Key words thorax - CT-quantitative - tracheobronchial tree - MR-functional imaging - fibrosis,
cystic
Introduction
Cystic fibrosis (CF) remains the most common lethal hereditary disease among white
populations. Progressive lung disease determines more than 90 % of morbidity and mortality,
but improvements in diagnostics and therapy have given rise to prolonged survival
of CF patients in the past, averaging around 40 years [1 ]. Implementation of screening programs in specialized centers in Germany and other
Western countries has led to earlier diagnosis [2 ], thus enabling treatment in a pre-symptomatic stage. Pulmonary function testing
underestimates the early stages of CF lung disease and has limited predictive value
in pulmonary exacerbations [3 ]
[4 ]. Imaging provides regional information on the distribution and severity of the different
components of CF lung disease. The hallmarks of the CF lung are bronchiectasis as
one early sign of lung damage, airway wall thickening, consolidations and atelectasis,
as well as emphysema in advanced stages of lung disease. Mucus plugging as well as
air trapping and perfusion impairment are linked to basic pathophysiology and are
potentially reversible under therapy. Bronchiectatic destruction of lung lobes, dilatation
of bronchial arteries and pulmonary hemorrhage are sequelae, which may require invasive
treatment and ultimately, lung transplantation. Originally, chest X-ray was employed
to depict morphological changes in the CF lung [5 ]. It has often been replaced by computed tomography (CT) at specialized centers,
because of its higher sensitivity for early and subtle changes in the CF lung [6 ]
[7 ]
[8 ]. However, the use of CT for short-term follow-up in infants and preschool children
as well as lifelong longitudinal monitoring are accompanied by an accumulation of
radiation dose [9 ]
[10 ]. Most recently, magnetic resonance imaging (MRI) has emerged as a radiation-free
technique for assessing the CF lung [11 ]
[12 ]. Besides morphological information comparable to CT, MRI can depict several components
of lung function, i. e. respiratory movements, ventilation and perfusion. CXR, CT
and MRI each have intensively studied individual strengths and drawbacks. Based on
the experience at our center, we intend to give an overview of the presentation of
CF lung disease in the different imaging techniques, their current status regarding
their application in the clinical routine, and to provide the reader with a rationale
to decide on the appropriate modality tailored to the individual clinical question.
Profound knowledge of the Fleischner Society’s terminology for airway disease is pivotal
[13 ]
[14 ]. A look at future MRI applications is given to conclude this review.
Technical aspects and requirements
Technical aspects and requirements
Chest X-Ray (CXR)
A posterior-anterior as well as a lateral view is recommended in adolescents and adults.
A study employing systematic scoring could show that the lateral view does not contain
relevant additional information and may be omitted in young children [15 ].
Computed Tomography (CT)
German and international guidelines on CT protocols for CF are missing, and many different
acquisition techniques for different age groups have been discussed in the past decade.
Some authors have suggested using limited slice sampling to restrict radiation exposure
[8 ]
[16 ]. Non-contrast-enhanced multidetector CT with full volume coverage and reconstructed
overlapping slice thicknesses of preferably 1.5 mm or less has the highest sensitivity
for morphological changes and should be given preference over incremental high-resolution
CT [17 ]
[18 ]. These datasets not only allow for exact comparison of follow-up exams, multiplanar
reformats and maximum intensity projections (MIP) for better identification of airway
changes, but also enable dedicated post-processing with advanced software tools [17 ]
[19 ]. Age-adapted low-dose acquisitions with an effective radiation dose of less than
2 mSv even in adults are sufficient for the evaluation of morphological changes including
ground glass opacities and mosaic perfusion [20 ]. A combined protocol of end-inspiratory with end-expiratory scans is generally recommended
to enhance the sensitivity for small airway obstruction [21 ]
[22 ], and both acquisitions may be performed with similar exposure settings, but added
radiation dose ([Table 1 ]). At the same time, all technical potential available for dose reduction must be
exploited, such as reduction of overbeaming, automatic tube current modulation, iterative
reconstruction, etc. [23 ]
[24 ]. [Table 1 ] seeks to summarize the most important protocol components for CT.
Table 1
Overview of CT acquisition parameters. Tab. 1 Überblick über die CT-Akquisitionsparameter.
0 – 5 years
6 – 18 years
≥ 18 years
detector lines
≥ 16
≥ 16
≥ 16
acquisition
volumetric
volumetric
volumetric
tube potential (kV)
80 – 100
80 – 100
120
effective tube current (mAs)
≤ bodyweight (kg) + 5 [69 ]
≤ bodyweight (kg) + 5 [69 ]
“low-dose”[1 ]
automatic current modulation
yes
yes
yes
reconstruction kernel
sharp, medium soft
sharp, medium soft
sharp, medium soft
iterative reconstruction
yes
yes
yes
reconstructed slice thickness
≤ 1.5 mm
≤ 1.5 mm
≤ 1.5 mm
reconstruction increment
≥ 25 % overlap
≥ 25 % overlap
≥ 25 % overlap
high-pitch mode
fixation, no sedation
yes
if dyspnoeic
sedation
if no high-pitch mode
no
no
expiratory scan
study conditions, intubation required [25 ]
yes
yes
1 An exact definition of low-dose is currently missing. Typical effective mAs is 20 – 70 mAs,
adapted to bodyweight.
