Key words thorax - CT - infection - sinusitis - pneumonia - respiratory syncytial virus (RSV)
Introduction
An unusually high number of nosocomial as well as community-acquired respiratory syncytial
virus (RSV) infections of the respiratory tract in adult patients particularly with
a hematological or autoimmune primary disease was seen in the winter of 2011/2012 at
a hematology and transplant center [1 ]
[2 ]. RSV infection is known primarily as an inflammatory change of the lung in children.
This pathogenic agent has been less frequently described as a community-acquired infection
in adults. RSV pneumonia is more frequently seen in patients who have undergone stem
cell transplantation and have a malignant hematological primary disease or in the
case of immunosuppression [3 ]
[4 ]
[5 ]. Mild to moderate inflammation of the upper respiratory tract is typical in immunocompetent
adults. However, in immunosuppressed or older patients with preexisting conditions,
an RSV infection can result in severe pneumonia with a fatal outcome in some cases.
Therefore, it is important for the radiologist to know the immune status of the patient
being examined [6 ].
The RS virus is a pneumovirus and is part of the paramyxoviridae family. The mumps
virus and parainfluenza virus also belong to the paramyxoviridae family. RSV is an
enveloped virus with negative single-strand RNA. It can be differentiated into subtypes
A and B. RSV is transmitted primarily via smear infection. It can live on surfaces
for several hours and be incorporated by large drops or droplet infection. Therefore,
hands should be disinfected and masks should be worn for primary prevention. Initial
infection with RSV usually occurs in children. Due to incomplete immunity, reinfection
often occurs but has a milder clinical course. The following clinical symptoms are
seen in decreasing order of frequency: rhinorrhea, whistling respiratory sound, dyspnea,
productive coughing, fever, and myalgia. RSV replicates in the ciliated epithelial
cells of the mucous membranes of the respiratory tract. The infection can last from
6 days to 8 weeks. There are four established diagnostic procedures for detecting
RSV including virus culture, antigen determination via immunofluorescence assay (IFA)
or enzyme immunoassay (EIA), RNA determination via RT-PCR (real-time polymerase chain
reaction) and serology. With 60 – 90 % PCR has the highest sensitivity. Adult patients
are typically treated based on symptoms. A bacterial superinfection that complicates
the course of the RSV infection occurs in 10 – 30 % of cases. It was shown in vitro
that a superinfection in the case of an existing RSV infection is most frequently
associated with Haemophilus influenzae and Streptococcus pneumoniae [7 ]. Ribavirin is a broad-spectrum antiviral medication with a controversial therapeutic
effect. It is used in children and older patients. There is currently no approved
RSV vaccine [6 ].
The CT-morphological signs of RSV pneumonia are not frequently described in the current
literature and are often underdiagnosed or misinterpreted in the clinical routine,
e. g. due to similar morphologies confused with fungal infections [8 ]. Fungal infections, primarily Aspergillus infections, are a complication with a
potentially fatal outcome in patients with malignant hematological diseases or immunosuppression.
Therefore, diagnosis (microbiology and primarily computed tomography) plays a major
role in differentiation and treatment planning [9 ].
The goal of this analysis is to identify characteristic morphologies in chest CT in
patients with RSV pneumonia in order to make CT-morphological diagnosis of RSV pneumonia
more probable.
Materials and Methods
All patients of medical clinic V between November 2011 and July 2012 who tested RSV-positive
and underwent a chest CT (median 1 day after [30 days before, 19 days after] a positive
test) were included in the retrospective analysis approved by the ethics commission.
The significant variability was the result of the fact that the epidemic (November
2011) was initially not recognized as such and thus serial RSV testing was not performed
until December 2011. The first chest CT in each case was performed within the first
24 hours after the start of clinical symptoms. Samples were acquired via throat swab
(n = 447), sputum (n = 7), tracheal secretion (n-14), or bronchoalveolar lavage (BAL)
(n = 62) and examined via PCR [10 ]. The tests were thus performed in all patients several times with different sample
materials and analysis techniques over the course of the disease. Bronchoalveolar
lavage was performed in patients (n = 20) who had already undergone chest CT showing
pronounced radiological changes.
