Materials and Methods
A multicentre prospective study of consecutive cases was conducted across seven centres
in two large cities in Western India- Mumbai and Ahmedabad.
Two of the five centres in Mumbai were large hospitals, one private and one public
hospital, and the other three were stand alone diagnostic centres, the two Ahmedabad
centres were stand alone diagnostic centres. This provided a wide spectrum of patients.
All CT studies suggestive of COVID-19 pneumonia based on well-documented features
in literature were collated. These were evaluated independently by three radiologists
with 30 years, 15-year and 8-year experience to confirm the imaging diagnosis of COVID
pneumonia. The RT-PCR values were retrospectively obtained. Based on the RT-PCR values
the CT studies were grouped into three groups – Positive, Negative and unknown. The
unknown group were individuals who did not want to reveal their RT-PCR status or self-isolated
based on CT results.
The CT features from all three groups were compared to evaluate any communality or
discordance.
The CT features evaluated were as follows:
-
Lung opacity type:
-
Ground glass[[5], [6], [7], [8]]
-
Consolidation[[6], [7]]
-
Mixed pattern of consolidation/ground glass
-
Location of abnormality:
-
Subpleural, peribronchovascular, mixed[[9]]
-
Anterior, posterior[[10]]
-
Upper lobe, lower lobe, multi-lobar.
-
Morphology of opacity:
-
Crazy paving[[5], [11], [12]]
-
Atoll sign[[13], [14]]
-
Halo sign[[8], [15]]
-
Vacuolar sign[[16]]
-
Prominent vessels in opacity[[7], [8]]
-
Dilated bronchioles in opacity[[7], [17], [18]]
-
Subpleural curvilinear fibrosis.[[13]]
Results
Patient characteristic
Totally, 2,581 patients with positive CT findings for COVID pneumonia were collated.
Of them 825 were females and 1,756 were male patients. The age group included patients
from 28 years to 90 years; mean age of 60 +/-30 years.
CT evaluation in RT PCR Subgroups: [[Table 1]]
Table 1
Subgroup features
|
Group A (n=1445)
|
Group B (n=477)
|
Group C (n=659)
|
Total (n=2581)
|
Statistical Significance P
|
|
-
|
56%
|
18.5%
|
25.5%
|
-
|
-
|
|
Mixed (Peripheral + Central)
|
923 63.9%
|
254 53.2%
|
349 53%
|
1526 59.12%
|
|
|
Peripheral
|
499 34.5%
|
211 44.2%
|
285 43.2%
|
995 38.55%
|
<0.001
|
|
Central
|
23 1.6%
|
12 2.5%
|
25 3.8%
|
60 2.33%
|
|
|
Posterior
|
133092%
|
438 91.8%
|
607 92.1%
|
2375 92.0%
|
|
|
Anterior
|
115 7.9%
|
39 8.1%
|
52 7.9%
|
206 8.0%
|
0.001
|
|
GGO (Pure GGO)
|
1 105 76.5%
|
392 82.2%
|
546 82.8%
|
2043 79.15%
|
|
|
Consolidations
|
75 5.2%
|
10 2.1%
|
9 1.4%
|
94 3.64%
|
<0.001
|
|
Mixed (GGO + Consolidations)
|
265 18.3%
|
75 15.7%
|
104 15.8%
|
444 17.21%
|
|
|
Upper
|
43 3%
|
17 3.5%
|
36 5.4%
|
96 3.7%
|
|
|
Lower
|
246 17%
|
169 35.4%
|
242 36.7%
|
657 25.4%
|
<0.001
|
|
Multilobar
|
115680%
|
291 61%
|
381 57.9%
|
1828 70.9%
|
|
Group A - 1445 (56%) positive RT PCR.
Group B - 477 (18.5%) negative initial RT PCR.
Group C - 659 (25.5%) unknown RT PCR results.
CT Features across subgroups: [[Table 2]]
Table 2
CT features across subgroups
|
Pattern
|
Percentage n=2581
|
|
Predominant GGO
|
94.8%
|
|
Predominant Peripheral distribution
|
38.55%
|
|
Mixed (Central and peripheral) distribution
|
59.12%
|
|
Subpleural lines
|
45.3%
|
|
Crazy Paving
|
17.7%
|
|
Halo/reverse halo sign
|
17.9%
|
|
Prominent vessels
|
24.7%
|
|
Atoll sign
|
5%
|
|
Vacuolar sign
|
14.4%
|
Predominant CT features observed in all the patients were Ground glass densities (94.8%),
in mixed distribution: peripheral and central (59.12%), in the posterior segments
(92%) and multilobar involvement in 70.9% patients.