In young children unable to cooperate, CT scanning may require sedation. High-end
CT scanners provide a high-pitch mode that delivers nearly artifact-free images even
in free-breathing children without the need for sedation ([Table 1 ]). To acquire paired inspiratory and expiratory scans in uncooperative children,
usually anesthesia, intubation and controlled ventilation are necessary [25 ]. However, this would rather be applied for research than for clinical imaging.
The i. v. application of iodinated contrast agents in CF is restricted to specific
situations, such as pulmonary emergencies including pulmonary arterial embolism and
hemorrhage. In advanced CF lung disease, hypertrophy of bronchial arteries frequently
occurs and can be identified by CT and MR angiography alike. Currently, CT angiography
is recommended to identify and delineate the course of dilated bronchial arteries
when embolization procedures are planned [26 ] ([Table 1 ]).
Magnetic Resonance Imaging (MRI)
Lung proton MRI sequences with dedicated protocols are now readily provided by all
large vendors [27 ]. Depending on patient size and ability to breath-hold, it is useful to prepare three
separate protocols ([Fig. 1 ], [Table 2 ]) [11 ]
[28 ]
[29 ]. Each should start with balanced steady-state free precession (bSSFP) sequences.
Acquired in free-breathing, a negative distance factor (-50 % slice thickness) provides
an overview of respiratory movements. Airway changes are assessed using spoiled gradient
echo sequences (GRE). In children unable to breath-hold, a T1-weighted fast spin echo
(FSE) sequence and averaging may be used. Mucus plugging within the large airways
is sensitively depicted by T2-weighted sequences, for example a half-Fourier single
shot fast spin echo acquisition. A four-dimensional dynamic contrast-enhanced perfusion
study (spoiled GRE) at high temporal resolution (1.5 s per lung volume with 20 – 30
consecutive acquisitions) with intravenous application of gadolinium-based contrast
by a power injector is recommended [27 ]. The common side effects of i. v. contrast injection, dose as well as national prescription
regulations need to be considered with respect to patient age. For a quick review
of these large datasets, perfusion maps with subtraction of the pre-contrast series
from the series with the highest parenchymal enhancement are very helpful. Multi-phasic
MR angiography at high spatial resolution can be added for the confident identification
of dilated bronchial arteries, for which the perfusion study may serve to determine
circulatory time (contrast volume may be split into doses of 20 – 50 % for perfusion
imaging and 50 – 80 % for angiography). In case of incorrect timing of contrast bolus
or image deterioration due to coughing or patient movement, recirculating vessel contrast
is still sufficient to acquire additional T1-weighted images with reasonable angiographic
quality. The overall room time for this imaging protocol approximates 30 min. The
standard protocol ([Fig. 1 ]) may be further extended to the specific needs, e. g. by adding further functional
studies and cardiac sequences [30 ]. Moreover, ultra-short echo time (UTE) sequences as introduced recently offer a
potentially high parenchymal signal and may produce CT-like images of the CF lung,
but their added value compared to the established CF MRI protocols has not yet been
assessed [31 ].
Fig. 1 MRI protocol options. Three separate basic MRI protocols should be kept ready to
use, optimized to the patient’s ability to breath-hold and comply with the procedure.
cor = coronary plane, tra = transverse plane, sag = sagittal plane, bSFFP = balanced
steady-state free-precession sequence; 50 % slice overlap should be used. FSE = fast
spin echo sequence; for T1-weighted acquisitions averaging 3 – 4x should be used to
compensate for breathing artifacts; for T2-weighted acquisitions a half-fourier single
shot technique or rotating phase encoding should be used. nav = navigator techniques.
3 D GRE = three-dimensional gradient echo sequence; echo-sharing should be used for
perfusion imaging. ce = contrast-enhanced.Abb. 1 Optionen für MRT-Protokolle. Es ist empfehlenswert drei separate MRT-Protokolle bereitzuhalten,
die an die individuelle Fähigkeit des Patienten zur Kooperation und zum Atemanhalt
angepasst sind. cor = koronare Schicht, tra = transversale Schicht, sag = sagittale
Schicht, bSFFP = Balanced Steady-State Free-Precession-Sequenz; 50 % Schichtüberlappung
sollte gewählt werden. FSE = Fast-Spin-Echo-Sequenz; für T1-gewichtete Akquisitionen
sollten 3 – 4 Mittelungen als Kompensation für Atembewegungen gewählt werden; für
T2-gewichtete Akquisitionen sollte eine Half-Fourier Single-Shot-Technik oder rotierende
Phasenkodierung gewählt werden. nav = Navigatortechnik. 3D-GRE = dreidimensionale
Gradientenechosequenz; Echo-Sharing sollte für die Perfusionsmessung verwendet werden.
ce = kontrastmittelverstärkt.
