Abstract. Pneumothorax is reported to be a more common complication of lung biopsy performed under computed tomography (CT) than under fluoroscopic ...
The British Journal of Radiology, 72 (1999), 1160±1163
E
1999 The British Institute of Radiology
Pneumothorax post CT-guided lung biopsy: a comparison between detection on chest radiographs and CT 1
H K BUNGAY, MRCP, FRCR, 2J BERGER, MRCP, FRCR, 1Z C TRAILL, MRCP, FRCR and 1 F V GLEESON, FRCP, FRCR 1
Department of Radiology, Churchill Hospital, Old Road, Headington, Oxford OX3 7LJ and 2Department of Radiology, Barnet General Hospital, Wellhouse Lane, Barnet EN5 3DJ, UK
Abstract. Pneumothorax is reported to be a more common complication of lung biopsy performed under computed tomography (CT) than under ¯uoroscopic guidance. This may simply re¯ect the greater sensitivity of CT over chest radiographs (CXRs) in the detection of small pneumothoraces. This study aimed to determine the incidence of pneumothorax detected by CXR and by CT after CT-guided biopsy of non-pleurally based pulmonary masses, and to compare these incidences with previous reports in the literature of pneumothorax incidence post ¯uoroscopic biopsy. 88 consecutive CT-guided lung biopsies of masses not abutting the pleural surface were included. Immediate post-biopsy CT images, and 1 and 4 h CXRs were assessed independently by two observers for the presence and size of pneumothorax. 72 biopsies were ®ne needle aspirations (FNAs) performed with 22 G spinal needles only, seven were cutting needle biopsies (CNBs) performed with 18 G cutting needles only, and nine were both. 37 patients (42%) developed a pneumothorax. 35 were detected on CT (40%) and 22 on CXR (25%). None required tube drainage. Of the patients in whom CT demonstrated a pneumothorax, the average depth of this was signi®cantly greater for those in whom CXR also detected a pneumothorax compared with those in whom CXR was negative (7.3 mm versus 3.4 mm, p,0.05). The incidence of pneumothorax detected on CXR post CT-guided biopsy is similar to the reported incidence post ¯uoroscopic biopsy. Image-guided biopsy of thoracic lesions can be performed using ¯uoroscopic, computed tomography (CT) or ultrasound guidance. Fluoroscopy is suitable for lesions visible in two planes, which tend to be large lesions, ultrasound for pleurallybased lesions, and CT for smaller lesions not visible on lateral chest radiographs (CXRs) or closely related to major vessels [1]. The most frequent complication of thoracic biopsy is pneumothorax, with a reported incidence varying from 8 to 61% [1]. The risk of pneumothorax and the need for subsequent chest tube drainage may vary according to the size, nature and depth of the mass, pulmonary function, the number of passes and the post-biopsy procedure adopted, such as puncture site down positioning [1±11]. The published ®gures suggest an increased risk of pneumothorax post CT-guided biopsy (43±46%) [2, 12, 13] as opposed to ¯uoroscopically-guided biopsy (25±27%) [4, 8, 14, 15]. This may simply re¯ect increased sensitivity in detection of small pneumothoraces on CT compared with CXR, but Received 3 June 1999 and in revised form 18 August 1999, accepted 25 August 1999. Address correspondence to Dr F V Gleeson. 1160
two recent studies suggest comparable sensitivities [2, 12]. We performed a retrospective analysis of CTguided lung biopsies of masses not abutting the pleural surface in order to determine and compare the incidence of pneumothorax as detected by CT and by CXR independently. We also compared these frequencies with previously published ®gures for ¯uoroscopically-guided biopsy. Masses abutting the pleura have a lower incidence of pneumothorax post biopsy [3], so only masses distant from the pleura were included.
