| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Departments of Radiology (M.B.G.), Thoracic Imaging Section (M.B.G., S.K.D., D.S., G.P.R., W.R.W.), Pulmonary and Critical Care Medicine (J.A.G.), and Surgery (F.M.K.), San Francisco General Hospital, University of California, San Francisco, 1001 Potrero Ave, Rm 1X 55A, Box 1325, San Francisco, CA 94110. Received January 30, 2001; revision requested March 6; revision received April 3; accepted May 2. Address correspondence to M.B.G. (e-mail: michael.gotway@radiology.ucsf.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thin-section CT studies from 64 lung transplant recipients were retrospectively reviewed. CT studies were temporally correlated with various grades of biopsy-proved acute rejection (n = 34); 30 other CT studies were from a control group with no histopathologic evidence of acute rejection. Acute rejection was diagnosed as present or absent, and the diagnostic was calculated.
RESULTS: The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of CT for the diagnosis of acute rejection were as follows: 35%, 73%, 60%, 50%, 53%, respectively. No individual CT finding was significantly associated with acute rejection. The sensitivity of CT for the detection of various grades of acute rejection was 17% for grade A1, 50% for grade A2, and 20% for grade A3. The combination of volume loss and septal thickening, with or without pleural effusion, was never seen in the absence of acute rejection.
CONCLUSION: Thin-section CT has limited accuracy for the diagnosis of acute rejection following lung transplantation, and no individual CT finding is significantly associated with this diagnosis.
Index terms: Computed tomography (CT), thin-section, 60.12118 • Lung, transplantation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Histopathologic Sampling
Surveillance (asymptomatic patients) or diagnostic (symptomatic patients) bronchoscopy was performed in all patients in the study group, typically within 1 or 2 days following the thin-section CT examination. The lung transplant surveillance regimen consisted of postsurgical follow-up with clinical evaluation, pulmonary function testing, bronchoscopy, and thin-section CT studies performed three times at 2-week intervals, three times at 1-month intervals, four times at 3-month intervals, and every 6 months thereafter. Fiberoptic bronchoscopy was performed in all patients and included both bronchoalveolar lavage (BAL) and transbronchial biopsy (TBB) procedures. Thin-section CT studies were reviewed with the bronchoscopist before the procedure, and both BAL and TBB procedures were directed at abnormal areas on the CT scan when present. A minimum of six TBB procedures were performed in each of at least two segments in all patients. BAL specimens were analyzed for typical and atypical infections, including bacteria, acid-fast bacilli, Pneumocystis carinii, and viruses. TBB specimens were examined with various histopathologic methods for the presence of acute and chronic rejection, preservation injury, reimplantation response, infections, and malignancy. The histopathologic diagnosis of acute rejection depended on the demonstration of perivascular and interstitial mononuclear cell infiltrates (16). The presence and degree of acute rejection was scored according to the criteria set forth by the Lung Rejection Study Group (16): grade A0, no rejection; grade A1, minimal rejection; grade A2, mild rejection; grade A3, moderate rejection; grade A4, severe rejection.
Only a very small proportion of the total number of the TBB specimens in this population were read as histopathologically indeterminate for acute rejection. Thirty-four patients with acute rejection were included in the analysis; the remainder consisted of patients with grade A0 biopsy results (no histopathologic evidence of acute rejection). Among patients with histopathologic evidence of acute rejection, 11 had grade A1 acute rejection, 16 had grade A2, and seven had grade A3. The mean time from surgery to the TBB procedure that temporally corresponded to the thin-section CT scan reviewed for the study was 57.2 days (range, 21–180 days).
Thin-Section CT Techniques
All patients underwent thin-section CT studies as part of a postoperative surveillance regimen. All studies were performed at full inspiration with 1-mm collimation at 2-cm intervals (scanning was performed with model 9800 Quick and HiSpeed Advantage, or CT/i scanners, or at 1.25-mm collimation with LightSpeed QX/i helical CT scanners; GE Medical Systems, Milwaukee, Wis) in the supine and prone positions. Three postexpiratory views were routinely obtained at the level of the aortic arch, the carina, and just above the diaphragm. Intravenous contrast material was not administered. A high-spatial-frequency reconstruction algorithm was employed, and scans were reviewed in soft-tissue (level, 40 HU; width, 440 HU) and lung (level, -700 HU; width, 1,000–1,500 HU) windows.
