Fax: (205) 975-7284
Fine-needle aspiration biopsy of hepatic lesions
Computerized tomographic-guided versus endoscopic ultrasound-guided FNA
Article first published online: 21 APR 2006
Copyright © 2006 American Cancer Society
Volume 108, Issue 3, pages 180–185, 25 June 2006
How to Cite
Crowe, D. R., Eloubeidi, M. A., Chhieng, D. C., Jhala, N. C., Jhala, D. and Eltoum, I. A. (2006), Fine-needle aspiration biopsy of hepatic lesions. Cancer, 108: 180–185. doi: 10.1002/cncr.21912
- Issue published online: 12 JUN 2006
- Article first published online: 21 APR 2006
- Manuscript Revised: 7 OCT 2005
- Manuscript Accepted: 5 OCT 2005
- Manuscript Received: 1 JUL 2005
- fine-needle aspiration;
- liver lesions;
Computerized tomographic (CT)-guided fine-needle aspiration (FNA) cytology is a well-established tool in the diagnosis of hepatic lesions. Endoscopic ultrasound-guided FNA (EUS-FNA), developed recently and used predominantly in evaluating mediastinal and pancreatic lesions, provides access to a significant portion of the liver and to perihepatic structures not readily accessible by a percutaneous approach.
A recent experience (1997-2002) with CT-guided FNA of liver lesions at the University of Alabama Birmingham (UAB) was compared with the first 2.5 years of EUS-FNA experience (2000-2002). Cases were identified using a SNOMED search and all reports and cytologic slides were retrieved for review.
In 6 years, 34 percutaneous CT-FNA liver biopsies were performed at UAB; in approximately 2.5 years, 16 EUS-FNA liver biopsies were done. In both groups the primary clinical indication was suspected metastatic carcinoma (CT, 41% of cases vs. EUS, 56%). The 2 techniques yielded a similar range of benign, atypical, and malignant diagnoses (CT: 26%, 18%, and 56% vs. EUS: 19%, 25%, and 56%). Because of the clinical setting in which EUS-FNA is usually performed, a much narrower range of neoplasms was sampled by EUS-FNA. Benign gastrointestinal epithelial cells were identified in 60% of the EUS-FNA specimens.
Early experience suggests EUS-FNA is comparable to CT-FNA in terms of diagnostic utility for hepatic lesions. Anatomy limits EUS-FNA to only a fraction of the hepatic parenchyma, but that fraction includes the hilum and left lobe of the liver and the proximal biliary tract. The gallbladder, extrahepatic biliary system, and perihilar lymph nodes are readily accessible. Proximate high-resolution ultrasound imaging and cytopathologist involvement in the EUS-FNA process are further advantages. Awareness of artifacts inherent in EUS-FNA sampling (i.e., gut epithelial cells) can minimize a potential diagnostic pitfall. Cancer (Cancer Cytopathol) 2006. © 2006 American Cancer Society.
Percutaneous aspiration of the liver has been part of the diagnostic and therapeutic armamentarium since the mid-19th century, when palpable hydatid cysts and abscesses were drained in this manner.1 During the mid-20th century the parallel development of fine-needle aspiration (FNA) cytology and computerized tomographic (CT) and ultrasonographic (US) imaging technology have made percutaneous aspiration of liver lesions a viable diagnostic modality as well. During the past decade, endoscopic ultrasonography coupled with fine needle aspiration cytology (EUS-FNA) has been demonstrated to be a highly useful means of evaluating a host of intrathoracic and intraabdominal lesions.2–4 Early reports of experience with EUS-FNA show that the majority of procedures evaluating abdominal lesions are aimed at the pancreas and intraabdominal lymph nodes, with considerably less emphasis on examination of liver lesions. In this report we compare our experience with CT-guided FNA of liver lesions with our early experience with EUS-FNA of the liver.