Methods and materials 88 consecutive CT-guided lung biopsies of masses not abutting the pleural surface were included for analysis. CT guidance in our institution is chosen for lesions that are not easily visible on the lateral CXR, not abutting the chest wall or are inaccessible to ultrasound guidance, adjacent to bullae on CT or close to large vessels, when cavitation is present, and in cases when CT enables biopsy to be performed without crossing of ®ssures. All biopsies were performed on a GE Pace CT unit (General The British Journal of Radiology, December 1999
Pneumothorax after lung biopsy
Electric Medical Systems, Paris, France). Biopsies were performed under local anaesthesia, without sedation, using 22 G spinal needles and/or 18 G cutting needles. Cutting needles were selected for larger masses not close to major vessels, if there was clinical and/or radiological suspicion of noncarcinomatous disease. A single CT image through the level of the biopsy was performed after each pass. If a pneumothorax developed, no further passes were attempted. A 5 mm CT section was performed at the level of the puncture immediately post biopsy for all patients, and inspiratory CXRs supine at 1 h and erect at 4 h. Post biopsy, all patients were positioned puncture site down and instructed not to move, talk or cough until after the 1 h CXR. The post biopsy CT scans were reviewed by two observers who reached a consensus opinion as to the presence and depth from the parietal pleura of any pneumothorax. 3 months later the same observers reviewed the 1 and 4 h CXRs, without reference to the CT results, and recorded the presence and approximate percentage size of any pneumothorax. Discordance between the 1 and 4 h CXRs with respect to the presence or absence of a pneumothorax was recorded. Proportions of pneumothoraces detected by CXR and by CT were compared using the x2 statistic. The 4 h result was used for comparison with CT when there was discordance between the 1 and 4 h CXRs, on the assumption that the negative 1 h supine CXR was a false negative. The Mann±Whitney U-test was used to compare pneumothorax size between groups.
Table 1. Comparison of the detection of pneumothorax on computed tomography (CT) and chest radiograph (CXR)
Results
The apparent increased risk of pneumothorax reported in the literature for CT-guided as opposed to ¯uoroscopically-guided thoracic biopsy [2, 4, 8, 12±16] may be a genuine ®nding. The reasons for this may be an increased needle ``dwell'' time for biopsies performed under CT, an increased number of passes, or differences in the depth and size of lesions. We could ®nd no direct comparative studies of ``dwell'' time under CT compared to ¯uoroscopic guidance, nor of the number of passes performed using the two techniques. However, despite some reports of increased risk of pneumothorax with increased number of passes [8, 17, 18], several other studies have found no such relationship [2, 5, 7, 10, 15, 19, 20]. In our study the average number of passes was higher for the group with no pneumothorax. This issue is complicated by the fact that, having caused a pneumothorax, many operators will not perform further passes. A similar explanation could account for the apparently reduced risk of pneumothorax for cutting needle as opposed to ®ne needle biopsy in our study. Cutting needle biopsies are usually performed for larger lesions
88 biopsies were performed in 86 patients, 58 males and 28 females, median age 68 years (range 24±86). 72 biopsies were ®ne needle aspirations (FNAs) performed with 22 G spinal needles only, seven were cutting needle biopsies (CNBs) with 18 G cutting needles only, and nine were both FNAs and CNBs. The number of passes performed was documented in 79 of the 88 cases, the mean passes per procedure being 1.85. 37 patients developed a pneumothorax (42%). 35 were detected on CT (40%) and 22 on CXR (25%). There was discordance as to the presence or absence of a pneumothorax on the 1 and 4 h CXRs in only ®ve cases, with no visible pneumothorax on the 1 h ®lms but a visible pneumothorax at 4 h. Table 1 demonstrates the detection rate on CT as compared to CXR. Signi®cantly more pneumothoraces were detected by CT (p,0.001, x2 test with Yates' correction). Two pneumothoraces were detected by CXR in patients in whom the post biopsy CT had been negative. The British Journal of Radiology, December 1999
CXR
CT
Positive Negative
Positive
Negative
20 2
15 51
Pneumothoraces were usually small. Only seven cases were .10% of the hemithorax on CXR, the largest being 30%, and no chest tube drainage was required. In patients in whom CT demonstrated a pneumothorax, the average depth of the pneumothorax was signi®cantly greater (p,0.05, Mann±Whitney U-test) for those in whom subsequent CXR also detected a pneumothorax (1± 20 mm, average 7.3 mm, SD 5.2 mm) compared with those in whom subsequent CXR was negative (1±10 mm, average 3.4 mm, SD 2.4 mm). Table 2 demonstrates the number of passes performed with 22 G spinal needles and 18 G cutting needles for patients with and without subsequent pneumothorax for the 79 cases in whom these data were available. It shows that slightly more passes were performed for patients without a pneumothorax (p.0.1, ns). Considering the patients who had one or more cutting needle biopsy, the rate of pneumothorax was only 3/16 (19%) compared with 34/72 (47%) of those who had only a ®ne needle aspiration.
