- Email: [email protected]
Percutaneous Transthoracic Needle Biopsy of the Lung: Review of 612 Lesions
Visceral Intervention
Percutaneous Transthoracic Needle Biopsy of the Lung: Review of 612 ~esions' James L. Swischuk, MD Flavio Castaneda, MD Jitendra C. Patel, MD Ruizong Li, MD Kenneth W. Fraser, MD Terrence M. Brady, MD Raymond E. Bertino, MD
Index terms: Fine needle aspiration, Lung, biopsy Pneumothorax
JVIR 1998; 9:347-352 Abbreviation: FNAB piration biopsy
=
fine-needle as-
PURPOSE: The results and complications of 651 pulmonary fineneedle aspiration biopsies (FNABs) were reviewed. The number of needle passes and needle size were correlated to pneumothorax and chest tube placement rates. MATERIALS AND METHODS: FNAB of the lung was performed on 651 occasions in 612 patients with 18- to 22-gauge Franseen needles. Diagnostic rates were calculated. The number of needle passes performed and needle size used were evaluated for their association with pneumothorax and subsequent chest tube placement. RESULTS: Diagnostic accuracy was 94%with sensitivity for malignancy of 95%. Positive and negative predictive values were 99.5% and 90%,respectively. Pneumothorax occurred in 26.9% of patients with 9.2% requiring chest tube placement. Increasing numbers of needle passes and larger needle sizes did not increase the rates of pneumothorax or chest tube placement. CONCLUSIONS: FNAB of the lung has excellent diagnostic rates and remains the procedure of choice for diagnosing pulmonary lesions. This large study contradicts perceptions that pneumothorax and chest tube placement rates decrease with thinner needles and fewer passes.
From the Departments of Radiology (J.L.S., F.C., R.L., K.W.F., T.M.B., R.E.B.) and Pathology (J.C.P.) University of Illinois College of Medicine a t Peoria, Peoria, Illinois. Received July 2, 1997; revision requested August 4; revision received and accepted September 9. Presented a t the 1997 SCVIR annual meeting. Address correspondence to J.L.S., Department of Radiology 530 NE Glen Oak, Peoria, IL 61637. O SCVIR. 1998
ALTHOUGH the first reports of needle biopsy of the lung were published in the late 1800s (1,2), it was not until Nordenstrom (3) introduced the technique of fine-needle aspiration biopsy (FNAB) of the lung in 1965 that this technique first became accepted as a useful and safe diagnostic tool in the evaluation of suspicious intrathoracic lesions. Since then, improving technologies in needle design and radiologic equipment have helped secure the place of FNAB in the evaluation of suspicious pulmonary lesions. The diagnostic sensitivity for malignant lesions has been documented repeatedly at over 90% (4-9), but the diagnostic usefulness of FNAB in the evaluation of benign pulmonary disease is less impressive,
with sensitivities ranging from 9% to 91% (4,5,8-12). In addition, the diagnostic rates for FNAB of the lung have been shown to be more sensitive than those for bronchoscopic techniques (5,13,14). The procedure is most often performed on an outpatient basis, and the inclusion of a cytopathologist to the biopsy team has further increased the utility of the procedure (15). Complications include pneumothorax, hemoptysis, air embolism, needle tract seeding with tumor, and death; however, pneumothorax remains, by far, the most pervasive. In addition, hemoptysis is seldom significant, and air embolism, needle track tumor seeding, and death remain extremely rare (1620).
348
Percutaneous Transtha~racicNeedle Biopsy of the Lung
March-April 1998 JVIR
I MATERIALS AND METHODS From January 1990 to July 1996, 651 lung biopsies were performed on 612 patients with suspicious intrathoracic lesions documented on plain film radiographs and computed tomography (CT) of the chest. Fifty-seven percent of the patients were men and 43% were women who ranged in age from 19 to 96 years. Most of the lesions sampled were noncalcified pulmonary nodules. Previous malignancy was documented in 116 patients. Thirty-five patients underwent more than one biopsy, with the procedure repeated once in 32 patients, twice in two patients, and three times in one patient. Lesion size ranged from 3 mm to 10 cm. Standard CT and fluoroscopic techniques were used (17-20) and all procedures were performed with use of Franseen needles (Cook, Bloomington, IN) of either 18, 20, or 22 gauge. Larger diameter needles were often used when samples obtained with smaller diameter needles were insufficient for interpretation. Biopsies were performed by one of four experienced interventional radiologists, and a cytopathologist was present during all procedures. Cytologic study was performed on all specimens, and often there was sufficient material for histologic examination. The biopsy procedure was continued until a sufficient sample was obtained as determined by the cytopathologist, or until samples could not be obtained because of complications or an uncooperative patient. Before the biopsy procedure, all patients underwent posteroanterior and lateral chest radiography and a complete chest CT. The use of either fluoroscopic or CT guidance was dependent on how well the lesion in question was seen on plain films. Most procedures were performed on an outpatient basis unless the patient was previously admitted to the hospital or subsequent complications required hospitalization. The only absolute contraindication for FNAB of the lung was the inability of a patient to cooperate
Table 1 Results of FNAB Final FNAB Diagnosis
Final Lesion Diagnosis Malignant (n
=
450)
Benign ( n = 140)
Unconfirmed (n
=
22)
Positive for malignancy Suspicious for malignancy False-negative for malignancy Inadequate specimen Negative for malignancy: (specific benign: 55) (nonspecific benign: 75) False-positive for malignancy Inadequate specimen Follow-up < 2 years Inadequate specimen
during the procedure. After the biopsy procedure, immediate and 2-4-hour delayed chest radiographs were obtained in the inspiratory and expiratory phase. For those patients with documented ~neumothorax, the indication for ccest tube placement was either shortness of breath or an enlarging pneumothorax estimated at 30% lung volume. Initial treatment for pneumothorax consisted of an 8-12-F catheter (Cook) that was placed on either a Heimlich valve or wall suction. In the past, a surgical chest tube was placed if the initial small-bore chest tube was unsuccessful in re-expanding the lung; however, our current practice is to simply exchange the initial chest tube over a guide wire for a larger one. A lesion was determined to be malignant if a specific cellular diagnosis was made at FNAB that correlated well with radiologic findings and subsequent clinical course, or if the diagnosis was proven a t surgery or autopsy. A lesion was proven benign if follow-up radiologic studies showed a shrinking lesion or a stable lesion during a 2-year interval. Also, some benign lesions were proven a t open biopsy. A lesion was considered unconfirmed if radiologic follow-up of a suspected benign lesion was less than 2 years, or if the patient was lost to follow-up. FNAB results were considered positive for malignancy if the specimen revealed a specific cellular diagnosis. Biopsy results that were suspicious for malignancy were not
