- Email: [email protected]
CT-guided liver biopsy: An update
CLINICAL
ED LIV
radiology;
CT biopsies
From the Department of Radiology, Kangnam St. Mary’s Hospita: Seo Cho Ku. Seoul, Korea J3.K.H); Department of Radiology (P.B.S., !R.H.\. and Department of Pathology (F.A.K.), University kospitais of Cleve!and, Cleveland, Ohio. _Address reorint reaiiests to: lohn R. Haaaa, M.D., Director, Division of Diagnostic kadiology: University Hospitals of Cleveland, 2574 Abington Road. Cleveland, OH 44106. Received August 1990; revised February 1991. 7: ‘~21 hi’ i;lset-ier Science Publishing Co., Inc. 853 Awmue
of the Americas,
0899’7071;92i$3.50
New
York,
,VY 10010
?991;:.3:99-104
99
IOP
Computed tomography (CT)-guided liver biopsy has become an accepted part ofthe diagnostic evaluation of patients with benign and malignant hepatic disof imaging guidance over ease (X,2]. The advantages b!ind percutaneous biopsy are now well established m ?he literature (3, 4, 5).Controversy continues to exist, howeverV over the safety and accuracy of the procedure, as weii as over the proper choice of needle: large-caliber cutting needle versus small caliber aspiration needle. In I ~84, ourgroup reported the results in a preliminary set of 180 &patients in whom CT-guided biopsy of focal liver lesions was performed. The study indicated that cutting-needle sampling could provide in:reased diagnostic accuracy over fine-needle aspiration in both benign and malignant hepatic disease without a significant increase in complications [I]. This paper reports the results in a series of 179 new _pa!ients who/ underwent CT-guided liver biopsy using a variety of needles. Comparison is again made between fine-needle and cutting-needle biopsy with respect !o diagnostic accuracy and safety. A detailed discussion of the proper choice of biopsy needle as weli as of the causes of both biopsy failure and complications is provided.
Liver biopsy; Entcrve:ltiuna;
iMAGING
This study retrospectively reviewed !. 79 consecutive patients who underwent computed tomography (CT)guided liver biopsy for suspected focal benign or maIignant hepatic disease with either e or cutting needle between October 1983, and Ju 19@3, atour institution. Patients ranged in age from 1 to 90 years with a mean age of 60. Eighty-three patients were men and 96 were women. Forty patients underwent zutting-needle biopsy alone, 78 patients had only fine-needle aspiration and 49 underwent both fineand cutting-needle biopsy. Twelve patients were exeluded from the analysis because of absent or incomplele clinical data. All patients underwent a diagnosric contrast-enhanced scan at some time prior to the planned biopsy on either a Technicare Quantum scanner or a Siemens Somatom 19 scanner. The prothrombin time (PT), partial thromboplastin time [PTT), and platelet count were measured before the procedure. Fine-needle aspiration was undertaken regardless of the values, whiie cutting-needle biopsy was performed only if the PT was within z set of control, the PTT was less than two times normal, and the platelet count was greater than 50,000.At the time of the biopsy, preliminary scams were obtained 10 localize the lesion. Al! patients then received a 301~3 injection of 50 ml of Renografin-76. Lesions were considered “vascular” if they enhanced to the same degree as the aorta or had an appearance typical 3f hemangioma. Such lesions were excluded from cutting-needie biopsy, but aspirations were performed. Cutting needle biopsies were performed with either a 14 gauge Travenol Tru-Cut or 18 gauge Mengh:ni need!e. Fine-needle aspirations were performed with either 19 gauge Longdwei? (for suspected abscesses), 7.0gauge Chiba or F;aaga, or 2% gauge spinal needles. Trajectory was chosen so that normal liver was interposed between the lesion and the surface of
100
HA
ST AL.
FIGURE I. CT scan showing biopsy of liver kth needle with cutting gap.
Tru cut
the liver if at all possible. Between 1 and 4 passes were made per patient, with an average of 1.24 passes per patient with the cutting needles and 1.32 passes per patient with the aspiration needles. Cutting needle specimens were immediately fixed in formalin and sent to the Pathology laboratory. Aspirated material was smeared and immediately placed in absolute alcohol for cytologic evaluation. The needle was then rinsed in Saccamanno fixative and cytocentrifuge slides were prepared. Larger pieces of aspirated material, when present, were placed in 10% Formalin for routine histologic preparation. The specimens for patients undergoing biopsy with more than one needle type were reviewed separately. Results for malignancy were considered carrect if the report read “malignant cells present” or a specific tissue diagnosis was made. For benign disorders two types of results were considered. Generally, results were considered correct if the specimen was satisfactory for interpretation and the report read “no malignant cells present” or a specific diagnosis was provided. In patients with malignant disease, results were confirmed by clinical follow-up in 78 out of aOQ cases and surgery or autopsy in 22 of 100 cases. In patients with benign disease, results were confirmed by clinical follow-up (including drainage results) in 57 of 67 and by surgery or autopsy in 10 of 67 cases. Follow-up varied from 3 months to 4 years.