Table 2
Suggested imaging scheme according to experience in Heidelberg for life-long imaging
surveillance of CF patients starting at birth.Tab. 2 Heidelberger Schema als Vorschlag zum longitudinalen bildgebenden Monitoring der
CF ab-Geburt.
CXR
CT
MRI
diagnosis, screening < 1 year
X
X (no CM)
diagnosis ≥ 1 year
X
X
annual follow-up < 18 years
X
X
annual follow-up ≥ 18 years
X
X
clinical exacerbation
X
(X)
emergency, hemorrhage
X (CM)
CM = contrast material.
Routine sedation is usually necessary in preschool children (< 6 years). For propofol
an incidence of up to 42 % for atelectasis may mask or even simulate relevant pathology
[11 ]
[32 ]. Chloral hydrate or phenobarbital have been reported to produce less atelectasis
[33 ], and chloral hydrate, administered rectally or orally under monitoring by a pediatrician,
has been used at our institution as the preferred medication with satisfactory results
for the past 10 years [11 ].
Morphological changes of the CF lung
Morphological changes of the CF lung
Airways
Characteristic airway abnormalities in CF are mucus plugging together with inflammatory
airway wall thickening and progressive bronchiectasis ([Fig. 2 ], [3 ], [4 ]) that usually appear in heterogeneous combinations of different severity [25 ]
[34 ]. Recent CT and MRI studies in infants and young children with CF also demonstrated
high variability and regional heterogeneity of early lesions throughout the lung without
predilection for a specific region that, especially in early disease, cannot be captured
by global measures, such as spirometry, due to functional compensation by structurally
normal areas [8 ]
[11 ]
[18 ]. Bronchiectasis is considered one of the earliest irreversible structural abnormalities
detected by morphologic imaging even in asymptomatic infants identified by newborn
screening, and also correlates with disease severity and exacerbation rate [6 ]
[25 ]. Bronchiectasis may appear as superimposed line shadows and ring shadows on CXR,
depending on the course of the airway in relation to the image plane ([Fig. 3 ], [4 ]) [13 ]
[35 ]. Affection of the small airways, which are usually not visualized by CXR, may lead
to visibility of grouped mottled shadows. CXR has the lowest sensitivity for early
changes in the CF lung, whereas CT is considered the reference standard because of
its high isotropic resolution. Multiplanar reformats help to identify central to peripheral
bronchiectasis. However, even in MDCT, the visualization of small airways is precluded
by the system-inherent resolution of 200 – 300 µm [17 ]. If small airways (by convention smaller than 1 mm in diameter) are affected by
wall thickening, mucus plugging or bronchiectasis (usually a combination of all three),
they may increase in size over the resolution threshold and become visible as centrilobular
nodules, often grouped with a tree-in-bud appearance. In more advanced disease, sacculations,
or cystic bronchiectasis, may be observed, which ultimately may lead to the destruction
of a whole lung lobe.
Fig. 2 Typical CT appearance of CF lung disease. a, b This adolescent male CF patient shows typical bronchiectasis (black arrow) together
with wall thickening and mucus plugging (white arrow). A consolidation and pleural
thickening (asterisk) is present in the middle lobe. c In the left lower lobe, bronchoceles are present (black arrowhead), embedded in an
area of reduced lung density (mosaic perfusion). d Minimum intensity projections (MinIP) enhance the detection of such low-density areas,
(white arrowhead) which are due to hyperinflation and reduced perfusion.Abb. 2 Typische Merkmale der CF-Lungenerkrankung im CT. a, b Typische Bronchiektasen (schwarzer Pfeil) bei einem männlichen Jugendlichen mit CF
begleitet von Bronchialwandverdickungen sowie Mukoidimpaktion (weißer Pfeil). Eine
Konsolidierung sowie angrenzende pleurale Verdickung (Sternchen) ist im Mittellappen
erkennbar. c Der linke Unterlappen weist Bronchozelen auf (schwarzer Pfeilkopf), welche in ein
Lungenareal reduzierter Dichte eingelagert sind (Mosaikperfusion). d Minimum-Intensitäts-Projektionen (MinIP) können die Sichtbarkeit solcher Areale mit
reduzierter Dichte (weißer Pfeilkopf) verbessern, welche die Folge einer lokalen Überblähung
und reduzierter Perfusion sind.