426 4 15 5 130 2 8 16 6
considered true positives. The FNAB results were considered negative for malignancy if the specimen yielded a specific benign diagnosis (granuloma, hamartoma, cores of fibrosis), or specimens showed nonspecific benign changes (giant cells, leukocytes, histocytes, fragments of fibrosis). When a specific benign diagnosis correlated well with radiologic studies, these lesions were treated as such, and radiologic follow-up was performed. Alternatively, for those patients with nonspecific benign changes, the first course of action was to perform multiple needle passes during the biopsy procedure. If multiple passes showed reproducible nonspecific benign changes, these were correlated to radiologic studies and clinical presentation. If all suggested benign disease, the patients were followed conservatively with radiologic studies. If the biopsy results showed inconsistent or scant nonspecific material, and the clinical suspicion for malignancy remained moderate to high despite FNAB results, repeated FNAB or open biopsy was recommended. If a negative result was accepted and no surgical biopsy was planned, radiologic follow-up included plain film chest radiographs at 3, 6, 12, 18, and 24 months for those lesions visible on plain films. For those lesions visible on CT only, repeated CT of the chest was performed a t identical intervals. If the lesion was seen to shrink, no further radiologic follow-up was performed.
Swischuk et a1
349
Volume 9 Number 2
Table 2 Pneumothorax Rate* and Subsequent Chest Tube Placement Ratet Related to Number of Needle Passes No. of Passes
Bx ( n = 651)
Patients with PTX (n = 175)
Patients without PTX (n = 476)
Patients with PTX and CT (n = 60)
Patients with PTX and without CT (n = 115)
1 2 3 4 5 6
192 225 153 67 10 4
58 64 37 13 1 2
134 161 116 54 9 2
24 23 10 1 0 2
34 41 27 12 1 0
* 2 = 6.37, P = 0.27. t 2 = 10.67, P = .06. 71% of patients underwent multiple needle passes PTX = pneumothorax; Bx = number of biopsy procedures; CT
The final lesion diagnosis was then compared with the final biopsy diagnosis, and diagnostic rates were rendered. Also, the type and frequency of complications were documented. In addition, needle size and number of needle passes were evaluated as to whether they individually contributed to pneumothorax and chest tube placement rates. Specifically, the test of independence was used to determine whether these two variables independently contributed to significantly higher rates of pneumothorax or chest tube placement (in those patients with pneumothorax). In addition, the t test of independent groups was performed to see if there were significant differences in needle size used or number of needle passes performed in those patients with pneumothorax and in those with subsequent chest tube placement.
2
RESULTS
There were 450 malignant, 140 benign, and 22 unconfirmed lesions (Table 1). Of the 450 confirmed malignant lesions, a final FNAB diagnosis was shown to be positive for malignancy in 426 patients (true positives), yielding a sensitivity of 95%. Thirteen patients with a final biopsy diagnosis of malignancy had a previous biopsy, that was considered negative. Nine of these biopsies yielded a sample that was nondiagnostic or insufficient, and the
=
chest tube.
decision to repeat the biopsy was obvious. The other four of these revealed a mix of inflammatory cells. Of the 140 confirmed benign lesions, a final FNAB diagnosis was found to be a specific benign diagnosis (hamartoma, granuloma, cores of fibrosis) in 55 patients, and a nonspecific benign diagnosis in 75 patients (a total of 130 true negatives), yielding a sensitivity for benign disease of 93%. Of the 75 patients with nonspecific benign changes, seven had a repeated biopsy confirming the previous results. Of the 22 patients with an unconfirmed final diagnosis, 16 had a FNAB diagnosis that is negative for malignancy; however, radiologic follow-up has been for less than 2 years. The overall diagnostic accuracy for both malignant and benign lesions combined is 94% with positive and negative predictive values of 99.5% and 90%, respectively. Complications included pneumothorax with and without chest tube placement, hemoptysis, and transient hypotension. Pneumothorax was seen in 175 (26.9%) patients with chest tube placement required in 60 (9.2%)of these patients. Of the patients with hemoptysis (3.1%) and transient hypotension (0.1%), none required additional treatment, transfusions, or hospitalization. Table 2 summarizes the number of needle passes performed with subsequent pneumothorax and chest tube placement rates. When these data were evaluated using the 2
test, no significant increase in the pneumothorax rate was seen in patients undergoing an increasing number of needle Dasses. Furthermore, in those patients with pneumothorax, an increasing number of needle passes did not contribute to an increase in the chest tube placement rate. Patients receiving four or more passes were grouped together because of the smaller sample size. The needle size used was then evaluated using the same statistical analysis. Specifically, needle size was correlated to pneumothorax and chest tube placement rates (Table 3). This shows that there was a significant increase in the rate of pneumothorax (P = .04) in those patients biopsied with 22guage needles; however, in patients with pneumothorax, no increase in chest tube la cement rate was seen with differint needle sizes. In addition, the same data were evaluated with use of the t test. In this instance, instead of looking for significant differences in pneumothorax rates, patients with pneumothorax and chest tube la cement were evaluated for any significant difference in needle size used or number of passes performed. Table 4 evaluates those patients with pneumothorax and compares them to those without pneumothorax. Once again, significantly smaller diameter needles were used when pneumothorax occurred (P = .012), corroborating previously mentioned observations. No significant differu
Percutaneous Transthoracic Needle Biopsy of the Lung
350
March-April 1998 JVIR
I
Table 3 Pneumothorax Rate* and Subsequent Chest Tube Placement Ratet Related to Needle Size Needle Size (gauge) 18 20 22
Bx (n
=
651)
Patients with PTX (n = 175)
Patients without PTX ( n = 476)
Patients with FTX and CT (n = 60)
Patients with PTX and without CT (n = 115)
2 66 107
21 205 250
0 18 42
2 48 65
23 271 357
* For 22 gauge versus 20 and 18 gauge 2 = 6.485, P = .04. 2 = 3.65, P = .16. Bx
=
number of biopsy procedures; PTX
ence in the number of needle passes was seen between those patients with and without pneumothorax. Table 5 evaluates those patients with pneumothorax and subsequent chest tube placement and shows that, unlike the previously mentioned data, significantly smaller needles were used in those patients with chest tube placement; however, the level of significance is a t a threshold value ( P = .05). No significant difference in the number of needle passes was seen between those patients with and without chest tube placement.