A total of 100 patients with malignant hepatic disease were biopsied. Sixty-one cutting and 67 aspiration biopsies were obtained. The overall diagnostic accu-
racy in malignant disease was 91.8% for cutting biopsy and 80.6% for aspiration biopsy. A specific diagnosis was obtained in 90.2% of cutting-needle biopsies and 77.6% of aspiration biopsies. Table 3 outlines these results and provides details on the specific entities encountered. A total of 67 benign lesions were biopsied. Twentysix cutting needle and 33 aspiration biopsies were obtained. The overall diagnostic accuracy was 100% for cutting biopsies, and a specific diagnosis was obtained in 88.5%. For aspiration biopsy the diagnostic: accuracy was 87.9%, although a specific diagnosis was made in only 39.4%, as reported in Table 2. We were able to collect a subset of 40 patients iwho underwent both aspiration and cutting needle biopsy and these results are tabulated in Tables 3, 4, and 5. ?‘or malignant disease the overall diagnostic accuracy was 87.1% for cutting biopsy (all of which represented a specific diagnosis) and 77.5% for aspiration biopsy (with a specific diagnosis obtainable in 71%). ?;or benign disease, the correct diagnosis was made in 100% of patients who underwent cutting biopsy {in 88.9% a specific diagnosis was made). In this subset of patients, aspiration biopsy was 88.9% accurate (a specific diagnosis was made in only 44.4%). We were able to evaluate the role of histologic versus cytologic specimens in patients undergoing fine-needle aspiration. In malignant disease, histology alone was correct in 7 of 27 patients, cytology alone in 5 of 27, and both were accurate in 12 of 27 3ases. Thus, cytology had an overall accuracy of 55% and histology of Sl%, with a combined accuracy of 77.4%. In benign entities, histology alone was correct in 1 of 9 patients, cytology alone in 6 of 9, and both in 1 of 9 patients. Therefore, the accuracy of cytology was 78% versus 22% for histology with a combined accuracy of 89%. These results are summarized in Table 4. Careful review of the clinical and radiologic rezords has allowed us to evaluate the cause of biopsy failure (Table 5). Utilizing 14-gauge nee opsies failed because of the very small size (less t 5 mm) of the lesions resulting in sampling error one failed because of poor technique. In addition poor technique in 2 cases and small lesion size in 2, 20-gauge aspiration failed because of a bloody sample rendering the sample inadequate for interpretation in 3 cases and insufficient tissue obtained in 7. In 9 zases of aspiration biopsy in which the pathologist reported a bloody sample (Table 6), cytology was adequate for interpretation in 3, istology was mde-, quate in 7, but the correct diagnosis was made only in those 3 patients who had adequate cytology specimens.
APRIL-TUNE
1991
kE 1. Ma!ipnani
Lesion
in Patients
with
Single Cutting
Biopsy Biopsy
Aspiration
____-
Correct
Correct Nonspecific
Specific
Zliagnosis Metastasis (83) Hepatoma (9) Zymphoma(4) Uemangioendothelioma(2) hlesenchymal Sarcoma(l) Cholangioca (1) Subtotal (100) Total
Ii47
42147
incorrect
Specific
214 l/l
414
l/2
l/2
212 l/l
l/l
55161
(90.2%)
56161
(91.8%)
1161 (1.6%)
5161 (8.2%)
52167
(77.6%)
5i61
58167
(80.6%)
(8.2%)
Necrosis infarct
Specific
biopsy
Aspiration
(19.4%
biopsy
Specific
1
Nonspecific
2
3
Incorrect
4 Al
2
3 2
(2) 2 2
3
1
3
1
3n
(6)
Normai (3) Kemangioma Cyst (111”
Subtotal Total
Incorrect
(2)
Lesion
(3.0%)
Correct Nonspecific
8
Primary Biliary Cirrhosis (2) F;brosis in Wedge Resection Site (l] Nonspecific Benign
Abscess
2/67
:n 132 cases with no major complications and three minor complications (2.3%), In one patient, surgery was required following biopsy but bleeding was pri_narily due to spontaneous rupture of a hepatoma. One patient developed a small subcapsular hemacorna and one developed abdominal and right shoulder pain, none of which required treatment. There were no deaths associated with CT-guided liver biopsy. A number of high-risk patients had biopsies. Sixteen patients with coagulopathies underwent CTguided biopsy following treatment with a Factor VIII,
Correct
(11)
13167 (19.4%
with Single Biopsy Cutting
Fatty Metamorphosis Liver Cirrhosis (4) Hepatitis (31) Focal I-Iepatoceliuiar
Incorrect 8159
Z/59
2i4
l/7
Lesion in Patients
Diagnosis
Nonspecific
49159
4i47
617
7 summarizes the complications seen in this Tabie study population. A major complication was defined as one requiring further therapy or intervention. Only (l.l%), one patient developed a major complication 6 hepattc hematoma that developed four days following biopsy and on readmission required transfusion. One patient developed a small subcapsular hematoma, and one a small extraperitoneal hematoma, both v?rere asymptomatic and neither required transfusion Two patients experienced a transient drop in blood pressure requiring no additional therapy. Passes were made with 20- or Z-gauge aspirations
IGABLE %. Benign ---
Biopsy
(6)
1
1
23126
(88.5%)
3
3
26126
(100%
4
3
)
1
(16)“l
1 3126 (11.5%)
’ Four cases are unsatisfactory. ’ k’onspecific benign lesion represents the benign lesion in clinical follow-up LThree cases were performed with IS-gauge Longdwell needle. d Fourteen cases were performed with 19-gauge Longdwell needle.
13/33
(39.4%)
although specific diagnosis cannot
16133
(48.5%)
4133 (12.1%)
29133
(87.9%)
4/33
be made patfiologically.
(12.1%)
102
HA ET AL.