The patients included in the study primarily had hematological and autoimmune primary
diseases, e. g. multiple myeloma or acute myelocytic leukemia ([Table 1 ]).
Table 1
Primary diseases of the patients included in the evaluation (n = 51).
malignant primary disease
frequency
multiple myeloma
14
acute myelocytic leukemia (AML)
13
acute lymphatic leukemia (ALL)
3
diffuse large B-cell lymphoma (DLCBL)
3
B-non-Hodgkin lymphoma
2
chronic lymphatic leukemia (CLL)
2
myelodysplastic syndrome (MDS)
2
T-lymphoblastic leukemia (T-LBL)
2
chronic myelocytic leukemia (CML)
1
mantel cell lymphoma (MCL)
1
Hodgkin’s disease
1
small lymphocytic leukemia (SLL)
1
T-non-Hodgkin lymphoma
1
autoimmune primary disease
amyloidosis
2
scleroderma
1
vasculitis
1
other primary disease
kidney transplant in diabetic nephropathy
1
total
51
The CT examinations of the thorax (n = 132) and in some cases also of the paranasal
sinuses (n-19) were primarily performed using a 128-row iCT (Philips, Hamburg). In
addition, individual chest examinations were performed using other devices (Somatom
Definition Flash, Sensation 16, Volume Zoom and Definition AS40, Siemens Medizintechnik,
Forchheim). The technique was usually native (n = 106/132, with a slice thickness
of 3 – 5 mm (n = 124 with 3 mm, n = 5 with 4 mm, n = 3 with 5 mm) always in spiral
technique during an inspiratory breath hold with dose modulation. The images were
interpreted in consensus by two experienced chest radiologists on a PACS workstation
(Synapse 3.2.1, Fujifilm, Düsseldorf) in accordance with DIN 6868 – 57 with two category
A monitors using all of the usual image viewing techniques.
In the case of follow-up examinations, each examination was evaluated individually
as well as in comparison to determine whether worsening or improvement could be detected.
A visually noticeable increase in pathologies or the occurrence of new pathologies
was evaluated as worsening of the findings. An improvement was defined as a visually
detectable relevant decrease in pathologies.
To evaluate characteristic early signs of RSV pneumonia, the particular CT was evaluated
at the presumable start of the disease. Even patients without a follow-up examination
were included. A further inclusion criterion was a low number of pathologies (fewer
than two at the same time). Exclusion criteria were more than two simultaneously present
pathologies or large-area changes (> 25 % of the lung parenchyma).
A tabular schema based on the current literature was used for the qualitative and
quantitative evaluation of the lung parenchyma [3 ]. Bronchial wall thickening, tree-in-bud, nodules, halo, ground-glass opacity, consolidations,
and pleural effusions were evaluated with respect to their anatomical distribution
(bi-/unilateral, basal/apical, central/peripheral, peribronchovascular/lobar/(non-)segmental/random)
and if possible were quantified (longitudinal diameter and number (< 5, 5 – 10, > 10)
of nodules, percentage estimation of the proportion of ground-glass opacity or consolidations
in relation to the total lung parenchyma, ventrodorsal diameter of the pleural effusions).
The individual lung parenchymal changes were preferentially defined according to the
criteria of the “Fleischner Society’s glossary of terms” [11 ]: Nodules of up to 3 cm were included and classified as consolidations. An increase
in the density of the lung parenchyma with still recognizable vascular structures
was defined as a ground-glass change. Consolidations were then documented when increases
in the lung parenchyma density did not have recognizable vascular structures. These
were then additionally differentiated as respiratory tract consolidations and parenchyma
consolidations.
In individual patients (19/51), the skull was examined via CT or MRI for neuroradiological
purposes. These images were only used in the present analysis to evaluate the paranasal
sinuses. Air-fluid levels in the individual paranasal sinuses were evaluated as acute
sinusitis. If this was found in all sinuses, pansinusitis was diagnosed. Any osseous
arrosion was noted.
