Lawrence M. Weiss, M.D.
Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors
ABSTRACT A series of 43 adrenocortical tumors was analyzed using nine histologic features. Mitotic activity, especially with atypical forms, and venous invasion cor- related best with metastasizing or recurring tumors; however, no single criterion was useful alone. The com- bination of the following nine criteria was most useful in distinguishing malignant from benign tumors: nuclear grade III or IV; mitotic rate greater than 5/50 high-power fields; atypical mitoses; clear cells comprising 25% or less of the tumor; a diffuse architecture; microscopic necrosis; and invasion of venous, sinusoidal, and capsular struc- tures. None of the 24 tumors with two or less of these criteria metastasized or recurred, while all but one of the 19 tumors with four or more of these criteria either re- curred or metastasized.
Am J Surg Pathol 8: 163-169, 1984.
From the Department of Pathology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachu- setts.
INTRODUCTION
Adrenocortical tumors are rare and only a few large series have been reported. The distinction between benign and malignant adrenocortical tumors can often be difficult, and the diagnosis is usually made after consideration of a combination of clinical, gross, and microscopic features. The significance of microscopic features has been questioned,(9-11) and one major textbook of surgical pathology(14) states that microscopic findings alone do not provide a reliable index of survival. It would be extremely useful, however, to be able to determine malignant potential from microscopic features alone. With the advent of modern noninvasive imaging techniques, adrenocortical tumors are being discovered earlier and at a smaller size, increasing the need for more accurate microscopic diagnosis. Also, on occasion, an adrenocortical tumor is first discovered at surgery and extensive preoperative hormonal evaluation is not available. The purpose of this investigation was to analyze a series of adrenocortical neoplasms to ascertain whether one or a combination of micro- scopic features could be used as an effective pre- dictor of carcinoma.
MATERIALS AND METHODS
Sixty-five cases coded as adrenal cortical adenoma and carcinoma were retrieved from the pathology files of the Peter Bent Brigham and Brigham & Women’s Hospitals from the period 1959-1980. Cases chosen for study met the following criteria: 1) absence of coexistent visceral carcinoma at time
of diagnosis; 2) at least 5-year follow-up or docu- mentation of metastasis or local recurrence; and 3) tumor definitely an adrenocortical tumor upon re- view of slides and clinical information. Ten tumors were discarded from the study due to coexistent carcinoma; 11 tumors (nine adenomas and two carcinomas) were discarded due to inadequate fol- low-up, and one tumor was discarded due to un- certainty of primary site. Forty-three tumors meeting the criteria were studed microscopically. Nine features were carefully assessed: 1) nuclear grade; 2) mitotic rate; 3) atypical mitoses; 4) char- acter of cytoplasm; 5) architecture of tumor cells; 6) necrosis; 7) invasion of venous structures; 8) invasion of sinusoidal structures; and 9) invasion of capsule of tumor. Nuclear grade was assessed from grade I to IV according to the criteria of Fuhrman et al.(1) used in the evaluation of renal cell carcino- ma. Mitoses were evaluated by counting 10 random high-power fields in the area of the greatest numbers of mitotic figures on the five slides with the greatest numbers of mitoses. If less than five slides were available for a case, a correspondingly greater number of fields per slide were used, to make a total of 50 high-power fields. The counts were performed on an American Optical microscope using a 45X objective with a 0.47-mm-diameter field. Mitoses were regarded as atypical when they definitely showed an abnormal distribution of chromosomes or an excessive number of mitotic spindles. The cytoplasm was evaluated for the percentage of clear or vacuolated cells resembling the normal zona fasciculata. The architecture was assessed as diffuse if greater than one-third of the tumor formed pat- ternless sheets of cells, and was regarded as non- diffuse when two-thirds of the tumor showed orga- nization into patterns, whether nesting, trabecular, or of another type. Necrosis was regarded as present when occurring in at least confluent nests of cells. Invasion of sinusoidal and venous structures and capsule was only accepted when unequivocal. A sinusoid was arbitrarily defined as an endothelial- lined vessel in the adrenal gland with little sup- porting tissues and a vein was defined as an endo- thelial-lined vessel with smooth muscle as a com- ponent of the wall. Invasion of capsule was accepted as present when nests or cords of tumor extended into or through the capsule, with a corresponding stromal reaction.