The ancillary features like Crazy Paving, Atoll sign, halo sign, prominent vessel,
prominent bronchiole and vacuolar signs were seen in smaller groups of patients; Collectively
seen in 17-45% of the patients.
Discussion
COVID-19 is a highly infectious disease caused by a single-stranded RNA corona virus-
SARS COV 2.[[19]] The main routes of spread are human to human via droplets as well as surface contamination.
The key to control of COVID-19 is to break the human to human contact chain. To achieve
this, early detection and prompt isolation is imperative. RT-PCR is the current gold
standard to detect SARS-CoV-2; however, the false negative rate ranges from 30 to
40%.[[1], [2]] There are numerous reasons for this high false negative rate, these are related
to sampling, transportation and processing errors. RT-PCR may also be negative if
the viral load is low, a second/third or fourth RT-PCR is positive.[[3], [4]] Sensitivity of RT-PCR is maximum between 5th and 7th day, lower before day 5, and peters off after day 7 of contracting the infection.[[9], [20]] Additionally in resource constrained regions RT-PCR may not be freely available,
or if available with a significant turnaround time spanning from 24 to 48 hrs. The
accuracy may be improved in certain centres with better training and facilities; however,
COVID-19 being a global pandemic it is difficult to ensure uniform quality. Thus,
the main concerns are the high false negative rate, in resource constrained environments
where RT-PCR is not freely available or with long turn around times. Undetected individuals
are infectious and unless isolated, will be mediums for transmission of SARS COV 2,
thus perpetuating the pandemic.[[21], [22], [23]]
Symptomatology is also not a criteria. Numerous studies have documented asymptomatic
individuals as well as symptomatic patients in the pre symptomatic phase are known
to transmit the infection.[[24]] CT screening of asymptomatic RT-PCR positive individuals on the diamond princess
cruise ship showed findings of pneumonia in 54%.[[25]] Numerous other studies have also supported this finding of asymptomatic with positive
CT features.[[26]]
In view of these limitations of RTPCR there is need to increase the accuracy of RTPCR
or supplement with another diagnostic technique to reduce the false negative rate
thus increasing the accuracy of detection of SARS COV-2.
Corona viruses are characterised by spike proteins which are optimised to engage human
ACE 2 receptors. Gaining entry into the cell via proteolytic action and membrane fusion.
ACE 2 receptors are in abundance in type 2 alveolar epithelial cells, GI tract, heart,
endothelium and kidney. The lungs are the most vulnerable because of their large surface
area, as well as type 2 alveolar cells act as a reservoir for viral replication. After
gaining entry into the cell, viral genome replication occurs triggering apoptosis,
release of pro inflammatory cytokines, exudation into alveolar space with associated
diffuse alveolar damage.[[19], [27], [28]] These appear as ground glass densities on imaging.[[5], [6], [7], [8]] With increasing body immune response there is increasing exudation resulting in
consolidation.[[6], [7], [27]] As a result a mixed appearance of ground glass densities and consolidation may
be seen on imaging.[[29], [30], [31]] As the infectious process and body responses mount different features of progression
and regression, a variety of internal appearances may be seen on imaging. The internal
contents may be visualised such as central clearing of ground glass, known as Atoll
sign.[[13], [14]]
Progression of inflammation along the periphery of consolidation is seen as ground
glass on the periphery of consolidation - Halo sign.[[8], [15]] There may be interlobular septal thickening due to prominence off lymphatics superimposed
on ground glass densities, appearing as a crazy paving[[5], [11], [12]] appearance.[[32]] Presence of ACE 2 receptors in endothelium result in inflammation of the vessels,
vessels appear prominent in the affected areas.[[19]] There may be intravascular thrombosis with resultant hemorrhagic infarction. Most
cases resolve with fibroblast proliferation which may leave a residue of organising
pneumonia[[33]] or fibrosis particularly in the subpleural regions.[[13]]
In a small percentage the alveolar cell damage progresses to acute respiratory distress
syndrome resulting in a white out appearance to the lungs.[[27]]
CT is extremely sensitive to early pathological changes in the alveoli as demonstrated
with HRCT in interstitial lung disease over the last few decades. This high sensitivity
of CT with relatively typical features for COVID-19 has been borne out by numerous
studies.[[1], [2], [5], [34], [35], [36]]
In our study we found similar findings. In a large cohort of 2,581 patients ground
glass densities/consolidations in a posterior, peripheral location often with multilobar
involvement were found in COVID 19 pneumonia regardless of the PCR status, positive,
negative and unknown [[Figures 1], [2], [3]].
