Eur. Radiol. (2001) 11: 982±989 DOI 10.1007/s003300100818

C. P. Heussel B. Hafner J. Lill W. Schreiber M. Thelen H.-U. Kauczor

Received: 4 September 2000 Revised: 12 December 2000 Accepted: 15 December 2000 Published online: 23 February 2001  Springer-Verlag 2001

)

C. P. Heussel ( ) ´ W. Schreiber ´ M. Thelen ´ H.-U. Kauczor Department of Radiology, Johannes Gutenberg University, Langenbeckstrasse 1, 55131 Mainz, Germany E-mail: [email protected] Phone: +49-61-31 17 51 96 Fax: +49-61-31 17 34 23 B. Hafner Department of Ear-Nose-Throat Surgery, Johannes Gutenberg University, Langenbeckstrasse 1, 55131 Mainz, Germany J. Lill Department of Pneumology, Internal Medicine III, Johannes Gutenberg University, Langenbeckstrasse 1, 55131 Mainz, Germany

CHEST

Paired inspiratory/expiratory spiral CT and continuous respiration cine CT in the diagnosis of tracheal instability

Abstract In tracheo- and bronchomalacia, localisation and determination of collapse is necessary for planning of surgical procedure. We compared inspiratory and expiratory spiral CT, cine CT, bronchoscopy, exemplary cine MR, and evaluated the clinical relevance. Twenty-nine patients (2 follow-ups; mean age 61 years, age range 27±85 years) with suspected or verified tracheal stenosis or collapse (post-tracheotomy: n = 17; neoplasm: n = 5; other: n = 7) underwent paired breathhold inspiratory and expiratory spiral CT. Forty-five additional cine CT were performed at 1±4 levels (mean 1.5) during continuous respiration (increment 100 ms) to clarify respiratory collapse. The tracheal crosssectional diameters of both techniques were calculated. Comparison with bronchoscopy, follow-up, and influence upon therapy were evaluated retrospectively. Exemplary comparison with cine MR (8 frames/ s) was done in 3 cases. In addition to bronchoscopy, further information concerning localisation, extent, collapse, stability of the tracheal wall,

Introduction Instability of the respiratory tract occurs in diseases such as tracheo- and bronchomalacia, chronic-obstructive lung disease [1, 2], post-tracheotomy [3], infection due to destruction of cartilage (e.g. polychondritis) or after surgery for neoplasms or post-lung transplantation

distal portions of the stenosis and extraluminal compressions were obtained. A significantly higher degree and more pathological collapses (> 50 %) were seen using cine CT (38 %) compared with paired spiral CT (13 %; degree: p < 0.0001; number: p < 0.001). The findings changed the further therapeutic procedure in 16 of 29 patients. Further stenoses were excluded and bronchoscopy was verified in another 13 of 29. Temporal resolution of cine CT and cine MR is sufficient; however, spatial resolution of cine MR is inferior. Paired inspiratory and expiratory spiral CT localises tracheal stenoses and demonstrates relevant extraluminal compression. Significantly improved evaluation of respiratory collapse and further information of localised tracheal instability is obtained by cine CT. Cine MR promises more functional information especially due to free choice of imaging plane. Key words Cine CT ´ Helical CT ´ Image display and recording ´ Trachea ´ Stenosis or obstruction

[3, 4]. The destruction of the cartilage leads to a respiratory collapse of the tracheal or bronchial wall especially during expiration. Adequate technique for determination of the length of stenosis as well as extent and kind of collapse is necessary for the selection of the appropriate therapy. A primary ventral collapse might be treated by implanta-

