Key words: bronchiolitis obliterans; high-resolution CT scan; lung CT scan; lung transplantation; quantitative CT scan

Transcription

1 Dynamic High-Resolution Electron- Beam CT Scanning for the Diagnosis of Bronchiolitis Obliterans Syndrome After Lung Transplantation* Friedrich D. Knollmann, MD, PhD; Susanne Kapell, MD; Hans Lehmkuhl, MD; Bernhard Schulz, MD; Heidi Böttcher, MD; Roland Hetzer, MD, PhD; and Roland Felix, MD, PhD Purpose: To determine the diagnostic capabilities of dynamic high-resolution electron-beam (HREB) CT scanning for diagnosing bronchiolitis obliterans syndrome (BOS) in lung transplant recipients. Materials and methods: At the time of follow-up examinations after lung transplantation, 52 patients were examined by dynamic HREB CT scan. Visual signs of small airway disease were assessed and compared with lung function. For numerical analysis, the mean lung attenuation and its SD were determined and compared with the course of lung function tests. Results: On visual analysis, significant parenchymal attenuation inhomogeneities were present in eight of nine patients with manifest BOS, and in two of four patients who developed BOS during follow-up. Thirteen of 20 patients with persistent normal lung function displayed homogeneous lung attenuation. On numerical analysis, mean lung attenuation was significantly lower in patients who developed BOS during follow-up than in patients with persistent normal lung function (both in expiration and inspiration, p < ). With an optimal threshold, the sensitivity was 100% (4 of 4 patients) and the specificity was 90% (19 of 20 patients). In patients with BOS at the time of the CT scan examination, parenchymal attenuation was less homogeneous than in patients with persistent normal lung function (p < ). With an optimal threshold, the sensitivity was 78% (7 of 9 patients) and the specificity was 85% (17 of 20 patients). Conclusions: Dynamic HREB CT of lung transplant recipients correlates well with lung function criteria of BOS at the time of the CT examination and with the subsequent progression to BOS. (CHEST 2004; 126: ) Key words: bronchiolitis obliterans; high-resolution CT scan; lung CT scan; lung transplantation; quantitative CT scan Abbreviations: ANOVA analysis of variance; BOS bronchiolitis obliterans syndrome; DUHR dynamic ultrafast high-resolution; HREB - high-resolution electron-beam; HU Hounsfield units; ISHLT International Society of Heart and Lung Transplantation; pbos potential bronchiolitis obliterans syndrome; ROC receiver operating characteristic The early diagnosis of bronchiolitis obliterans in lung transplant recipients remains a formidable challenge, because in vivo histology does not reliably establish the diagnosis. 1 The course of lung function *From the Department of Radiology (Drs. Knollmann and Felix), Charité, Campus Virchow-Klinikum, Humboldt-University, Berlin, Germany; and the Department of Cardiothoracic and Vascular Surgery (Drs. Kapell, Lehmkuhl, Schulz, and Böttcher, and Hetzer), German Heart Institute, Berlin, Germany. Manuscript received September 25, 2003; revision accepted January 26, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( permissions@chestnet.org). Correspondence to: Friedrich D. Knollmann, MD, PhD, Klinik für Strahlenheilkunde, Charité, Campus Virchow-Klinikum, Augustenburger Platz 1, Berlin, Germany; Friedrich. Knollmann@charite.de is now widely used as a surrogate indicator, but the early detection of bronchiolitis obliterans has evaded secure recognition so far. 2 This has led to a proposed revision 2 of the lung function criteria that have been published by the International Society for Heart and Lung Transplantation, 3 which includes the use of expiratory CT scanning as a surrogate marker for bronchiolitis obliterans. In one investigation 4 using spirometrically gated high-resolution CT scans of lung transplant recipients, a much lower sensitivity for diagnosing bronchiolitis obliterans syndrome (BOS) by visual analysis was found at 20% of vital capacity than in earlier investigations 5,6 without spirometric gating. One potential method to ensure that the endexpiratory position is depicted is the use of dynamic CHEST / 126 / 2/ AUGUST,