Results
The main scan period of the patients in this evaluation was from December 2011 to
March 2012. 51 patients with a positive RSV test were examined in a total of 132 chest
CT examinations. 6 additional patients had a positive RSV test but only had minor
clinical symptoms so that no radiological imaging was performed. Therefore, these
6 patients were not included in the present study. The average age was 58 years. 75 %
of the patients were examined multiple times. A pathological lung finding in terms
of the study evaluation was found in 86 % of the examined patients. As a result, 14 %
of the patients had a normal finding in this regard despite clinical symptoms of pneumonia.
55 % of the patients required oxygen and 29 % required intensive care and ventilation.
27 % of the patients suffered an infection-associated death ([Table 2 ]).
Table 2
Overview of basic data of all included patients.
basic data
patients: n = 51
chest CT scans: n = 132
female
43 % (22/51)
male
57 % (29/51)
age
Ø 58 [22 – 78]
multiple examinations
75 % (38/51) [2 – 9 times]
pathological finding
86 % (44/51)
79 % (104/132)
no pathological finding
14 % (7/51)
21 % (28/132)
multiple pathologies at the same time
80 % (41/51)
72 % (95/132)
requiring oxygen
55 % (28/51)
intensive care and ventilation
29 % (15/51)
infection-associated death
27 % (14/51)
germ detected in addition to RSV
10 % (5/51)
14 % of the patients did not show any pathology on the chest CT scans despite clinical
symptoms of pneumonia. 28 of the 132 CT scans were performed in these patients. All
of these CT scans were reconstructed with a slice thickness of 3 mm. 4 of these patients
underwent a single examination. Two patients underwent one follow-up examination and
one patient underwent three follow-up examinations. The average age of these patients
was also 58 years (from 24 to 76 years) and the primary diseases also varied. All
of these CT scans did not show any pathological change, also not over the course of the
disease, i. e., these patients were RSV-positive and had clinical symptoms of pneumonia
but lung parenchymal changes were not seen over the entire period.
In general, the following pathological CT changes were seen in descending order of
frequency: ground-glass opacity (95 % bilateral, average 18 % of the lung parenchyma),
consolidations (average 11 % of the lung parenchyma, 23 % with positive bronchoaerogram),
nodules, tree-in-bud, bronchial wall thickening ([Table 3 ]). These pathologies were associated with pleural effusions (average 2 cm) in 63 %
of the patients. Nodules with an average maximum diameter of 8 mm (from 3 to 25 mm)
typically occurred in high numbers (14 %, < 5, 13 % 5 – 10, 74 % > 10 nodules) and
in a centrilobular position (58 %) (often also random (29 %)) and had a halo in 81 %
of the patients ([Table 3 ], [Fig. 1 ]).
Table 3
Pathological CT changes.
pathological CT change
patients: n = 51
chest CT scans: n = 132
bronchial wall thickening
37 % (19/51)
27 % (35/132)
tree-in-bud
51 % (26/51)
36 % (47/132)
nodules
61 % (31/51)
55 % (72/132)
halo around nodules
81 % (25/31)
71 % (51/72)
ground-glass opacity
76 % (39/51)
64 % (85/132)
consolidations
71 % (36/51)
56 % (74/132)
pleural effusion
63 % (32/51)
44 % (58/132)
Fig. 1 This CT image of a 45-year-old female patient with ALL was acquired on the same day
as the positive RSV test. A few centrilobular nodules in segment 2 of the right upper
lobe of the lung surrounded by ground glass. This pattern was often seen. 6 days before
and 14 days after this CT scan, there was no pathological result in the CT examinations.
Multiple radiomorphological lung parenchymal changes with the following distribution
in the lung parenchyma in descending order were detected in 72 % of all CT examinations:
30 % random, 22 % segmental, 22 % basal, 12 % apical, and 12 % lobar. In addition
to the positive RSV test, HSV (herpes simplex virus) was found in BAL in two cases
and HSV as well as Aspergillus spec. in an additional two cases. Pseudomonas aeruginosa
was found in BAL immediately prior to death in one case. Thus, a second pathogenic
agent was found in 5/51 (10 %) patients.