Three thoracic radiologists (S.K.D., G.P.R., and W.R.W.) reviewed a single thin-section CT study from each patient. The thin-section CT scan selected for the study was identified by one thoracic radiologist (M.B.G.) not participating in the retrospective CT scan review. The radiologists reviewing the scans were aware only of the context of the study and the sex of the patient. Scans were evaluated for the presence of findings that have been associated with acute rejection in prior work: ground-glass opacity, septal thickening, pleural effusions, volume loss, and consolidation. Other patterns of abnormality readily apparent with thin-section CT, including peribronchovascular thickening, bronchial wall thickening, nodules and masses, reticulation, fissure thickening, bronchiectasis, mosaic perfusion, air trapping, and adenopathy, were also evaluated. The presence and degree of air trapping was scored with use of a previously published method (2,17). Briefly, the pulmonary parenchyma was visually evaluated at three levels for failure of the normal increase in lung attenuation expected with postexpiratory imaging, or a frank decrease in lung attenuation on postexpiratory scans. Air trapping was scored on a five-point scale at each level as follows: 0, no air trapping visible; 1, 1%–25% of the cross-sectional area of the lung affected; 2, 26%–50% affected; 3, 51%–75% affected; and 4, more than 75% affected. The total score was calculated (maximum, 24); the score for single-lung transplant patients was multiplied by two. A score of more than 3 was considered abnormal. The remaining findings were scored as present or absent, and the degree and distribution of the abnormalities were noted.
On the basis of available descriptions of acute rejection on thin-section CT scans (1,8–15,18), reviewers scored acute rejection as either present or absent. Reviewers scored the CT scans together, and diagnoses were reached by consensus. Because there are little data in the literature regarding the appearance of acute rejection at thin-section CT in patients with transplants, a grading system for the diagnosis of acute rejection could not be employed. Rather, reviewers diagnosed the presence or absence of acute rejection on the basis of their subjective assessments of the CT scan findings.
CT Scan Review
The CT scans chosen for review were generally obtained 1–6 months following transplantation. We avoided reviewing scans obtained earlier than 1 month postoperatively to minimize the possibility of reimplantation response confounding the CT scan analysis. Because acute rejection is most prevalent within the first 180 days following surgery (4,18), scans beyond 6 months following surgery were not reviewed. The CT scan chosen for review was temporally related to successful BAL and TBB procedures; all reviewed CT scans were obtained no more than 2 days before bronchoscopy. CT scans that temporally corresponded to differing grades of acute rejection were reviewed. A number of patients in whom acute rejection was excluded by means of TBB were included in the analysis as a control group (n = 30). TBB procedures that yielded insufficient material to diagnose or exclude acute rejection were excluded. In addition, TBB or BAL procedures with histopathologic evidence of preservation injury, reimplantation response, alveolitis not otherwise specified, hemorrhage, or infections were excluded from analysis. Patients were never treated empirically with corticosteroids or other immunosuppressives for presumed acute rejection without first undergoing TBB procedures. The radiologists reviewing the CT scans were not aware of this exclusion criterion.
Statistical Methods
We analyzed the presence or absence of each of the above thin-section CT scan findings alone and in combination for association with the presence or absence of acute rejection by using the Fisher exact test. A P value of .05 or less was considered to indicate a statistically significant association. The sensitivity, specificity, positive and negative predictive values, and accuracy of thin-section CT for the diagnosis of acute rejection were calculated. The strength of association of the diagnosis of acute rejection at thin-section CT with the various grades of acute rejection was also noted.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The strengths of association of thin-section CT scan findings, previously considered suggestive of acute rejection, with acute rejection both alone and in combination, are given in the Table. No finding, either alone or in combination, was significantly associated with acute rejection. The combination of volume loss and septal thickening, with or without pleural effusion, was not seen in the absence of biopsy-proved acute rejection (Fig 1). Five cases in which the CT scans were read as showing no findings were positive for acute rejection at biopsy (Figs 2, 3). False-positive CT examinations, with no histopathologic correlate to the observed CT scan findings, also were noted (Fig 4).