2 high-resolution electron-beam (HREB) CT scanning. 7 The aim of our study was to determine whether dynamic HREB CT scanning has potential for the early diagnosis of BOS. Materials and Methods Table 1 Visual Classification of Air Trapping and Mosaic Attenuation Level Definition 0 Homogeneous lung attenuation 1 Minimal inhomogeneity 2 Attenuation inhomogenities comprise less than one third of lung parenchyma 3 Attenuation inhomogeneities comprise more than one third, but less than two thirds of lung parenchyma 4 Attenuation inhomogeneities comprise at least two thirds of lung parenchyma The study group included 52 lung transplant recipients (24 women; mean [ SD] age, years; age range, 13 to 72 years). Three other patients were excluded from the study because of a decrease of lung function that was ascribed to tracheobronchial collapse based on fiberoptic bronchoscopy, which precluded a classification of lung function tests as the standard of reference for BOS. In these instances, it was deemed improper to use lung function parameters as a reference standard for diagnosing bronchiolitis obliterans. Lung function was abnormal in these instances, too, and one could not be certain whether the abnormal lung function was due to BOS, tracheobronchial collapse, or both. Double-lung transplantation was performed in 32 subjects, heart-lung transplantation was performed in 14 patients, and single-lung transplantation was performed in 6 patients. At the time of study inclusion, a mean duration of months (range, 3 to 97 months) had elapsed since surgery. Patients were included in a prospective, blinded diagnostic trial. The criteria for inclusion were a period of at least 3 months after transplantation and the exclusion of acute rejection or infectious lung disease that was based on clinical evaluations and test results. These inclusion criteria allowed for the classification of BOS status according to the guidelines of the International Society of Heart and Lung Transplantation (ISHLT). According to these guidelines, BOS denotes a diagnosis of chronic transplant rejection without pathologic confirmation, based on a decrease in FEV 1. 3 Institutional review board approval and informed consent were obtained for all procedures. Patients were examined in the supine position with an electron-beam CT scanner (model C-150XP; GE Imatron; South San Francisco, CA) that was equipped with a high-resolution detector system. At each of five image positions, 10 consecutive images were acquired with a scan duration of 100 ms. Scans were equally distributed over a 6-s period. During these 6 s, patients performed a complete forced inspiration and expiration cycle. This maneuver was trained before the actual scan until perfect synchronization with the image acquisition cycle was ascertained. The slice thickness was 1.5 mm, and a high-resolution reconstruction algorithm was used. Images were acquired at the level of the carina, and 2.5 and 5 cm both above and below the carina. This technique also has been termed dynamic ultrafast high-resolution (DUHR) CT scanning. 8 Images were read by a thoracic radiologist on hard copy with a window center of 600 Hounsfield units (HU) and a window width of 1,600 HU. For the HREB CT scan diagnosis of bronchiolitis obliterans, the presence of airtrapping and mosaic attenuation were visually assessed on a five-level scale (Table 1) that was modified from earlier investigations 5,9 for the optimal differentiation of BOS. 5 Airtrapping was defined by areas of subnormally increased attenuation in expiratory CT scan images. 8 Mosaic attenuation was defined by inhomogeneous attenuation in inspiratory CT scan images. 8 The presence of BOS was thought to be suggested if the second degree of either feature was surpassed. In addition, the presence of bronchiectasis also was classified on a five-level scale, since bronchiectasis has been described as a CT scan sign of BOS in some instances in the literature. 10,11 This feature was not used to determine the diagnosis of BOS, however. With this classification of bronchiectasis, grade 0 was defined by a bronchial transsectional area not exceeding the transsectional area of an adjacent artery, and by the absence of bronchial wall thickening, peripheral visibility of the airways, or contour abnormalities. 8 Grade 1 was defined by bronchial transsectional areas that exceeded the adjacent vessel transsectional area by no more than 25%, grade 2 was defined by bronchial transsectional areas that exceeded the adjacent vessel transsectional area by no more than 50%, grade 3 was defined by bronchial transsectional areas that exceeded the adjacent vessel transsectional area by 50% (but in no more than three of the five image positions), and grade 4 was defined by bronchial transsectional areas that exceeded the adjacent vessel transsectional area by 50% in more than three image positions and involved the peripheral third of the transsectional lung area. For numerical analysis, images representing the maximum inspiration and expiration were visually selected for the level of the carina, and for 5 cm above and below the carina, and were electronically transferred to a separate computer workstation (Magicview 1000; Siemens Medical Solutions; Erlangen, Germany). Then, the contours of the lung graft were manually traced, and the lung parenchymal attenuation and its SD were calculated by including pixels with an attenuation between 1,000 and 500 HU only. Within this range, lung attenuation data typically follow a normal distribution. 