DISCUSSION Our overall accuracy (94%), sensitivity for malignancy (95%), and sensitivity for benign disease (93%) are comparable to those of other studies (4,6-9,12,21). Positive and negative predictive values are 99.5% and 90%, respectively. This confirms the efficacy of FNAB as the primary diagnostic tool in evaluating suspicious pulmonary lesions. Although its usefulness in detecting malignancy is widely accepted, the same cannot be said for its role in detecting benign disease. This is evident in the wide range (9%-91%) of benign diagnostic rates reported in the literature (4,5,8-12) and is most likely not due to differences in technique or ability but rather in how negative biopsy results have been interpreted. Lower diagnostic rates will invariably be based on FNAB results yielding a specific benign diagnosis, whereas higher rates will most often include
=
pneumothorax; CT
=
chest tube.
all results whether specific or nonspecific that are nonmalignant. The decision to include nonspecific benign results while calculating accuracy rates for benign disease undoubtedly will rely on the confidence that an adequate nonspecific negative biopsy will accurately predict benign disease. This confidence can be enhanced bv the use of various techniques inciuding the more frequent use of CT for small lesions (4,10), the presence of a cytopathologist (15,22), and repeated needle passes duplicating findings (6,7,9,21-24). In other words, a radiologically demonstrated, wellplaced needle yielding similar nonspecific benign cells as interpreted by an onsite cytopathologist can provide a result that reliably predicts the absence of malignancy. Therefore, it seems reasonable to conclude, as our data would support, that FNAB of the lung performed under current standards is also quite useful in the prediction of benign disease. This is a significant point considering it is the ability to accurately predict benign disease that most often obviates the need for surgery (6, 7, 22).
Complications encountered in our patient population included pneumothorax, hemoptysis, and one episode of transient hypotension. As expected, pneumothorax was by far the most common. Air embolism, biopsy tract tumor seeding, and death were not encountered. The frequency of pneumothorax (26.9%) and chest tube placement (9.2%) are within the expected range when compared with other large studies (4,7,9,12,25-28). The literature cites many factors contributing to an increasing pneumothorax rate including lesion size, lesion depth, number of needle passes, pleural surfaces crossed, and pulmonary function studies (16-19,25-30). Small lesion size, abnormal pulmonary function tests, and increasing lesion depths seem to be the most substantiated predictors of pneumothorax (25,26,28-30). When we evaluated the association between the number of needle passes with pneumothorax and chest tube placement rates, we found that no significant correlation exists. In reviewing the literature specifically for instances where formal statistical analysis was used in evaluating this associa-
Table 4 Significance of Needle Size and Number of Passes Related to Pneumothorax Variable
Patients with FTX ( n = 175)
Needle size gauge No. of passes
21.2 + 1.03 2.07 + 1.04
Note.-PTX
=
pneumothorax, NS
=
Patients without PTX (n = 476)
not significant.
21.0 2.26
+ 1.16 ?
1.07
P Value .012 NS
Swischuk et a1
351
Volume 9 Number 2
Table 5 Significance of Needle Size and Number of Passes Related to Chest Tube Placement in Patients with Pneumothorax Variable
Patients with CT (n = 60)
Paitents without CT (n = 115)
P Value
Needle size gauge Number of Passes
21.4 5 .92 1.93 + 1.13
21.0 5 1.07 2.25 +- 1.01
.05 NS
Note.-CT
=
chest tube, NS
=
7.
8.
not significant. 9.
tion, there is a single early report by Sinner et a1 (28), a study involving well over 2,000 patients, showing a significant increase in the pneumothorax rate when a second pass was performed within a 2-hour period. One shortcoming of this study, however, is that the exact number of patients represented by their statistics is uncertain. We have found no additional reports statistically supporting this positive correlation. In fact, since then there have been multiple reports, including ours, involving 1,828 patients showing that no correlation exists between the number of needle passes and pneumothorax rate (2426,29,30). We then evaluated needle size for any association with increasing pneumothorax and chest tube placement rates. Surprisingly, we found that there was a significant positive correlation between pneumothorax and possibly chest tube placement with the use of small diameter needles (22 gauge vs 20-18 gauge). Initially, this seems to make little sense; however, it is important to realize that although statistical analysis shows a correlative relationship, it does not necessarily prove cause and effect. In explaining this positive correlation, we do not necessarily believe that smalldiameter needles cause pneumothorax, but rather that this most likely indicates a bias or preference on our part for using smaller gauge needles for so-called "difficult" lesions. These would include smaller d e e ~ e lesions r in ~ a t i e n t swith known obstructive pulmonary disease, conditions which, as previously mentioned, are known to in-
crease the risk of pneumothorax. Perhaps a more practical conclusion is that the use of larger (ie, 18gauge) needles does not necessarily increase the risk of pneumothorax. In reviewing the literature, we found little information on the relationshi~between needle size and pneumothorax. In fact, only a single report by Sinner et a1 (28) shows no statistically significant correlation between needle size and pneumothorax rate. In conclusion, the data support the use of FNAB of the lung as the diagnostic procedure of choice in evaluating both malignant and benign disease of the lung. In addition, while striving to obtain an adequate and diagnostic FNAB specimen, multiple passes with various-sized needles (18-22 gauge) should not be cause for increased apprehension.