983

tion of an external ring for stabilisation. Treatment of a dorsal collapse is more difficult and comparable to a circular or diffuse collapse. Either resection of the tracheal level or stent implantation are frequently performed in these cases; thus, a functional investigation of the tracheal wall during respiration is necessary for localisation of the collapse and detailed evaluation of its nature. This is usually done with bronchoscopy during continuous respiration and provocation (coughing). Local irritations of the sensitive trachea are possible by this invasive procedure. In dyspnoeic patients, lifethreatening airway obstruction and oedema might be triggered [3]. Additionally, narrow stenoses cannot be passed with the bronchoscope and the investigation is limited to the upper levels of stenoses. Lower levels may remain uninspected by bronchoscopy. Therefore, crosssectional techniques, such as CT, are useful in imaging the tracheobronchial tree, because the upper and lower levels of a stenosis are shown equally well. Furthermore, the relationship of extraluminal structures, such as aortic arch or thyroid gland, are depicted at CT complementing bronchoscopy in contrast to fluoroscopy which is limited by superimposition [3, 5]. Thus, CT is the most widespread technique in imaging the tracheobronchial tree [3, 6, 20]. Post-processing of volume datasets for multiplanar reconstruction (MPR) or surface-shaded reconstruction of airways based on spiral CT yield particularly high-quality images [3, 7]. Functional information concerning localisation and extent of collapse are obtained by a comparison of scans acquired during an inspiratory and an expiratory breath-hold [3, 8]. However, a breath-hold investigation does not represent physiological conditions of continuous respiration [8]. Moreover, the patient closes the glottis for breath-holding, and the thoracic pressure is increased which might prevent or at least reduce a potential collapse of the trachea; thus, the true maximal movement of the tracheal wall is not depicted by breath-hold acquisitions and false-negative finding may be the result [3, 5, 9]. Fast imaging during continuous respiration promises improvement in visualisation of the maximal movement of the tracheal wall and, therefore, of the true extent of tracheal collapse. The aim of this study was to compare paired inspiratory and expiratory spiral CT of the trachea with cine CT which was acquired during continuous respiration. Furthermore, clinical relevance of the whole CT investigation was evaluated. Due to the drawbacks of cine CT (radiation dose, temporal overlap of images), exemplary cine MR was planned as an alternative approach.

Methods Patients Twenty-nine patients (median age 61 years, age range 27±82 years; 19 women and 10 men; two follow-ups) with suspicion of or previously bronchoscopically verified tracheal stenosis or collapse underwent CT for localisation of stenosis, evaluation of collapse, and relationship to extraluminal compression. Tracheal disease resulted from tracheotomy during long-term mechanical ventilation (n = 16), neoplasm (n = 5) or other causes (n = 6). Uncertain external findings were checked in two additional cases. Due to results published previously [20], several levels were scanned with cine CT if the stenosis was longer than approximately 2 cm. Thus, 45 cine CT were performed at 1±4 levels per patient (mean 1.5). The influence of CT upon the selection of therapeutic procedure (e.g. external anterior stabilisation, resection of tracheal level, stent implantation, PEEP mask, conservative) was evaluated retrospectively by the referring physician. Spiral CT Spiral CT was performed from carina to glottis at Siemens Somatom scanner (plus series, Siemens, Erlangen, Germany). After careful patient instruction, two identical spiral-CT scans (slice thickness 4±5 mm, table feed 8 mm, increment 4 mm, 120 kV, 53±145 mAs) were performed in supine position, one during an inspiratory and one during an expiratory breath-hold, each lasting for approximately 20 s. Image reconstruction was done in an edgeenhancing (high-resolution) algorithm, a slim z-profile interpolation, with a field of view of 200 mm. Cine CT After localisation of the collapse region, dynamic multiscan CT (cine CT without table feed) with a slice thickness of 2 mm and continuous data acquisition (no interscan delay) was performed for 10±12 s at this table position (120 kV, 53±145 mAs). Temporal reconstruction increment was 100 ms. The patient was instructed to breath deeply and slowly. Cine MR Exemplary scans were performed on a Siemens Magnetom Vision 1.5 T (Siemens Medical Systems, Erlangen, Germany) in supine position using the neck coil. An ultrafast spoiled gradient-echo pulse sequence with spin-density weighting was used during deep respiration with a TR 2.4 ms, TE 1.2 ms, flip angle 8 , field of view 300 mm, matrix 128 pixels, slice thickness 8 mm, resulting in a frame time of 140±150 ms for ~13 s. Images were acquired consecutively without interscan delay and with 54±64 phase-encoding steps and 128 data points in readout direction. Before reconstruction, raw data were filtered using a ªHamming Filterº, zero filled to a 256 ” 256 raw matrix, and then reconstructed using a half Fourier algorithm. Evaluation For measurements, a post-processing workstation (Magic-View, Siemens Medical Systems, Erlangen, Germany) was used with a high zoom factor and a uniform window setting (2000/±400 HU).