12 The SD of lung attenuation was used as a numerical indicator of parenchymal attenuation homogeneity. The results were expressed as the mean and SEM. Bullous disease was not encountered in any of the lung transplant patients. The standard of reference was lung function testing. All patients were repeatedly assessed at variable intervals after surgery. Lung function testing used a standard spirometer (Master Screen; Jäger; Würzburg, Germany). For the classification of lung function, the system of the ISHLT and its 2001 update were used. According to this system, BOS is defined by a decrease in FEV 1 of 20% compared to the postoperative maximum value. With the 2001 update, a potential BOS (pbos) stage was introduced that comprises a decrease of FEV 1 by 10 to 19% of the postoperative maximum and/or a 25% decrease in the midexpiratory forced expiratory flow rate. The course of lung function in our patients was classified as either normal lung function throughout follow-up, pbos that occurred during follow-up, pbos already present at the time of the CT scan examination, progression to BOS during follow-up, and BOS already present at the time of the CT scan examination. For the classification of lung function at the time of the CT scan examination, lung function testing was performed on the same day. Lung function then was followed up until March The date of the last follow-up examination was noted, with a mean follow-up period of days (range, 134 to 1,592 days). One patient was excluded because no follow-up could be obtained (56 patients underwent the CT scan examination). 448 Clinical Investigations

3 For statistical analysis, the visual signs of BOS were compared with the course of lung function using 2 statistics. The visual signs of BOS then were compared with the numeric characteristics of lung attenuation using factorial analysis of variance (ANOVA). Pairwise comparisons employed the Scheffé correction. To assess the diagnostic value of attenuation measurements, the mean parenchymal attenuation and its SD were compared with the course of lung function using factorial ANOVA. The level of statistical significance was 5%. For an assessment of the diagnostic qualities of lung attenuation measurements for detecting and predicting BOS, an analysis of the receiver operating characteristics (ROCs) was performed using a computer program (LABROC, for MacIntosh; Charles E. Metz, MD; University of Chicago; Chicago, IL; 1991). With this software, the relationship between the true-positive fraction (ie, sensitivity) and the falsepositive fraction (ie, specificity) were computed, and critical test result values and their corresponding operating points on the fitted binormal ROC curve were estimated. Results On visual evaluation, inspiratory abnormalities were found in 20 of 52 patients, with 18 instances of class 1 mosaic attenuation, 1 instance of class 2 mosaic attenuation, and 1 instance of class 4 mosaic attenuation (Fig 1 3). Expiratory parenchymal inhomogeneity was detected in 49 patients, with 26 instances of class 1 airtrapping, 10 instances of class 2 airtrapping, 7 instances of class 3 airtrapping, and 6 instances of class 4 airtrapping. Thus, 23 patients displayed visual CT scan signs that were suggestive of BOS. Inspiratory images did not contribute any suggestions of BOS beyond those detected in expiratory images. Bronchiectasis was found in 18 patients, with 12 instances of class 1 bronchiectasis, 6 instances of class 2 bronchiectasis, and no instances of class 3 or 4 bronchiectasis. On numerical analysis of lung attenuation, the mean attenuation was HU (range, 957 to 614 HU), and the attenuation SD for each image averaged HU (range, 52 to 166 HU). Both anatomic slice position and level of inspiration heavily influenced parenchymal attenuation (p [ANOVA]). On pairwise comparison, attenuation at the most caudal slice position (ie, HU) exceeded that at the carina ( HU; p 0.04) and that at the most apical position ( HU; p ). Expectedly, expiratory attenuation ( HU) exceeded inspiratory attenuation ( HU; p ). Comparing the visual severity of small airway disease with numerical lung attenuation, it was found that a mosaic attenuation pattern correlated with an increase of inspiratory mean attenuation (p [ANOVA]) [Table 2], but not with inspiratory attenuation homogeneity (p 0.95). Airtrapping was associated with increased expiratory lung attenuation (p [ANOVA]) [Table 2], and decreased