Needle Biopsy of the Adrenal Gland: Retrospective Review of 54 Cases
David Dusenbery, M.D., and Andrew Dekker, M.D.
A retrospective review of 54 cases of adrenal gland needle biopsy in 53 patients is presented. The cases included 43 fine-needle aspirations (FNA), six core-needle biopsies, and five cases in which both types of needle biopsy were done. Clinical or histologic follow- up was available in 28 of the 36 specimens deemed adequate for evaluation. Metastatic malignancies represented the largest group of cases (19), with lung being the most common primary site (8). The series included four cases of metastatic hepatocellular carci- noma. Probable adrenocortical adenoma was the next most com- mon diagnosis (12 cases).
Cases causing diagnostic confusion included a case of meta- static well-differentiated hepatocellular carcinoma which was ini- tially confused with an adrenocortical adenoma and a probable adrenocortical adenoma which was mistaken for a metastatic small round cell malignancy. (This case is unproved because of lack of clinical or histologic follow-up). By evaluating only those cases with histologic confirmation or clinical follow-up greater than one year, the sensitivity of needle biopsy for the presence of malignancy was 95% and the specificity was 100%.
The cytologic findings are described with attention to the poten- tial problem of confusing primary adrenocortical neoplasms with metastases from hepatocellular carcinoma. Diagn Cytopathol 1996;14:126-134. @ 1996 Wiley-Liss, Inc.
Key Words: Needle biopsy; Adrenal gland; Metastatic neoplasm; Adrenocortical neoplasm; Hepatocellular carcinoma
The adrenal glands are a frequent site of metastatic tu- mors.1 Additionally, benign adrenal masses are found commonly as incidental findings in patients with cancer undergoing staging work-ups.2-4 Since needle biopsy is the only currently available non-surgical means of obtain- ing a diagnosis in patients with adrenal masses and be- cause of the increased use of sensitive imaging techniques, the adrenal glands are a frequent target of needle biopsies.
Others have described the cytologic findings in adrenal needle biopsy. 5-9 In this report, we present our findings in 54 adrenal needle biopsies to contribute to the cytologic characterization of primary adrenal lesions as well as to discuss the differential diagnosis of potentially confusing metastatic lesions.
Materials and Methods
Fifty-four needle biopsies of the adrenal gland from 53 patients obtained between May of 1985 and January of 1994 were retrospectively reviewed. These percutaneous needle biopsies were all performed by radiologists under computed tomographic or sonographic guidance. Most biopsies were fine-needle aspirations (FNA) alone (43). In six cases, cutting-type needles were used to obtain cores of tissue. And in five cases, both FNA and core-needle biopsy (CNB) were performed.
In most instances, direct cytologic smears were made by cytotechnologists in attendance at the time of the proce- dure and stained by the Papanicolaou method. Occasion- ally, rapid hematoxylin and eosin and LeukoStat (Fisher Scientific, Pittsburgh, PA) stains were performed. In the cases of CNB, touch preparations were made prior to fixing the cores of tissue in 10% neutral buffered formalin. Cell block material was also obtained where possible and was fixed in 10% neutral buffered formalin (NBF). Cell block material and CNB-obtained tissue was processed and em- bedded in paraffin, sectioned and stained with hematoxylin and eosin. In rare instances, immunohistochemical and/or histochemical stains were performed on sections from the cell block or CNB. In one case, tissue was retrieved from the cell block for ultrastructural examination.
In the few instances when cytotechnologists were not present to make smears, the needles were rinsed in Muco- lexx (Lerner Laboratories, Pittsburgh, PA) and forwarded to the laboratory where filter and cell block preparations were made. Immediate on-site cytologic evaluation of the sample by a pathologist was performed inconsistently. Medical records and histologic material, when available,
were reviewed to assess accuracy of the needle biopsy diagnosis.
Results
The patients ranged in age from 34 yr to 79 yr with a mean age of 61 yr. There were 28 males and 25 females. Thirty- three biopsies came from the left side and 21 came from the right.
Eighteen of the 54 biopsies (33%) were, on review, insufficient for diagnosis. These cases were excluded from the subsequent calculations and the remaining 36 cases form the basis of this report. The review diagnoses on the 36 cases are presented in Table I.