LE 3. Patients
with Cutting
and Aspiration
Biopsies Cutting
biopsy
Aspiration
Correct Specific
Metastasis (26) Hepatoma (21 Lymphoma (1) Hemangiomaendothelioma Normal (11 Fatty infiltration (4) Liver cirrhosis (2) Hemangioma (1)
Unknown Benign Lesion (1) Subtotal Malignant (31) Total Subtotal Benign (9) Total Benign n In “Incorrect,”
unsatisfactory
Correct Ncnspecific
Incorrect
23 1 1 2 1 4 2 1
(2)
3 1
IncorrecP
1 1
4 1
1 1
1 1 1 2
3 1
27131 (87.1%) 819 (88.9%) 9/s (100%)
219 (11.1%)
4131 (12.9%)
22131 (71.0%) 419 (44.4%) 819 (88.4%)
2131 ( 6.5%) 419 (44.4%)
1 7131 (22.5%) 219 (11.1%) 219 (11.1%)
are included.
data support the contention of several previous authors that hemangioma can be safely sam fine needle (8, 9, 10).Six patients unde needle sampling of hemangioma without complications. In one patient in whom contrast ~~~a~~e~e~t was missed on the prebiopsy bolus study, a small subcapsular hematoma developed r sampling with a 14gauge needle (although no tional treat.. ment was required). In patients with known primary hepaiic or other lignancies and suspected he atic metastases, ever, the current data support t e contention that fine-needle aspiration is essentially comparable to cutting-needle biopsy in both accuracy and specifkity in diagnosing malignancy with the exception of lymphoma. In considering all patients studied an that subset of patients who e sampled with bot types of needle aspiration fine needle provide an accurate diagnosis of malignancy in apprsximately 80% of patients, nearly all specific, versus 90% with cutting-needle biopsy. This represents a considerable increase in diagnostic accuracy for fineneedle aspiration over our previous series, in which an accurate diagnosis -\~iasonly made in fewer than
of Cytology
and Histology
in -40 Palients
Who Had Both Biopsies Aspiration
Cutting biopsy
Benignancy
Nonspecific
1
In our initial series comparing fine-with cutting-needle biopsy of focal hepatic disease, we advocated the nse of cutting needles in both benign and malignant processes, allowing for increased diagnostic accuracy without a significant increase in complications. (1). This represented a departure from previous reports, which favored fine-needle sampling for all focal hepatic disease (2, 6, 7). The current data clearly support the use of cutting needles in unknown or suspected benign disease (i.e., patients without previously documented malignancy). Although aspiration biopsy resulted in an accurate diagnosis of benignity in 44.4% of patients, a specific diagnosis of benignity was made in 100% of patients and a specific diagnosis was made in nearly 90% of these. Thus, the increased diagnostic accuracy and specificity provided in benign disease is well worth the minimal increase in complications. The obvious exception to this is in patients with suspected hemangioma. These
Accuracy rate Malignancy
Specific 21
cryoprecipitate, fresh frozen plasma, platelets or a combination thereof and none of these patients expe-
LE 4. Comparison
biopsy
Histology (no. of correct cases)
Histology
27131 (87.1%) 919 (100%)
19131 (61.3%) 219 (22.2%)
biopsy
Cytology 17131 719
(54.8%) (77.8%)
Total [no. of cases] 24131 8i9
(77.4%) (08.9%)
X 5
TA
Caims
of Biopsy Failures lb-gauge needle
Cause E!ocdy sample Poor localizaion techniaue Biologic samplixg error Insufficient tissue TGta! patients
1 2 3
. Cytologic Z&gauge needle
Succsssfcl cytology
6 2 2 7 17
15).
of Needle
No. of needle passes” 1
Passes and
319
-.
lbgauge
x
Correct final diagnosis
719
Complications needle
20-gauge ard 22-gauge needle”
61 9 3 1 92
2X 3X /-x Tom! passes
3f “Bloodiy”
zutting or Sne needle) as well as ill Ihe su patients who underwent biopsy with both fine and cutting needles. Although there is no significant difference between the two groups, the subset of patients allows a unique opportunity to evaluate the role of both cytology and histology in fine-needle aspiration. The data strongly support the contention that both istologic specimens are crucial to ac‘iagnosis using skinny edEes (X3,17: l&19). isease, the combination of cytologic and histologic specimens signifizantly improved the diagnostic accuracy over either one alone. The complication rate reported ere in comparable with that reported by ’ authors (2, 6, 7). Only 1 patient out of 179 die ed a major complication hat required only tra n for stabilization. Previous authors have discussed the role of larger needle size and number of needle passes in predisposing to complications (4, 20). There was no identifiable correlation between the number of needle passes and the incidence of complications, as one-half of the 3mplications were seen in patients undergoing a single pass. Additional factors may include the vaszularity of the lesion, the patient’s condition, and the technique used by the operator. One of the minor complications seen in the e group was due 10 erroneous use of that n a patient with a hemangioma. One delayed c ication (subcapsuYarhematoma occurring fo ostbic3ps y requir-
We were able to evaluate the data in our overall group of patients [some of whom were biopsied with
&E 5. Number
XesiAts
Sxcessful histology
319
70”/. d patiems. We believe that this improvement is the result of three factors: increased experience on the part of the operators; increased experience on the part of the cytologist; increased use of the 20-gauge er than the 22-gauge needle. Our group bas previously reported the improved sampling characteristics provided by the 20-gauge needle as comTherefore, in patients with pared to the 22-gauge (11). a known primary mahgnancy and suspected hepatic metastates, aspiration biopsy should be the initial rocedure of choice. Only if aspiration cutting-needle biopsy be undertaken. The cutting needle provides several advantages cpiration in selected circumover fine-needle isease it provides the increased stances. In benign specificity of diagnosis required for proper clinical care. In patients with suspected lymphoma it allows for cell subtyping (12, IS]. In patients undergoing long-term chemotherapy, lesions often undergo central cystic change, necrosis, or fibrosis and specimens from the periphery or wall of the lesion are often necessary for diagnosis. As the l&gauge needle can obtain a specimen 2 cm in length, it provides a more Be because it can be used to traverse the ab~or~a~i~y, the transition zone, and normal liver (14,
and Yistologic
Samples
71 27 1 1 132
CompXcatior, major
1
XLlRO~
Complication
(1.1%) (4.3%) 5/92 (5.4%]
0 (0.0%) 3 (2.3%) 31132 (2.3%)
4
rate
o The number of patienb
using 22g. needle are 7 cases. “The SUCCBSS rates associated with needle pass (1 X, 2X, 3X, and 4 x 1 are 59161 (96.7%j, 59171 (83.i%J,21127 (77.6%).l/l (100%) and l/l (100%) in ZO-gaugeand 22.gauge needle.