Discussion
1161
H K Bungay, J Berger, Z C Traill and F V Gleeson Table 2. Number of passes performed using 22 G spinal needle (F) and 18 G cutting needle (C) for patients subsequently developing or not developing a pneumothorax
1F 2F 3F 1F+1C 2F+1C 3F+1C 1F+2C 1C 2C Average No. of passes
No pneumothorax
Pneumothorax
10 19 4 2 2 1 3 2 2
12 17 3 0 0 0 0 2 0
2.0
1.7
distant from major vessels, which would be easier to biopsy and so perhaps at lower risk of a complicating pneumothorax. Other authors have also found no increased risk from larger needle sizes [8, 19]. Alternatively, smaller, more dif®cult lesions may be attempted under CT guidance. However, although some authors have found an increased risk of pneumothorax for smaller lesions [2, 6±8, 10], others have found no relationship [5, 15, 21]. In fact, the apparent increased risk of pneumothorax for CT-guided biopsy is likely to be artefactual. Smaller pneumothoraces may be more easily detected on post-biopsy CT than on a post ¯uoroscopically-guided biopsy CXR, although two studies suggest that this is not the case. Kazerooni et al [2] found that post CT-guided biopsy CXRs detected slightly more pneumothoraces than were detected on CT during the procedure. They do not state whether this difference was signi®cant, nor do they state that post-biopsy CT to exclude pneumothorax was performed in all cases. Murphy et al [12] found no signi®cant difference in the proportion of post CT-guided biopsy pneumothoraces visualized by a post-biopsy CT or CXR. However, the CT image in their study was obtained either through the level of the puncture or at the level of the nipples. The latter may miss small pneumothoraces at the biopsy site. Both of these studies included pleurally-based lesions, which have a lower risk for pneumothorax [3], particularly if CT guidance is used, when a route not crossing aerated lung may be chosen. In addition, neither study stated whether the CT and CXR were reviewed in a blind fashion. Our study has avoided these problems. We included only pulmonary parenchymal lesions not abutting the pleura, and so at increased risk of pneumothorax. All patients had an immediate post-biopsy CT at the level of the puncture, where 1162
the earliest air leakage can be assumed to occur. The CTs and CXRs were reviewed blind, 3 months apart. We demonstrated that signi®cantly more pneumothoraces were detectable on the post-biopsy CT than CXRs (40% versus 25%). The two cases in our series where the post-biopsy CT scan was negative but the CXRs demonstrated a pneumothorax probably represent delayed onset pneumothoraces rather than false negative CT. One potential criticism of our study would be that we are comparing the incidence of pneumothorax on an immediate CT image with that detected on CXR delayed by 1 or 4 h, and that resorption during this delay explains the lower rate of pneumothoraces on the CXRs. However, it is extremely unlikely that a pneumothorax could resorb within 1±4 h. Once the source of air leakage has sealed, re-expansion takes place at a rate of only 1.25% of the volume of the hemithorax per day [22, 23]. It is also unlikely that the supine posture of the 1 h CXR explains the lower rate of pneumothorax on CXRs compared to CT: again, the slow rate of resorption should mean that the pneumothorax was at least visible on the 4 h erect CXR, and the 1 and 4 h CXRs were analysed together. Our decision to analyse the 1 and 4 h CXRs together and compare these with the CTs could also be criticized. However, there was discordance between the 1 and 4 h CXRs in only 5 of the 88 cases. It is generally recognized that detection of small pneumothoraces on supine CXRs is dif®cult and it can reasonably be assumed that measurement of their size would be unreliable. We therefore used measurement of size on the 4 h erect CXR for the comparison with CT. It is our personal practice to perform both 1 and 4 h CXRs despite their high concordance. This is because detection of an enlarging pneumothorax, implying the presence of a continued air leak, combined with the patient's clinical condition, is used to make an informed decision regarding the timing of patient discharge. It is of note that our ®gures of 40% patients with a pneumothorax visible on CT and 25% visible on CXR post CT-guided biopsy correspond to the published ®gures for CT-guided biopsies (43±46%) [2, 12, 13] and ¯uoroscopicallyguided biopsies (25±27%) [4, 8, 14, 15]. We believe that the increased sensitivity of CT for detecting small pneumothoraces accounts for the apparent increased incidence of this complication after lung biopsy under CT guidance.