Originally, 22 cases were diagnosed as malignant and 14 as benign (an additional case was interpreted as consistent with adenoma, but on review it showed normal hepato- cytes and was reclassified as insufficient; therefore, this case has not been included in subsequent calculations). On review, one case, originally called benign, was reclassified as malignant and one case, originally called malignant, was reclassified as benign.
Clinical and/or histologic follow-up was available in 28 of the 36 cases (histologic confirmation in six and clinical follow-up in 22). Of the 14 cases originally classified as benign, nine had clinical follow-up only, one had histo- logic confirmation, and four had no follow-up. Clinical follow-up in these cases ranged from 1 mo to 60 mos in length with a mean follow-up period of 16 mo. Only three of the nine cases with clinical follow-up had follow-up greater than 1 yr. The single case which was reclassified from benign to malignant was originally considered to be consistent with an adrenocortical adenoma, but, upon re- view, was reclassified as metastatic well-differentiated
Table I. Review Diagnoses in 36 Cases of Adrenal Gland Needle Biopsies Considered Adequate for Evaluation
Metastatic neoplasms (19) Pulmonary primary (8)
Non-small cell carcinoma (7) Small cell carcinoma (1)
Hepatocellular primary (4)
Unknown primary (3)
Malignant melanoma (1)
Renal cell carcinoma (1)
Transitional cell carcinoma (1)
Adrenocortical carcinoma-metastatic from contralateral adrenal (1)
Probable adrenocortical adenoma (12) Adrenocortical carcinoma (2) Primary malignancy (1)
Largely necrotic tumor suspicious for recurrence (1)
Lymphoma (1)
Myelolipoma (1)
Miscellaneous (1)
“Splenosis” (1)
hepatocellular carcinoma. Histologic confirmation was available in this case.
Of the 22 cases originally diagnosed as malignant, 13 had clinical follow-up, five had histologic confirmation, and four had no follow-up. The five cases having histo- logic confirmation were: three cases of adrenocortical car- cinoma (one primary, one recurrent, and one metastatic from the contralateral side), one case of metastatic hepato- cellular carcinoma and one case of metastatic non-small cell carcinoma which was confirmed at autopsy. The cases with clinical follow-up only were either cases with known, proven primaries elsewhere and strong clinical and radio- logic evidence of metastasis; overwhelming clinical evi- dence of metastatic disease; or proven by positive mucicar- mine staining on cell block material (one case). The single case, originally diagnosed as malignant, which was reclas- sified as benign on review was the case of a 69-yr-old man with a lung tumor and an enlarged left adrenal gland. A fine-needle aspirate of the lung tumor revealed a carcinoid tumor. The aspirate from the adrenal gland was inter- preted as metastatic small cell malignancy, consistent with metastasis from pulmonary carcinoid, but on review was felt to be consistent with adrenocortical adenoma. Unfor- tunately, histologic confirmation in this possibly falsely positive case is lacking. The patient subsequently died of cardiac causes and an autopsy was not performed.
If the possible false-positive case discussed above is ex- cluded from consideration and if only those negative cases with clinical follow-up greater than 1 yr are considered true negatives, the sensitivity of needle biopsy for the diag- nosis of malignancy in this series was 95%, the specificity was 100%, the predictive value of a positive test was 100%, the predictive value of a negative test was 75%, and the overall efficiency of the test was 95%. If the uncon- firmed false positive result is included in the calculations, the specificity falls to 75%, the predictive value of a posi- tive test falls to 95%, and the overall efficiency of the test falls to 91%.
Core-needle biopsy alone was done in six of the 54 cases. CNB and FNA were done in five cases. FNA alone was done in the remaining 43 cases. Of the six cases with CNB alone, two were insufficient (both showed only normal hepatic tissue). However, in one of the four cases consid- ered adequate for diagnosis, the diagnosis was made on cytologic material obtained via the core biopsy needle; no solid tissue being obtained (this was a case of myelolipoma). Of the five cases in which CNB as well as FNA were done, the diagnosis was made by CNB alone in two cases and by both CNB and FNA in the other three. The addition of the FNA procedure in these five cases did not improve the diagnostic yield. The insuffi- ciency rate for those cases in which a CNB was performed was 18% (2 of 11), while the insufficiency rate for FNA alone was 37% (16 of 43).