984

Cine CT

Fig. 1 Decrease of the cross-sectional tracheal area, normalised to the individual maximal value Due to spatial effects at spiral CT (slice thickness 4 mm, effective slice thickness ~6 mm), and cine CT (slice thickness 2 mm), crosssectional diameters were measured in anatomically corrected positions at comparable table positions and simplified cross-sectional area was calculated as product of anterior±posterior by lateral diameter. Collapse was calculated in percent of the maximal crosssectional area. For demonstration of dynamic tracheal changes, cross-sectional area was evaluated planimetrically with a double threshold setting (CT: ±1024/±400 HU, MR: 0/4”signal intensity within tracheal lumen) for each cine image in 3 exemplary patients. Comparison with bronchoscopy was done visually together with the referring physician.

The respiratory phase is not clearly defined during continuous respiration of cine CT. Hence, an accurate assignment to an inspiratory or expiratory collapse is not possible. Direction and type of collapse was comparable for cine CT as well as for spiral CT, only the extent was different. A mean decrease of cross-sectional lumen from 16 ” 12 to 11 ” 8 mm (range 3±29 ” 3-19 to 2±23 ” 2-19 mm) was measured; thus, the mean simplified tracheal area decreased from 178 to 97 mm2 (range 15±551 to 6±288 mm2; p < 0.0001; Fig. 1). The minimal lumen was 56 % of the maximal tracheal lumen, which is significantly smaller as compared with spiral CT (p < 0.0001). The visual comparison of bronchoscopy correlated better with findings derived from cine CT. A pathological collapse of more than 50 % of the maximal cross-sectional area was seen in 17 of 45 levels (38 %), which is significantly more frequent when compared with spiral CT (p < 0.001). Cine MR Exemplary comparison of cine CT and cine MR revealed very similar time curves of tracheal cross-sectional area (Fig. 2a). Due to lower spatial resolution of cine MR, the information concerning tracheal wall stability is lacking.

Results

Consequences

The CT data of all 29 patients could be obtained completely. Nasal oxygen application was performed in 2 cases to facilitate investigation in the supine position. Nevertheless, breathing artefacts occurred in 7 of 29 patients.

Complementary to bronchoscopy, localisation and extent of stenosis and collapse were estimated and verified in 13 of 29 patients. Computed tomography was necessary for exclusion of previously unknown distal collapses and therefore influenced treatment by confirming the planned therapeutic procedure. An extension of the previously planned treatment resulted from CT findings in 4 of 29 cases. This led to two additional thyroidectomies, one laser surgery of a neoplasm, and one extensive surgical sublaryngeal widening. A relevant change of the initially planned treatment was necessary in 12 of 29 patients. Due to anterior collapse, an external stabilisation was preferred in 3 of 12, stent placement in 3 of 12, change in surgical procedure in 3 of 12, jet ventilation in 3 of 12, and initiation of additional immunosuppressive therapy due to tracheal manifestation of Wegener's granulomatosis in 1 of 12.

Spiral CT Smallest cross-sectional area was seen at 40 of 45 levels during expiration and the remaining 5 of 45 at inspiration. Major direction of respiratory collapse was lateral in 25 of 45 levels, anterior±posterior in 19 of 45 (markedly in anterior tracheal portions) and irregular in 1 of 45. A mean decrease of cross-sectional lumen from 16 ” 14 to 10 ” 9 mm (range 3±31 ” 3±26 to 3±26 ” 2±20 mm) was measured. Thus, the mean simplified tracheal area decreased from 174 to 132 mm2 (range: 15±400 to 12±360 mm2; p < 0.0001; Fig. 1). The minimal lumen was 72 % of the maximal tracheal lumen. A pathological collapse of more than 50 % of the maximal cross-sectional area was seen in 6 of 45 levels (13 %).