719 (7?.8%),
313 (100%)
and I/I (loo%] in 14-gauge needle, and
104
KAET
AL.
CLINICAL IMAGING '"JOL.15, X0. 2
ingsurgery)
did occur, a problem previously reported in blind percutaneous liver biopsy (21).Thus, patients should be followed closely for a period after biopsy. Additionally, 16 patients with coagulopathies were safely biopsied following plasma replacement. Our analysis of the reasons for biopsy failure suggests that this is a product of insufficient tissue, bloody samples, and small lesion size (~5 mm) causing sampling error in fine-needle aspiration. The “bloody sample” continued to be a hindrance to accurate evaluation of fine-needle specimens. In 9 such cases, a bloody sample was obtained and in only 3 of 9 was the cytologic sample adequate for interpretation. Although Saccamanno fixative provided an adequate histologic specimen in 4 additional cases, t ogy did not permit a correct diagnosis to be made. The use of heparinized skinny needles may, in fact, improve diagnostic accuracy as suggested by Kasugai (16).In cutting-needle biopsies, the small size resulting in sampling error seem.s to be the relevant factor. Small lesions are difficult to sample and multiple passes are still prone to fail and are perhaps unwarranted. Interestingly, however, our data suggest that additional passes do not improve the outcome. Although the numbers are small, the success rate with one pass was 97% and with two passes 78% for X&gauge needles and 83% and 78% respectively for fine needles. With careful preparation and using the guidelines outlined above, CT-guided liver biopsy can provide extremely accurate and safe diagnosis of both benign and malignant focal hepatic disease.
4. Alspaugh
J” Bernardino M, Seweli C, Sones P, Be&man W’: Price R: CT directed hepatic biopsies: increased diagnostic accuracy with low patient risk, J Comput Assist Tomogr 1983: 7:1012-1017,
5. Haaga J, Vanek J: Computed tomographic guided iiv~eerbiopsy using the Menghini needle, Radiology 1979; :33:405406. 6. Isler R, Ferrucci J, Wittenberg J, et al.: Tissue core biopsy of abdominal tumors with a 22 gauge cutting needle, AJR 1981; 136:727-728. 7. FerrucciJ, Wittenberg J, Mueller P, et al.: Diagnosis of abdominal malignancy by radiologic fine-needle aspiration biopsy, AJK 1979; 134:323-330. 8. Cronan J, Esparza A, Dertman G; Ridlen M, Paolelia L: Cavern QUS hemangioma of the liver: role of percutaneous biopsy, Radiolgy 1988; 166:135-138. 9 Solbati L, Lirraghi T, De Pra L, Ieraco T, Masciadr N, Ravetto C: Fine-needle biopsy of hepatic hemangioma with sonographic guidance, AJR 1985; 144:474-477. 10. Whitlach S, Nunez C, Pitlik D: Fine needle aspiration of the liver, Acta Cytol 1984; 28:719-725. 11.
biopsy
AdrioIeJ, Haaga J, Adams R, Nunez C: Biopsy needle characteristics assessed in the laboratory, Radiology 1983; 148:659-662.
12. Hajdu S, Melamed M: Limitations of aspiration cytology in the diagnosis of primary neoplasms, Acta Cytol 1984; 28:337345. 13. Lieberman R. Hafez G, Crummy A: Histology from aspiration biopsy: Turner needle experience, 1982; 1~~561-564. 14. Bateson M, Hopwood D, Duguid H, Boucher I: A comparative trial of liver biopsy needles, J Clin Path01 iaao; 33:131133. 15. Littlewood E, Gilgore I, Murray-Lyon I, Stephens K, Paradinas F: Comparison of the Tru-Cut and Sure-Cut live; biopsy needles, J Clin Path01 1982; 35:761-763. 16. Kasugai H, Yamamoto lR, Tatsuta M, et al.: Value of iheparinized fine-needle aspiration biopsy in liver malignancy, AJR 1985; 144:243-244. 17. Bell D, Carr C, Szyfelbein W: Fine needle aspiration of focal liver lesions, Acta Cytol 1986; 30:397-402,
cytology
18. Atterbury C, Enriquez R, Desoto-Nagy 6, Conn H: Comparison of the histologic and cytologic diagnosis of liver biopsies in hepatic cancer, Gastroenterology 1979; 76:1352-1357.
Martin0 6, Haaga J, Bryan I’, LiPuma J, El Yousef S, Alfidi R: CT-guided liver biopsies: eight years’ experience, Rediology 1984;152:755-757.
19. Tatsuta M, Yamamoto R, Kasugai H, et al.: Cytohistologic diagnosis of neoplasms of the liver by ultrasonical!y guided fineneedle aspiration biopsy, Cancer 1984; 54:16821686.
1983;
20. Ho C, McLaughlin M, Tao L, Blendis L, Evans W: Guided percutaneous fine-needle aspiration biopsy of the liver, Cancer 1981;47:1701-1783.
Wlitmire L, Galambos J, Phillips V, et al.: Imaging guided percutaneous hepatic biopsy; diagnostic accuracy and safety, J Glin Gastroenterol 1985; 7:511-515.
21. Reichert C, Weisenthal L, Klein H: Delayed hemorrhage, after percutaneous liver biopsy, J Clin Gastroenterol 1983; 5:263-266.