RESULTS
The clinical and gross pathologic features of the 43 cases are summarized in Table 1. The original
| Metasta- sizing/ Recurring Tumors | Nonmetasta- sizing/ Nonrecurring Tumors | Total | |
|---|---|---|---|
| Age | |||
| Median | 45 yrs. | 50 yrs. | 49 yrs. |
| Range | 26-70 yrs. | 20-64 yrs. | 20-70 yrs. |
| Sex | |||
| Male | 7 | 10 | 17 |
| Female | 11 | 15 | 26 |
| Side | |||
| Left | 9 | 15 | 24 |
| Right | 9 | 10 | 19 |
| Clinical Syndrome | |||
| Conn's | 2 | 15 | 17 |
| Cushing's | 4 | 9 | 13 |
| Virilization | 6 | 1 | 7 |
| Gynecomastia | 2 | 0 | 2 |
| No endocrine | 4 | 0 | 4 |
| syndrome | |||
| Size (Greatest Diameter) | |||
| Median | 14 cm | 2 cm | 3.5 cm |
| Range | 3.5-22 cm | 1-21 cm | 1-22 cm |
| Weight | |||
| Median | 510 g | 10 g | 18 g |
| Range | 38-1582 g | 1-1458 g | 1-1582 g |
diagnoses were 25 adenomas and 18 carcinomas. Twenty-five tumors had no evidence of metastasis or recurrence with a median follow-up time of 11 years (range: 5-23 years). Seventeen tumors me- tastasized, with metastases becoming evident at a median time of 2 months (range: time of surgery to 4 years). One tumor recurred at 11 years postoper- atively (in the contralateral pelvis).
The microscopic findings are summarized in Table 2. The most useful criteria in distinguishing malignant tumors were a mitotic rate greater than 5/50 high-power fields, the presence of atypical mitoses, and the presence of venous invasion (Figs. 1 and 2). Less absolute, but also very helpful, pre- dictors of carcinoma were the presence of necrosis, the presence of invasion of sinusoids and tumor capsule, the presence of a diffuse architecture, and a nuclear grade of III or IV. Unfortunately, there were exceptional cases in which nonmetatasizing lesions had one or more of these latter features (Figs. 3-6). The least useful criterion was the character of the cytoplasm. However, nonmetastasizing and nonrecurring tumors were generally found to have a more conspicuous clear cell component than the metastasizing or recurring tumors which were pre- dominantly nonclear cell type.
10
| Metastasizing/ Recurring Tumors | Nonmetasta- sizing/ Nonrecurring Tumors | |
|---|---|---|
| Nuclear Grade | ||
| I | 0 | 14 |
| HI | 4 | 8 |
| III | 6 | 2 |
| IV | 8 | 1 |
| Mitotic Rate (per 50 hpf) | ||
| 0 | 2 | 23 |
| 1-5 | 2 | 2 |
| 6-10 | 3 | 0 |
| 11-50 | 8 | 0 |
| Greater than 50 | 3 | 0 |
| Atypical Mitoses | ||
| Absent | 11 | 25 |
| Present | 7 | 0 |
| Cytoplasm | ||
| 76-100% clear | 2 | 18 |
| 51-75% clear | 0 | 0 |
| 26-50% clear | 1 | 3 |
| 0-25% clear | 15 | 4 |
| Architecture of Tumor | ||
| Nondiffuse | 5 | 23 |
| Diffuse | 13 | 2 |
| Necrosis | ||
| Absent | 1 | 23 |
| Present | 17 | 2 |
| Invasion of Venous | ||
| Structures | ||
| Absent | 9 | 25 |
| Present | 9 | 0 |
| Invasion of Sinusoidal | ||
| Structures | ||
| Absent | 8 | 23 |
| Present | 10 | 2 |
| Invasion of Capsule of | ||
| Tumor | ||
| Absent | 8 | 24 |
| Present | 10 | 1 |
FIGURE 2 Venous invasion is demonstrated, with a plug of tumor present in an endothelial-lined vessel with a wall which on the left includes smooth muscle. Metastasizing tumor.
As can be seen from Table 2, no one criterion alone could be used to distinguish all malignant from benign tumors. An attempt was then made to de-
| No. of Criteria Presentª | Metastasizing/ Recurring Tumors | Nonmetastasizing/ Nonrecurring Tumors |
|---|---|---|
| 0 | 0 | 16 |
| 1 | 0 | 6 |
| 2 | 0 | 2 |
| 3 | 0 | 0 |
| 4 | 4 | 0 |
| 5 | 3 | 1 |
| 6 | 3 | 0 |
| 7 | 5 | 0 |
| 8 | 1 | 0 |
| 9 | 2 | 0 |
a Nuclear grade III or IV, mitotic rate greater than 5/50 hpf, atypical mitoses, clear cells comprising 25% or less of the tumor, diffuse ar- chitecture, necrosis, invasion of venous structures, invasion of si- nusoidal structures and invasion of tumor capsule.
termine if use of all nine features in combination could separate the two classes. Each tumor was graded from 0 to 9 according to how many of the following features were present or absent: 1) nuclear grade III or IV; 2) mitotic rate greater than 5/50 high-power fields; 3) atypical mitoses; 4) clear cells comprising 25% or less of the tumor; 5) diffuse ar- chitecture; 6) necrosis; 7) invasion of venous struc- tures; 8) invasion of sinusoidal structures; and 9) invasion of tumor capsule. These results are sum- marized in Table 3. As seen in this table, all but one of the nonmetastasizing and nonrecurring tumors had two or less of the nine criteria, while all metas- tasizing or recurring tumors had four or greater of the nine criteria.