Five of the 54 cases showed hepatic tissue or cells. All of these came from the right side. In four of these cases the hepatic cells or tissue was the only material obtained, and, therefore, the samples were deemed insufficient. In the remaining case, the liver cells were a minor component in smears from a probable adrenocortical adenoma.
Probable Adrenocortical Adenoma
Twelve cases were reviewed. All cases had FNA and one also had a CNB performed. The smears ranged in cel- lularity from low to extremely high with most being of moderate cellularity. The cytologic pattern was that of mostly dissociated cells with occasional syncytial groups. Many, if not most, of the nuclei were stripped of cyto- plasm (Fig. 1). Focally, the nuclei were grouped, occasion- ally giving the impression of some nuclear molding. Where cytoplasm was visible, it was finely vacuolated or finely granular and usually delicate and frayed in appear- ance. The nuclei in most cases were monotonous and round with occasional conspicuously larger nuclei pre- sent. Other cases demonstrated a moderate amount of
anisonucleosis. The background of the smear was gener- ally bloody with a foamy, lipid-rich appearance. This smear pattern of small round stripped nuclei lying in a foamy, lipid-rich background (Fig. 2) recapitulated the histologic appearance seen in CNBs (Fig. 3). The chroma- tin was finely distributed. Small nucleoli were occasionally seen, but generally they were inconspicuous. A rare intra- nuclear pseudoinclusion was identified. The case which was originally misclassified as metastatic small cell malig- nancy differed in no respect cytologically from the other cases. This case was compared with the lung aspirate from this patient which was diagnosed as carcinoid tumor. Al- though the two cases bore a striking cytologic similarity, careful retrospective comparison revealed several recog- nizable differences. Both cases had small nuclei of similar size, but the adrenal tumor had more anisonucleosis in general and had occasional nuclei which were much larger. The chromatin pattern differed in that in the lung tumor it was distinctly speckled as is typical of neuroendo- crine tumors (Fig. 4) while in the adrenal tumor it was more uniform and occasionally smudgy. Cytoplasmic dif- ferences were evident as well. The adrenal aspirate re- vealed more stripped nuclei whereas in the lung aspirate,
most nuclei had a thin rim of delicate cytoplasm. In the adrenal aspirate, in the few areas where cytoplasm was intact, it was abundant and vacuolated (inset, Fig. 1). This was in marked contrast to the cells of the lung aspirate which had very high nucleocytoplasmic ratios and non- vacuolated cytoplasm.
Histologic confirmation on this group of lesions was lacking. Clinical follow-up was available in eight of the 12 cases ranging in length from 1 to 60 mo with a mean of 10 mo.
Adrenocortical Carcinoma
Three cases were reviewed. One case consisted of a CNB with accompanying touch preparations. The other two were FNA specimens. One of the FNAs consisted primar- ily of necrotic cellular debris with only rare atypical viable cells identified in the cell block preparations. The touch preparation from the CNB and the remaining FNA speci- mens were similar in cytologic appearance. Both samples were highly cellular and were composed of large poly- gonal cells which were almost totally dissociated from each other. The cells had abundant cytoplasm and eccen-
trically placed nuclei (Fig. 5). The cytoplasm appeared dense and very finely granular; for the most part, it had well-defined margins. The nuclei were mostly round with occasional grooves and notches and obviously malignant irregularly distributed chromatin. Nucleoli were obvious and often irregular in contour. In the touch preparation, numerous abnormal mitotic figures were present and a neutrophilic infiltrate surrounded the tumor cells. Occa- sional tumor giant cells were seen (Fig. 6). These findings were confirmed in the histologic sections of the CNB. A rare intranuclear pseudoinclusion was seen.
Histologic confirmation was present in all three of these cases.