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
There are few data regarding the utility of thin-section CT for the early detection of acute rejection following lung transplantation. Human studies detailing the accuracy of imaging techniques for the detection of acute rejection following lung transplantation either have been primarily concerned with chest radiographic findings of acute rejection (8,10,11) or have included as many, if not more, heart-lung transplant recipients as lung-only transplant recipients (8,12,15) in the study populations. Radiographic studies have described the typical appearance of acute rejection in lung transplant recipients as fine reticulation and consolidation, sometimes with basilar or perihilar predominance (10,11). In a series of 16 heart-lung transplant recipients, Bergin and colleagues (8) found that the combination of septal thickening and pleural effusions was 68% (17 of 25) sensitive and 90% (52 of 58) specific for the diagnosis of acute rejection. These investigators differentiated acute rejection from volume overload in their patient population by noting the absence of an increase in cardiac size, vascular pedicle width, lack of vascular redistribution, and stable or decreasing patient weight in patients with septal thickening and pleural effusions and biopsy-proved acute rejection. However, the chest radiograph (particularly portable radiography) is not necessarily a very sensitive indicator of a patient’s volume status, and occasionally cardiac dysfunction (eg, diastolic dysfunction) may not present with an easily perceptible change in cardiac size.
Ikonen et al (13) studied this issue in an experimental porcine model that controlled for superimposed infection and reimplantation response. These authors found an overall CT sensitivity and specificity for acute rejection of 86.7% (52 of 60) and 85.6% (83 of 97), respectively. In this study, CT findings associated with acute rejection included ill-defined micronodules, patchy ground-glass opacity, bronchial wall thickening, and reduced volume of the graft (13). These authors also found that CT sensitivity and specificity improved with worsening acute rejection histopathologic grade. Another investigation primarily focusing on acute rejection in a porcine lung transplantation model (9) found that early, severe, untreated acute rejection was characterized by peripheral alveolar opacities. In this same study, in immunosuppressed animals acute rejection was detected as a densitometrically measured increase in attenuation over the whole graft.
Loubeyre et al (15) reported a large study of the utility of thin-section CT for the diagnosis of acute rejection in lung transplant recipients. Their patient population consisted of 17 heart-lung, one double-lung, and 14 single-lung transplant patients. These investigators found that ground-glass opacity on thin-section CT scans had a sensitivity of 65% for the detection of acute rejection, and that CT performed more reliably after the 1st month following surgery. Similar to the porcine model of Ikonen et al (13), Loubeyre et al (15) noted a tendency toward more extensive CT findings with worsening histopathologic grade of acute rejection.
Using the conclusions of these studies as criteria for the diagnosis of acute rejection, we retrospectively reviewed the accuracy of thin-section CT for the diagnosis of acute rejection following lung transplantation. To avoid the confounding influence that heart transplantation may have on any observed pulmonary findings, heart-lung transplant recipients were purposefully excluded. The investigators were blinded to the fact that patients with clinical evidence of volume overload or histopathologic evidence of infection were excluded, thus the possibility that any observed CT findings were related to acute rejection was maximized. Nevertheless, using the primary findings of ground-glass opacity, septal thickening, pleural effusion, and volume loss within the graft as indicators of acute rejection, we achieved only 35% sensitivity and 73% specificity for the detection of acute rejection. None of the individual findings observable with thin-section CT, alone or in combination, were significantly associated with the presence of acute rejection at biopsy. Furthermore, in five patients, the thin-section CT study was read as normal despite biopsy-proved histopathologic evidence of acute rejection. In addition, thin-section CT findings considered suggestive of acute rejection were observed frequently in patients with no histopathologic evidence of acute rejection at TBB. The most useful observation resulting from our data was that the combination of volume loss in the graft and septal thickening (whether or not pleural effusion is considered) were never found in patients without histopathologic evidence of acute rejection.
In contrast to the studies of Ikonen et al (13) and Loubeyre et al (15), we did not find that higher histopathologic grades of acute rejection were associated with more extensive CT abnormalities. Why our conclusions differ somewhat from previous studies is unclear. One difference between our study and the study of Loubeyre et al (15) is that our study population consisted of both surveillance and symptomatic patients; the study by Loubeyre et al consisted of symptomatic patients. Perhaps symptomatic patients are more likely to have disease that is visible with CT. Another difference between the two studies is that our study included each patient only once. This was done purposefully because including the same patient at different time points would require a far more complex analysis to account for the possibility that one episode of acute rejection may influence the likelihood and course of subsequent episodes. Another difference between the two studies is that we made every attempt to review the thin-section CT scan that corresponded to the first biopsy-proved episode of acute rejection. Again, this was done purposefully to minimize any bias that could result from the possibility that repeated episodes of acute rejection may be more easily detected than isolated or index episodes of acute rejection, a situation that could falsely elevate the sensitivity and positive predictive value of thin-section CT.