10.
11.
12.
13.
14.
15.
References 1. Leydon 00. Uber infektiose pneumonic. Dtsch Med Wochenschr 1883; 952. 2. Menetrier P. Cancer primitif du poumon. Bull Soc Anat Paris 1886; 11:643. 3. Nordenstrom B. A new technique for transthoracic biopsy of lung changes. Br J Radiol 1965; 38550553. 4. Westcott JL, Rao N, Colley DP. Transthoracic needle biopsy of small pulmonary nodules. Radiology 1997; 202:97-103. 5. Gasparini S, Ferretti M, Secchi EB, Baldelli S, Zuccatosta L, Gusella P. Integration of transbronchial and percutaneous approach in the diagnosis of peripheral pulmonary nodules or masses: experience with 1,027 cases. Chest 1995; 108:131-137. 6. Khouri NF, Stitik FP, Erozan YS, e t al. Transthoracic needle aspi-
16.
17. 18. 19. 20. 21.
ration biopsy of benign and malignant lung lesions. AJR 1985; 144: 281-288. Westcott JL. Direct percutaneous needle aspiration of localized pulmonary lesions: results in 422 patients. Radiology 1980; 137:31-35. Klein JS, Salomon G, Stewart EA. Transthoracic needle biopsy with a coaxially placed 20-gauge automated cutting needle: results in 122 patients. Radiology 1996; 198:715-720. Stanley JH, Fish GD, Andriole JG, e t al. Lung lesions: cytologic diagnosis by fine needle biopsy. Radiology 1987; 162:389-391. vansonnenberg E, Casola G, Ho M, et al. Difficult thoracic lesions: CT guided biopsy experience in 150 cases. Radiology 1988; 167:457-461. Stevens GM, Jackman RJ. Outpatient needle biopsy of the lung: its safety and utility. Radiology 1984; 151:301-304. Sargent EN, Turner AF,Gordonson J , Schwinn CP, Pashky 0. Percutaneous pulmonary needle biopsy: report of 350 patients. Am J Roentgenol Radium Ther Nucl Med 1974; 122:758-768. Wallace JM, Deutsch AL. Flexible fiberoptic bronchoscopy and percutaneous needle lung aspiration for evaluating the solitary pulmonary nodule. Chest 1982; 81: 665-671. Mark JB, Marglin SI, Castellino RA. The role of bronchoscopy and needle aspiration in the diagnosis of peripheral lung masses. J Thorac Cardiovasc Surg 1978; 76:266-268. Austin JH, Cohen MB. Value of having a cytopathologist present during percutaneous fine-needle aspiration biopsy of lung: report of 55 cancer patients and metaanalysis of the literature. AJR 1993; 160:175-177. Shepard JO. Complications of percutaneous needle aspiration biopsy of the chest: prevention and management. Semin Intervent Radiol 1994; 11:3. Weisbrod GL. Transthoracic needle biopsy. World J Surg 1993; 17:705711. Sanders C. Transthoracic needle aspiration. Clin Chest Med 1992; 13:ll-16. Westcott JL. Percutaneous transthoracic needle biopsy. Radiology 1988; 169593-601. Weisbrod GL. Transthoracic percutaneous lung biopsy. Radiol Clin North Am 1990; 28:647-655. Gobien RP, Valicenti J F , Paris BS, Daniel1 C. Thin-needle aspiration
352
Percutaneous Transthoracic Needle Biopsy of the Lung
March-April 1998 JVIR
biopsy: methods of increasing the accuracy of a negative prediction. Radiology 1982; 145:603-605. 22. Conces DJ, Schwenk GR, Doering PR, Glant MD. Thoracic needle biopsy: improved results utilizing a team approach. Chest 1987; 91:813-816. 23. Lalli AF, McCormack U,Zelch M, Reich NE, Belovich D. Aspiration biopsies of chest lesions. Radiology 1978; 127:35-40. 24. Williams AJ, Santiago S, Lehrman S, Popper R. Transcutaneous needle aspiration of solitary pulmonary masses: how many passes? Am Rev Respir Dis 1987; 136:452-454.
25. Miller KS, Fish GB, Stanley JH, Schabel SI. Prediction of pneumothorax rate in percutaneous needle aspiration of the lung. Chest 1988; 93:742-745. 26. Fish GB, Stanley JH, Miller KS, Schabel SI, Sutherland SE. Postbiopsy pneumothorax: estimating the risk by chest radiography and pulmonary function tests. AJR 1988; 150:71-74. 27. Berquist TH, Bailey PB, Cortese CA, Miller WE. Transthoracic needle biopsy: accuracy and complications in relation to location and type
of lesion. Mayo Clin Proc 1980; 55: 475-481. 28. Sinner WN. Complications of percutaneous transthoracic needle aspiration biopsy. Acta Radio1 [Diagnl 1976; 17:813-828. 29. Kazerooni EA, Lim FT, Mikhail A, Martinez FJ. Risk of pneumothorax in CT-guided transthoracic needle aspiration biopsy of the lung. Radiology 1996; 198:371-375. 30. Poe RH, Kallay MC, Wicks CM, Odoroff CL. Predicting risk of pneumothorax in needle biopsy of the lung. Chest 1984; 85:232-235.