Fig. 3 Severe exacerbation of CF lung disease on CXR and CT. This 18-year-old female was
hospitalized for intravenous antibiotic therapy for severe pulmonary exacerbation.
a, b, d, e She presented with more extensive bronchiectasis than the patient in [Fig. 2 ], with partly thin-walled bronchiectasis (black arrow), but also in combination with
wall thickening and mucus plugging (white arrow). It is clear that a layer of mucus
on the airway surface cannot be distinguished from inflammatory wall-thickening by
CT. Especially the middle lobe shows destructive cystic bronchiectasis (black arrowhead),
which is typical for CF (sometimes also called sacculations). The right superior lobe
as well as the left lung were affected by patchy consolidations (asterisk), the latter
accompanied by a volume loss of the left hemithorax. Here, remaining aerated lung
areas showed little structure such as airways or vasculature. c, f Small nodules (white arrowhead) of the right inferior lobe were found to have a centrilobular
distribution (tree-in-bud pattern) as evidenced by a maximum intensity projection
(MIP 10/2 mm, c , and are thus related to small airways disease. Note the sparing of the subpleural
space.Abb. 3 Schwere Exazerbation der CF-Lungenerkrankung in der CXR und der CT. Die 18-jährige
Patientin wurde zur intravenösen antibiotischen Therapie bei schwerer pulmonaler Exazerbation
stationär aufgenommen. a, b, d, e Im Vergleich zum Patienten aus [Abb. 2 ] wies sie stärker ausgeprägte Bronchiektasen auf, welche teilweise dünnwandig (schwarzer
Pfeil) und teilweise in Kombination mit Wandverdickung und Mukoidimpaktionen (weißer
Pfeil) imponierten. Dieses Beispiel verdeutlicht nochmals, das mittels CT ein oberflächlicher
Mukusbelag nicht von einer entzündlichen Wandverdickung unterschieden werden kann.
Insbesondere der Mittellappen zeigte destruktive zystische Bronchiektasen (schwarzer
Pfeilkopf), welche typisch für die CF sind (auch Sakkulationen genannt). Der rechte
Oberlappen sowie die gesamte linke Lunge waren von fleckigen Konsolidierungen betroffen
(Sternchen), letztere begleitet von einer Volumenminderung des linken Hemithorax.
Verbliebene belüftete Lungenanteile zeigten wenig strukturelle Merkmale wie Atemwege
oder Gefäße. c, f Die Mikronoduli (weißer Pfeilkopf) im rechten Unterlappen wiesen ein zentrilobuläres
Muster auf (sog. Tree in bud-Muster), wie die Maximum-Intensitäts-Projektion (MIP
10/2 mm, c belegt, und sind daher durch eine Erkrankung der kleinen Atemwege verursacht. Beachte
die Aussparung des unmittelbaren Subpleuralraumes.
Fig. 4 Longitudinal surveillance with CXR and MRI. This female patient with CF participates
in our surveillance imaging program with annual routine follow-up. Irreversible bronchiectasis
(black arrow) could be identified in the right superior lobe from age 5. These airways
showed mucus plugging (black arrowhead) at age 5 and 6, which was reversible thereafter.
Around this area as well as in the left superior lobe reduced parenchymal signal on
T2 (white arrowhead) similar to mosaic perfusion on CT was present, with different
severity over time. Perfusion abnormalities (white arrow) were detected at all ages,
but with significantly different severity. Note that perfusion abnormalities also
correlate with the areas of reduced signal on T2. At age 6, the subject had a pulmonary
exacerbation, which was evidenced on imaging by the presence of a right superior lobe
consolidation (circle) with adjacent pleural reaction (asterisk) and more severe perfusion
abnormalities. One year later (therapy was enacted immediately), mucus plugging, consolidation
and perfusion abnormalities were alleviated.Abb. 4 Longitudinale CXR und MRT zum Krankheitsmonitoring. Diese weibliche CF-Patientin
nimmt an dem lokalen Programm mit jährlicher CXR und MRT zum Monitoring der Krankheitsaktivität
teil. Irreversible Bronchiektasen (schwarzer Pfeil) sind bereits ab dem 5. Lebensjahr
sichtbar. Diese Atemwege zeigten im Alter von 5 und 6 Jahren Mukoidimpaktionen (schwarzer
Pfeilkopf), welches hiernach reversibel war. Um dieses Areal ebenso wie im linken
Oberlappen fand sich ein reduziertes Parenchymsignal in der T2-Wichtung (weißer Pfeilkopf),
vergleichbar der Mosaikperfusion in der CT, mit unterschiedlicher Ausprägung zu den
unterschiedlichen Zeitpunkten. Perfusionsstörungen (weißer Pfeil) zeigten sich zu
allen Zeitpunkten, jedoch mit deutlich variabler Ausprägung. Beachte, dass das reduzierte
Parenchymsignal in T2-gewichteten Aufnahmen mit den Perfusionsstörungen korreliert.