Pagani J: Biopsy 147:673-675.
of focal hepatic
lesions,
Radiology
ED LIV
radiology;
CT biopsies
From the Department of Radiology, Kangnam St. Mary’s Hospita: Seo Cho Ku. Seoul, Korea J3.K.H); Department of Radiology (P.B.S., !R.H.\. and Department of Pathology (F.A.K.), University kospitais of Cleve!and, Cleveland, Ohio. _Address reorint reaiiests to: lohn R. Haaaa, M.D., Director, Division of Diagnostic kadiology: University Hospitals of Cleveland, 2574 Abington Road. Cleveland, OH 44106. Received August 1990; revised February 1991. 7: ‘~21 hi’ i;lset-ier Science Publishing Co., Inc. 853 Awmue
of the Americas,
0899’7071;92i$3.50
New
York,
,VY 10010
?991;:.3:99-104
99
IOP
Computed tomography (CT)-guided liver biopsy has become an accepted part ofthe diagnostic evaluation of patients with benign and malignant hepatic disof imaging guidance over ease (X,2]. The advantages b!ind percutaneous biopsy are now well established m ?he literature (3, 4, 5).Controversy continues to exist, howeverV over the safety and accuracy of the procedure, as weii as over the proper choice of needle: large-caliber cutting needle versus small caliber aspiration needle. In I ~84, ourgroup reported the results in a preliminary set of 180 &patients in whom CT-guided biopsy of focal liver lesions was performed. The study indicated that cutting-needle sampling could provide in:reased diagnostic accuracy over fine-needle aspiration in both benign and malignant hepatic disease without a significant increase in complications [I]. This paper reports the results in a series of 179 new _pa!ients who/ underwent CT-guided liver biopsy using a variety of needles. Comparison is again made between fine-needle and cutting-needle biopsy with respect !o diagnostic accuracy and safety. A detailed discussion of the proper choice of biopsy needle as weli as of the causes of both biopsy failure and complications is provided.
Liver biopsy; Entcrve:ltiuna;
iMAGING
This study retrospectively reviewed !. 79 consecutive patients who underwent computed tomography (CT)guided liver biopsy for suspected focal benign or maIignant hepatic disease with either e or cutting needle between October 1983, and Ju 19@3, atour institution. Patients ranged in age from 1 to 90 years with a mean age of 60. Eighty-three patients were men and 96 were women. Forty patients underwent zutting-needle biopsy alone, 78 patients had only fine-needle aspiration and 49 underwent both fineand cutting-needle biopsy. Twelve patients were exeluded from the analysis because of absent or incomplele clinical data. All patients underwent a diagnosric contrast-enhanced scan at some time prior to the planned biopsy on either a Technicare Quantum scanner or a Siemens Somatom 19 scanner. The prothrombin time (PT), partial thromboplastin time [PTT), and platelet count were measured before the procedure. Fine-needle aspiration was undertaken regardless of the values, whiie cutting-needle biopsy was performed only if the PT was within z set of control, the PTT was less than two times normal, and the platelet count was greater than 50,000.At the time of the biopsy, preliminary scams were obtained 10 localize the lesion. Al! patients then received a 301~3 injection of 50 ml of Renografin-76. Lesions were considered “vascular” if they enhanced to the same degree as the aorta or had an appearance typical 3f hemangioma. Such lesions were excluded from cutting-needie biopsy, but aspirations were performed. Cutting needle biopsies were performed with either a 14 gauge Travenol Tru-Cut or 18 gauge Mengh:ni need!e. Fine-needle aspirations were performed with either 19 gauge Longdwei? (for suspected abscesses), 7.0gauge Chiba or F;aaga, or 2% gauge spinal needles. Trajectory was chosen so that normal liver was interposed between the lesion and the surface of
100
HA
ST AL.
FIGURE I. CT scan showing biopsy of liver kth needle with cutting gap.
Tru cut
the liver if at all possible. Between 1 and 4 passes were made per patient, with an average of 1.24 passes per patient with the cutting needles and 1.32 passes per patient with the aspiration needles. Cutting needle specimens were immediately fixed in formalin and sent to the Pathology laboratory. Aspirated material was smeared and immediately placed in absolute alcohol for cytologic evaluation. The needle was then rinsed in Saccamanno fixative and cytocentrifuge slides were prepared. Larger pieces of aspirated material, when present, were placed in 10% Formalin for routine histologic preparation. The specimens for patients undergoing biopsy with more than one needle type were reviewed separately. Results for malignancy were considered carrect if the report read “malignant cells present” or a specific tissue diagnosis was made. For benign disorders two types of results were considered. Generally, results were considered correct if the specimen was satisfactory for interpretation and the report read “no malignant cells present” or a specific diagnosis was provided. In patients with malignant disease, results were confirmed by clinical follow-up in 78 out of aOQ cases and surgery or autopsy in 22 of 100 cases. In patients with benign disease, results were confirmed by clinical follow-up (including drainage results) in 57 of 67 and by surgery or autopsy in 10 of 67 cases. Follow-up varied from 3 months to 4 years.