Percutaneous Transthoracic Needle Biopsy of the Lung: Review of 612 ~esions' James L. Swischuk, MD Flavio Castaneda, MD Jitendra C. Patel, MD Ruizong Li, MD Kenneth W. Fraser, MD Terrence M. Brady, MD Raymond E. Bertino, MD
Index terms: Fine needle aspiration, Lung, biopsy Pneumothorax
JVIR 1998; 9:347-352 Abbreviation: FNAB piration biopsy
=
fine-needle as-
PURPOSE: The results and complications of 651 pulmonary fineneedle aspiration biopsies (FNABs) were reviewed. The number of needle passes and needle size were correlated to pneumothorax and chest tube placement rates. MATERIALS AND METHODS: FNAB of the lung was performed on 651 occasions in 612 patients with 18- to 22-gauge Franseen needles. Diagnostic rates were calculated. The number of needle passes performed and needle size used were evaluated for their association with pneumothorax and subsequent chest tube placement. RESULTS: Diagnostic accuracy was 94%with sensitivity for malignancy of 95%. Positive and negative predictive values were 99.5% and 90%,respectively. Pneumothorax occurred in 26.9% of patients with 9.2% requiring chest tube placement. Increasing numbers of needle passes and larger needle sizes did not increase the rates of pneumothorax or chest tube placement. CONCLUSIONS: FNAB of the lung has excellent diagnostic rates and remains the procedure of choice for diagnosing pulmonary lesions. This large study contradicts perceptions that pneumothorax and chest tube placement rates decrease with thinner needles and fewer passes.
From the Departments of Radiology (J.L.S., F.C., R.L., K.W.F., T.M.B., R.E.B.) and Pathology (J.C.P.) University of Illinois College of Medicine a t Peoria, Peoria, Illinois. Received July 2, 1997; revision requested August 4; revision received and accepted September 9. Presented a t the 1997 SCVIR annual meeting. Address correspondence to J.L.S., Department of Radiology 530 NE Glen Oak, Peoria, IL 61637. O SCVIR. 1998
ALTHOUGH the first reports of needle biopsy of the lung were published in the late 1800s (1,2), it was not until Nordenstrom (3) introduced the technique of fine-needle aspiration biopsy (FNAB) of the lung in 1965 that this technique first became accepted as a useful and safe diagnostic tool in the evaluation of suspicious intrathoracic lesions. Since then, improving technologies in needle design and radiologic equipment have helped secure the place of FNAB in the evaluation of suspicious pulmonary lesions. The diagnostic sensitivity for malignant lesions has been documented repeatedly at over 90% (4-9), but the diagnostic usefulness of FNAB in the evaluation of benign pulmonary disease is less impressive,
with sensitivities ranging from 9% to 91% (4,5,8-12). In addition, the diagnostic rates for FNAB of the lung have been shown to be more sensitive than those for bronchoscopic techniques (5,13,14). The procedure is most often performed on an outpatient basis, and the inclusion of a cytopathologist to the biopsy team has further increased the utility of the procedure (15). Complications include pneumothorax, hemoptysis, air embolism, needle tract seeding with tumor, and death; however, pneumothorax remains, by far, the most pervasive. In addition, hemoptysis is seldom significant, and air embolism, needle track tumor seeding, and death remain extremely rare (1620).
348
Percutaneous Transtha~racicNeedle Biopsy of the Lung
March-April 1998 JVIR
I MATERIALS AND METHODS From January 1990 to July 1996, 651 lung biopsies were performed on 612 patients with suspicious intrathoracic lesions documented on plain film radiographs and computed tomography (CT) of the chest. Fifty-seven percent of the patients were men and 43% were women who ranged in age from 19 to 96 years. Most of the lesions sampled were noncalcified pulmonary nodules. Previous malignancy was documented in 116 patients. Thirty-five patients underwent more than one biopsy, with the procedure repeated once in 32 patients, twice in two patients, and three times in one patient. Lesion size ranged from 3 mm to 10 cm. Standard CT and fluoroscopic techniques were used (17-20) and all procedures were performed with use of Franseen needles (Cook, Bloomington, IN) of either 18, 20, or 22 gauge. Larger diameter needles were often used when samples obtained with smaller diameter needles were insufficient for interpretation. Biopsies were performed by one of four experienced interventional radiologists, and a cytopathologist was present during all procedures. Cytologic study was performed on all specimens, and often there was sufficient material for histologic examination. The biopsy procedure was continued until a sufficient sample was obtained as determined by the cytopathologist, or until samples could not be obtained because of complications or an uncooperative patient. Before the biopsy procedure, all patients underwent posteroanterior and lateral chest radiography and a complete chest CT. The use of either fluoroscopic or CT guidance was dependent on how well the lesion in question was seen on plain films. Most procedures were performed on an outpatient basis unless the patient was previously admitted to the hospital or subsequent complications required hospitalization. The only absolute contraindication for FNAB of the lung was the inability of a patient to cooperate
Table 1 Results of FNAB Final FNAB Diagnosis
Final Lesion Diagnosis Malignant (n
=
450)
Benign ( n = 140)
Unconfirmed (n
=
22)
Positive for malignancy Suspicious for malignancy False-negative for malignancy Inadequate specimen Negative for malignancy: (specific benign: 55) (nonspecific benign: 75) False-positive for malignancy Inadequate specimen Follow-up < 2 years Inadequate specimen
during the procedure. After the biopsy procedure, immediate and 2-4-hour delayed chest radiographs were obtained in the inspiratory and expiratory phase. For those patients with documented ~neumothorax, the indication for ccest tube placement was either shortness of breath or an enlarging pneumothorax estimated at 30% lung volume. Initial treatment for pneumothorax consisted of an 8-12-F catheter (Cook) that was placed on either a Heimlich valve or wall suction. In the past, a surgical chest tube was placed if the initial small-bore chest tube was unsuccessful in re-expanding the lung; however, our current practice is to simply exchange the initial chest tube over a guide wire for a larger one. A lesion was determined to be malignant if a specific cellular diagnosis was made at FNAB that correlated well with radiologic findings and subsequent clinical course, or if the diagnosis was proven a t surgery or autopsy. A lesion was proven benign if follow-up radiologic studies showed a shrinking lesion or a stable lesion during a 2-year interval. Also, some benign lesions were proven a t open biopsy. A lesion was considered unconfirmed if radiologic follow-up of a suspected benign lesion was less than 2 years, or if the patient was lost to follow-up. FNAB results were considered positive for malignancy if the specimen revealed a specific cellular diagnosis. Biopsy results that were suspicious for malignancy were not