Im Alter von 6 erlebte die Patientin eine pulmonale Exazerbation, welche in der Bildgebung
durch eine Konsolidierung im rechten Oberlappen (Kreis) mit angrenzender Pleuraverdickung
(Sternchen), und stark ausgeprägten Perfusionsstörungen auffiel. Im folgenden Jahr
(eine Therapie wurde umgehend eingeleitet) waren Mukoidimpaktionen, Konsolidierung
und Perfusionsstörungen wieder deutlich gebessert.
As expected from the higher spatial resolution, MDCT is superior to MRI in the depiction
of small peripheral airways. However, the aforementioned pathological changes of the
CF lung represent high signal components against the black background of healthy lung
tissue (“plus pathologies”). This facilitates detection and results in a comparably
high sensitivity of MRI for most pathologies as with CT. Recent CT studies reported
bronchiectasis in approx. 30 % at the age of 3 months, and progression to approx.
60 % at the age of 3 years [8 ]
[25 ]. The aforementioned pathological changes of the CF lung represent high signal components
against the black background of healthy lung tissue (“plus pathologies”). This facilitates
detection and results in a comparably high sensitivity for MRI as with CT for most
pathologies [29 ]. MRI detected a similar overall prevalence of approx. 90 % in patients aged 0 to
6 years (mean age: 3.1 years) [11 ]. In a direct comparison, MRI showed a high correlation with CT-diagnosed structural
abnormalities in a CF population aged 7 – 42 years (mean age: 16.7 years) [29 ]. Even key features such as the tree-in-bud pattern could be observed.
Mucus plugging is linked to the basic ion-transport defect and constitutes the second
most frequent morphological abnormality ([Fig. 2 ], [3 ], [4 ]) [11 ]
[36 ]. Whereas mucus plugging received little attention in recent CT studies, MRI detected
a high overall prevalence of mucus plugging of approx. 63 % of cases, making it the
second most frequent morphological abnormality in clinically stable infants and preschool
children with CF (mean age: 3.1 years) [11 ]. Neither CT nor CXR can distinguish mucus on the airway surface from inflammatory
wall thickening of larger airways ([Fig. 2 ], [3 ]). The possibility for different tissue contrasts in combination with contrast enhancement
is a clear advantage of MRI. Wall thickening due to edema will lead to high signal
intensity on T2-weighted images, reflecting active inflammation ([Fig. 4 ]). Contrast enhancement of the airway wall on T1-weighted sequences is also a marker
of inflammation, whereas intraluminal fluid will show a low signal. Importantly, mucus
plugging may become a useful outcome measure in early CF lung disease as a potentially
reversible abnormality [11 ]
[12 ]
[36 ].
Parenchyma
Consolidations are typical signs of infection and are found in pulmonary exacerbations
in CF. In many cases, an atelectasis with reduced volume and displacement of the pulmonary
fissures occurs in exacerbation, unlike typical lobar pneumonia in otherwise healthy
patients [37 ]. CXR usually has the lowest sensitivity, while CT and MRI perform equally well ([Fig. 2 ], [3 ], [4 ], [5 ]) [38 ]. On MRI, consolidations stand out brightly on T2-weighted sequences ([Fig. 4 ]). In case of a destroyed segment or lobe, bronchiectasis embedded in persistent
consolidation and volume loss are evident. In a group of 10 patients with pulmonary
exacerbations (age range: 0 – 6 years, mean age: 3.7 years), consolidations on MRI
were more frequent than in a comparable group in a clinically stable situation [11 ]. Moreover, they were alleviated under antibiotic therapy, making it a potentially
reversible abnormality. However, mucus plugging and perfusion abnormalities seemed
to play a greater role in exacerbation and were more responsive to treatment than
consolidations [11 ]. Peripheral consolidations may lead to pleural thickening and enhancement of the
adjacent pleura ([Fig. 2 ], [4 ]) [11 ]
[39 ]. Recent work using quantitative CT has confirmed earlier histopathological descriptions
that adolescent and adult CF patients develop emphysema (age range: 7 – 66 years,
median age: 20.1 years) [40 ]
[41 ]. This is also supported by a mouse model showing that emphysema formation in advanced
CF is pathophysiologically linked to emphysema in COPD [42 ]
[43 ]
[44 ].
