Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved Original Report R. C. Gilkeson 1 Leslie M . Ciancibello 1 Rana B. Hejal 2 Hugo D. M ontenegro 2 Paul Lange 2 Tracheobronchomalacia: D ynamic Airway Evaluation with Multidetector CT OBJECT IVE. The objective of our study was to evaluate the role of dynamic inspiratory– expiratory imaging with multidetector CT in patients with suspected tracheobronchomalacia. CON CLUSION . Multidetector CT with inspiratory–expiratory imaging is a promising method in the evaluation of patients with dynamic airway collapse. In our study, the degree of dynamic collapse correlated well with bronchoscopic results. Dynamic expiratory multidetector CT may offer a feasible alternative to bronchoscopy in patients with suspected tracheobronchomalacia. A Received M arch 14, 2000; accepted after revision June 20, 2000. 1 Department of Radiology, University Hospitals of Cleveland, Case Western Reserve University School of M edicine, 11100 Euclid Ave., Cleveland, OH 44106. Address correspondence to R. C. Gilkeson. 2 Division of Pulmonary and Critical Care M edicine, University Hospitals of Cleveland, Cleveland, OH 44106. AJR 2001;176:205–210 0361–803X/01/1761–205 © American Roentgen Ray Society AJR:176, January 2001 lthough tracheobronchomalacia is traditionally viewed as a disease of infants and neonates, recent literature suggests that there is greater recognition of this airway disorder by physicians treating adult patients with respiratory symptoms. Several large studies suggest that the incidence of tracheobronchomalacia is 5–10% in patients presenting with pulmonary complaints and that the incidence increases with advanced age [1, 2]. Equally impressive are recent data that dynamic airway collapse is seen in 10–15% of patients referred to a pulmonologist for evaluation of chronic cough [3]. Although historically diagnosed with cine fluoroscopy [4] and more recently with bronchoscopy, several articles have documented the value of dynamic expiratory CT, and particularly of electron beam CT [5], in the diagnosis of tracheobronchomalacia. Early reports suggest that multidetector CT provides greater anatomic accuracy than conventional single-slice CT [6]. In devising this study, we postulated that the decreased imaging time enabled by multidetector CT would allow volumetric evaluation of the central airways during dynamic expiration, a technique shown to be more sensitive in airway obstruction than conventional end expiratory techniques [7]. This report presents our early experience with the use of multidetector CT in the evaluation of patients with suspected tracheobronchomalacia. Subjects and Methods Thirteen patients with suspected tracheobronchomalacia were evaluated by multidetector CT. Clinical histories were reviewed in all patients. The patients included seven males and six females who ranged in age from 14 to 88 years (mean age, 49 years). Three patients had a clinical history of asthma, and one patient was a smoker. Six patients were referred with a history of chronic cough, whereas seven patients presented with unexplained dyspnea. Of the 13 patients, 12 underwent pulmonary function testing including flow-volume loops, and six patients underwent correlative fiberoptic bronchoscopy. Chest radiographs were available for 12 of the 13 patients. All patients were imaged on a multidetector CT scanner (Picker MX 8000; Marconi Medical Systems, Cleveland, OH). Initial topographic images were obtained first to determine the coverage of the trachea and central bronchi, which generally corresponded to an imaging range of 10–12 cm. CT parameters included a slice thickness of 2.5 mm with a reconstruction interval of 1.25 mm. A single-slice pitch equivalent of 1.5–1.75 was chosen to ensure both thin collimation and adequate coverage of the trachea and central bronchi. With multidetector CT technology, four contiguous 2.5-mm slices are obtained simultaneously. With a gantry rotation speed of 500 msec, the effective z-axis coverage is 3 cm/sec 205 Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved Gilkeson et al. at 2.5-mm collimation, for a single-slice pitch equivalent of 12. Given these parameters, 10–12 cm of the central tracheobronchial tree can be imaged in 4–6 sec, with all slices having an effective thickness of 2.5 mm. Because of an effective 