985

a

b

c

d

h

i

j

k

l

m

e

f

g

Fig. 2 a±k A 72-year-old patient after emergency tracheotomy due to anaphylaxis. Dyspnoea increased during the last years. Finally, the collapse extended to the whole trachea but the patient refused surgery. After decompensation of stridor, she agreed to stent placement. a,b Emergency fluoroscopy demonstrates the extent of the lateral collapse during respiration (arrows). c Best impression of collapse can be obtained in animation. For planning of treatment, paired spiral CT was performed (surface-shaded reconstruction inspiration: d (anteroposterior); f (lateral); expiration: e (anteroposterior); g (lateral); h curved multiplanar reconstruction inspiration; i expiration). Cine CT was done to evaluate the collapse in detail. A higher collapse down to 18 % of maximal cross-sectional area was detected at l,m cine CT compared with j paired inspiratory and k expiratory spiral CT (44 %) at the same position

986

a

b

c

d

e Fig. 3 a±e A 75-year-old patient after tracheotomy. Cine CT was done to evaluate the collapse in detail. A pathological collapse down to 24 % of maximal cross-sectional area was detected at cine CT (c,d), whereas paired inspiratory (a) and expiratory (b) spiral CT revealed a non-pathological collapse down to 87 %. e The dynamic collapse is best seen in animation

1. In younger patients or cases with problems particularly at night, a PEEP mask might be the instrument of choice. 2. Surgical stabilisation by implantation of an anterior external ring is done in cases with short (one to two cartilaginous rings) and predominately anterior collapse taking place in a lateral direction. This is performed usually after tracheotomy to inflate the tracheal lumen and stabilise the tracheal wall. The posterior portions of the trachea should remain in their original shape. An anterior±posterior collapse as well as a soft or mobile stenosis indicates the patient for open surgery or stent placement. 3. In curative tumour resections, laser surgery or open tracheal wall resection are alternatives. Stability of contralateral and surrounding tracheal wall is essential for this decision. Cine CT provides further functional information in special cases as demonstrated in Fig. 3. 4. Stent placement is frequently used in cases with a long collapse or in supportive tumour therapy [22]. 5. Surgical resection of a tracheal level is a very invasive procedure which is only done at specialised centres; however, the exact length of the collapse has to be known. Thus, characterisation of the direction (anterior±posterior or lateral), and the length of the collapse are indispensable contributions for these decisions. Especially interventional and surgical procedures require an accurate localisation of stenosis and collapse. Since the collapse is a fast functional finding, its evaluation should be done using a real-time technique. Bronchoscopy is the method of choice for this purpose; however, view is limited to the upper tracheal levels. Further stenoses or collapses of lower tracheal levels are hidden, if the stenosis cannot be passed by the bronchoscope. Furthermore, in dyspnoeic patients, local or psychic irritations may limit this method.

Discussion

Technique

Stenoses of the trachea and the main stem bronchi are divided into diffuse or focal forms, depending on their extent and quality [12]. Aetiology might either be instability of the tracheal cartilage (tracheo- or bronchomalacia, polychondritis, post-tracheotomy, postlung transplantation), or external compression, for instance, by the aortic arch, aneurysm or goiter [1, 2, 3, 4, 13], because, due to a stenosis, an additional respiratory collapse occurs frequently and indicates instability of the tracheal wall. Dyspnoea and stridor are the major clinical findings in these patients. Detailed diagnostics of the respiratory collapse is essential for selection of the adequate therapeutic option:

Several non-invasive imaging techniques are available for the detection of tracheal stenosis. At chest X-ray, approximately two of three of tracheal stenosis are already discernible (Fig. 4) [12]. Cross-sectional imaging techniques, especially CT, displays the tracheobronchial tree without superimposition. By this, stenosis is demonstrated in its whole extent, including the unaffected upper and lower tracheal sections in combination with neighbouring potentially compressing organs or structures [3]. Additional post-processing of volume data sets using MPR, or surface-shaded reconstruction, yields illustrative information for invasive procedures [3]. To obtain the anatomical relationship, spiral CT should