Figure 1. HREB CT scan images of a 39-year-old double-lung transplant recipient 5 months after undergoing surgery 5 cm above the carina. During 537 days of follow-up, lung function remained stable. Top: inspiratory image with homogeneous parenchymal attenuation. Bottom: expiratory image displaying minimal inhomogeneity of lung attenuation (grade 1 expiratory inhomogeneity). expiratory homogeneity (p [ANOVA]) [Table 2]. There was no correlation of the severity of bronchiectasis with mean lung attenuation (p 0.11) [Table 2]. The presence of bronchiectasis correlated with decreased lung homogeneity (p ) [Table 2]. During follow-up, 20 patients had persistently normal lung function, 19 patients qualified as pbos CHEST / 126 / 2/ AUGUST,

4 Figure 2. HREB CT scan images of a 62-year-old double-lung transplant recipient 55 months after undergoing surgery at the level of the carina. At the time of the CT scan examination, lung function classified the patient as a BOS case. Top: inspiratory image with homogeneous parenchymal attenuation. Bottom: expiratory image displaying marked inhomogeneity of lung attenuation (grade 4 expiratory inhomogeneity). Figure 3. HREB CT scan images of a 36-year-old heart-lung transplant recipient 54 months after undergoing surgery at the level of the carina. The patient acquired BOS status 362 days after the CT scan examination. Top: inspiratory image with homogeneous attenuation of the lung parenchyma. Bottom: expiratory image displaying grade 3 inhomogeneity of lung attenuation. as per the revised ISHLT classification, 4 patients progressed to BOS, and 9 patients qualified as having BOS at the time of the CT scan examination. The length of the follow-up period was not associated with the course of lung function during that time (p 0.7 [ANOVA]). On correlation of the visual signs of small airway disease with the clinical course during follow-up, only advanced degrees of expiratory inhomogeneity (ie, greater than class 2) were significantly associated with BOS (p 0.02) [Table 3]. Using this criterion, eight of nine patients with BOS at the time of the examination were detected, and two of four patients who progressed to BOS during follow-up were detected. The specificity was 13 of 20 patients (65%) with persistent normal lung function as normal. 450 Clinical Investigations

5 Table 2 Correlation of Visual CT Scan Signs of Small Airway Disease With Measurements of Lung Attenuation* Mean Parenchymal Attenuation, HU Attenuation Homogeneity, HU Variables 5 cm Above Carina At Carina 5 cm Below Carina 5 cm Above Carina At Carina 5 cm Below Carina Mosaic attenuation Air trapping Bronchiectasis *Values given as mean SD. Statistically significant difference as compared to class 0. On correlation of lung attenuation with the clinical course during follow-up, it was found that patients who progressed to BOS during follow-up displayed a reduced attenuation when compared to all other courses of lung function (p [ANOVA]) [Fig 4], and that patients with manifest BOS had a higher attenuation than all other courses (p vs p , respectively) [Fig 4]. For the detection of patients who would progress to BOS, differences were most marked in inspiration 5 cm above the carina. The lung homogeneity of patients who either progressed to BOS or had manifest BOS at the time of the CT examination was significantly lower than that in patients who either remained stable or acquired pbos status only (p 0.04 vs p , respectively), while patients with initial BOS had no different attenuation homogeneity than did patients who progressed to BOS during follow-up (p 0.2) [Fig 5]. The best differentiation of patients with BOS from patients without BOS was achieved in expiration at the most caudal slice position (Fig 5). Analysis of the ROC values for predicting progression to BOS during follow-up from mean lung attenuation 5 cm above the carina at inspiration displayed a mean area under the ROC curve of Az (Fig 6). The sum of sensitivity and specificity was greatest with a diagnostic threshold of 880 HU, which resulted in a sensitivity of 100% (4 of 4 patients) and a specificity of 90% (1 of 20 patients). Table 3 Correlation of Visual Signs of Small Airway Disease With the Course of Lung Function Variables Degree Persistent Normal Lung Function Potential BOS During Follow-up New-onset BOS During Follow-up BOS at the Time of the CT Scan Mosaic attenuation Airtrapping Bronchiectasis CHEST / 126 / 2/ AUGUST,