DISCUSSION
Although there are excellent reports of series of adrenocortical carcinomas,(4,5,7) there are few pathologic studies devoted to a comparison of benign
and malignant adrenocortical tumors. Heinbecker et al.,(2) in 1958, compared 10 malignant with six benign tumors and concluded that there were his- tologic criteria that tended to differentiate the two classes. They stated that carcinoma is usually as- sociated with pleomorphism, mitoses (including atypical forms), hemorrhage, necrosis, and calcifi- cation; absolute criteria were considered to be invasion of veins or capsule, and distant metastases. However, their study size was limited, their data were qualitative rather than quantitative, and their follow-up times were not optimal.
In 1966, Neville and Symington(11) reported on 18 patients with primary aldosteronism, two of whom had malignant tumors. They found that the diagnosis of a malignant tumor could not be made solely on morphologic criteria since mitotic activity and vascular invasion frequently were absent. They did find that an alveolar pattern and prominent thick-walled vascular channels were a common feature of the carcinomas.
In 1968, Schteingart et al.(12) reported a series of 12 cases of Cushing’s syndrome; five were adenomas
and seven were carcinomas. They concluded that histologic distinction between adenoma and carci- noma was possible and that the most reliable criteria of carcinoma were capsular invasion, numerous and atypical mitoses, and large cells with pleomorphic nuclei and irregular distribution of nuclear chro- matin. They also believed that blood vessel invasion was seen only in carcinomas, but was not a common feature.
Lewinsky et al.,(8) in 1974, reported 20 cases of nonfunctional adrenocortical tumors and reviewed the literature. They concluded that reliable indica- tions of the metastatic potential of these tumors could not be made from histologic examination, and recommended that all these tumors be viewed as malignant. They found that capsular invasion or vascular invasion were independent of the potential behavior.
Tang and Grey,(6) in 1975, reported a comparison study on 16 adrenocortical carcinomas and 23 ad- enomas. Unfortunately, 12 of the adenomas were found at autopsy and thus follow-up was impossible. In their study, they found tumor weight as the fea-
ture most consistently correlated with outcome. Microscopic criteria which they found useful in the diagnosis of carcinoma included architectural dis- array, pleomorphism, increased mitotic activity, vascular invasion, hemorrhage, and necrosis. They concluded, however, that criteria aside from histo- pathology were important in predicting prognosis.
In 1976, Kay(6) reported a descriptive study of adrenal pathology which included 10 cortical car- cinomas and 16 cortical adenomas. Although the study does not mention any follow-up, this author stressed poor differentiation, abundant mitoses, bizarre nuclei, and spindling of cells in addition to evidence of capsular or venous invasion as criteria for carcinoma. He noted that cellular pleomorphism may occur in adenomas and should not be used alone as an indicator of carcinoma.
Hough et al.,(3) in 1979, published the largest and perhaps most careful study on prognostic factors in adrenocortical tumors. In a statistical analysis, this group studied 41 patients retrospectively with a minimum follow-up time of 5 years on all patients. Multiple histologic and nonhistologic criteria were studied. They found that no single criterion sepa- rated benign from malignant tumors, but that the summation of 12 histologic and nonhistologic cri- teria using a statistical weighting system enabled total separation between metastasizing and non- metastasizing tumors. The major problems with this study are that it may be difficult to apply this system to individual cases and that a criterion such as weight loss (their best single predictor) may not be readily available.
The current study was designed to address some of the problems encountered in these previous studies, particularly those of small study size, lack of adequate follow-up, and difficulty in application of criteria. The present study defines the value of nine histologic criteria in analyzing the behavior of adrenal cortical tumors, providing a simple method applicable by practicing surgical pathologists. Certain features such as mitotic activity, especially with atypical forms, and venous invasion were especially important in the evaluation, but all the criteria studied were helpful. The use of the nine criteria in combination permitted separation of the tumors into two groups. In the first group of 24 tu- mors with two or less of the nine criteria, there were no metastases or recurrences. In the second group of 19 tumors with four or more of the nine criteria, there was only one tumor without either metastasis or recurrence. This tumor was 1458 g, 21 × 13 × 11 cm, and partially necrotic. It occurred in a 33- year-old female presenting with virilization, and an 8-year follow-up was available. This tumor, which
would have been considered malignant on combined clinical, gross, and microscopic features, may indeed have been a malignant tumor and its potential may have been realized had a longer follow-up time been available.