987

a

b

c

d

e

f

g

h

Fig. 4 a±h A 51-year-old patient suffering from tracheal neoplasm. A pathological collapse a,b down to 40 % of maximal cross-sectional area was detected at spiral CT and c,d down to 30 % at cine CT. The major finding was the protrusion of the right-sided tumour mass, whereas the contralateral left-sided tracheal wall was unaffected and stable; thus, laser surgery of the tumour without further tracheal stabilisation was successfully performed. The dynamic collapse is best seen in animation (g). e,f Cine MR was performed to demonstrate its capability in tracheal cine imaging. The principal finding is demonstrated in a comparable way to CT; however, the lower spatial resolution limits the identification of the tumour protrusion and stability of the contralateral left tracheal wall. h The dynamic collapse is best seen in animation

cover at least the region from the glottis to the carina. A slice thickness of 4±5 mm and a pitch of ~2 appeared to be a useful compromise in dyspnoeic patients taking volume of interest, time of breath-hold, and radiation dose into account. However, lumen measurement at spiral CT is limited by spatial volume effects and, therefore, planimetry of the tracheal lumen would not be comparable to cine CT done with 2-mm slices (Figs. 3a, 5). The irregular lumen in post-tracheotomy and tumour patients requires an individual threshold for planimetry of spiral CT (Figs. 3a,b, 5). Since this results in non-comparable and erroneous values, we decided to measure the transversal diameters of the tracheal lumen manually in spiral CT and cine CT using anatomical landmarks. Furthermore, information concerning change of the tracheal shape can only be quantified by this manual procedure. Only an intra-individual comparison of the tracheal area within a single cine scan can be obtained using planimetry (Fig. 2). Valuable functional information for assessment of the collapse results from comparison of identical paired spiral-CT scans, one acquired during an inspiratory breath-hold, and the second during an expiratory breath-hold [3, 4, 10, 20]. However, false-negative findings concerning tracheal collapse are known to occur especially in cases with a low degree of stenosis [3, 9, 14, 20]. Real-time or time-resolved imaging techniques during continuous or forced respiration might surmount this limitation [5, 6, 10, 20]. Time-resolved image reconstructions can be obtained using a spiral-CT scanner with continuous data acquisition (as in spiral CT) without table feed. The time per frame depends on the rotation time (500±1000 ms) and the interpolation algorithm. The temporal increment of 100 ms results from an overlap of frames, which had a acquisition time of approximately 250±750 ms depending on the scanner used [4]. This temporal resolution is sufficient for investigation of normal respiration in adults [6, 9, 15, 20], which is confirmed by comparison of cine CT and cine MR (Fig. 2a). The real time frame of cine MR using 140 ms without temporal overlap did not reveal any advantage in the three studied patients. However, forced respiration and coughing, which was investigated in a different study, led to inadequate motion artefacts. Shorter acquisition time is necessary for these provocation tests. This limitation might be surmounted by the use of electron-beam tomography (EBT). Frames with 50±100 ms can be acquired; however, actual studies allow 500 ms interscan delay between single frames, and thus continuous data cannot be acquired [21]. The resulting image stacks are best viewed using a powerful workstation in a fast loop with a high zoom factor. This is the appropriate way for evaluation of maximal and minimal collapse [6, 9, 15, 20]. After export of registered image stacks, compilation into video

988

Fig. 5 a,b Exemplary demonstration of dynamic change in crosssectional area of 3 patients. The values are normalised on the maximal and minimal individual area. a Dynamic changes in a patient with a tracheal neoplasm are demonstrated (same as in Fig. 3). A very continuous variation in tracheal cross-sectional area is shown by CT (black line). Maximum is reached in steps, which is a result of wobbling polypoid tumour mass within the tracheal lumen. Magnetic resonance imaging (grey line) shows variation in tracheal cross-sectional area is less continuous than in CT. A lack of temporal overlap in combination with lower spatial resolution might be causal. Stepwise increase up to maximum is just as good as in CT of this patient. b Dynamic changes of tracheal area is shown for a patient with tracheal malacia (same as in Fig. 4). A very smooth and continuous curve results from adequate time resolution and temporal overlap of the images. Only mild fluctuations occur prior to the maximum area being reached