50% reconstruction interval, multiplanar and virtual endoscopic imaging were performed in all patients. With the use of these imaging parameters, an initial CT scan was obtained during full inspiration; the scan time was 4–6 sec. All patients were scanned in the craniocaudal direction. After the patient was coached for a short period, the onset of the patient’s expiratory effort was synchronized with the onset of the expiratory phase of the CT examination. These scans were also obtained in 4–6 sec and were successfully obtained during the expiratory phase in all patients. In patients with suspected laryngomalacia, a second imaging volume was appropriately chosen to cover the upper airways. Images were reviewed at both mediastinal windows (level, 40 H; width, 340 H) and lung parenchymal windows (level, –500 H; width, 1500 H). Inspiratory and dynamic expiratory images were visually inspected for evidence of airway collapse. In the regions of maximal expiratory collapse, cross-sectional area measurements of the tracheobronchial tree were performed using standard software available on the Voxel Q workstation (Marconi Medical Systems). Airway collapse was then classified as 50–75% collapse, 75–100%, and 100% collapse. Axial images were volumetrically reconstructed on a workstation (Voxel Q or OmniPro; Marconi Medical Systems), and virtual bronchoscopic images were acquired using software (Voyager; Marconi Medical Systems). All studies were reviewed with the patient’s pulmonologist and were correlated with the bronchoscopic results if bronchoscopy had been performed. from 23% to 76% of predicted values, whereas three patients had a normal FEV1 value. There was little correlation between the degree of tracheobronchomalacia and the degree of obstruction indicated by the by FEV1 value. Flattening of the expiratory limb of the flow volume curve is a finding highly suggestive of tracheobronchomalacia and was seen in six of 12 patients who underwent pulmonary function tests. Although the three patients with significant tracheobronchomalacia seen on dynamic expiratory CT had a normal FEV1 value, flattening of the expiratory limb of the flow-volume loop was seen in all three patients. All patients showed evidence of airway collapse on inspiratory–dynamic expiratory CT. These data are summarized in Table 1. Of the 13 patients, dynamic expiratory CT showed complete collapse (100%) in three patients, collapse of greater than 75% in seven patients, and 50–75% collapse in three patients. The appearance and extent of tracheobronchomalacia varied significantly. In eight of the 13 patients, involvement of the central tracheobronchial tree was diffuse. In six of these patients, the crescentic form of tracheobronchomalacia (Figs. 1–3) was seen, whereas two patients exhibited the T ABLE 1 Seven patients presented with unexplained dyspnea, whereas six patients presented with chronic cough. In the group of six patients with chronic cough, sinus disease was excluded on the basis of CT findings and results of a barium swallow, which showed no evidence of gastroesophageal reflux. Asthma was excluded because of negative findings on a methacholine challenge test in five of six patients, whereas one patient who had a history of asthma had a persistent cough despite therapy. Chest radiographs were interpreted as showing normal findings in 10 of the 13 patients, whereas two patients exhibited diffuse intrathoracic tracheal narrowing without evidence of focal parenchymal disease. Pulmonary function testing with determinations of flow volume was performed before inspiratory–expiratory multidetector CT in 12 of 13 patients. Data are summarized in Table 1. Nine of the 13 patients showed evidence of an obstructive impairment, with values for forced expiratory volume in 1 sec (FEV1) ranging 206 Discussion Tracheobronchomalacia is characterized by weakness of the tracheal walls and supporting cartilage. Pathologically, tracheobronchomalacia is thought to result from a weakening of the cartilage and hypotonia of the posterior membranous trachea, with degeneration and atrophy Patient Characteristics, Pulmonary Function T est Results, and Imaging Findings Patient Characteristics Patient No. Age Sex (yr) Results “saber sheath” form. Of the four patients exhibiting focal forms of airway collapse, CT showed focal