Two or more CT examinations were performed in the course of the disease in 38/51 (75 %)
patients, thus making it possible to evaluate the radiomorphological development of
the RSV pneumonia in a relevant patient collective. The median time between the first
and second follow-up examination was 11 days. With 79 % (30/38) ground-glass opacity
was the most common pathological sign on the first CT scan of patients with later
follow-up examinations ([Fig. 2 ]). With 61 %, nodules were the second most common pathological sign ([Fig. 3 ]). With 87 % nodules were the most frequent pathological sign in the first follow-up
examination and with 65 % ground-glass opacity was the second.
Fig. 2 There are discrete centrilobular nodular changes with ground glass in the lower lobes
and the lingular segment of the lung in a 58-year-old woman with a positive RSV test
and B cell lymphoma (a). The proportion of ground glass and nodules with halo increased
and new consolidations were seen 14 days later (b). Exitus letalis on day 17.
Fig. 3 CT examinations of a 42-year-old male aplastic patient with ALL. Only one small nodule
in segment 2 of the right upper lobe was seen in CT on the day of the positive RSV
test a . There were new pleural fluids and bipulmonary ground glass nine days after the previous
CT b . The pleural fluids decreased but the ground glass increased on day 15 after a positive
RSV test c . The next CT scan on day 25 showed new bipulmonary consolidations with a positive
air bronchogram as a sign of organizing pneumonia d . At the same time the ground glass and pleural fluids decreased d . The last examination on day 38 showed recurrent consolidations e . Hematologic reconstitution began on day 15 and advanced until the last day of CT
examination.
To evaluate the characteristic early changes of RSV pneumonia, the relevant CT scan
was assessed at the presumed start of the disease. However, as already described above,
the positive RSV test varied with respect to time since the epidemic was not initially
recognized as such.
34/51 (67 %) patients were evaluated in this regard. Nodules were found in 53 % (18/34)
of patients, bronchial wall thickness in 41 % (14/34), tree-in-bud pattern in 32 %
(11/34), pleural effusions in 21 % (7/34), ground-glass opacity only in 3 % (1/34)
and consolidations in no patients. It is noteworthy that bronchial wall thickening
occurred in combination with nodules or the tree-in-bud pattern in 79 % (11/14) of
the cases. An exemplary but very typical image is shown in [Fig. 1 ].
The median time between CT examinations with progression of the findings was 7 days
and 22 days in the case of regression of the findings. A radiological increase in
findings was diagnosed in 45 % of patients ([Fig. 3b-d ]), regression in 33 % ([Fig. 3e ]) and no significant change in the findings in the remaining 22 % in the follow-up
examinations. The ground-glass opacity increased by 21 % and the consolidations by
12 % in the examinations with progression of the findings. New nodules and the tree-in-bud
pattern occurred in 22 % of the cases ([Fig. 4 ]).
Fig. 4 67-year-old female patient with CLL and a positive RSV test. There were pleural fluids
and only a few bronchial wall thickenings in the bronchi of the middle and the right
lower lobe a . Bronchial wall thickening increased and was seen bipulmonary 10 days later b . Furthermore, there were new centrilobular nodules in both lower lobes and consolidations
in the right lower lobe b . Consolidations in the right lower lobe increased and showed a positive air bronchogram
14 days later as a sign of organizing pneumonia. There was also progress of the ground
glass and centrilobular nodules in the left lower lobe c . These findings increased on day 20 again. Exitus letalis on day 28.
The paranasal sinuses were evaluated via CT or MRI in 37 % (19/51) of the patients.
In 63 % (12/19) of these examinations, air-fluid levels and thus sinusitis were able
to be diagnosed. All paranasal sinuses were affected in 26 % (5/19) of the cases,
i. e. pansinusitis was diagnosed. In addition to these findings, further intracranial
pathologies (otitis media (2 cases), mastoiditis (4 cases), subdural hematoma (3 cases),
subdural hygroma (1 case)) were found in 42 % (8/19) of the patients but no osseous
arrosions were found. 90 % (11/12) of the patients with sinusitis or pansinusitis
also had pneumonic lung infiltrates and 50 % (6/12) required ventilation.