position with tilt gantry allows for coronal acquisition [4, 16]. This is usually not possible in dyspnoeic patients. Furthermore, a relatively high radiation dose is applied within a small organ volume. Unfortunately, the relatively sensitive thyroid gland is within the investigation range. To minimise the exposed organ volume, a thin slice was chosen for cine CT (2 mm) in comparison with spiral CT (4±5 mm). We measured a surface dose of 90 mGy in a 12-s cine CT with the oldest available scanner (plus) with unfavourable parameters of 137 kV and 145 mAs to check the worst case of radiation dose. With an actual scanner (Somatom plus 4) with solidstate detector system, 120 kV and 53 mAs for a 10-s cine CT results in a dose of 47 mGy. The volume of interest is very small in comparison with routine techniques, and only a very small tissue volume is irradiated with this dose. The high radiation dose of this technique must lead to restriction of indications for cine CT, and presurgical determination of relevant functional data, especially in elderly patients (mean age in our study was 61 years). Cine MR surmounts both limitations of CT and furthermore improves temporal resolution. Real frame times of 140 ms are possible without sophisticated sequences. The planimetry of the tracheal lumen derived from cine CT and cine MR is comparable. However, spatial resolution is inferior when compared with CT. Furthermore, cine MR offers the possibility of multiplanar image acquisition, e.g. in the coronal or sagittal plane. This promises additional interesting functional information. This technique was already established in dynamic studies of tracheal and bronchial stenosis, as well as in cardiac examinations, and revealed favourable results [17, 18, 19]. Collapse

files (e.g. AVI, MPG, animated GIF) enables viewing at PC platforms (see http://www.uni-mainz.de/FB/Medizin/Radiologie). One disadvantage of the used technique is limitation to a single scan level per acquisition. In fact, respiratory z-axis movement of the extrathoracic trachea can be neglected [5]. But intrathoracic trachea and main stem bronchi are relevant scan regions and respiratory motion interferes with detailed investigation of the tracheal wall movement. Several cine-CT scans can be performed side by side (Fig. 4). But due to the necessity of successive scanning, respiratory cycles are not synchronous. Multi-row detector scanners will enable to scan neighbouring levels, at least within a limited region. Two further disadvantages of cine CT need to be mentioned: CT is mainly limited to the transverse plane. Only in very cooperative patients, scanning in prone

A decrease to 64±68 % of the maximal cross-sectional area is described as the physiological end-expiratory condition [5, 6, 10]. It is caused mainly by a shifting of the relaxed paries membranaeca [5, 6, 10]. A collapse of more than 50 % is described in the literature as pathological [5, 6, 10]. Even so, no therapeutic consequences should be based upon this finding alone [6, 9, 10, 11]. Decreases to 21 % have been previously described in tracheomalacia and were confirmed bronchoscopically [6, 11]. Using paired spiral CT, a collapse of at least 50 % was found at significantly fewer levels when compared with cine CT (13 vs 38 %; p < 0.001). Furthermore, the median collapse at paired cine CT is significantly higher compared with spiral CT (59 vs 77 %; p < 0.0001; Fig. 1). This is probably due to the higher airway pressure during a breath-hold (regardless whether inspiratory or expiratory) with closed glottis compared with the physio-

989

logical behaviour of the tracheal wall during continuous respiration in cine CT. The false-positive cases described in the literature can possibly be explained by this fact. Cine CT provides additional relevant information concerning the respiratory change of tracheal shape. Since a ventral collapse is preferably treated with external stabilisation and a dorsal or diffuse collapse requires tracheal level resection or stent placement, this information has a high clinical impact. Thus, cine CT is a suitable tool for functional imaging of extent (percent of maximal cross-sectional area) and type (e.g. anterior±posterior, lateral) of tracheal collapse.

Conclusion The following conclusions were reached as a result of this study:

1. Paired inspiratory and expiratory spiral CT localises and determines extent of tracheal stenosis. This procedure enables differentiation of focal and diffuse tracheomalacia. 2. Complementary cine CT determines extent of tracheal collapse more reliably. Results of cine CT correlate better with bronchoscopy. 3. Relevant functional information is obtained in patients suffering from tracheal stenoses and collapses of different causes by combined use of paired spiral CT and cine CT. Due to their influence upon therapeutic options, complementary use with bronchoscopy allows for a complete assessment of tracheal stenosis and collapse. 4. Due to the relatively high radiation dose even with modern scanners, cine CT has to be restricted to dedicated functional questions. Acknowledgement We thank Mrs. P. Meinhardt for proofreading and language editing.