426 4 15 5 130 2 8 16 6
considered true positives. The FNAB results were considered negative for malignancy if the specimen yielded a specific benign diagnosis (granuloma, hamartoma, cores of fibrosis), or specimens showed nonspecific benign changes (giant cells, leukocytes, histocytes, fragments of fibrosis). When a specific benign diagnosis correlated well with radiologic studies, these lesions were treated as such, and radiologic follow-up was performed. Alternatively, for those patients with nonspecific benign changes, the first course of action was to perform multiple needle passes during the biopsy procedure. If multiple passes showed reproducible nonspecific benign changes, these were correlated to radiologic studies and clinical presentation. If all suggested benign disease, the patients were followed conservatively with radiologic studies. If the biopsy results showed inconsistent or scant nonspecific material, and the clinical suspicion for malignancy remained moderate to high despite FNAB results, repeated FNAB or open biopsy was recommended. If a negative result was accepted and no surgical biopsy was planned, radiologic follow-up included plain film chest radiographs at 3, 6, 12, 18, and 24 months for those lesions visible on plain films. For those lesions visible on CT only, repeated CT of the chest was performed a t identical intervals. If the lesion was seen to shrink, no further radiologic follow-up was performed.
Swischuk et a1
349
Volume 9 Number 2
Table 2 Pneumothorax Rate* and Subsequent Chest Tube Placement Ratet Related to Number of Needle Passes No. of Passes
Bx ( n = 651)
Patients with PTX (n = 175)
Patients without PTX (n = 476)
Patients with PTX and CT (n = 60)
Patients with PTX and without CT (n = 115)
1 2 3 4 5 6
192 225 153 67 10 4
58 64 37 13 1 2
134 161 116 54 9 2
24 23 10 1 0 2
34 41 27 12 1 0
* 2 = 6.37, P = 0.27. t 2 = 10.67, P = .06. 71% of patients underwent multiple needle passes PTX = pneumothorax; Bx = number of biopsy procedures; CT
The final lesion diagnosis was then compared with the final biopsy diagnosis, and diagnostic rates were rendered. Also, the type and frequency of complications were documented. In addition, needle size and number of needle passes were evaluated as to whether they individually contributed to pneumothorax and chest tube placement rates. Specifically, the test of independence was used to determine whether these two variables independently contributed to significantly higher rates of pneumothorax or chest tube placement (in those patients with pneumothorax). In addition, the t test of independent groups was performed to see if there were significant differences in needle size used or number of needle passes performed in those patients with pneumothorax and in those with subsequent chest tube placement.
2
RESULTS
There were 450 malignant, 140 benign, and 22 unconfirmed lesions (Table 1). Of the 450 confirmed malignant lesions, a final FNAB diagnosis was shown to be positive for malignancy in 426 patients (true positives), yielding a sensitivity of 95%. Thirteen patients with a final biopsy diagnosis of malignancy had a previous biopsy, that was considered negative. Nine of these biopsies yielded a sample that was nondiagnostic or insufficient, and the
=
chest tube.
decision to repeat the biopsy was obvious. The other four of these revealed a mix of inflammatory cells. Of the 140 confirmed benign lesions, a final FNAB diagnosis was found to be a specific benign diagnosis (hamartoma, granuloma, cores of fibrosis) in 55 patients, and a nonspecific benign diagnosis in 75 patients (a total of 130 true negatives), yielding a sensitivity for benign disease of 93%. Of the 75 patients with nonspecific benign changes, seven had a repeated biopsy confirming the previous results. Of the 22 patients with an unconfirmed final diagnosis, 16 had a FNAB diagnosis that is negative for malignancy; however, radiologic follow-up has been for less than 2 years. The overall diagnostic accuracy for both malignant and benign lesions combined is 94% with positive and negative predictive values of 99.5% and 90%, respectively. Complications included pneumothorax with and without chest tube placement, hemoptysis, and transient hypotension. Pneumothorax was seen in 175 (26.9%) patients with chest tube placement required in 60 (9.2%)of these patients. Of the patients with hemoptysis (3.1%) and transient hypotension (0.1%), none required additional treatment, transfusions, or hospitalization. Table 2 summarizes the number of needle passes performed with subsequent pneumothorax and chest tube placement rates. When these data were evaluated using the 2
test, no significant increase in the pneumothorax rate was seen in patients undergoing an increasing number of needle Dasses. Furthermore, in those patients with pneumothorax, an increasing number of needle passes did not contribute to an increase in the chest tube placement rate. Patients receiving four or more passes were grouped together because of the smaller sample size. The needle size used was then evaluated using the same statistical analysis. Specifically, needle size was correlated to pneumothorax and chest tube placement rates (Table 3). This shows that there was a significant increase in the rate of pneumothorax (P = .04) in those patients biopsied with 22guage needles; however, in patients with pneumothorax, no increase in chest tube la cement rate was seen with differint needle sizes. In addition, the same data were evaluated with use of the t test. In this instance, instead of looking for significant differences in pneumothorax rates, patients with pneumothorax and chest tube la cement were evaluated for any significant difference in needle size used or number of passes performed. Table 4 evaluates those patients with pneumothorax and compares them to those without pneumothorax. Once again, significantly smaller diameter needles were used when pneumothorax occurred (P = .012), corroborating previously mentioned observations. No significant differu
Percutaneous Transthoracic Needle Biopsy of the Lung
350
March-April 1998 JVIR
I
Table 3 Pneumothorax Rate* and Subsequent Chest Tube Placement Ratet Related to Needle Size Needle Size (gauge) 18 20 22
Bx (n
=
651)
Patients with PTX (n = 175)
Patients without PTX ( n = 476)
Patients with FTX and CT (n = 60)
Patients with PTX and without CT (n = 115)
2 66 107
21 205 250
0 18 42
2 48 65
23 271 357
* For 22 gauge versus 20 and 18 gauge 2 = 6.485, P = .04. 2 = 3.65, P = .16. Bx
=
number of biopsy procedures; PTX
ence in the number of needle passes was seen between those patients with and without pneumothorax. Table 5 evaluates those patients with pneumothorax and subsequent chest tube placement and shows that, unlike the previously mentioned data, significantly smaller needles were used in those patients with chest tube placement; however, the level of significance is a t a threshold value ( P = .05). No significant difference in the number of needle passes was seen between those patients with and without chest tube placement.