MATERIALS AND METHODS
In order to obtain sufficient material to be of interest when comparing experiences of similar age, we reviewed our most recent 6-year experience with CT-guided FNA of the liver (1997-2002) in conjunction with the first 2.5 years experience accumulated by our EUS-FNA service (2000-2002). Cases were identified by a SNOMED search of the archives at the University of Alabama Birmingham (UAB) Department of Pathology and all reports and cytologic slides were retrieved for review.
The CT-guided FNA biopsies were performed in the Department of Radiology at UAB by various interventional radiologists with a cytotechnologist present for slide preparation and appropriate specimen handling. Immediate cytopathologic interpretations were not performed.
One attending gastroenterologist (M.A.E.) performed all EUS-FNAs. Standard EUS was performed under conscious sedation using a curvilinear echoendoscope (UC-30P; Olympus, Lake Success, NY). Color flow and Doppler sonography were performed to identify intervening vascular structures and to choose a vessel-free needle track. All FNAs were performed using a 22G needle (Echotip; Wilson-Cook, Winston Salem, NC) inserted through the working channel of the echoendoscope. Once the tip of the catheter was visualized, the needle was advanced from the catheter sheath, through the wall of the gastrointestinal (GI) tract, and into the target lesion under ultrasound guidance. The stylet was removed and the initial passes were performed by moving the needle back and forth within the target lesion for 15 to 20 seconds. No suction was applied during biopsy unless the biopsy did not yield any material or the lesion was cystic.
The aspirate was placed on glass slides. Both air-dried and alcohol-fixed smears were prepared. Air-dried smears were stained with Diff-Quik (Baxter, McGraw Park, IL) and reviewed immediately by a cytopathologist on-site to ensure specimen adequacy. A preliminary diagnosis was rendered when possible. Alcohol-fixed smears held for subsequent staining with Papanicolaou stain and aspiration material was prepared for cell block specimens and appropriate ancillary studies. At least 3 passes were obtained for each target lesion, unless cytologic evaluation performed on-site confirmed the presence of malignant cells in the first 2 passes. Immediate cytologic diagnoses were categorized as positive for a neoplasm, suspicious for a neoplasm, nonneoplastic/reactive process, or nondiagnostic. Nondiagnostic specimens were defined either as aspirates in which cellularity was inadequate to characterize the lesion or in which the available material was thought to be unrepresentative of the target lesion.
The distribution of cytologic diagnoses between CT-guided vs. EUS-guided FNA was compared using the Fisher exact test.
In 6 years, 34 CT-guided FNA liver biopsies were performed at our institution; in 2.5 years, 16 EUS-FNA liver biopsies were done. Demographic data, available clinical histories, and diagnoses are shown in Table 1. There was no significant difference in age or gender distribution of the 2 groups of patients. Review of the cytologic material yielded no significant diagnostic discrepancies between the review and original diagnoses.
|Age range, y||34-79||18-89|
|Male: Female ratio||7:9||10:9|
|Negative for malignancy||3||9|
|Atypical (reactive or NOS)||2||2|
|Positive for malignancy||9||19|
|Malignant/ suspicious lesions|
|Squamous cell carcinoma||0||1|
|Non-small cell carcinoma||2||9|
In both groups of patients the primary clinical indication for FNA was suspected metastatic carcinoma (CT, 41% of cases vs. EUS, 56%). When diagnoses are categorized as benign, atypical, or malignant, the results for CT-guided FNA are 26%, 18%, and 56%, respectively. For EUS-FNA, they are 19%, 25%, and 56%. None of these categories demonstrated a significant difference between the 2 groups (P >.5).
There is a wide variety of malignant neoplasms in this relatively small series of cases. On review, benign GI epithelial cells were identified in 60% of the EUS-FNA specimens (Fig. 1); none were present in the CT-guided specimens.