The histologic data in this study are essentially in agreement with those of Hough et al.,(3) with the exception that capsular invasion was much less frequent in nonmetastatic tumors in the present series as compared to Hough’s cases. This may be explained by the difference in definition of capsular invasion, which in Hough’s study was defined as “nests of tumor present in capsule.” In the present study, a corresponding stromal reaction was an additional requirement in order to exclude “pseu- doinvasion,” the phenomenon described by Schteingart et al.(12) where in compressed residual adrenal tissue interposed between tumor and capsule of the gland may simulate capsular invasion in ad- enomas.
Several caveats should be mentioned. Although minimum follow-up was 5 years and the median follow-up was 11 years on nonmetastasizing tumors, on occasion adrenocortical tumors can recur or metastasize in a longer period of time (e.g., a re- currence at 11 years in one patient in this series). Also, it is acknowledged that there must be rare adrenocortical tumors that fall within the border zone of this classification system. In these instances, as in all cases, consideration of clinical and gross pathologic features outlined in Table 1 is important, and ultimately, rigorous follow-up studies on these borderline tumors will elucidate their biology. ☐
References
1. Fuhrman, S.A., Lasky, L.C., and Limas, C .: Prognostic significance of morphologic parameters in renal cell carci- noma.
Am J Surg Pathol 6: 655-663, 1982.
2. Heinbecker, P., O’Neal, L.W., and Ackerman, L.V .: Functioning and nonfunctioning adrenal cortical tumors. Surg Gynecol Obstet 105: 21-33, 1957.
3. Hough, A.J., Holifield, J.W., Page, D.L., and Hartmann, W.H .: Prognostic factors in adrenal cortical tumors. A’ mathematical analysis of clinical and morphologic data. Am J Clin Pathol 72: 390-400, 1979.
4. Huvos, A.G., Hajdu, S.I., Brasfield, R.D., and Foote, F.W .: Adrenal cortical carcinoma. Clinicopathologic study of 34 cases. Cancer 25: 354-361, 1970.
5. Ibanez, M.L .: The pathology of adrenal cortical carcinoma. A study of 22 cases. In Endocrine and Non-endocrine Hormone-Producing Tumors. Year Book Medical Pub- lishers, Chicago, 1971, pp. 231-239.
6. Kay, S .: Hyperplasia and neoplasia of the adrenal gland. Pathol Annu 11: 103-139, 1976.
7. King, D.R., and Lack, E.E .: Adrenal cortical carcinoma. A clinical and pathologic study of 49 cases. Cancer 44: 239-244, 1979.
8. Lewinsky, B.S., Grigor, K.M., Symington, T., and Neville, A.M .: The clinical and pathologic features of “non-hor- monal” adrenocortical tumors. Cancer 33: 778-790, 1974.
9. Lipsett, M.B., Hertz, R., and Ross, G.T .: Clinical and pathophysiologic aspects of adrenocortical carcinoma. Am J Med 35: 374-383, 1963.
10. Macfarlane, D.A .: Cancer of the adrenal cortex: The natural history, prognosis, and treatment in a study of fifty-five cases. Ann R Coll Surg Engl 23: 155-186, 1958.
11. Neville, A.M., and Symington, T: Pathology of primary al- dosteronism. Cancer 19: 1854-1868, 1966.
12. Schteingart, D.E., Oberman, H.A., Friedman, B.A., and Conn, J.W .: Adrenal cortical neoplasms producing Cushing’s syndrome. A clinicopathologic study. Cancer 22: 1005-1013, 1968.
13. Tang, C.T., and Gray, G.F .: Adrenocortical neoplasms. Prognosis and morphology. Urology 5: 691-695, 1975.
14. Waisman, J .: The adrenal glands. In Surgical Pathology, W.F. Coulson, Ed. J.B. Lippincott, Philadelphia, 1978, pp. 862-878.
Acknowledgments
The author thanks Joseph M. Corson, M.D., for his help in all stages of this work; Nancy J. Hunter, M.D., and Michael J. Warhol, M.D., for their critical review of the manuscript; and Anthony Merola for his excellent pho- tographic assistance.
Write for reprints to: Lawrence M. Weiss, M.D., De- partment of Pathology, Stanford University Medical Center, Stanford, California 94305.