tracheobronchomalacia in two patients and localized bronchomalacia in two patients (Fig. 4). One patient underwent inspiratory–expiratory CT before and after tracheobronchial stenting, which depicted persistent bronchomalacia and air-trapping distal to the endobronchial stents (Fig. 5). Fiberoptic bronchoscopy was performed in six of the 13 patients, with two patients undergoing fiberoptic bronchoscopy before CT, and four patients undergoing bronchoscopy after inspiratory–expiratory CT. Bronchoscopy correlated well with the pattern and distribution of airway collapse seen on CT (Table 1), although dynamic expiratory CT underestimated the degree of collapse in one patient. 1 2 14 20 M M 3 28 F 4 5 6 7 8 9 35 38 52 56 51 55 F F M M M F 10 56 F 11 72 12 13 82 88 Clinical History Cough, dyspnea Dyspnea, Hunter’s syndrome Chronic cough, asthma, right aortic arch Asthma, sarcoid Asthma Chronic cough Chronic cough Chronic cough Dyspnea, chronic obstructive pulmonary disease Dyspnea, suspected congenital lobar emphysema F M M Pulmonary Function Test Results FEV1 Flattening of Predicted Flow -Volume (%) Loop Percentage of Collapse Seen on CT Bronchoscopy 98 23 Present Absent 75–100 75–100 75–100 NP NP Present 50–75 NP 63 62 107 72 36 50 Present Absent Present Present Present Absent 75–100 75–100 50–75 75–100 75–100 50–75 NP NP NP NP 75–100 NP 73 Absent 100 100 Dyspnea 67 Absent 100 100 Chronic cough Dyspnea 102 NP Absent Absent 75–100 100 100 100 Note.— FEV1 = forced expiratory volume in 1 sec, M = male, NP = not performed, F = female. AJR:176, January 2001 Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved Multidetector CT of T racheobronchomalacia of the longitudinal elastic fibers. Primary tracheobronchomalacia is a congenital weakness that presents at birth, whereas secondary or acquired tracheobronchomalacia is associated with prior intubation or radiation and a history of tumors and chronic obstructive pulmonary disease. Previous surgery and a history of tracheoesophageal fistula are also recognized risk factors for the development of tracheobronchomalacia [8]. The incidence of tracheobronchomalacia has been studied in several bronchoscopic series. In a series of more than 2000 bronchoscopies described by Jokinen et al. [1], 4.5% of the patients were found to have tracheobronchomalacia. Other large series using videobronchoscopy have cited an incidence of 15%, with incidence increasing with advancing patient age [2]. Recent clinical data suggest that in patients presenting with chronic cough, tracheobronchomalacia is the third most common cause after asthma and gastroesophageal reflux [3]. The diagnosis of tracheobronchomalacia is often delayed, particularly in adults. Symptoms are nonspecific; include cough, wheezing, and dyspnea; and are often misinterpreted as asthma [9]. Unfortunately, as witnessed in our patient population, there is poor correlation between the degree of tracheobronchomalacia and the severity of obstruction indicated by results of pulmonary function tests. Although most of our patients showed evidence of obstructive lung disease, two of the most pronounced cases of tracheobronchomalacia had normal FEV1 values, which were measured spirometrically. Although flattening of the flow-volume loop is highly suggestive of upper airway collapse, this finding may be difficult to isolate in the setting of severe obstruction. Indeed, although this finding was helpful in our study population, it was present in only 50% of the cases. The body of literature defining the criteria for tracheobronchomalacia is significant. In one of the original articles about tracheobronchomalacia, Johnson et al. [4] used fluoroscopy and cine evaluation to define the increasing severity of tracheomalacia. In their study, Johnson et al. found that patients often had associated chronic obstructive pulmonary disease and that the degree of tracheomalacia correlated with the severity of chronic obstructive pulmonary disease [4]. Videobronchoscopic studies of healthy children showed that the ratio of the expiratory–inspiratory cross-sectional area was 0.82. Children with tracheobronchomalacia had a expiratory–inspiratory ratio of 0.35 [10]. Recent literature has focused on the use of dynamic CT to evaluate tracheomalacia. Stern et al. [5] used electron beam CT to evaluate the dy- AJR:176, January 2001 A B Fig. 1.