen demand, a follow-up CT of the thorax
was performed on day 6 after admission ([Fig. 2]). The images showed significantly increasing densities, whereby in contrast to the
initial CT, consolidations now dominated the image. Still, the periphery and the base
of the lungs were most affected. On day 7 a nasal high flow therapy was initiated
for severe respiratory insufficiency. By day 16 the patient had improved clinically
and was ready for discharge. No more follow-up CT examination was performed.
Fig. 2 Non-contrast enhanced thin slice Volume-CT 5 days later: MPRs in axial a–c), coronal d and sagittal e orientation. The opacities have markedly increased in size. Now, consolidation is
the predominant feature. Still mainly the lower lobes and the periphery of the lungs
are involved. Images courtesy PD Dr. med W. Kurre.
Materials and Methods
A literature search was conducted in the “pubmed” database with the keywords “COVID-19
OR SARS-CoV 2”. As of March 10, 2020, 637 articles were identified. Based on these
articles, all studies were identified that included at least 10 patients with proven
COVID-19 who had received a chest CT and whose characteristics had been systematically
evaluated. Fourteen studies met the inclusion criteria [5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18], describing a total of 1966 patients. In addition to the data, images of all studies
were reviewed.
Results
HRCT signs of SARS CoV-2 pneumonia
The most common manifestation of pneumonia caused by SARS-CoV-2 is ground glass opacities,
followed by a mixed picture of ground glass opacities and consolidation, followed
by consolidation alone ([Table 1]). Nodules are rare, tree-in-bud was not mentioned in any study. Crazy paving was
seen, but was present in highly variable frequency. However, the phrasing in some
studies suggested that crazy paving may have been present (“reticulations”), but this
was not explicitly described as such. These cases were not included in the overview
table. Two studies reported that the lesions were round (in 33 % and 54 % of cases,
respectively). Three studies investigated vessel width and reported dilatation of
vessels within the opacifications (82 %, 71 % and 59 % of cases). Reversed halo was
observed in approx. 5 % of patients. Pleural effusions or lymphadenopathy occurred
in only a limited number of patients. cavitations were not observed. In most of the
studies, lesion boundary was not an evaluation criterion, therefore the published
image material was reviewed by an author (OWH). Ground glass opacities (with or without
crazy paving) and consolidation were often at least partially sharply delineated and
geographically configured.
Table 1
Overview of CT-morphology of SARS-CoV 2 pneumonia (according to all studies including
more than 10 patients, as of March 10th 2020).
|
Bernheim
[6]
|
Pan (early stage: day 0–4) [7]
|
Chung
[8]
|
Ng
[9]
|
Song
[10]
|
Ai
[5]
|
Shi
[11]
|
Li
[13]
|
Li
[12]
|
Chen
[14]
|
Huang
[15]
|
Zhao
[16]
|
Bai
[17]
|
Xiong
[18]
|
|
epidemiology
|
|
number of patients
|
121
|
21
|
21
|
21
|
51
|
1014
|
81
|
83
|
51
|
99
|
41
|
101
|
219
|
42
|
|
male
|
61
|
6
|
13
|
13
|
25
|
467
|
42
|
44
|
28
|
67
|
30
|
56
|
119
|
25
|
|
middle age
|
45
|
40
|
51
|
56
|
49
|
51
|
50
|
46
|
58
|
56
|
49
|
44
|
45
|
50
|
|
CT signs
|
|
CT normal [%]
|
22
|
17
|
|
10
|
|
12
|
0
|
|
4
|
|
0
|
|
|
|
|
milk glass [%]
|
34
|
75
|
57
|
86
|
77
|
46
|
65
|
98
|
35
|
|
|
86
|
91
|
|
|
consolidation [%]
|
2
|
42
|
0
|
62
|
55
|
50
|
17
|
64
|
6
|
|
|
44
|
69
|
55
|
|
milk glass and consolidation [%]
|
41
|
|
29
|
|
59
|
|
|
|
55
|
|
|
64
|
|
|
|
crazy paving [%]
|
5
|
25
|
19
|
|
75
|
|
10
|
36
|
71
|
|
|
|
5
|
|
|
vasodilation in solidifications [%]
|
|
|
|
|
|
|
|
|
82
|
|
|
71
|
59
|