Metastatic Hepatocellular Carcinoma
Four cases were reviewed. One was a CNB with touch preparations and the remaining three were FNA speci- mens. Interestingly, the CNB was the false negative case which was originally misinterpreted as adrenocortical ad- enoma. In one of the FNA specimens, the majority of the diagnostic material was in the cell block preparation and the diagnosis was essentially a histologic one (Fig. 7). The
DUSENBERY AND DEKKER
remaining two FNAs had material available on the smears for review. One case was highly cellular with an obvious trabecular pattern recognizable on low power examina- tion. The cells were mostly cohesive with relatively high nucleocytoplasmic ratios and prominent nucleoli. They resembled normal hepatocytes except for their slightly smaller size, higher nucleocytoplasmic ratio, and uniform- ity. Smears from the other FNA showed a dissociated cell pattern with numerous stripped nuclei. Nucleoli were ob- vious. Once again, compared to normal hepatocytes, the cells were relatively small with increased nucleocytoplas- mic ratios.
Histologic confirmation was available in two of the four cases, strong clinical confirmatory evidence was available in one, and one case was lost to follow-up. In both of the cases lacking histologic confirmation, abun- dant material was available for evaluation in the cell block preparations.
Other Metastatic Malignancies
Fifteen cases were reviewed (this group does not include the four cases of metastatic hepatocellular carcinoma
discussed above). Two of the cases had CNB alone, three had CNB and FNA, and the remainder had FNA alone. The most common primary site of malignancy was the lung with eight cases coming from that site. Also present were three cases of metastatic carcinoma of unknown primary, one malignant melanoma, one renal cell carcinoma, one transitional cell carcinoma, and one metastatic adrenocortical carcinoma from the centralateral side (discussed above). The cytologic diag- noses in these cases were not problematic, with the ex- ception of the renal cell carcinoma, the cytologic find- ings being representative of the primary tumor in each case. The renal cell carcinoma was diagnosed by the aid of ultrastructural examination of material retrieved from the cell block.
Of these 15 cases, two had histologic confirmation and 11 others had strong clinical confirmatory evidence, his- tology of the primary tumor being available in many in- stances. In one case, the diagnosis of metastatic adeno- carcinoma was confirmed by positive mucicarmine staining on the cell block tissue.
Miscellaneous Lesions
The three remaining lesions were: one case of malignant lymphoma, one myelolipoma, and one case thought to represent “splenosis.”
The diagnosis of malignant lymphoma was made in a 78-yr-old man who developed testicular enlargement and was found to have an adrenal mass. A diagnosis of diffuse large cell lymphoma was established by histologic exami- nation of the orchiectomy specimens just prior to the adre- nal cytologic diagnosis. The FNA smears from the adrenal gland were cellular with a dispersed cell pattern. The cells had high nucleocytoplasmic ratios, the chromatin was ir- regularly clumped, and one to multiple nucleoli were pre- sent. Lymphoglandular bodies were present in the smear background.
The diagnosis of myelolipoma was made on cytologic preparations made in conjunction with a CNB procedure. Solid tissue was not obtained on the CNB pass corre- sponding to the cytologic smears, thus this diagnosis was based on the cytologic findings alone. The smears showed the full range of normal hematopoietic elements (Fig. 8).
This finding in conjunction with the radiologic finding of a low density mass lesion resulted in the diagnosis of myelolipoma. No follow-up is available in this case.
The diagnosis of “splenosis” on the remaining case was based on the finding of what appeared to be normal splenic tissue on the cell block preparation from an FNA. The patient was a 47-yr-old man who had undergone liver transplantation for alpha-one antitrypsin deficiency. He was status post-splenectomy and the computed tomo- graphic impression was “splenosis.” The aspirate was identified as coming from the left adrenal gland. The smears showed lymphoid cells in a background of blood admixed with scattered connective tissue elements, but it was the cell block which was most convincing. Clinical follow-up of 46 mo duration is available on this patient.