— 35-year-old w oman w ith history of sarcoidosis and asthma w ho presented w ith persistent dyspnea. Pulmonary function tests show ed flattening of expiratory limb of flow -volume loop, suggestive of upper airw ay collapse. A, Axial CT scan of trachea obtained during inspiration show s normal-caliber trachea. B, Axial CT scan of trachea obtained during dynamic expiration show s crescentic bow ing of posterior membranous trachea consistent w ith tracheobronchomalacia (arrow s ). A B Fig. 2.— 12-year-old boy w ith history of recurrent childhood infections and persistent barking cough. Pulmonary function tests show ed expiratory flattening of flow -volume loop, suggestive of upper airw ay collapse. A, Axial CT scan obtained at level of carina during inspiration show s normal caliber of mainstem bronchi. B, Axial CT scan obtained at level of carina during dynamic expiration show s marked narrow ing (arrow s ) of trachea and mainstem bronchi. namic range of normal tracheal diameters during inspiration and expiration. Similar to the findings in the bronchoscopy literature, Stern et al. reported that healthy volunteers showed a mean decrease of 35% in the cross-sectional area and that one patient with tracheomalacia showed a decrease of 82%. Ultrafast CT has been successfully used in infants with tracheomalacia associated with tracheoesophageal fistulas [11]. In these studies, dynamic axial images were obtained during inspiration and expiration at preselected levels in the trachea, and the maximal decrease in tracheal caliber was seen late but it was seen before end expiration. Other researchers have studied airway dynamics using electron beam CT and have been particularly successful in the evaluation of patients with obstructive sleep apnea. Reports have described the use of cine MR imaging as a particularly sensitive method in the evaluation of tracheomalacia [12]. The 50- to 100-msec imaging time, which was allowed by cine evaluation, of tracheal collapse during coughing is thought to be the most physiologically sensitive indicator of tracheomalacia. Although multidetector CT is an exciting new application and was successfully used in 207 Fig. 3.— 19-year-old man w ith Hunter’s syndrome, cough, and recurrent dyspnea in w hom indirect laryngoscopy (not show n) suggested laryngomalacia w ith soft-tissue infiltration of upper airw ays. Because of patient’s clinical status, family refused bronchoscopy. A, Axial CT scan of trachea obtained during expiration show s marked crescentic narrow ing of tracheal lumen. Note soft-tissue infiltration of mediastinum and trachea (arrow s ), consistent w ith mucopolysaccharide deposition. B, Shaded-surface display image of central airw ays in posterolateral projection show s diffuse narrow ing of trachea and bronchi (arrow s ). Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved Gilkeson et al. A B A B Fig. 4.— 57-year-old w oman w ith suspected congenital lobar emphysema of right lung. A, Axial CT scan obtained during dynamic expiration at level of bronchus show s extensive emphysematous change w ithin right lung, w ith extensive air-trapping and mediastinal shift due to hyperinflated right lung. B, Virtual bronchoscopic image obtained at level of bronchus intermedius during full inspiration show s mildly narrow ed but patent right middle (M ) and low er (L) lobe bronchi. C, Virtual bronchoscopic image obtained during dynamic expiration show s marked narrow ing of right middle lobe bronchus (straight arrow ) w ith complete collapse of low er lobe orifice (curved arrow ). These findings w ere not clearly appreciated on axial CT images (not show n) but w ere confirmed on bronchoscopy. C 208 AJR:176, January 2001 Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved Multidetector CT of T racheobronchomalacia Fig. 5.