|
|
round configuration [%]
|
54
|
|
33
|
|
|
|
|
|
|
|
|
|
|
|
|
reversed halo [%]
|
2
|
|
|
|
|
|
|
|
4
|
|
|
|
5
|
|
|
pleural effusion [%]
|
1
|
|
0
|
0
|
8
|
|
5
|
8
|
2
|
|
|
14
|
4
|
12
|
|
lymphadenopathy [%]
|
0
|
|
0
|
0
|
6
|
|
6
|
8
|
0
|
|
|
|
3
|
29
|
|
distribution of CT signs
|
|
emphasized periphery [%]
|
52
|
54
|
33
|
86
|
86
|
|
54
|
|
|
|
|
87
|
80
|
29
|
|
right upper lobe [%]
|
44
|
|
67
|
67
|
84
|
|
19
|
65
|
|
|
|
6
|
|
|
|
left upper lobe [%]
|
48
|
|
67
|
76
|
|
20
|
86
|
|
|
|
|
|
|
middle lobe [%]
|
41
|
|
57
|
48
|
59
|
|
10
|
74
|
|
|
|
|
|
|
|
right lower lobe [%]
|
65
|
|
76
|
76
|
90
|
|
27
|
94
|
|
|
|
55
|
|
|
|
left lower lobe [%]
|
63
|
0
|
67
|
81
|
|
24
|
96
|
|
|
|
|
|
|
unilateral [%]
|
|
42
|
|
|
14
|
|
21
|
5
|
|
25
|
2
|
10
|
19
|
|
|
bilateral [%]
|
60
|
42
|
76
|
|
86
|
|
79
|
95
|
|
75
|
98
|
82
|
75
|
|
Distribution of parenchymal changes
The above described CT signs were found in the axial plane predominantly in the periphery
of the lung. The lower lobes were mainly affected in the craniocaudal plane. The vast
majority of patients showed bilateral changes.
Temporal changes of CT morphology
Follow-up CTs were performed in some studies, allowing an assessment of the course
of parenchymal changes [6]
[7]
[9]
[10]
[11]
[18]. The authors unanimously reported an increasing expansion of the lesions with, in
particular, increased consolidation and crazy paving. The maximum finding was seen
approximately on day 10 after the onset of symptoms. Afterwards there was a decrease
of the parenchymal changes, but they were not completely regressive in one work describing
up to day 26 after the onset of symptoms.
Correlation between CT findings and clinical symptoms
The extent of parenchymal changes correlated with the severity of the clinical symptoms
[7]
[13]
[15]
[16].
Thoracic CT versus RT-PCR
Four studies compared the sensitivities of CT and RT-PCR [4]
[5]
[6]
[19].
Fang et al. report an initial RT-PCR sensitivity of 71 % (36/51) and sensitivity of
initial CT (“CT compatible with viral pneumonia”) of 98 % (50/51); the interval between
RT-PCR and CT was ≤ 3 days. Of patients with abnormal CT, 72 % (36/50) exhibited “typical”
CT morphology (lung periphery predominantly involved, ground glass, consolidation)
[4]. Bernheim et al. report on 121 patients, 120 of whom had a positive RT-PCR, the
interval between RT-PCR and CT is not mentioned in the study [6]. On day 0–2 after onset of symptoms, 56 % of patients exhibited inconspicuous CT
findings, 9 % on day 3–5 and 4 % on day 6–12. Ai et al. report that 60–93 % of patients
had an abnormal CT before the initial positive PCR; the median time between RT-PCR
and CT was 1 day (0–7 days) [5]. Xie et al. report that in 3 % (5/167) of patients the RT-PCR was initially negative,
but the CT showed signs of viral pneumonia. A follow-up PCR 2–8 days after a positive
CT confirmed SARS-CoV2 infection. In 4 % (7/167) of patients the CT was initially
negative, but the PCR was positive.
Differentiation of SARS CoV-2 pneumonia and other viral pneumonias
A study by Bai et al. evaluated whether pneumonia caused by SARS-CoV2 (n = 219) can
be differentiated from viral pneumonia of a different etiology (n = 205) on the basis
of CT findings [17]. The sensitivity of 7 radiologists was between 67 % and 97 %, the specificity for
6 radiologists lay between 93 and 100 %; one radiologist, who had received no training
regarding the CT morphology of COVID-19 pneumonia, showed a specificity of 7 %. Findings
indicative of COVID-19 pneumonia were findings predominantly in the periphery of the
lung (80 % versus 57 %), the presence of ground glass opacities (91 % versus 68 %),
reticulations (56 % versus 22 %), and vasodilation (59 % versus 22 %).