Discussion
Previous reports of adrenal needle biopsy have shown a high degree of accuracy and a high yield of diagnostic material obtained. 5,6,10-16 The use of different methods of reporting results and varying degrees of patient follow-up make calculation of sensitivity and specificity for the pro- cedure difficult. An attempt was made to determine these
| Senior author | No. of patients | Immediate evaluation by pathologist | Specimen adequacy (%) | Sensitivity (%)b | Specificity (%)b |
|---|---|---|---|---|---|
| Zornoza, 198114 | 21 | N.S. | N.S. | 75 | 100 |
| Heaston, 198215 | 14 | Yes | 93 | 100 | 100 |
| Berkman, 198416 | 16 | N.S. | 94 | 100 | 100 |
| Katz, 19845 | 22 | N.S. | 91 | 85 | 100 |
| Bernardino, 198513 | 53 | No | 83 | 100 | 100 |
| Karstrup, 199112 | 28 | N.S. | 100 | 89 | 89 |
| Tikkakoski, 199111 | 56 | N.S. | 96 | 78 | 97 |
| Wadih, 19926 | 48 | Yes | 88 | N.A. | 100 |
| Silverman, 199310 | 97 | N.S. | 86 | 93 | 100 |
ªN.S. = not stated; N.A. = not available.
“The terms sensitivity and specificity are in reference to the diagnosis of malignancy. Biopsies classified as insufficient by the authors were not included in these calculations.
parameters, as well as adequacy rates, from published series of adrenal gland needle biopsies. This information is presented in Table II. Despite these encouraging results, diagnostic problems exist such as distinction of adrenocor- tical adenoma from well-differentiated adrenocortical car- cinoma 5,6,9 and potential confusion of certain metastatic lesions, such as renal adenocarcinoma, with primary adre- nal neoplasms.7,9 The technique is generally acknowl- edged to perform well in the identification of metastatic tumors, the rare confusing metastatic tumor not with- standing. The results of the present study, in general, agree with those previously published, but several aspects warrant further discussion.
The relatively low adequacy rate in this series warrants explanation. Adequacy rates previously reported for adre- nal gland needle biopsy are generally higher than 80% (Table II). Our adequacy rate of 67% is decidedly low. One possible explanation for this is that pathologist assess- ment of the specimen was performed inconsistently. Im- mediate assessment of the sample is one possible way of improving the chances of obtaining an adequate sample. If the pathologist determines that the sample is insuffi- cient, a repeat sample can be advised. It is conceivable that, in some instances, radiologists chose to terminate the procedure despite the pathologist’s recommendation that another needle pass be made. Although we believe imme- diate cytologic assessment of the specimen is beneficial, some investigators have achieved high adequacy rates without it. 13 Another variable which may have an effect on adequacy rates is the size of the mass being aspirated. This variable was not recorded in the present study. Ulti- mately, accuracy of needle placement is the best deter- minator of specimen adequacy with poor operator tech- nique cited as the major cause of failed attempts. 13,17
Although the number of patients undergoing needle biopsy with the cutting type of needle was small in this study, it enabled us to make some comparisons of this technique with the FNA method. In the cases where CNB was part of the procedure or the sole procedure, the ade- quacy rate was 82%. In the cases where FNA alone was
employed, the adequacy rate was 63%. These numbers must be evaluated in light of the above discussion of as- sessment of specimen adequacy. There is sufficient evi- dence from other series that the FNA technique provides the diagnosis in a high proportion of cases. 5,6,10-13 Accu- racy of needle placement is undoubtedly more important than choice of needle type.
Also related to the topic of needle placement was the finding of hepatic cells or tissue in five of our cases. Two of the cases were CNBs with touch preparations and three were FNAs. In one instance an FNA containing benign liver was interpreted as consistent with an adrenocortical adenoma. This case, on review, was deemed insufficient and was excluded from calculations of accuracy since it represented neither a false positive or a false negative diagnosis, but was rather non-representative due to sam- pling error. It does, however, point to another potential source of confusion in the interpretation of needle biopsies from the adrenal gland, particularly from the right side, that being possible confusion with hepatic processes.
Our series was somewhat unique in having four exam- ples of hepatocellular carcinoma metastatic to the adrenal gland. In addition to the above-mentioned case of benign hepatic tissue misinterpreted as adrenocortical adenoma, we also identified a case of well-differentiated hepatocellu- lar carcinoma metastatic to the adrenal gland which was originally misdiagnosed as a probable adrenocortical ade- noma. This is a pitfall which has not received discussion in the literature, although the converse situation of adre- nal cells being misinterpreted as coming from a clear cell type of hepatocellular carcinoma has been reported. 12 Confusion of hepatocytes and adrenocortical cells is a recognized pitfall, 9 therefore, it logically follows that well- differentiated hepatocellular carcinoma would enter into the cytologic differential diagnosis as well. This possibility must be kept in mind especially if the patient has a history of primary hepatic malignancy.