A total of 100 patients with malignant hepatic disease were biopsied. Sixty-one cutting and 67 aspiration biopsies were obtained. The overall diagnostic accu-
racy in malignant disease was 91.8% for cutting biopsy and 80.6% for aspiration biopsy. A specific diagnosis was obtained in 90.2% of cutting-needle biopsies and 77.6% of aspiration biopsies. Table 3 outlines these results and provides details on the specific entities encountered. A total of 67 benign lesions were biopsied. Twentysix cutting needle and 33 aspiration biopsies were obtained. The overall diagnostic accuracy was 100% for cutting biopsies, and a specific diagnosis was obtained in 88.5%. For aspiration biopsy the diagnostic: accuracy was 87.9%, although a specific diagnosis was made in only 39.4%, as reported in Table 2. We were able to collect a subset of 40 patients iwho underwent both aspiration and cutting needle biopsy and these results are tabulated in Tables 3, 4, and 5. ?‘or malignant disease the overall diagnostic accuracy was 87.1% for cutting biopsy (all of which represented a specific diagnosis) and 77.5% for aspiration biopsy (with a specific diagnosis obtainable in 71%). ?;or benign disease, the correct diagnosis was made in 100% of patients who underwent cutting biopsy {in 88.9% a specific diagnosis was made). In this subset of patients, aspiration biopsy was 88.9% accurate (a specific diagnosis was made in only 44.4%). We were able to evaluate the role of histologic versus cytologic specimens in patients undergoing fine-needle aspiration. In malignant disease, histology alone was correct in 7 of 27 patients, cytology alone in 5 of 27, and both were accurate in 12 of 27 3ases. Thus, cytology had an overall accuracy of 55% and histology of Sl%, with a combined accuracy of 77.4%. In benign entities, histology alone was correct in 1 of 9 patients, cytology alone in 6 of 9, and both in 1 of 9 patients. Therefore, the accuracy of cytology was 78% versus 22% for histology with a combined accuracy of 89%. These results are summarized in Table 4. Careful review of the clinical and radiologic rezords has allowed us to evaluate the cause of biopsy failure (Table 5). Utilizing 14-gauge nee opsies failed because of the very small size (less t 5 mm) of the lesions resulting in sampling error one failed because of poor technique. In addition poor technique in 2 cases and small lesion size in 2, 20-gauge aspiration failed because of a bloody sample rendering the sample inadequate for interpretation in 3 cases and insufficient tissue obtained in 7. In 9 zases of aspiration biopsy in which the pathologist reported a bloody sample (Table 6), cytology was adequate for interpretation in 3, istology was mde-, quate in 7, but the correct diagnosis was made only in those 3 patients who had adequate cytology specimens.
APRIL-TUNE
1991
kE 1. Ma!ipnani
Lesion
in Patients
with
Single Cutting
Biopsy Biopsy
Aspiration
____-
Correct
Correct Nonspecific
Specific
Zliagnosis Metastasis (83) Hepatoma (9) Zymphoma(4) Uemangioendothelioma(2) hlesenchymal Sarcoma(l) Cholangioca (1) Subtotal (100) Total
Ii47
42147
incorrect
Specific
214 l/l
414
l/2
l/2
212 l/l
l/l
55161
(90.2%)
56161
(91.8%)
1161 (1.6%)
5161 (8.2%)
52167
(77.6%)
5i61
58167
(80.6%)
(8.2%)
Necrosis infarct
Specific
biopsy
Aspiration
(19.4%
biopsy
Specific
1
Nonspecific
2
3
Incorrect
4 Al
2
3 2
(2) 2 2
3
1
3
1
3n
(6)
Normai (3) Kemangioma Cyst (111”
Subtotal Total
Incorrect
(2)
Lesion
(3.0%)
Correct Nonspecific
8
Primary Biliary Cirrhosis (2) F;brosis in Wedge Resection Site (l] Nonspecific Benign
Abscess
2/67
:n 132 cases with no major complications and three minor complications (2.3%), In one patient, surgery was required following biopsy but bleeding was pri_narily due to spontaneous rupture of a hepatoma. One patient developed a small subcapsular hemacorna and one developed abdominal and right shoulder pain, none of which required treatment. There were no deaths associated with CT-guided liver biopsy. A number of high-risk patients had biopsies. Sixteen patients with coagulopathies underwent CTguided biopsy following treatment with a Factor VIII,
Correct
(11)
13167 (19.4%
with Single Biopsy Cutting
Fatty Metamorphosis Liver Cirrhosis (4) Hepatitis (31) Focal I-Iepatoceliuiar
Incorrect 8159
Z/59
2i4
l/7
Lesion in Patients
Diagnosis
Nonspecific
49159
4i47
617
7 summarizes the complications seen in this Tabie study population. A major complication was defined as one requiring further therapy or intervention. Only (l.l%), one patient developed a major complication 6 hepattc hematoma that developed four days following biopsy and on readmission required transfusion. One patient developed a small subcapsular hematoma, and one a small extraperitoneal hematoma, both v?rere asymptomatic and neither required transfusion Two patients experienced a transient drop in blood pressure requiring no additional therapy. Passes were made with 20- or Z-gauge aspirations
IGABLE %. Benign ---
Biopsy
(6)
1
1
23126
(88.5%)
3
3
26126
(100%
4
3
)
1
(16)“l
1 3126 (11.5%)
’ Four cases are unsatisfactory. ’ k’onspecific benign lesion represents the benign lesion in clinical follow-up LThree cases were performed with IS-gauge Longdwell needle. d Fourteen cases were performed with 19-gauge Longdwell needle.
13/33
(39.4%)
although specific diagnosis cannot
16133
(48.5%)
4133 (12.1%)
29133
(87.9%)
4/33
be made patfiologically.
(12.1%)
102
HA ET AL.