A search of computerized hospital records and anatomic pathology division records indicated that 9 of the 34 (26%) CT-guided FNA procedures and 1 of the 16 (6%) EUS-FNA procedures were associated with concurrent or subsequent biopsies yielding relevant histologic specimens. Of these 10 cases, histologic diagnoses confirmed all 4 cytologic diagnoses of “positive for malignancy” (100%). All 3 cases interpreted cytologically as “suspicious for malignancy” were confirmed as malignant in the histologic material. Two of the negative cases were also histologically confirmed outright. A third case, negative on FNA (CT-guided), involved a patient with a history of melanoma who had multiple liver lesions. Three months after the FNA, 3 open surgical wedge biopsies of the liver showed 1 benign lesion, a bile duct hamartoma, and 2 foci of metastatic melanoma. Review of the aspiration smears showed numerous benign hepatocytes but no significant population of biliary ductal epithelial cells and no malignant cells, suggesting that neither benign nor malignant focal lesions had been adequately sampled by the FNA. Because of the very limited number of cases in which cytologic/histologic correlation was available, receiver operator statistics were not calculated.
Ten of the 16 EUS-FNA procedures (62%) involved concurrent sampling of other sites in addition to the liver. There were 9 pancreatic EUS-FNAs and 1 aorto-pulmonary (AP) window lymph node EUS-FNA. All 10 of these specimens were positive for malignancy. Of the 9 hepatic EUS-FNA specimens associated with positive pancreatic EUS-FNAs, 6 were positive for metastatic carcinoma, 2 were atypical, and 1 was negative. The AP window lymph node EUS-FNA and associated hepatic EUS-FNA were both positive for malignancy, a cytokeratin 7-positive, nonsmall cell carcinoma that was clinically thought to be of pulmonary origin.
Because of the limited amount of directly correlative histologic material, clinical follow-up was also examined in computerized hospital records of all patients. Additional data on patients who had undergone EUS-FNA were obtained from the files of the endoscopist (M.A.E.). In addition to the patient with metastatic melanoma discussed above, the only other apparent clinical false-negative was a patient with pancreatic carcinoma who had multiple suspicious liver lesions but a negative EUS-FNA of the liver. Clinically, this patient had disease progression on her last known follow-up visit 8 months after her biopsy. No other patients in either the CT-guided FNA or the EUS-FNA group exhibited any clinical findings or radiological evidence incompatible with their original cytologic diagnoses.
FNA biopsy and cytologic diagnosis of liver lesions have become a standard part of the diagnostic armamentarium in most modern medical centers. Although there are reports of FNA evaluation of palpable liver masses,5 currently the vast majority of hepatic lesions are sampled in radiology suites under CT or US guidance. The usual indication for FNA of the liver is evaluation of a localized solid or cystic lesion with at least the potential for malignancy. In several reported large center series, malignant diagnoses are established in the majority of biopsies (50% to 93%)1–4, 6, 7; a preponderance of these cases are metastatic carcinoma.1, 6, 7 Correlation of FNA biopsy findings with appropriate histopathologic material and clinical follow-up has demonstrated the high sensitivity and specificity of cytopathologic diagnoses.1–4 Indeed, some institutions have found FNA to perform slightly better than surgical biopsy in these performance measures.3 Our experience seems to confirm this. Although histologic correlation was available in only 10 (20%) of our 50 total cases, we established only 2 known clinically false-negative cases (see Results) and found no confirmed or suspected false-positive cases.
The problems that persist in cytopathologic diagnosis, i.e., diagnosis of well-differentiated hepatocellular carcinoma,5, 8 diagnosis of cholangiocarcinoma and other biliary tract tumors,9, 10 distinguishing hepatocellular carcinoma and cholangiocarcinoma from metastatic neoplasms,8, 11, 12 and establishing the diagnosis of unusual neoplasms,13–15 are very similar to those confronting the diagnostic surgical pathologist. Strategies for resolving them depend on many of the same ancillary techniques, such as immunohistochemical staining.12, 14, 15
From the perspective of the cytopathologist, the diagnostic material presented by EUS-FNA is very similar to that yielded by traditional FNA procedures. All of the same diagnostic criteria apply. There are 2 caveats, however. In the setting in which the cytopathologist is in the endoscopy suite providing immediate information to the endoscopist, all of the material he or she sees at the time will be air-dried smears stained with a rapid Romanowsky stain such as Diff-Quik. Cytopathologists who are more comfortable dealing with alcohol-fixed material stained with Papanicolaou stain or hematoxylin and eosin will need to accustom themselves to dealing with air-dried material. Second, because the sampling needle in EUS-FNA must traverse the gut wall on its path to the liver, a significant number of cases (60% in this series) can be expected to include benign gut epithelial cells from the gastric antrum or the duodenum. Recognition of the fact that these cells are likely to be present reduces the possibility that they will be erroneously interpreted as well-differentiated adenocarcinoma.