Comparison of our two cases of metastatic hepatocellu- lar carcinoma which had cellular material available on the smears with our cases of probable adrenocortical adenoma
revealed a few cytologic differences which may prove use- ful in the distinction of these lesions. One of the hepatocel- lular carcinoma cases revealed an obvious trabecular pat- tern on the smear. This pattern was lacking in the adrenocortical adenomas. This is a pattern which has been described in FNA smears of primary well-differentiated hepatocellular carcinomas 18,19 and the pattern observed in the metastases is identical. Sinusoidal lining cells sur- rounding the trabecular cell groups might be further evi- dence of hepatocellular origin, however, this feature was not observed in our cases. Presence of intranuclear pseu- doinclusions would appear to be of little value since they were occasionally observed, albeit infrequently, in pri- mary adrenocortical neoplasms as well as in metastatic hepatocellular carcinoma. Nucleoli were prominent in the hepatocellular carcinoma cases, but inconspicuous in the adrenocortical adenoma cases. Although this feature may be useful in the distinction of metastatic hepatocellular carcinoma from adrenocortical adenoma, obviously, it may be of dubious value in separating metastatic hepato- cellular carcinoma from some cases of adrenocortical car- cinoma since both of our cases of adrenocortical carci- noma had prominent nucleoli. Eccentricity of the nucleus was marked in our cases of adrenocortical carcinoma, but was not observed in the hepatocellular carcinoma cases, so this feature may be of some discriminating value.
Cell block material is of obvious value in separating metastatic hepatocellular carcinoma from adrenocortical neoplasms since it affords the opportunity of performing immunoperoxidase stains. The usual patterns of im- munoreactivity in hepatocellular carcinoma have been previously reported. 20,21
Another lesion acknowledged to occasionally cause problems in the differential diagnosis of primary adrenal lesions is renal cell carcinoma. 9 We had one such case in our series, in which the diagnosis of metastatic renal cell carcinoma was made only after ultrastructural examina- tion of material retrieved from the cell block preparation.
Although the cytology of our cases of adrenocortical carcinoma differed significantly from what we observed in cases of adrenocortical adenoma, this distinction may not always be straightforward. 5,6,9 This difficulty is intuitive, given the problems which sometimes arise in this determi- nation based on surgical pathology material. Size and functional status have been offered as useful criteria since adrenocortical carcinomas are frequently functioning and of large size. 5,9 When the tumors are poorly differentiated, such as those in this series were, the diagnosis of malig- nancy will be easy, however, determination of whether the lesion is primary to the adrenal or metastatic to the adre- nal will be problematic. The feature of nuclear eccentric- ity, although by no means specific, was observed in both of our cases of adrenocortical carcinoma which had viable cellular material available for review.
We encountered one case, originally interpreted as met- astatic small cell malignancy, possibly metastatic carci- noid tumor from the lung, which on review was felt to be more likely consistent with adrenocortical adenoma. Un- fortunately, histologic proof in this case is lacking and clinical follow-up was not possible since the patient died of cardiac causes shortly after the cytologic diagnosis was made. This potential pitfall has been addressed by Min et al. 22 and the findings in their case mirror those of our case.
In conclusion, if the pathologist is forearmed with knowledge of the diagnostic pitfalls possible in the inter- pretation of needle biopsies of the adrenal gland, an accu- rate diagnosis can often be made. Consistent immediate assessment of the sample for adequacy may result in a higher adequacy rate. The technique seems particularly suited for the diagnosis of the most common metastatic tumors.
Acknowledgment
The authors thank Diana Winters and Lisa Miller for secretarial support and Etna L. Reyna, C.T. (A.S.C.P.), C.T. (I.A.C.) for technical assistance.
References
1. Page DL, DeLellis RA, Hough AJ. Tumors of the adrenal-atlas of tumor pathology. Fascicle 23, second series. Washington: Armed Forces Institute of Pathology, 1985:150.