Fig. 5 Functional CT and MRI. This figure refers to the same patient as [Fig. 3a, b ]. Coronary reconstructions of the acquisition in end-inspiration assist in depicting
the course of cystic bronchiectasis (black arrow). Mucus plugging (black arrowhead)
was present especially in the right inferior and patchy consolidations (asterisk)
in the left inferior lobe. Mosaic perfusion may be suspected in both lungs. c, d End-expiratory acquisitions assist in identifying areas of air-trapping as a cause
of mosaic perfusion by increasing the density of normal lung tissue able to exhale
normally. Areas of air-trapping do not significantly increase in density in expiration
and show a reduction of vascularity. Please note the volume loss and limited diaphragmatic
movement of the left lung as indexed by the black line. e, f Perfusion MRI revealed an altogether inhomogeneous lung perfusion (compare examples
in [Fig. 4 ]) as well as areas of complete perfusion loss (white arrow) nicely matching air-trapping
on expiratory CT.Abb. 5 Funktionelle CT und MRT. Selbe Patientin wie in [Abb. 3a, b ]. Koronore Rekonstruktionen der Akquisition in End-Inspiration sind hilfreich, um
den Verlauf der zystischen Bronchiektasen (schwarzer Pfeil) nachzuvollziehen. Mokoidimpaktionen
(schwarzer Pfeilkopf) fanden sich vor allem im rechten und fleckige Konsolidierungen
(Sternchen) im linken Unterlappen. Eine Mosaikperfusion ist angedeutet beidseits erkennbar.
c, d End-exspiratorische Aufnahmen dienen dem Nachweis von Air-Trapping als Ursache der
Mosaikperfusion durch Zunahme der Dichte von Lungengewebe aus dem in Ausatmung die
Luft normal entweichen kann. Lungengewebe mit Air-Trapping nimmt dagegen in Exspiration
nicht an Dichte zu und weißt reduzierte Gefäßkaliber auf. Beachte die Volumenreduktion
und verminderte Zwerchfellbeweglichkeit der linken Lunge, markiert durch die schwarze
Hilfslinie. e, f In der Perfusions-MRT zeigte sich eine insgesamt inhomogene Perfusion (vergleiche
Beispiele in [Abb. 4 ]) sowie flächige Perfusionsausfälle (weißer Pfeil), welche sehr gut mit dem Air-Trapping
korrelierten.
Functional imaging – air trapping and lung perfusion
Functional imaging – air trapping and lung perfusion
Small airway obstruction prevents air from being exhaled from lung volumes the size
of a lobule to whole lobes. These have a reduced alveolar oxygen level and may be
hyperinflated. The physiological effect called hypoxic pulmonary vasoconstriction
(HPV, formerly “Euler-Lilljestrand-Reflex”) leads to reduced perfusion to such lung
areas in order to prevent intrapulmonary shunting. In airway diseases such as CF,
airway obstruction frequently occurs and thus leads to a redistribution of the pulmonary
blood volume. On inspiratory CT scans, the reduced capillary blood content may be
detected by a reduced parenchymal density in Hounsfield Units (HU). It is often surrounded
by and sharply delineated against normal lung and shows reduced vessel numbers and
calibers also. Such an appearance was termed mosaic perfusion [13 ]
[14 ]
[45 ]. Its visual perception may be enhanced by end-expiratory CT acquisitions: During
normal expiration, lung volume as well as the amount of air per voxel decreases, thus
leading to an increase of its density value on CT. As compared to normal, lung areas
with small-airway obstruction do not significantly change volume or increase in density
on expiratory acquisitions. Thus, the density difference between areas of airway obstruction
and normal lung is expanded, increasing the sensitivity of CT for detection. If a
mosaic of different densities is seen on expiratory CT, it is generally termed “air
trapping” [13 ]
[14 ]. Using expiratory CT, air trapping has been described in approx. 70 % of newborns,
infants and preschool children with CF (age range: 0 – 5 years) ([Fig. 2 ], [5 ]) [8 ]
[18 ]
[25 ].
Similar to mosaic perfusion, areas of lower signal intensity may also be visible on
T2-weighted as well as post-contrast T1-weighted sequences with MRI due to the higher
parenchymal signal of normal lung, but the sensitivity may be lower than with CT ([Fig. 4 ]). The effect of HPV implies that imaging of lung perfusion approximates lung ventilation
[46 ]. Thus, perfusion MRI should – in theory – identify identical areas of pathology
as air trapping on expiratory CT. However, data on direct comparison is missing. Typical
patchy or wedge-shaped perfusion defects occur on MRI and it was shown that these
areas of hypoperfusion correlate with the degree of parenchymal changes in pediatric
(age range: 0 – 6 years, mean age: 3.1 years) and adolescent (age range: 11 – 19 years,
median age: 16 years) CF patients ([Fig. 4 ], [5 ]) [11 ]
[47 ]. Abnormal perfusion on MRI was already detected in the first year of life, with
an overall prevalence of 85 % in preschool children, comparable to the aforementioned
prevalence of air trapping [11 ]. Maybe more importantly, perfusion alterations occurred even without detectable
parenchymal changes [11 ]. This suggests that air trapping and perfusion abnormalities may be the earliest
signs of disease detectable in the CF lung, even before morphological changes become
visible. Air trapping/perfusion abnormalities may reflect reversible disease and hold
the possibility for therapy monitoring [11 ], but may become fixed in advanced CF with extensive parenchymal damage.