LE 3. Patients
with Cutting
and Aspiration
Biopsies Cutting
biopsy
Aspiration
Correct Specific
Metastasis (26) Hepatoma (21 Lymphoma (1) Hemangiomaendothelioma Normal (11 Fatty infiltration (4) Liver cirrhosis (2) Hemangioma (1)
Unknown Benign Lesion (1) Subtotal Malignant (31) Total Subtotal Benign (9) Total Benign n In “Incorrect,”
unsatisfactory
Correct Ncnspecific
Incorrect
23 1 1 2 1 4 2 1
(2)
3 1
IncorrecP
1 1
4 1
1 1
1 1 1 2
3 1
27131 (87.1%) 819 (88.9%) 9/s (100%)
219 (11.1%)
4131 (12.9%)
22131 (71.0%) 419 (44.4%) 819 (88.4%)
2131 ( 6.5%) 419 (44.4%)
1 7131 (22.5%) 219 (11.1%) 219 (11.1%)
are included.
data support the contention of several previous authors that hemangioma can be safely sam fine needle (8, 9, 10).Six patients unde needle sampling of hemangioma without complications. In one patient in whom contrast ~~~a~~e~e~t was missed on the prebiopsy bolus study, a small subcapsular hematoma developed r sampling with a 14gauge needle (although no tional treat.. ment was required). In patients with known primary hepaiic or other lignancies and suspected he atic metastases, ever, the current data support t e contention that fine-needle aspiration is essentially comparable to cutting-needle biopsy in both accuracy and specifkity in diagnosing malignancy with the exception of lymphoma. In considering all patients studied an that subset of patients who e sampled with bot types of needle aspiration fine needle provide an accurate diagnosis of malignancy in apprsximately 80% of patients, nearly all specific, versus 90% with cutting-needle biopsy. This represents a considerable increase in diagnostic accuracy for fineneedle aspiration over our previous series, in which an accurate diagnosis -\~iasonly made in fewer than
of Cytology
and Histology
in -40 Palients
Who Had Both Biopsies Aspiration
Cutting biopsy
Benignancy
Nonspecific
1
In our initial series comparing fine-with cutting-needle biopsy of focal hepatic disease, we advocated the nse of cutting needles in both benign and malignant processes, allowing for increased diagnostic accuracy without a significant increase in complications. (1). This represented a departure from previous reports, which favored fine-needle sampling for all focal hepatic disease (2, 6, 7). The current data clearly support the use of cutting needles in unknown or suspected benign disease (i.e., patients without previously documented malignancy). Although aspiration biopsy resulted in an accurate diagnosis of benignity in 44.4% of patients, a specific diagnosis of benignity was made in 100% of patients and a specific diagnosis was made in nearly 90% of these. Thus, the increased diagnostic accuracy and specificity provided in benign disease is well worth the minimal increase in complications. The obvious exception to this is in patients with suspected hemangioma. These
Accuracy rate Malignancy
Specific 21
cryoprecipitate, fresh frozen plasma, platelets or a combination thereof and none of these patients expe-
LE 4. Comparison
biopsy
Histology (no. of correct cases)
Histology
27131 (87.1%) 919 (100%)
19131 (61.3%) 219 (22.2%)
biopsy
Cytology 17131 719
(54.8%) (77.8%)
Total [no. of cases] 24131 8i9
(77.4%) (08.9%)
X 5
TA
Caims
of Biopsy Failures lb-gauge needle
Cause E!ocdy sample Poor localizaion techniaue Biologic samplixg error Insufficient tissue TGta! patients
1 2 3
. Cytologic Z&gauge needle
Succsssfcl cytology
6 2 2 7 17
15).
of Needle
No. of needle passes” 1
Passes and
319
-.
lbgauge
x
Correct final diagnosis
719
Complications needle
20-gauge ard 22-gauge needle”
61 9 3 1 92
2X 3X /-x Tom! passes
3f “Bloodiy”
zutting or Sne needle) as well as ill Ihe su patients who underwent biopsy with both fine and cutting needles. Although there is no significant difference between the two groups, the subset of patients allows a unique opportunity to evaluate the role of both cytology and histology in fine-needle aspiration. The data strongly support the contention that both istologic specimens are crucial to ac‘iagnosis using skinny edEes (X3,17: l&19). isease, the combination of cytologic and histologic specimens signifizantly improved the diagnostic accuracy over either one alone. The complication rate reported ere in comparable with that reported by ’ authors (2, 6, 7). Only 1 patient out of 179 die ed a major complication hat required only tra n for stabilization. Previous authors have discussed the role of larger needle size and number of needle passes in predisposing to complications (4, 20). There was no identifiable correlation between the number of needle passes and the incidence of complications, as one-half of the 3mplications were seen in patients undergoing a single pass. Additional factors may include the vaszularity of the lesion, the patient’s condition, and the technique used by the operator. One of the minor complications seen in the e group was due 10 erroneous use of that n a patient with a hemangioma. One delayed c ication (subcapsuYarhematoma occurring fo ostbic3ps y requir-
We were able to evaluate the data in our overall group of patients [some of whom were biopsied with
&E 5. Number
XesiAts
Sxcessful histology
319
70”/. d patiems. We believe that this improvement is the result of three factors: increased experience on the part of the operators; increased experience on the part of the cytologist; increased use of the 20-gauge er than the 22-gauge needle. Our group bas previously reported the improved sampling characteristics provided by the 20-gauge needle as comTherefore, in patients with pared to the 22-gauge (11). a known primary mahgnancy and suspected hepatic metastates, aspiration biopsy should be the initial rocedure of choice. Only if aspiration cutting-needle biopsy be undertaken. The cutting needle provides several advantages cpiration in selected circumover fine-needle isease it provides the increased stances. In benign specificity of diagnosis required for proper clinical care. In patients with suspected lymphoma it allows for cell subtyping (12, IS]. In patients undergoing long-term chemotherapy, lesions often undergo central cystic change, necrosis, or fibrosis and specimens from the periphery or wall of the lesion are often necessary for diagnosis. As the l&gauge needle can obtain a specimen 2 cm in length, it provides a more Be because it can be used to traverse the ab~or~a~i~y, the transition zone, and normal liver (14,
and Yistologic
Samples
71 27 1 1 132
CompXcatior, major
1
XLlRO~
Complication
(1.1%) (4.3%) 5/92 (5.4%]
0 (0.0%) 3 (2.3%) 31132 (2.3%)
4
rate
o The number of patienb
using 22g. needle are 7 cases. “The SUCCBSS rates associated with needle pass (1 X, 2X, 3X, and 4 x 1 are 59161 (96.7%j, 59171 (83.i%J,21127 (77.6%).l/l (100%) and l/l (100%) in ZO-gaugeand 22.gauge needle.