6 Figure 4. Mean lung attenuation (in Hounsfield units) as a function of anatomic slice position (levels 1 to 5) and inspiration for patients with persistent normal lung function (0), new-onset pbos during follow-up (1), new-onset BOS (3), and BOS at the time of the CT scan examination (4). Level 1, 5 cm above the carina; level 3, at the carina; level 5, 5 cm below the carina. Patients who progressed to BOS during the follow-up period could be discerned by decreased parenchymal attenuation in all instances. in full inspiration; ex full expiration. For the diagnosis of BOS at the time of the CT scan examination from lung homogeneity measurements 5 cm below the carina during full expiration, the mean area under the fitted ROC curve was determined as The sum of sensitivity and specificity for diagnosing BOS was greatest at an attenuation SD of HU, which resulted in a sensitivity of 78% (7 of 9 patients) and a specificity of 85% (3 of 20 patients). Discussion The findings of this investigation indicate that dynamic HREB CT scanning of lung transplant recipients can be used to both diagnose and predict BOS. Interestingly, visual analysis of dynamic HREB CT scan images allowed for the diagnosis of BOS only, which significantly correlated with the numerically determined inhomogeneity of lung attenuation, while numerically low lung attenuation predicted the progression to BOS. Visually, we have no evidence that the progression to BOS can be predicted with similar accuracy, which makes numerical analysis of parenchymal attenuation an important part of the test. Using numerical analysis of lung attenuation, the confounding effect of display window settings in visual analysis 13 can be eliminated. Our findings confirm those of an earlier report 4 on spirometrically gated high-resolution CT scanning, in which the progression to BOS was predicted by pulmonary hyperinflation. In this earlier investigation, however, BOS at the time of the CT scan could not be diagnosed reliably, and the visual signs of small airway disease did not correlate well with lung 452 Clinical Investigations

7 Figure 5. Homogeneity of lung density (in Hounsfield units) as a function of anatomic slice position (levels 1 to 5) and inspiratory level (percent vital capacity) for patients with persistent normal lung function (0), progression to pbos during follow-up (1), new-onset BOS during follow-up (3), and BOS at the time of the CT scan (4). Level 1, 5 cm above the carina; level 3, at the carina; level 5, 5 cm below the carina. Patients with manifest BOS had less homogeneous lung attenuation than those with persistent normal lung function. See the legend of Figure 4 for abbreviations not used in the text. homogeneity measurements. This apparent discrepancy is explicable by differences in the examination technique. The earlier report used an expiratory level of 20% of vital capacity only, which may not suffice for the detection of airtrapping. Using dynamic HREB CT scanning, the full end-expiratory position is depicted, while the spirometrically gated method loses its reproducibility near full expiration. 14,15 The most accurate CT scan signs of manifest bronchiolitis obliterans can be detected near full expiration only, 9 which explains why the nongated dynamic HREB CT scan technique achieved diagnostic powers beyond those of the spirometrically gated conventional CT scan approach. A combination of spirometric gating with dynamic HREB CT scanning is feasible, however. 16 One benefit of using HREB CT scanning is that it effectively eliminates the effect of cardiac motion on lung attenuation measurements, even without ECG-gated data acquisition. The image acquisition time of 100 ms is also essential for eliminating breathing artifacts in a dynamic scan acquisition. Conventional CT scan techniques typically have longer image acquisition times, which require a breath hold. In patients with lung disease, however, the individual s ability to sustain a breath-hold maneuver is notoriously variable, and even a breath-hold period of only a few seconds may be unattainable just at the diagnostically most important end-expiratory position. Although additional research on the reproducibility of lung attenuation measurements with the dynamic HREB CT scanning method is necessary, our results indicate that the method has sufficient test quality to be appear to be of greater clinical value CHEST / 126 / 2/ AUGUST,