Discussion
The seasonal occurrence of RSV infections is known and is also reflected in the main
scan period of the patients in this analysis [3 ]
[5 ]. Compared to other studies, this study is the first to address the CT-morphological
changes of RSV pneumonias over the course of the disease and to additionally analyze
its association with sinusitis. 51 patients and 132 chest CT examinations were included.
In these CT examinations ground-glass opacity was found most frequently with a rate
of 64 %, followed by consolidations with 56 %, and nodules with 50 % (in 74 % > 10).
The most common pathological morphologies in patients with a follow-up examination
were ground-glass opacity and nodules (primarily centrilobular) in this order of frequency.
This was reversed in the first follow-up examination and nodules were seen more frequently
than ground-glass opacity. The other detected pathologies were rarer. The clinical
spectrum ranged from mild clinical infection constellation to intensive care with
organ failure [12 ].
The above range of pulmonary changes as seen for example in H3N2 influenza infections
generally supports the diagnosis of viral pneumonia. However, the occurrence of multiple
pulmonary changes is only suggestive but not specific for viral pneumonia as in H3N2,
for example [13 ]. Other studies that have addressed CT changes in immunosuppressed patients with
RSV pneumonia obtained similar results:
The three most common pathological signs in descending order of frequency according
to Franquet et al. were ground-glass opacity, nodules, and bronchial wall thickening.
Of 130 patients who underwent hematopoietic stem cell transplantation, BAL, and chest
CT, 26 patients with proven RSV pneumonia were included in this retrospective evaluation
[3 ].
In a study by Gasparetto et al., 20 bone marrow transplant patients with nosocomial
RSV pneumonia underwent CT examination within 24 hours of the start of symptoms. RSV
was diagnosed via BAL, nasopharyngeal swab, or nasal lavage. Nodules, consolidations,
and ground-glass opacity were found in this order of frequency [4 ].
Escuissato et al. included 111 bone marrow transplant recipients from a cohort of
774 patients who underwent chest CT within 24 hours of the start of symptoms. All
of these patients had symptoms of pneumonia. However, RSV pneumonia was verified in
only 30 patients. Primarily nodules, consolidations, and ground-glass opacity were
found in patients with RSV pneumonia in this study [14 ].
Miller et. al [15 ] who retrospectively examined the CT-morphological characteristics of 115 community-acquired
bacterial and viral (19 RSV) infections of the lower respiratory tract found the tree-in-bud
pattern and bronchial wall thickening most frequently in association with RSV.
It should be mentioned that there were only a limited number of RSV-positive patients
with no pathological changes in the CT examinations despite symptoms of pneumonia.
In contrast, the percentage of such cases was 14 % in the study by Ko et al. [16 ] and in the present study and 20 % in the studies by Escuissato et al. [14 ] and Gasparetto et al. [4 ]. All negative CT scans in the present study were reconstructed with a slice thickness
of 3 mm. It is therefore not assumed that significant findings were overlooked but
rather that these CT scans in fact did not show any pathologies.
In light of the fact that nodules were the second most frequent finding with a rate
of 50 % in this analysis and 70 % of these had a halo, it is useful to point out that
this halo is a non-specific morphological change that can be caused by a hemorrhage
or local infiltration. However, the halo sign is often found in association with a
pulmonary fungal infection [11 ]
[17 ].
Fungal infections, primarily Aspergillus infections, are a complication with a potentially
fatal outcome in patients with malignant hematological diseases or immunosuppression.
Therefore, diagnosis (microbiology, computed tomography) plays a major role in differentiation
and treatment planning [9 ].