— 52-year-old man w ith idiopathic tracheobronchomalacia and persistent cough after undergoing stenting of mainstem bronchi. A, Fiberoptic bronchoscopic image obtained before stenting show s marked expiratory collapse of central airw ays (arrow s ), w hich is consistent w ith tracheobronchomalacia. B, Axial CT scan obtained at level of upper lobe bronchus during inspiration show s persistent narrow ing of proximal portion of upper lobe bronchus (arrow ) distal to endobronchial stent. Note position of bronchial stents. Position of bronchial stents are identified by arrow heads. C, Axial CT scan obtained during dynamic expiration show s focal collapse of proximal right upper lobe bronchus (straight arrow ). Note hyperlucency of right upper lobe, consistent w ith air-trapping of affected lung (curved arrow s ). Arrow heads = bronchial stents. A B our patient population, several limitations in our study design should be recognized. Our cohort was a highly selected patient population without healthy control subjects, and the images were interpreted with the knowledge of the clinical history and pulmonary function tests. Airway physiology studies have shown that during expiration, the small airways generally collapse before the larger airways. Our scanning protocol means that we imaged the upper airway early during expiration, whereas the distal airways were imaged near or at end expiration. These factors suggest that our craniocaudal scanning technique may have underestimated the degree of tracheobronchial collapse in the proximal air- AJR:176, January 2001 way. Although the close correlation of findings in our small population undergoing multidetector CT and bronchoscopy is encouraging, further research to determine the role of multidetector CT in the evaluation of airway dynamics is clearly needed. Despite these limitations, our preliminary results suggest that the information provided by our technique was important to our referring clinicians. Because of the volumetric nature of scan acquisition, we were able to visualize focal malacic segments that would have potentially been missed if interpreting scans obtained at selected axial levels. Although studies have shown that the addition of virtual bronchoscopic images does not significantly alter diagnosis, these C images were preferred by many of our clinicians, and in several cases obviated further bronchoscopic evaluation in patients who refused bronchoscopy or whose clinical status precluded fiberoptic bronchoscopy. Although further studies need to be performed, we suspect future research will establish an important role for multidetector CT in the evaluation of patients with airway disease. References 1. Jokinen K, Palva T, Sutinen S, Nuutinen J. Acquired tracheobronchomalacia. Ann Clin Res 1997;9:52–57 2. Ikeda S, Hanawa T, Konishi T, et al. Diagnosis, incidence, clinicopathology and surgical treatment of acquired tracheobronchomalacia [in Jap- 209 Gilkeson et al. Downloaded from www.ajronline.org by 52.73.204.196 on 05/17/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved 3. 4. 5. 6. 210 anese]. Nihon Kyobu Shikkan Gakkai Zasshi 1992;30:1028–1035 Palombini BC, Villanova CA, Araujo E, et al. A pathologenic triad in chronic cough: asthma, post nasal drip, and gastroesophageal reflux disease. Chest 1999;116:279–284 Johnson TH, Mikita JJ, Wilson RJ, Feist JH. Acquired tracheomalacia. Radiology 1973;109:577–580 Stern EJ, Graham CM, Webb WR, Gamsu G. Normal trachea during forced expiration: dynamic CT measurements. Radiology 1993;187:27–31 Min J, Shiau M, Leung AN, Rubin G. Quantitation of improved pulmonary arterial visualization with multidetector-row CT. (abstr) Society of Thoracic Radiology Annual Meeting, San Diego, California, March 12–16, 2000:123 7. Gotway MB, Golden JA, Lee ES, Reddy GP, Webb WR. Low-dose dynamic expiratory HRCT of the lungs using a spiral CT scanner. (abstr) AJR 2000;174[American Roentgen Ray Society 100th Annual Meeting Program Book suppl]:38–39 8. Demajumdar R, Rajesh PB. Have we got the full picture? J Laryngol Otol 1998;112:788–789 9. Imaizumi H, Kaneko M, Mori K, Yamada K, Asakura K, Tamagawa M. Reversible acquired tracheobronchomalacia of a combined crescent type and saber-sheath type. J Emerg Med 1995;13:43–49 10. Rozycki HJ, Van Houten ML, Elliott GR. Quantitative assessment of intrathoracic airway collapse in infants and children with tracheobronchomalacia. Ped Pulmonol 1996;21:241–245 11. Kao SC, Smith WL, Sato Y, Franken EA Jr, Kimura K, Soper RT. Ultrafast CT of laryngeal and tracheobronchial obstruction in symptomatic postoperative infants with esophageal atresia and tracheoesophageal fistula. AJR 1990;154:345–350 12. Suto Y, Tanabe Y. Evaluation of tracheal collapsibility in patients with tracheomalacia using dynamic MR imaging during coughing. AJR 1998;171:393–394 AJR:176, January 2001