719 (7?.8%),
313 (100%)
and I/I (loo%] in 14-gauge needle, and
104
KAET
AL.
CLINICAL IMAGING '"JOL.15, X0. 2
ingsurgery)
did occur, a problem previously reported in blind percutaneous liver biopsy (21).Thus, patients should be followed closely for a period after biopsy. Additionally, 16 patients with coagulopathies were safely biopsied following plasma replacement. Our analysis of the reasons for biopsy failure suggests that this is a product of insufficient tissue, bloody samples, and small lesion size (~5 mm) causing sampling error in fine-needle aspiration. The “bloody sample” continued to be a hindrance to accurate evaluation of fine-needle specimens. In 9 such cases, a bloody sample was obtained and in only 3 of 9 was the cytologic sample adequate for interpretation. Although Saccamanno fixative provided an adequate histologic specimen in 4 additional cases, t ogy did not permit a correct diagnosis to be made. The use of heparinized skinny needles may, in fact, improve diagnostic accuracy as suggested by Kasugai (16).In cutting-needle biopsies, the small size resulting in sampling error seem.s to be the relevant factor. Small lesions are difficult to sample and multiple passes are still prone to fail and are perhaps unwarranted. Interestingly, however, our data suggest that additional passes do not improve the outcome. Although the numbers are small, the success rate with one pass was 97% and with two passes 78% for X&gauge needles and 83% and 78% respectively for fine needles. With careful preparation and using the guidelines outlined above, CT-guided liver biopsy can provide extremely accurate and safe diagnosis of both benign and malignant focal hepatic disease.
4. Alspaugh
J” Bernardino M, Seweli C, Sones P, Be&man W’: Price R: CT directed hepatic biopsies: increased diagnostic accuracy with low patient risk, J Comput Assist Tomogr 1983: 7:1012-1017,
5. Haaga J, Vanek J: Computed tomographic guided iiv~eerbiopsy using the Menghini needle, Radiology 1979; :33:405406. 6. Isler R, Ferrucci J, Wittenberg J, et al.: Tissue core biopsy of abdominal tumors with a 22 gauge cutting needle, AJR 1981; 136:727-728. 7. FerrucciJ, Wittenberg J, Mueller P, et al.: Diagnosis of abdominal malignancy by radiologic fine-needle aspiration biopsy, AJK 1979; 134:323-330. 8. Cronan J, Esparza A, Dertman G; Ridlen M, Paolelia L: Cavern QUS hemangioma of the liver: role of percutaneous biopsy, Radiolgy 1988; 166:135-138. 9 Solbati L, Lirraghi T, De Pra L, Ieraco T, Masciadr N, Ravetto C: Fine-needle biopsy of hepatic hemangioma with sonographic guidance, AJR 1985; 144:474-477. 10. Whitlach S, Nunez C, Pitlik D: Fine needle aspiration of the liver, Acta Cytol 1984; 28:719-725. 11.
biopsy
AdrioIeJ, Haaga J, Adams R, Nunez C: Biopsy needle characteristics assessed in the laboratory, Radiology 1983; 148:659-662.
12. Hajdu S, Melamed M: Limitations of aspiration cytology in the diagnosis of primary neoplasms, Acta Cytol 1984; 28:337345. 13. Lieberman R. Hafez G, Crummy A: Histology from aspiration biopsy: Turner needle experience, 1982; 1~~561-564. 14. Bateson M, Hopwood D, Duguid H, Boucher I: A comparative trial of liver biopsy needles, J Clin Path01 iaao; 33:131133. 15. Littlewood E, Gilgore I, Murray-Lyon I, Stephens K, Paradinas F: Comparison of the Tru-Cut and Sure-Cut live; biopsy needles, J Clin Path01 1982; 35:761-763. 16. Kasugai H, Yamamoto lR, Tatsuta M, et al.: Value of iheparinized fine-needle aspiration biopsy in liver malignancy, AJR 1985; 144:243-244. 17. Bell D, Carr C, Szyfelbein W: Fine needle aspiration of focal liver lesions, Acta Cytol 1986; 30:397-402,
cytology
18. Atterbury C, Enriquez R, Desoto-Nagy 6, Conn H: Comparison of the histologic and cytologic diagnosis of liver biopsies in hepatic cancer, Gastroenterology 1979; 76:1352-1357.
Martin0 6, Haaga J, Bryan I’, LiPuma J, El Yousef S, Alfidi R: CT-guided liver biopsies: eight years’ experience, Rediology 1984;152:755-757.
19. Tatsuta M, Yamamoto R, Kasugai H, et al.: Cytohistologic diagnosis of neoplasms of the liver by ultrasonical!y guided fineneedle aspiration biopsy, Cancer 1984; 54:16821686.
1983;
20. Ho C, McLaughlin M, Tao L, Blendis L, Evans W: Guided percutaneous fine-needle aspiration biopsy of the liver, Cancer 1981;47:1701-1783.
Wlitmire L, Galambos J, Phillips V, et al.: Imaging guided percutaneous hepatic biopsy; diagnostic accuracy and safety, J Glin Gastroenterol 1985; 7:511-515.
21. Reichert C, Weisenthal L, Klein H: Delayed hemorrhage, after percutaneous liver biopsy, J Clin Gastroenterol 1983; 5:263-266.
Pagani J: Biopsy 147:673-675.
of focal hepatic
lesions,
Radiology