8 Figure 6. ROC values for predicting progression to BOS during follow-up from DUHR CT scan measurements of mean lung density (true-positive fraction [TPF] or sensitivity), and for diagnosing manifest BOS at the time of the CT scan examination from lung homogeneity (TPF2). FPF falsepositive fraction (specificity). The diagnostic accuracy of predicting future BOS from mean lung attenuation was superior to that of diagnosing manifest BOS from lung homogeneity measurements. than the spirometrically gated conventional CT scan technique. 4 To improve the reproducibility of the test, we employed a modification of the original DUHR CT scan analysis 7,17 by using semiautomatic analysis of lung attenuation. 18 Although no data from normal individuals measured with this method have been published so far, a comparison with earlier data on healthy subjects using a spirometrically gated conventional CT scan technique revealed 14 no apparent discrepancies, a finding that needs to be interpreted with reservations concerning the differences in the techniques. With a spirometrically gated conventional CT scan technique, it has been demonstrated 19 that patient age may play a role in lung attenuation in healthy subjects. For the task at hand, however, a comparison with healthy subjects was not required, because it is usually known that a patient is a lung transplant recipient, and the diagnostic task is to diagnose or predict the progression to BOS. Thus, a reference to a normal group of nontransplant subjects may not add any useful diagnostic intelligence, and the role that lung transplantation in general has on lung attenuation is of little clinical import. Evidence for the greater value of the DUHR CT scan technique consist of the better correlation with lung function at the time of the CT scan examination and the greater ability to predict imminent BOS, as per the area under the respective ROC curve. In comparison to ungated conventional high-resolution CT scanning at full expiration, we found test qualities for the visual diagnosis of manifest BOS that were similar to those in earlier reports, 5,6,9,20 which offers additional evidence for our suspicion that spirometric gating at 20% of vital capacity may not suffice to reliably detect airtrapping. Whether ungated conventional high-resolution CT scan studies also might predict imminent BOS remains a matter of further research, since we are unaware of any such quantitative analysis in the literature. Still, the number of patients who progressed to BOS during the follow-up period was relatively small, despite the fact that the size of our entire patient group was larger than that reported in most other publications on this subject. 5,6,9,21 23 Considering both patients who presented with preexisting BOS and those who would progress to BOS after the CT scan examination as one group, however, both subgroups shared the feature of highly inhomogeneous lung attenuation on expiratory images. The different imaging features of patients with preexisting BOS and those who progress to BOS later suggest a pathophysiologically interesting course of lung function in lung transplant recipients. Apparently, 454 Clinical Investigations

9 patients have normal lung attenuation initially, then develop hyperinflation as an early sign of imminent BOS, which can be detected by dynamic HREB CT scan measurements, and then progress to clinically overt BOS, which is paralleled by the appearance of lung attenuation inhomogeneities that can be detected on dynamic HREB CT scan as airtrapping on visual analysis and decreased lung homogeneity on numerical analysis. This sequence confirms earlier observations 4 that the occurrence of airtrapping may represent an already advanced stage of chronic graft rejection in lung transplant recipients, and that the extent of airtrapping increases with the progression of BOS severity. 5 Of note, mean parenchymal attenuation was lower in patients with imminent BOS than in patients with manifest BOS at the time of the CT scan examination, implying that the progression to inhomogeneous lung attenuation may parallel the deterioration of lung function. This implication is in accordance with earlier concepts on the CT scan signs of obliterative bronchitis. 24 The potential clinical benefits of our findings are that the diagnosis of BOS may be improved using the DUHR CT scan method, and that patients with imminent BOS may be identified at an earlier time when intensified immunosuppressive therapy may still improve the course of lung function. A more secure diagnosis of BOS would be particularly helpful in patients with a history of normal lung function who are presenting with new signs of lung function impairment. Also, the reliable diagnosis of chronic rejection in single-lung transplant recipients remains problematic with lung function data alone, because FEV 1 cannot be determined for each lung separately. The newly established pbos stage in the classification of the ISHLT represents a dilemma since relatively small changes in lung function may well revert to normal, and additional diagnostic intelligence may greatly aid clinical decisions. In our cohort, 19 of 52 patients fell into the pbos stage and remained in this group throughout the follow-up period. Based on DUHR CT scan findings, these patients were