References 1. Shepard JO (1995) The bronchi: an imaging perspective. J Thorac Imaging 10: 236±254 2. Dunn WF, Hubmayr RD, Pairolero PC, Sheedy PF, Edell ES, Nelson SB (1990) The assessment of major airway function in a ventilator dependent patient with tracheomalacia. Chest 97: 939±942 3. Quint LE, Whyte RI, Kazerooni EA, Martinez FJ, Cascade PN, Lynch JP III, Orringer MB, Brunsting LA III, Deeb GM (1995) Stenosis of the central airways: evaluation by using helical CT with multiplanar reconstructions. Radiology 194: 871±877 4. Kauczor HU, Buchenroth M, Heussel CP, Mayer E (1996) Functional CT imaging of the lung in axial and coronal plane after single lung transplantation. Fortschr Roentgenstr 164: 441±444 5. Ell SR, Jolles H, Keyes WD, Galvin JR (1985) Cine CT technique for dynamic airway studies. AJR 145: 35±36 6. Stern EJ, Graham CM, Webb WR, Gamsu G (1993) Normal trachea during forced expiration: dynamic CT measurements. Radiology 187: 27±31 7. Holbert JM, Strollo DC (1995) Imaging of the normal trachea. J Thorac Imaging 10: 171±179 8. Kauczor HU, Heussel CP, Fischer B, Hast J, Mildenberger P, Thelen M (1998) Value of spirometry-gated high resolution computerized tomography of the lung during inspiration and expiration. Fortschr Roentgenstr 169: 658±661

9. Brody AS, Kuhn JP, Seidel FG, Brodsky LS (1991) Airway evaluation in children with use of ultrafast CT: pitfalls and recommendations. Radiology 178: 181±184 10. Blechschmidt R, Loercher U, Rust M (1997) Radiologische Studie der Tracheaquerschnitte im Verlauf des Atemzyklus am Beispiel der Tracheomalazie. Fortschr Roentgenstr 166 (Suppl):S197 11. Newth CJ, Lipton MJ, Gould RG, Stretton M (1990) Varying tracheal cross-sectional area during respiration in infants and children with suspected upper airway obstruction by computed cinetomography scanning. Pediatr Pulmonol 9: 224±232 12. Kwong JS, Adler BD, Padley SP, Muller NL (1993) Diagnosis of diseases of the trachea and main bronchi: chest radiography vs CT. AJR 161: 519±522 13. Stark P (1995) Imaging of tracheobronchial injuries. J Thorac Imaging 10: 206±219 14. Whyte RI, Quint LE, Kazerooni EA, Cascade PN, Iannettoni MD, Orringer MB (1995) Helical computed tomography for the evaluation of tracheal stenosis. Ann Thorac Surg 60: 27±31 15. Brasch RC (1988) Ultrafast computed tomography for infants and children. Radiol Clin North Am 26: 277±286 16. Kaiser MC, Veiga-Pires JA (1981) Sitting position variation of direct longitudino-axial (semi-coronal) mode in CTscanning. Fortschr Roentgenstr 134: 97±99

17. Hofmann U, Hofmann D, Vogl T, Wilimzig C, Mantel K (1991) Magnetic resonance imaging as a new diagnostic criterion in paediatric airway obstruction. Prog Pediatr Surg 27: 221±230 18. Hite SH, Neeson HR, Cunningham T, Remley KB, Stillman AE, Rimell FL (1997) Dynamic MR imaging of the paediatric trachea. Radiology 205 (P):696 19. Rominger MB, Bachmann GF, Geuer M, Puzik M, Boedeker RH, Ricken WW, Rau WS (1999) Accuracy of right and left ventricular heart volume and left ventricular muscle mass determination with cine MRI in breath holding technique. Fortschr Roentgenstr 170: 54±60 20. Heussel CP, Hafner B, Lill J, Thelen M, Kauczor H-U (2000) Diagnosis of tracheal instability: inspiratory and expiratory spiral-CT and cine-CT. Fortschr Roentgenstr 172: 115±121 21. Hein E, Rogalla P, Hentschel C, Taupitz M, Hamm B (2000) Dynamic and quantitative assessment of tracheomalacia by electron beam tomography: correlation with clinical symptoms and bronchoscopy. J Comput Assist Tomogr 24: 247±252 22. Hafner B, Mann W, Heussel CP, Scherhag A, Schlegel J, Kauczor HU (2000) Self-expanding nitinol stents in proximal tracheal stenosis. Laryngorhinootologie 79: 165±170