The current evaluation analyzes both the morphological criteria in each CT scan and
the course of the disease with respect to progression or regression. The following
parameters were taken into consideration: which change could be detected from the
beginning, which new parenchymal patterns occurred, what was the extent of any new
parenchymal patterns. The current literature includes a study by Ko et al. [16 ] who also used CT to examine the course of RSV pneumonias. The collective of 7 patients
in this study included lung transplant recipients. However, the follow-up evaluation
only relates to progression, regression, and stabilization. Concrete morphological
changes are not described.
It can be assumed that the course of RSV pneumonia includes different visible morphological
stages. The most characteristic is the initial stage since many different parenchymal
changes occur at the same time in the advanced stages. Nodules which are often seen
in combination with bronchial wall thickening occur in the greatest numbers and have
the most characteristic morphological appearance. The paranasal sinuses were also
evaluated in 19 of the 51 analyzed patients (37 %) with a positive RSV test. Acute
sinusitis or pansinusitis was diagnosed in 63 % of the patients and 90 % of these
had clinical and radiological pneumonia.
In viral rhinosinusitis, there is increased production of viscous material in the
sinus that accumulates there since the regular removal process is impaired by disrupted
and reduced cilia functionality [18 ]. This can facilitate a bacterial superinfection [18 ]. In addition, RSV plays different roles in the pathogenesis of the infection from
colonization and incidental co-infection, which acts as the precursor to other pathogens,
to possible major pathogens [19 ]
[20 ]
[21 ]. RSV-associated sinusitis or pansinusitis is rarely described in the current literature
[22 ]
[23 ]. New literature addressing both the CT-morphological changes of RSV pneumonia and
the presence of sinusitis also cannot be found. RSV sinusitis presumably arises from
RSV pneumonia. Therefore, it would be useful to test the upper respiratory tract separately
from the lower tract during pathogen diagnosis in order to be able to prove this in
possible further studies.
This study has the following limitations:
Firstly, it is a retrospective, descriptive analysis whose observed incidence was
presumably increased during RSV screening and mild clinical processes that would not
have been tested for RSV in a targeted manner under different circumstances were monitored.
This allowed a broader radiological characterization of RSV pneumonia.
Secondly, only chest CT scans with a slice thickness of 3 – 5 mm were available for
the evaluation. 1-mm slices were not available. It is possible that minor pathologies,
such as ground-glass opacities, may not have been correctly evaluated or detected
on the few CT scans with a slice thickness of 4 mm or 5 mm (n = 9).
Thirdly, a superinfection was detected via BAL in 5/51 patients. However, not every
patient underwent BAL and not every pathogenic agent detected via BAL causes pneumonia
but in some cases only colonization without a pathogenic effect [20 ]
[24 ]
[25 ]. For example, an RSV infection of the respiratory tract was detected in a prospective
study including 165 children on ventilation (median age 1.6 months old), 22 % of whom
had a verified bacterial co-infection and another 21 % had a possible bacterial co-infection
[19 ].
Fourthly, no special or systematic examination of the paranasal sinuses (e. g. nasopharyngeal
lavage, nasal swab) was performed in patients with sinusitis since this issue was
not the focus of the head examinations [23 ].
Conclusion
This analysis of chest CT scans shows that ground-glass opacity, consolidations, and
nodules with a halo were seen in cases of RSV pneumonia. Ground-glass opacity was
primarily seen in the initial stage, while centrilobular nodules were seen in the
advanced stages. In most cases, multiple pathological changes that developed primarily
due to worsening of the symptoms were present at the same time. However, the chest
CT scans acquired at the presumed start of the disease with nodules or tree-in-bud
often in combination with bronchial wall thickening were most characteristic.
Although these are not pathognomic lung parenchymal changes, they are characteristic
so that RSV-associated pneumonia should be considered with greater frequency on a
seasonal basis (winter months) and associated sinusitis should also be taken into
consideration.
Relevance of the study
Characterization of pneumonia to identify indicators of possible triggers
Targeted use of suitable screening tests for pathogen identification
Early detection of RSV pneumonia for early treatment planning primarily in immunosuppressed
patients
Prevention of unnecessarily broad and long-lasting empiric antibiotic regimens to
lower costs, toxicity, and resistance development