Discussion
In December 2019, cases of patients with acute pneumonia, the cause of which was initially
unclear, increased in Wuhan Province, China. Within a short time, a novel coronavirus
could be isolated as a trigger [1], which WHO officially named “SARS-CoV-2”. The disease itself has been designated
“COVID-19”.
The primary transmission route of SARS-CoV-2 is droplet infection. Smear infections
do not seem to play a role, but according to publications of the Robert Koch Institute
this cannot be ruled out at this point in time.
The clinical pattern initially shows the symptoms of an infection of the respiratory
tract, although the severity can be variable with unspecific, mild symptoms up to
severe disease manifestations. The most common symptoms are cough, fever, myalgia
and headaches [20]. According to the evaluation of 72 314 patients suffering from SARS-CoV-2 by the
Chinese Center for Disease Control and Prevention, 81 % of patients showed a mild
and 14 % a severe course. Five percent of patients were critically ill with severe
respiratory insufficiency [3].
To date, no proven specific antiviral therapy for COVID-19 is known, although a number
of trials with different substances (Camostat, Remedesivir, Lopinavir-Ribonavir or
hydroxychloroquine) have been conducted [21]. The primary therapy is therefore symptomatic; in the case of serious illnesses,
intensive medical treatment including respiration and ECMO therapy may be necessary
[22]. Depending on the country, the mortality rate of the disease is between 0.5 and
3.5 % (updated figures can be found under https://www.who.int).
Diagnostic measures depend on the symptoms and the current risk profile (visit to
a risk area, contact with an infected person). With the increasing spread of COVID-19,
however, symptomatology alone will decide whether further clarification is necessary.
Detection of SARS-CoV-2 is performed by RT-PCR from a throat wash or deep throat swab.
Despite the fact that COVID-19 has only been known since December 2019, numerous reports
on CT morphology are already available; in most cases these are case reports. We identified
14 studies in which the CT characteristics of at least 10 patients each had been systematically
evaluated. A limiting factor here is that the CT technique was not standardized among
the various studies and the radiological terminology was not uniformly employed. Nevertheless,
the picture emerges that, unlike most other pathogen-induced pneumonia, pneumonia
caused by SARS-CoV-2 in many patients seems to provide a CT image, especially initially,
that is suggestive for diagnosis. The cardinal signs are ground glass opacities, consolidation
and crazy paving. The opacification are typically round or geographically configured
and at least partially show a sharply delineated border to the surrounding healthy
lung. The vast majority of cases show multifocal, bilateral involvement. The periphery
of the lung and the lower lobes are predominantly affected. Pleural effusions or lymphadenopathy
are usually not present. Overall, the CT image initially resembles a cryptogenic organizing
pneumonia pattern [23]. As the disease progresses, the lesions, especially the consolidation and crazy
paving increase, and the picture changes into that of lobar pneumonia/bronchopneumonia
or even acute respiratory distress syndrome (ARDS). Keeping in mind the CT findings
described above, it seems possible, according to a recent study, to distinguish SARS-CoV-2
pneumonia radiologically from other viral pneumonia with moderate sensitivity but
high specificity [17]. Parameters that argued for COVID-19 pneumonia were a predominant involvement of
the lunge periphery, the presence of ground glass opacity, reticulation and vasodilation.
Despite these data, there is naturally a large overlap in CT morphology with other
viral pneumonias, so that RT-PCR is still the gold standard in diagnostics.
Summary
The number of patients with COVID-19 is rapidly increasing in Europe. The lethality
of the disease appears to be higher than that of seasonal flu, for example, especially
in older patients. RT-PCR is the gold standard for establishing a diagnosis. In many
patients, chest CT seems to provide an image that is suggestive for diagnosis. The
cardinal signs are ground glass opacities, consolidation and crazy paving, predominantly
located in the periphery of the lower lobes. This CT morphology can support the diagnosis
and differentiation from other viral pneumonia. This is particularly important because
RT-PCR can initially provide a false negative result.