Scoring and quantitative imaging
Scoring and quantitative imaging
To quantify disease severity and facilitate patient follow-up and monitoring of therapeutic
effects in CF, visual scoring systems have been developed for CXR (e. g. Chrispin-Norman
Score, Brasfield Score, Wisconsin Score) [15 ]
[35 ]
[48 ]
[49 ], CT (e. g. Bhalla Score, Helbich Score, Brody Score) [4 ]
[50 ]
[51 ], and more recently, MRI (Eichinger Score) [39 ]. These scoring systems are necessary because the described changes in the CF lung
show a heterogeneous distribution within one patient and between different patients,
and may intra-individually show a different course over time. Thus, the scores encompass
structural changes (CXR, CT, MRI) as well as functional changes (air trapping on CT,
perfusion abnormalities on MRI), and assign a numeric score to lung regions or lobes
depending on the severity of the individual pathology. A previous study reported that
the correlation of lung function parameters with CT was higher than with CXR, indicating
that CT provides a more precise grading than CXR [52 ]. Most importantly, CT scoring proved to be superior over pulmonary function testing
in detecting subtle disease progression [6 ], and has already been used to detect therapy response [7 ]. A more advanced approach uses a grid overlay on selected CT slices and allows a
reader to assign a pathology to each lung-containing square, leading to semi-automatic
scoring [53 ]. Still, automatic objective quantification of image information remains desirable.
There is high potential in the direct quantification of airway changes by generation-
and lobe-based quantitative post-processing of non-enhanced thin-slice CT datasets
([Fig. 6 ]). Putative imaging biomarkers such as wall thickness or airway diameter, air trapping,
and emphysema may be derived [19 ]
[40 ], but a high amount of automation is necessary to avoid any user interaction and
bias [19 ].
Fig. 6 Quantitative CT post-processing. a The initial step of automatic airway analysis is the segmentation of the whole airway
tree from the CT dataset. Bronchiectasis can nicely be seen as buddings at the end
of an airway branch on the 3 D volume rendering. b A centerline is then calculated for each airway segment, which represents the long
axis of each airway. c Subsequently, secondary reconstructions running perpendicular to the airway axis
(centerline) are produced, which show an axial view for each airway segment. d On these, the inner (green line) and outer (red line) airway wall may be detected
and measured by sophisticated algorithms. The yellow line marks the points of maximum
wall attenuation. Images by YACTA, programming by Oliver Weinheimer, Heidelberg.Abb. 6 Quantitative CT-Nachverarbeitung. a Als erster Schritt einer automatischen Atemwegsanalyse erfolgt die Segmentierung
des gesamten Atemwegsbaums aus einem CT-Datensatz. Bronchiektasen lassen sich leicht
als Knospungen an den Enden der Atemwege in der 3 D gerenderten Rekonstruktion erkennen.
b Hiernach wird eine Mittellinie (Centerline) berechnet, die der Längsachse eines jeden
Atemwegs folgt. c Sekundäre Rekonstruktionen senkrecht zur Atemwegsachse (Centerline) werden erstellt,
die nun eine axiale Sicht des jeweiligen Atemwegssegments erlauben. d Auf diesen kann nun die innere (grüne Linie) und äußere (rote Linie) Begrenzung der
Atemwegswand mittels spezieller Rechenalgorithmen detektiert werden. Die gelbe Linie
zeigt die Punkte der höchsten Dichte der Atemwegswand. Bilder erstellt mit YACTA,
programmiert von Oliver Weinheimer, Heidelberg.
Dedicated software tools for the quantification of MRI perfusion based on the indicator
dilution theory are already available, which can at least perform the initial step
of segmenting the lung from the 4 D perfusion dataset [54 ]
[55 ]. Four parameters have been developed to reflect the characteristics of pulmonary
hemodynamics: pulmonary blood flow (PBF), blood volume (PBV), mean transit time (MTT),
and time-to-peak (TTP) [56 ]. Using a modification of these parameters it could be shown that perfusion in the
CF lung may not only be reduced by peak quantity but also delayed [57 ]. It has been speculated that delayed perfusion may reflect increased bronchial arterial
supply in advanced lung disease. Because these receive blood from the systemic circulation,
increased flow will result in a left-to-left shunt, which is still of uncertain clinical
significance.