The fact that our study casts thin-section CT in a less favorable light than do previous experimental models (9,13) is not surprising. Experimental studies allow for more frequent monitoring with purposeful interruptions in immunosuppression, as has been observed by Ikonen et al (13), and would likely result in a more favorable impression of the diagnostic accuracy of thin-section CT.
Our study has several limitations. The most important limitation is the size of the population studied. It is quite possible that some findings observable at thin-section CT may predict acute rejection, but our sample size was simply too small to enable detection of the association. As indicated above, including more episodes of acute rejection in the same patients at different time points to increase this sample size would not have been a valid analytical method because the variables in our analysis must be independent of one another.
Another limitation is that our histopathologic criterion standard, TBB, is an imperfect one. Sampling error could easily miss the diagnosis of acute rejection when the latter is actually present. Furthermore, there is no guarantee that any observed thin-section CT finding actually represents histopathologically proved acute rejection. For example, although Hruban et al (12) have demonstrated the thin-section CT correlate (both in life and in inflation-fixed lungs) of the perivascular inflammatory cell infiltrate that characterizes acute rejection, these authors observed that the thin-section CT findings of acute rejection are surprisingly modest despite the presence of extensive histopathologic abnormalities.
Also limiting the study is the lack of independent interpretation of the CT scans, rendering it impossible to assess interobserver agreement among the reviewers.
A final limitation of our study is the relatively few cases of higher rejection grades (grades A3 and A4) included in the analysis. It is possible that, with larger numbers of higher grades of acute rejection, we may have eventually observed an overall increase in the sensitivity of thin-section CT as well as a relatively higher sensitivity for the detection of higher rejection grades, as other investigators have (13,15). However, it is likely that, owing to the frequent and intense surveillance programs that are typically employed for the evaluation of lung transplant recipients, lower grades of acute rejection may be more prevalent than higher grades, and including relatively more episodes of high-grade acute rejection may result in an inaccurate impression of the diagnostic accuracy of thin-section CT in this setting.
It should be noted that, although thin-section CT has limited accuracy for the diagnosis of acute rejection, CT remains a useful technique for the evaluation of the lung transplant recipient. CT, in particular thin-section CT, is highly useful for the detection and characterization of numerous operative and postoperative complications of lung transplantation, including bronchial anastomotic complications, pulmonary infections, malignancy, and chronic rejection (2,5,18,20). Thin-section CT is also very useful for directing tissue-sampling procedures. Finally, CT’s superiority to chest radiography for the evaluation of the lung transplant recipient has been well documented (9,10,13,15,18).
In conclusion, thin-section CT demonstrates limited sensitivity, specificity, positive and negative predictive values, and accuracy for the detection of acute rejection shortly following lung transplantation. None of the various findings visible with thin-section CT, including ground-glass opacity, pleural effusions, septal thickening, and volume loss in the graft, are significantly associated with histopathologic proof of acute rejection at TBB. In addition, the performance of thin-section CT did not improve with higher histopathologic grades of acute rejection. Because thin-section CT remains a useful technique for the evaluation of lung transplant recipients and will undoubtedly continue to be used in this setting, awareness of this limitation is important.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, M.B.G.; study concepts, D.S., M.B.G., W.R.W., J.A.G., S.K.D., G.P.R.; study design, M.B.G., W.R.W., D.S.; literature research, M.B.G.; clinical studies, F.M.K., M.B.G., J.A.G.; data acquisition, M.B.G., W.R.W., D.S., J.A.G., S.K.D., G.P.R.; data analysis/interpretation, M.B.G.; statistical analysis, M.B.G.; manuscript preparation, M.B.G., S.K.D.; manuscript definition of intellectual content, M.B.G., W.R.W.; manuscript editing, M.B.G., W.R.W., S.K.D., J.A.G.; manuscript revision/review, S.K.D., M.B.G., W.R.W.; manuscript final version approval, M.B.G., W.R.W., S.K.D.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. S. Krishnam, R. D. Suh, A. Tomasian, J. G. Goldin, C. Lai, K. Brown, P. Batra, and D. R. Aberle Postoperative Complications of Lung Transplantation: Radiologic Findings along a Time Continuum RadioGraphics, July 1, 2007; 27(4): 957 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Marom, S. M. Palmer, J. J. Erasmus, J. E. Herndon, C. Zhang, and H. P. McAdams Pleural Effusions in Lung Transplant Recipients: Image-guided Small-Bore Catheter Drainage Radiology, July 1, 2003; 228(1): 241 - 245. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
