Risk Assessment in 457 Adrenal Cortical Carcinomas: How Much Does Tumor Size Predict the Likelihood of Malignancy?
Cord Sturgeon, MD, Wen T Shen, MD, Orlo H Clark, MD, FACS, Quan-Yang Duh, MD, FACS, Electron Kebebew, MD, FACS
| BACKGROUND: | Laparoscopic adrenalectomy for tumors > 6 cm is controversial because of the risk of malig- nancy, but data to support this position are mostly from small series. The recent NIH consensus conference did not make a definitive recommendation about management of 4- to 6-cm nonfunctioning incidentalomas. |
| STUDY DESIGN: RESULTS: | Adrenocortical carcinomas (ACC) recorded in the Surveillance, Epidemiology, and End Results (SEER) database (1988 to 2000) were compared with benign functional or nonfunctional adrenal cortical adenomas (excluding aldosteronomas) operated on at our institution between January 1, 1993, and July 1, 2003. Data were compared using t-tests, chi-square tests, likelihood ratios, and receiver operating characteristic (ROC) curves. We identified 457 patients with ACC and 47 patients with adrenal cortical adenomas; 376 and 44 neoplasms, respectively, had tumor size data available. Tumor size was larger in ACC (12.0 ± 5.6 versus 4.2 + 1.9 cm, mean ± SD, p < 0.05). For ACC presenting with local disease, the sensitivity, specificity, and likelihood ratios of tumor size to predict malignancy were 96%, 52%, and 2.0, respectively, for tumors ≥ 4 cm; 90%, 80%, and 4.4 for tumors ≥6 cm; 77%, 95%, and 16.9 for tumors ≥ 8 cm; and 55%, 98%, and 24.4 for tumors ≥ 10 cm. Assuming a pretest probability of malignancy of 5%, the likelihood ratios derived from this study yield a posttest probability of 10%, 19%, and 47% for cancer in adrenal cortical tumors ≥ 4 cm, ≥6 cm, and ≥8cm, respectively. |
| CONCLUSIONS: | These data suggest that size is useful for predicting malignancy, and that at a size threshold of ≥ 4 cm, the likelihood of malignancy doubles (to 10%) and it is more than ninefold higher for tumors ≥8 cm (47%). (J Am Coll Surg 2006;202:423-430. @ 2006 by the American College of Surgeons) |
Despite the fact that laparoscopic adrenalectomy for ad- renal cortical neoplasms has been widely performed with excellent success over the past decade,1 application of
Competing Interests Declared: None.
Supported in part by The Robert Wood Johnson/Harold Amos Faculty De- velopment Program Fellowship, The Hellman Family Grant, The Albert Clark Family Foundation, The Helen and Sanford Diller Foundation, Friends of Endocrine Surgery, and Mt Zion Health Systems.
Received March 16, 2005; Revised October 24, 2005; Accepted November 4, 2005.
From the Department of Surgery, University of California, San Francisco, and UCSF Comprehensive Cancer Center at Mount Zion Medical Center, San Francisco, CA.
Correspondence address: Electron Kebebew, MD, UCSF/Mt Zion Medical Center, Department of Surgery, 1600 Divisadero St, Hellman Bldg, Room C3-47, San Francisco, CA 94143-1674.
this technique to potentially malignant lesions is contro- versial. Reports by Ross and Aron2 and others have sug- gested that for adrenal cortical neoplasms larger than 6 cm, the rate of malignancy is 35% to 98%. Laparo- scopic adrenalectomy for tumors larger than 6 cm is controversial because of this perceived increased risk of malignancy. Controversy about whether laparoscopic adrenalectomy is indicated continues even though there are no convincing data to support the assumption that, when performed correctly, laparoscopic complete adre- nalectomy is oncologically inferior to open adrenalec- tomy for malignant adrenal neoplasms.
Because up to 5% of incidentally discovered adrenal tu- mors are ultimately found to be malignant,3-5 and differen-
Abbreviations and Acronyms
ACC = adrenocortical carcinoma
FNA = fine-needle aspiration
LR = likelihood ratio
NPV = negative predictive value
PPV = positive predictive value
ROC = receiver operating characteristic
SEER = Surveillance, Epidemiology, and End Results
UCSF = University of California, San Francisco
tiation between benign and malignant cortical tumors of- ten is not possible preoperatively, a method of identifying malignant adrenal cortical lesions preoperatively would be beneficial. Currently, the choice of surgical approach is dic- tated by tumor size, signs of local invasion or metastases, and surgeon experience. When designing this study, we assumed that knowledge of a high likelihood of malignancy would lead to an open surgical approach.