Figure 1 (A-C): Typical bilateral peripheral /sub pleural and central ground glass
densities suggesting COVID-19 Pneumonia. These examples demonstrate communality of
appearances irrespective of RT PCR status
Figure 2 (A-C): Subtle bilateral peripheral/sub pleural ground glass densities suggesting
COVID-19 pneumonia in pandemic.Demonstrating communality of appearances in a pandemic
irrespective of RT PCR status
Figure 3 (A-C): Peri- bronchovascular ground glass densities indicative of atypical
appearance of COVID 19 pneumonia.These examples also demonstrate communality of appearances
irrespective of RT PCR Status
Indicating the specificity of CT features of COVID 19 in the setting of this pandemic.
As well as the positivity of CT in the setting of negative RT-PCR.[[36]]
Ancillary findings such as crazy paving, Atoll sign, Halo sign, prominent vessel,
prominent bronchiole and vacuolar sign as in other studies are seen in our study also
regardless of RT PCR status, however, are not specific for COVID 19.
This is not surprising as RTPCR and CT scan evaluate different aspects of COVID 19.
Then why is CT not used to reduce the false negative rate thus helping curb the spread
of SARS-CoV2.
Guidelines issued early during the pandemic dissuaded the usage of CT.[[37]] The main reasons for not advocating CT were[[37], [38], [39], [40]]
-
Low specificity of CT,
-
Advocacy of CT may overwhelm existing resources as well as may reduce access of non
covid patients to imaging suites,
-
CT may act as a potential disease transmitter via surface contamination, especially
exposure to imaging department staff
-
Utilisation of ionising radiation.
Most of these guidelines were constituted in the early part of the pandemic, in fact
coinciding with WHO declaration of a Pandemic. Five months have passed, the pandemic
has raged on with no sign of abatement. The number of cases has increased from hundred
thousand to 21.5 million- 250 x increase! As we reflect back, all these points of
concern can be addressed hopefully with a fresh view for the future.
The specificity of CT though has been questioned especially in its ability to differentiate
from other viral pneumonias and other chronic lung diseases such as small airway disease,
chronic eosinophilic pneumonia, hypersensitivity pneumonitis. There are numerous publications
which have helped differentiate between these different pathological processes.[[41], [42]] SARS, H1N1 and Covid-19 have similar specific appearances of multi focal areas
of ground glass density in a subpleural location with lower lobe preponderance.[[14], [32], [43], [44], [45], [46], [47], [48], [49], [50], [51]] Differentiation between these is difficult as the appearances overlap.
However, in a pandemic due to the sheer propensity of a number of cases, these typical
patterns as well as atypical patterns point towards COVID-19, other diseases recede
into the background due to sheer numbers. Sensitivity is the key, not specificity.[[40]] Social distancing has been advocated extensively through every medium possible,
health care establishments have become hotspots for COVID 19. This has resulted in
significant drop in non-Covid imaging volumes, thus lack of access to imaging suites
for non-Covid patients does not really arise. In fact imaging facilities are extremely
under utilised : in a recent study imaging volumes plummeted 75 to 90%. Imaging studies
done for emergency medical conditions such as stroke etc., also reduced significantly.[[52]] Diseases actually disappeared in the pandemic providing health care establishments
capabilities to cater to COVID 19 patients.[[52], [53], [54]]
Protocols for surface decontamination and infection control procedures are now very
well documented. Personal protective equipment, surface decontamination of CT gantry
and table as well as air exchanges[[55], [56]] to remove any aerolisation are required to be practiced by all imaging facilities
as asymptomatic COVID positive patients may be scanned for other symptoms, COVID pneumonia
being incidentally detected.
There are also numerous means to reduce the radiation and achieve low dose CT studies
minimising the utilisation of ionising radiation. Modulating tube current to body
habitus, increasing the slice thickness to 1.5 mm, increasing the pitch to 1.5, collimating
scan to cover apices to bases helps to reduce scan time and MA thus reducing MAs.
The KV may be reduced in thinner individuals to 100Kv. Iterative reconstructions further
help to reduce radiation dose. In a recent study utilising these parameters the CTDI
vol was reduced significantly from 3.4mGy to 0.4 mGy.[[57], [58]]
A number of studies from China, where the pandemic started, have advocated the utilisation
of CT as a tool to detect COVID-19, as well as also have alluded to the fact that
CT features are independent of the RT PCR status.[[1], [2], [44], [58], [59]] China, the most populous country,[[60]] where the pandemic started has reported only 84,000 infections out of a world total
of 21.5 million with a very low level of new infections.[[61]]
In this study we did not attempt to determine the sensitivity of CT vis a vis RT-PCR
as all patients with positive RT-PCR were not scanned. Negative CT studies with positive
RT-PCR is well documented. We also did not attempt to study the false positive rate
of CT though most RT-PCR negative patients underwent respiratory panel for atypical
pneumonia as co-infections between atypical pneumonia and SARS COV-2 has been documented.[[62]]