Close cooperation between the endoscopist and the cytopathologist in the endoscopy suite allows the diagnostic procedure to be tailored to the individual circumstances and also ensures that if special ancillary procedures (such as flow cytometry for immunophenotype analysis) are likely to be required, adequate material can be obtained and properly handled, reducing the likelihood that a second biopsy procedure will be necessary.2 For example, 1 of our clinical false-negative cases was a CT-guided FNA in a patient with multiple liver masses who was strongly suspected of having metastatic disease. In the EUS-FNA setting, an immediate diagnosis of “negative for malignancy” would likely have resulted in additional FNA passes being performed until the endoscopist was satisfied that the lesions had been adequately sampled. Conversely, our second clinical false-negative case involved an EUS-FNA procedure in which pancreatic carcinoma was established in the first set of aspirates and in which the initial sample from the liver was negative, with scant cellularity. Because definite cytologic confirmation of liver metastasis was not critical to making future therapeutic decisions, additional aspirates were not obtained.
Our data demonstrate the versatility of the EUS approach. Nine of 16 (56%) of our EUS-FNA of the liver cases involved concurrent sampling of pancreatic masses, i.e., the primary diagnosis of pancreatic carcinoma was established and meaningful staging data were obtained in a single diagnostic encounter. The case involving concurrent staging of pulmonary carcinoma by sampling an AP window lymph node and the liver during a single endoscopic procedure is another example of this versatility.
Early experience suggests that EUS-FNA is comparable to CT-FNA in terms of diagnostic utility for hepatic lesions. Both techniques are useful in evaluating a broad range of malignant and benign processes. Because EUS-FNA is often used to diagnose and stage pancreatic carcinoma, the range of metastatic tumors we have seen thus far is somewhat constrained, but as experience with EUS accumulates, we expect to see a broader array of lesions. Although anatomy limits EUS to only a portion of the liver, that portion includes the hilum, left lobe, and proximal right lobe including the major portion of the intrahepatic biliary tract. Offsetting the disadvantage of not providing access to the right lobe adjacent to the dome of the diaphragm and the lateral and inferior portions of the right lobe, EUS offers proximity and high-resolution imaging of that part of the liver that it does “see,” allowing detection of lesions smaller than those detectable by previous imaging technology.6, 8, 16 Also, the capability of examining and sampling adjacent structures such as regional lymph nodes and the gallbladder17 allows critical staging information to be obtained concurrently with establishing a primary diagnosis, thus expediting the management of a group of very difficult neoplasms.
Despite our enthusiasm for the EUS-FNA procedure, we do not intend to minimize the utility of percutaneous FNA of the liver. Algorithms for deciding which modality to use in a given patient are likely to be evolving for quite some time and will probably vary from institution to institution. Critical elements to be considered will include such patient-based facts as the location of the lesion in the liver, the presence of extrahepatic lesions, and even more general considerations such as whether there is some contraindication to sedation. Institutional resources will also influence the decision. Bringing together a team including gastroenterologists expert in EUS-FNA and cytopathologists skilled in rapid assessment of this material, together with the necessary ancillary personnel and specialized instrumentation for the procedure, involves a substantial commitment of resources that many institutions may be unwilling or unable to make. In our institution, EUS-FNA is preferred for patients with suspected pancreatic neoplasms or in whom additional staging of other suspected neoplasms is anticipated. Patients in whom the liver lesion is the object of primary concern or who have a peripheral lesion in the right lobe of the liver are much more likely to undergo percutaneous FNA.