References 1. Westcott JL. Percutaneous transthoracic needle biopsy. Radiology 1988;169:593±601. The British Journal of Radiology, December 1999
Pneumothorax after lung biopsy 2. Kazerooni EA, Lim FT, Mikhail A, Martinez FJ. Risk of pneumothorax in CT-guided transthoracic needle aspiration biopsy of the lung. Radiology 1996;198:371±5. 3. Haramati LB, Austin JHM. Complications after CT-guided needle biopsy through aerated versus nonaerated lung. Radiology 1991;181:778. 4. Moore EH, Shepard J-AO, McLoud TC, Templeton PA, Kosiuk JP. Positional precautions in needle aspiration lung biopsy. Radiology 1990;175:733±5. 5. Poe RH, Kallay MC, Wicks CM, Odoroff CL. Predicting risk of pneumothorax in needle biopsy of the lung. Chest 1984;85:232±5. 6. Anderson CLV, Crespo JCA, Lie TH. Risk of pneumothorax not increased by obstructive lung disease in percutaneous needle biopsy. Chest 1994; 105:1705±8. 7. Fish GD, Stanley JH, Miller KS, Schabel SI, Sutherland SE. Post biopsy pneumothorax: estimating the risk by chest radiography and pulmonary function tests. AJR 1988;150:71±4. 8. Sinner WN. Complications of percutaneous transthoracic needle aspiration biopsy. Acta Radiol 1976;17:813±28. 9. Berquist TH, Bailey PB, Cortese DA, Miller WE. Transthoracic needle biopsy: accuracy and complications in relation to location and type of lesion. Mayo Clin Proc 1980;55:475±81. 10. Miller KS, Fish GB, Stanley JH, Schabel SI. Prediction of pneumothorax rate in percutaneous needle aspiration of the lung. Chest 1988;93:742±5. 11. Cormier Y, Laviolette M, Tardif A. Prevention of pneumothorax in needle lung biopsy by breathing 100% oxygen. Thorax 1980;35:37±41. 12. Murphy FB, Small WC, Wichman RD, Chalif M, Bernardino ME. CT and chest radiography are equally sensitive in the detection of pneumothorax after CT-guided pulmonary interventional procedures. AJR 1990;154:45±6. 13. vanSonnenberg E, Casola G, Ho M, Neff CC, Varney RR, Wittich GR, et al. Dif®cult thoracic lesions: CT-guided biopsy experience in 150 cases. Radiology 1988;167:457±61.
The British Journal of Radiology, December 1999
14. Flower CDR, Verney GI. Percutaneous needle biopsy of thoracic lesionsÐan evaluation of 300 biopsies. Clin Radiol 1979;30:215±8. 15. Williams AJ, Santiago S, Lehrman S, Popper R. Transcutaneous needle aspiration of solitary pulmonary masses: how many passes? Am Rev Respir Dis 1987;136:452±4. 16. Fink I, Gamsu G, Harter LP. CT-guided aspiration biopsy of the thorax. J Comput Assist Tomogr 1982;6:958±62. 17. Johnsrude IS, Silverman JF, Weaver MD, McConnell RW. Rapid cytology to decrease pneumothorax incidence after percutaneous biopsy. AJR 1985;144:793±4. 18. Stevens GM, Weigen JF, Lillington GA. Needle aspiration biopsy of localised pulmonary lesions with ampli®ed ¯uoroscopic guidance. AJR 1968; 103:561±71. 19. Sakai T, Hayashi N, Kimoto T, Maeda M, Ishii Y, Murashima S, et al. CT-guided biopsy of the chest: usefulness of ®ne-needle core biopsy combined with frozen-section pathologic diagnosis. Radiology 1994;190:243±6. 20. Moore EH, LeBlanc J, Montesi SA, Richardson ML, Shepard J-AO, McLoud TC. Effect of patient positioning after needle aspiration lung biopsy. Radiology 1991;181:385±7. 21. Li H, Boiselle PM, Shepard J-AO, TrotmanDickenson B, McLoud TC. Diagnostic accuracy and safety of CT-guided percutaneous needle aspiration biopsy of the lung: comparison of small and large pulmonary nodules. AJR 1996;167:105±9. 22. Millard FJC, Pepper JR. Pleural disease. In: RAL Brewis, B Corrin, DM Geddes, GJ Gibson, editors. Respiratory medicine. London: WB Saunders, 1995:1569. 23. Kircher LT, Swartzel RL. Spontaneous pneumothorax and its treatment. JAMA 1954;155:24±9.
1163