2. Copeland PM. The incidentally discovered adrenal mass. Ann In- tern Med 1983;98:940-945.
3. Gajraj H, Young AE. Adrenal incidentaloma. Br J Surg 1993;80: 422-426.
4. Glazer HS, Weyman PJ, Sagel SS, Levitt RG, McClennan BL. Nonfunctioning adrenal masses: incidental discovery on computed tomography. Am J Radiol 1982;39:81-85.
5. Katz RL, Patel S, Mackay B, Zornoza J. Fine needle aspiration cytology of the adrenal gland. Acta Cytol 1984;28:269-282.
6. Wadih GE, Nance KV, Silverman JF. Fine-needle aspiration cytol- ogy of the adrenal gland: fifty biopsies in 48 patients. Arch Pathol Lab Med 1992;116:841-846.
7. Nguyen G-K. Percutaneous fine-needle aspiration biopsy cytology of the kidney and adrenal. Pathol Annu 1987;22 (part 1):163-191.
8. Tao L-C. Primary lesions of the adrenals. In: Tao L-C, ed. Transab- dominal fine-needle aspiration biopsy. New York: Igaku-Shoin Med- ical Publishers, 1990:218-248.
9. Katz RL. Kidney, adrenal and retroperitoneum. In: Bibbo M, ed. Comprehensive cytopathology. Philadelphia: WB Saunders, 1991: 771-805.
10. Silverman SG, Mueller PR, Pinkney LP, Koenker RM, Seltzer SE. Predictive value of image-guided adrenal biopsy: analysis of results of 101 biopsies. Radiology 1993;187:715-718.
11. Tikkakoski T, Taavitsainen M, Paivansalo M, Lahde S, Apaja-Sark- kinen M. Accuracy of adrenal biopsy guided by ultrasound and CT. Acta Radiol 1991;32 (fasc. 5):371-374.
12. Karstrup S, Torp-Pedersen S, Nolsoe C, Horn T, Hegedus L. Ul- trasonically guided fine-needle biopsies from adrenal tumors. Scand J Urol Nephrol Suppl. 1991;137:31-34.
13. Bernardino ME, Walther MM, Phillips VM, et al. CT-guided adre- nal biopsy: accuracy, safety, and indications. Am J Roentgenol 1985; 144:67-69.
DUSENBERY AND DEKKER
14. Zornoza J, Ordóñez N, Bernardino ME, Cohen MA. Percutaneous biopsy of adrenal tumors. Urology 1981;18:412-416.
15. Heaston DK, Handel DB, Ashton PR, Korobkin M. Narrow gauge needle aspiration of solid adrenal masses. Am J Roentgenol 1982; 138:1143-1148.
16. Berkman WA, Bernardino ME, Sewell CW, Price RB, Sones PJ. The computed tomography-guided adrenal biopsy: an alternative to sur- gery in adrenal mass diagnosis. Cancer 1984;53:2098-2103.
17. Koenker RM, Mueller PR, van Sonnenberg E. Interventional radiol- ogy of the adrenal glands. Semin Roentgenol 1988;22:314-322.
18. Cohen MB, Haber MM, Holley EA, Ahn DK, Bottles K, Stoloff AC. Cytologic criteria to distinguish hepatocellular carcinoma from non-neoplastic liver. Am J Clin Pathol 1991;95:125-130.
19. Bottles K, Cohen MB, Holley EA, et al. A step-wise logistic regres- sion analysis of hepatocellular carcinoma: an aspiration study. Can- cer 1988;62:558-563.
20. Johnson DE, Powers CN, Rupp G, Frable WJ. Immunocytochemi- cal staining of fine-needle aspiration biopsies of the liver as a diag- nostic tool for hepatocellular carcinoma. Mod Pathol 1992;5:117- 123.
21. Saul SH. Masses of the liver. In: Sternberg SS, ed. Diagnostic surgi- cal pathology. 2nd ed. New York: Raven Press, 1994:1537-1541.
22. Min K-W, Song J, Boesenberg M, Acebey J. Adrenal cortical nodule mimicking small round cell malignancy on fine needle aspiration. Acta Cytol 1988;32:543-546.