indistinguishable from those with persistent normal lung function. The discriminatory powers of DUHR CT scanning could thus be used to stratify the risk for BOS in lung transplant recipients. The clinical benefit of such stratification still needs to be confirmed in prospectively conducted investigations of intensified immunosuppressive treatment, and our diagnostic study may serve as a starting point for designing such research. The ISHLT committee on BOS has identified studies on surrogate markers of BOS to predict future decreases in lung function as a research priority, 2 and our DUHR CT scan data may provide important new insights into the time-course of lung disease following transplantation. We conclude that dynamic HREB CT scanning of lung transplant recipients may allow for an earlier diagnosis of BOS. Pulmonary hyperinflation indicates imminent BOS, while inhomogeneous attenuation is a common feature of both imminent and manifest BOS. The clinical benefit of an HREB CT scan diagnosis of early BOS warrants further investigation in a prospective, randomized, and blinded investigation. References 1 Arcasoy SM, Kotloff RM. Lung transplantation. N Engl J Med 1999; 340: Estenne M, Maurer JR, Boehler A, et al. Bronchiolitis obliterans syndrome 2001: an update of the diagnostic criteria. J Heart Lung Transplant 2002; 21: Cooper JD, Billingham M, Egan T, et al. A Working formulation for the standardization of nomenclature and for clinical staging of chronic dysfunction in lung allografts. J Heart Lung Transplant 1993; 12: Knollmann FD, Ewert R, Wündrich T, et al. Bronchiolitis obliterans syndrome in lung transplant recipients: use of spirometrically gated CT. Radiology 2002; 225: Bankier AA, Van Muylem A, Knoop C, et al. Bronchiolitis obliterans in heart-lung transplant recipients: diagnosis with expiratory CT. Radiology 2001; 218: Lee E-S, Gotway MB, Reddy GP, et al. Early bronchiolitis obliterans following lung transplantation: accuracy of expiratory thin-section CT for diagnosis. Radiology 2000; 216: Stern EJ, Webb WR, Gamsu G. Dynamic quantitative computed tomography: a predictor of pulmonary function in obstructive lung diseases. Invest Radiol 1994; 29: Webb WR, Müller NK, Naidich P. High-resolution CT of the lung. 2nd ed. Philadelphia, PA: Lippincott-Raven, 1996; , Leung AN, Fisher K, Valentine V, et al. Bronchiolitis obliterans after lung transplantation. Chest 1998; 113: Lentz D, Bergin CJ, Berry GJ, et al. Diagnosis of bronchiolitis obliterans in heart-lung transplantation patients: importance of bronchial dilatation on CT. AJR Am J Roentgenol 1992; 159: Loubeyre P, Revel D, Delignette A, et al. Bronchiectasis detected with thin-section CT as a predictor of chronic lung allograft rejection. Radiology 1995; 194: Rienmueller RK, Behr J, Kalender WA, et al. Standardized quantitative high resolution CT in lung diseases. J Comput Assist Tomogr 1991; 15: Hansell DM. Small-vessel diseases of the lung: CT-pathologic correlates. Radiology 2002; 225: Beinert T, Behr J, Mehnert F, et al. Spirometrically controlled quantitative CT for assessing diffuse parenchymal lung disease. J Comput Assist Tomogr 1995; 19: Lamers RJS, Kemerink GJ, Drent M, et al. Reproducibility of spirometrically controlled CT lung densitometry in a clinical setting. Eur Respir J 1998; 11: Robinson TE, Leung AN, Moss RB, et al. Standardized high-resolution CT of the lung using a spirometer-triggered electron beam CT scanner. AJR Am J Roentgenol 1999; 172: Webb WR, Stern EJ, Kanth N, et al. Dynamic pulmonary CT: findings in healthy adult men. Radiology 1993; 186: CHEST / 126 / 2/ AUGUST,

10 18 Kalender WA, Fichte H, Bautz W, et al. Semiautomatic evaluation procedures for quantitative CT of the lung. J Comput Assist Tomogr 1991; 15: Smit HJM, Golding RP, Schramel FMNH, et al. Lung attenuation measurements in healthy young adults. Respiration 2003; 70: Siegel MJ, Bhalla S, Gutierrez FR, et al. Post-lung transplantation bronchiolitis obliterans syndrome: usefulness of expiratory thin-section CT for diagnosis. Radiology 2001; 220: Worthy SA, Park CS, Kim JS, et al. Bronchiolitis obliterans after lung trans-plantation: high resolution CT indings in 15 patients. AJR Am J Roentgenol 1997; 169: Morrish WF, Herman SJ, Weisbrod GL, et al. The Toronto Lung Transplant Group: bronchioloitis obliterans after lung transplantation; findings at chest radiography and highresolution CT. Radiology 1991; 179: Ikonen T, Kivisaari L, Taskinen E, et al. High-resolution CT in long-term follow-up after lung transplantation. Chest 1997; 111: Hansell DM, Rubens MR, Padley SPG, et al. Oblitertive bronchiolitis: individual signs of small airway disease and functional correlation. Radiology 1997; 203: Clinical Investigations