Advanced ventilation and perfusion imaging with MRI
Advanced ventilation and perfusion imaging with MRI
An option for direct visualization of lung ventilation is the imaging of nuclei other
than 1 H, namely 3 He and 129 Xe [58 ]. By this approach, MRI after inhalation of the noble gas will display ventilated
airspace only. Hyperpolarized 3 He-MRI depicted a high number of ventilation defects in CF patients compared to healthy
volunteers, which correlated with a decrease in lung function [59 ]
[60 ], but showed poor correlation with chest X-ray scoring [61 ]. Ventilation defects are present even in CF patients with normal lung function testing
and may change after airway clearance treatment [62 ]. Sophisticated technical prerequisites and the price for noble gas isotopes make
this promising research tool expensive and rather unlikely to be introduced into routine
patient care.
A promising technique is the direct regional quantification of T1 relaxation times.
As a physical parameter, it is thought to provide an objective parameter for the characterization
of pulmonary tissue independent of scanner type or observer [63 ]. Preliminary results obtained in patients indicate that T1 relaxation time is significantly
shorter in lungs affected by emphysema or cystic fibrosis [64 ]. Furthermore, T1 mapping can be combined with oxygen-enhanced MRI, which exploits
the paramagnetic effect of molecular oxygen (O2 ) for the indirect assessment of lung ventilation. The slope of T1 decrease at different
oxygen levels correlated with perfusion abnormalities [65 ].
Another newly developed technique relies on the periodical signal changes of free-breathing
bSSFP sequences at high temporal resolution induced by respiration and pulsatory blood
inflow [66 ]. A mathematical Fourier decomposition separates these different frequency peaks
and allows for the calculation of ventilation and perfusion maps. Preliminary results
from patients with CF (age range: 0 – 30 years, median age: 4.1 years) show a good
agreement with contrast-enhanced perfusion imaging [54 ] ([Fig. 7 ]).
Fig. 7 Non-contrast enhanced combined ventilation and perfusion imaging with MRI. a, b Apart from areas with reduced parenchymal signal on T2-weighted imaging (white arrowhead)
this school-age female with CF in stable clinical condition showed few airway abnormalities.
Contrast-enhanced perfusion MRI revealed areas of reduced perfusion (white arrow,
b ), comparable to the aforesaid areas with reduced T2-signal. c, d Fourier-decomposition MRI detected nicely matching areas of reduced ventilation (black
arrow, c ) and perfusion (white arrow, d ) without the need for contrast material injection.Abb. 7 Kontrastmittelfreie kombinierte Bildgebung von Ventilation und Perfusion mittels
MRT. a, b Neben flächigen Signalminderungen des Parenchyms in der T2-Wichtung (weißer Pfeilkopf)
zeigte diese Patientin im Schulalter mit stabiler CF kaum Atemwegsveränderungen. Die
kontrastmittelverstärkte Perfusions-MRT deckte deutliche Areale mit reduzierter Perfusion
auf (weißer Pfeil, b ), vergleichbar zu den vorgenannten Arealen mit T2-Signalminderung. c, d Die Fourier-Dekompositions-MRT erlaubte die Detektion von hierzu gut korrelierenden
Arealen mit reduzierter Ventilation (schwarzer Pfeil, c ) und Perfusion (weißer Pfeil, d ) ohne die Notwendigkeit einer Kontrastmittelinjektion.
Summary and outlook
Although many authors advocate regular imaging studies at specialized CF centers,
data on the actual impact of imaging findings on treatment decisions and patient survival
is lacking. Therefore, German and international guidelines usually do not specify
at what age surveillance imaging of the CF lung should be started, or even which modality
should be employed [67 ]. Chest CT is superior to CXR due to higher sensitivity for morphological changes
in the CF lung, but routine surveillance CT acquisitions have subsequently led to
an increase in radiation exposure to CF patients, which may even rise further with
earlier diagnosis and prolonged survival [10 ]. A remaining role for CXR could be imaging at annual follow-up together with MRI
as a cross-sectional modality for use of CXR as a reference when it is repeated at
interim presentations between annual follow-up, for example in the case of exacerbation.
Recently, chest MRI has entered clinical routine practice in CF [12 ]. Thus, radiologists and clinicians now can opt for the optimal modality adapted
to the clinical context of their CF patients ([Table 2 ]). The risk of sedation in preschool children and allergies against MRI contrast
material must be weighed against the risk from radiation exposure [9 ]
[10 ]
[68 ]. Importantly, to use MRI in CF as a routine surveillance tool is not limited to
the depiction of structural information as with CT just using a radiation-free method.
MRI’s capability for combined morphological and functional imaging at sufficient spatial
and high temporal resolution to obtain information on regional lung function should
be taken into account as well. To appreciate its advantages over CT, a perfusion study,
which is available on most state-of-the-art MRI scanners already, should be included
in the MRI protocol ([Fig. 1 ]).