DISCUSSION Our overall accuracy (94%), sensitivity for malignancy (95%), and sensitivity for benign disease (93%) are comparable to those of other studies (4,6-9,12,21). Positive and negative predictive values are 99.5% and 90%, respectively. This confirms the efficacy of FNAB as the primary diagnostic tool in evaluating suspicious pulmonary lesions. Although its usefulness in detecting malignancy is widely accepted, the same cannot be said for its role in detecting benign disease. This is evident in the wide range (9%-91%) of benign diagnostic rates reported in the literature (4,5,8-12) and is most likely not due to differences in technique or ability but rather in how negative biopsy results have been interpreted. Lower diagnostic rates will invariably be based on FNAB results yielding a specific benign diagnosis, whereas higher rates will most often include
=
pneumothorax; CT
=
chest tube.
all results whether specific or nonspecific that are nonmalignant. The decision to include nonspecific benign results while calculating accuracy rates for benign disease undoubtedly will rely on the confidence that an adequate nonspecific negative biopsy will accurately predict benign disease. This confidence can be enhanced bv the use of various techniques inciuding the more frequent use of CT for small lesions (4,10), the presence of a cytopathologist (15,22), and repeated needle passes duplicating findings (6,7,9,21-24). In other words, a radiologically demonstrated, wellplaced needle yielding similar nonspecific benign cells as interpreted by an onsite cytopathologist can provide a result that reliably predicts the absence of malignancy. Therefore, it seems reasonable to conclude, as our data would support, that FNAB of the lung performed under current standards is also quite useful in the prediction of benign disease. This is a significant point considering it is the ability to accurately predict benign disease that most often obviates the need for surgery (6, 7, 22).
Complications encountered in our patient population included pneumothorax, hemoptysis, and one episode of transient hypotension. As expected, pneumothorax was by far the most common. Air embolism, biopsy tract tumor seeding, and death were not encountered. The frequency of pneumothorax (26.9%) and chest tube placement (9.2%) are within the expected range when compared with other large studies (4,7,9,12,25-28). The literature cites many factors contributing to an increasing pneumothorax rate including lesion size, lesion depth, number of needle passes, pleural surfaces crossed, and pulmonary function studies (16-19,25-30). Small lesion size, abnormal pulmonary function tests, and increasing lesion depths seem to be the most substantiated predictors of pneumothorax (25,26,28-30). When we evaluated the association between the number of needle passes with pneumothorax and chest tube placement rates, we found that no significant correlation exists. In reviewing the literature specifically for instances where formal statistical analysis was used in evaluating this associa-
Table 4 Significance of Needle Size and Number of Passes Related to Pneumothorax Variable
Patients with FTX ( n = 175)
Needle size gauge No. of passes
21.2 + 1.03 2.07 + 1.04
Note.-PTX
=
pneumothorax, NS
=
Patients without PTX (n = 476)
not significant.
21.0 2.26
+ 1.16 ?
1.07
P Value .012 NS
Swischuk et a1
351
Volume 9 Number 2
Table 5 Significance of Needle Size and Number of Passes Related to Chest Tube Placement in Patients with Pneumothorax Variable
Patients with CT (n = 60)
Paitents without CT (n = 115)
P Value
Needle size gauge Number of Passes
21.4 5 .92 1.93 + 1.13
21.0 5 1.07 2.25 +- 1.01
.05 NS
Note.-CT
=
chest tube, NS
=
7.
8.
not significant. 9.
tion, there is a single early report by Sinner et a1 (28), a study involving well over 2,000 patients, showing a significant increase in the pneumothorax rate when a second pass was performed within a 2-hour period. One shortcoming of this study, however, is that the exact number of patients represented by their statistics is uncertain. We have found no additional reports statistically supporting this positive correlation. In fact, since then there have been multiple reports, including ours, involving 1,828 patients showing that no correlation exists between the number of needle passes and pneumothorax rate (2426,29,30). We then evaluated needle size for any association with increasing pneumothorax and chest tube placement rates. Surprisingly, we found that there was a significant positive correlation between pneumothorax and possibly chest tube placement with the use of small diameter needles (22 gauge vs 20-18 gauge). Initially, this seems to make little sense; however, it is important to realize that although statistical analysis shows a correlative relationship, it does not necessarily prove cause and effect. In explaining this positive correlation, we do not necessarily believe that smalldiameter needles cause pneumothorax, but rather that this most likely indicates a bias or preference on our part for using smaller gauge needles for so-called "difficult" lesions. These would include smaller d e e ~ e lesions r in ~ a t i e n t swith known obstructive pulmonary disease, conditions which, as previously mentioned, are known to in-
crease the risk of pneumothorax. Perhaps a more practical conclusion is that the use of larger (ie, 18gauge) needles does not necessarily increase the risk of pneumothorax. In reviewing the literature, we found little information on the relationshi~between needle size and pneumothorax. In fact, only a single report by Sinner et a1 (28) shows no statistically significant correlation between needle size and pneumothorax rate. In conclusion, the data support the use of FNAB of the lung as the diagnostic procedure of choice in evaluating both malignant and benign disease of the lung. In addition, while striving to obtain an adequate and diagnostic FNAB specimen, multiple passes with various-sized needles (18-22 gauge) should not be cause for increased apprehension.
10.
11.
12.
13.
14.
15.
References 1. Leydon 00. Uber infektiose pneumonic. Dtsch Med Wochenschr 1883; 952. 2. Menetrier P. Cancer primitif du poumon. Bull Soc Anat Paris 1886; 11:643. 3. Nordenstrom B. A new technique for transthoracic biopsy of lung changes. Br J Radiol 1965; 38550553. 4. Westcott JL, Rao N, Colley DP. Transthoracic needle biopsy of small pulmonary nodules. Radiology 1997; 202:97-103. 5. Gasparini S, Ferretti M, Secchi EB, Baldelli S, Zuccatosta L, Gusella P. Integration of transbronchial and percutaneous approach in the diagnosis of peripheral pulmonary nodules or masses: experience with 1,027 cases. Chest 1995; 108:131-137. 6. Khouri NF, Stitik FP, Erozan YS, e t al. Transthoracic needle aspi-
16.
17. 18. 19. 20. 21.
ration biopsy of benign and malignant lung lesions. AJR 1985; 144: 281-288. Westcott JL. Direct percutaneous needle aspiration of localized pulmonary lesions: results in 422 patients. Radiology 1980; 137:31-35. Klein JS, Salomon G, Stewart EA. Transthoracic needle biopsy with a coaxially placed 20-gauge automated cutting needle: results in 122 patients. Radiology 1996; 198:715-720. Stanley JH, Fish GD, Andriole JG, e t al. Lung lesions: cytologic diagnosis by fine needle biopsy. Radiology 1987; 162:389-391. vansonnenberg E, Casola G, Ho M, et al. Difficult thoracic lesions: CT guided biopsy experience in 150 cases. Radiology 1988; 167:457-461. Stevens GM, Jackman RJ. Outpatient needle biopsy of the lung: its safety and utility. Radiology 1984; 151:301-304. Sargent EN, Turner AF,Gordonson J , Schwinn CP, Pashky 0. Percutaneous pulmonary needle biopsy: report of 350 patients. Am J Roentgenol Radium Ther Nucl Med 1974; 122:758-768. Wallace JM, Deutsch AL. Flexible fiberoptic bronchoscopy and percutaneous needle lung aspiration for evaluating the solitary pulmonary nodule. Chest 1982; 81: 665-671. Mark JB, Marglin SI, Castellino RA. The role of bronchoscopy and needle aspiration in the diagnosis of peripheral lung masses. J Thorac Cardiovasc Surg 1978; 76:266-268. Austin JH, Cohen MB. Value of having a cytopathologist present during percutaneous fine-needle aspiration biopsy of lung: report of 55 cancer patients and metaanalysis of the literature. AJR 1993; 160:175-177. Shepard JO. Complications of percutaneous needle aspiration biopsy of the chest: prevention and management. Semin Intervent Radiol 1994; 11:3. Weisbrod GL. Transthoracic needle biopsy. World J Surg 1993; 17:705711. Sanders C. Transthoracic needle aspiration. Clin Chest Med 1992; 13:ll-16. Westcott JL. Percutaneous transthoracic needle biopsy. Radiology 1988; 169593-601. Weisbrod GL. Transthoracic percutaneous lung biopsy. Radiol Clin North Am 1990; 28:647-655. Gobien RP, Valicenti J F , Paris BS, Daniel1 C. Thin-needle aspiration
352
Percutaneous Transthoracic Needle Biopsy of the Lung
March-April 1998 JVIR
biopsy: methods of increasing the accuracy of a negative prediction. Radiology 1982; 145:603-605. 22. Conces DJ, Schwenk GR, Doering PR, Glant MD. Thoracic needle biopsy: improved results utilizing a team approach. Chest 1987; 91:813-816. 23. Lalli AF, McCormack U,Zelch M, Reich NE, Belovich D. Aspiration biopsies of chest lesions. Radiology 1978; 127:35-40. 24. Williams AJ, Santiago S, Lehrman S, Popper R. Transcutaneous needle aspiration of solitary pulmonary masses: how many passes? Am Rev Respir Dis 1987; 136:452-454.
25. Miller KS, Fish GB, Stanley JH, Schabel SI. Prediction of pneumothorax rate in percutaneous needle aspiration of the lung. Chest 1988; 93:742-745. 26. Fish GB, Stanley JH, Miller KS, Schabel SI, Sutherland SE. Postbiopsy pneumothorax: estimating the risk by chest radiography and pulmonary function tests. AJR 1988; 150:71-74. 27. Berquist TH, Bailey PB, Cortese CA, Miller WE. Transthoracic needle biopsy: accuracy and complications in relation to location and type
of lesion. Mayo Clin Proc 1980; 55: 475-481. 28. Sinner WN. Complications of percutaneous transthoracic needle aspiration biopsy. Acta Radio1 [Diagnl 1976; 17:813-828. 29. Kazerooni EA, Lim FT, Mikhail A, Martinez FJ. Risk of pneumothorax in CT-guided transthoracic needle aspiration biopsy of the lung. Radiology 1996; 198:371-375. 30. Poe RH, Kallay MC, Wicks CM, Odoroff CL. Predicting risk of pneumothorax in needle biopsy of the lung. Chest 1984; 85:232-235.