The natural history of laparoscopic surgery for ma- lignant adrenal cortical neoplasms is unknown be- cause there is a paucity of data on the longterm fol- lowup of patients who have undergone laparoscopic adrenalectomy for adrenocortical carcinomas (ACC). Data about the surgical treatment of ACC are largely from the prelaparoscopic era. The 5-year actuarial survival for patients with ACC ranges from 16% to 60%.4,6 Recurrent and metastatic disease is common, even in patients who undergo complete open resec- tion. Approximately two-thirds of patients develop recurrence within 2 years, and approximately 85% eventually develop local recurrence or distant metas- tases.6,7 These dismal data from open resections should be kept in mind when interpreting outcome data. For this reason, it might be impossible to detect a difference in recurrence or survival with laparo- scopic adrenalectomy. In addition, a randomized pro- spective trial comparing the two methods is not fea- sible considering the rarity of ACC, the distinct advantage that laparoscopic adrenalectomy has in be- nign disease, and the difficulty in accurately discrim- inating malignant from benign adrenal cortical le- sions preoperatively.
Single institution databases are unlikely to be useful for statistical comparisons with benign adrenal neo- plasms because ACC is so rare. We hypothesized that an analysis of the clinical and demographic data on ACC from the National Cancer Institute’s Surveillance, Epi- demiology, and End Results (SEER) database would al-
low us to identify factors common to ACC. To establish differences between malignant and benign adrenal cor- tical neoplasms and to better characterize the relation- ship between tumor size and risk of malignancy, we compared the SEER data with our own surgical experi- ence with benign adrenal cortical adenomas (excluding aldosteronomas) over a similar time period.
METHODS
Approval for this project was obtained from the Univer- sity of California, San Francisco Committee on Human Research and is filed under approval number H5241- 24389 to 01.
The SEER database, compiled by the NCI, repre- sents one of the largest sources of information about cancer incidence and survival in the United States. Data in the SEER database are accrued from 14 re- gional cancer registries and include demographics, primary site, morphology, stage at diagnosis, first treatment, and followup for survival. Tumor size for malignant adrenal cortical neoplasms has been re- corded only since 1988. So we limited our review of the SEER database8 to patients treated for ACC be-
| SEER | UCSF |
|---|---|
| Age at diagnosis | Age at presentation |
| Diagnostic confirmation | Diagnosis |
| Gender | Race |
| Race/ethnicity | Gender |
| Tumor size (1988+) | Past medical history |
| Laterality | Pathology |
| Primary site (ICD-O-2) | Tumor size |
| Grade (ICD-O-2) | Presentation |
| AJCC stage 3rd edition (1988+) | Procedure details |
| Lymph nodes examined | |
| Extent of nodal disease | |
| First primary malignancy | |
| Radiation treatment and sequence | |
| Reason for no cancer-directed surgery | |
| Site-specific surgery | |
| Survival time (mo) | |
| Cause of death | |
| Type of reporting source | |
| Vital status | |
| Year of diagnosis |
AJCC, American Joint Committee on Cancer; SEER, Surveillance, Epidemi- ology, and End Results; UCSF, University of California San Francisco.
80
70
☐ localized
60
regional
☒ distant
50
N
40
51.5
49.6
50.3
30
*
20
T
I
T
10
**
11.2
11.5
13.8
5.2
4.7
4.2
0
Mean age (yrs)
Mean survival (mos)
Mean size (cm)
tween 1988 and 2000. Data recorded in the SEER database for ACC are listed in Table 1.
We reviewed our surgical database of patients who underwent adrenalectomy for benign adrenal cortical tumors at the University of California, San Francisco (UCSF) between 1993 and 2003. Patients with aldoste- ronomas were excluded because these tumors are uni- formly small and benign. Tumor size was determined from preoperative imaging studies. Data included in this database are also listed in Table 1.
Comparisons were made between patients with be- nign and malignant adrenal cortical tumors based on these databases. The major variables studied were those common to both databases: patient age at diagnosis, gender, race, and tumor size.
SEER data were stratified by stage at presentation (localized, regional, or distant disease; Fig. 1). Subgroup analysis was then performed to compare benign adrenal neoplasms with ACC presenting with localized disease only. This subgroup of localized ACC was studied be- cause these tumors are difficult to distinguish from be- nign tumors preoperatively because of the lack of re- gional invasion or distant metastases at the time of presentation.
Data from benign and malignant adrenal cortical neoplasms were grouped and then stratified based on
tumor size. Statistical analyses were performed using the two-tailed t-test, Mann-Whitney rank sum test, and chi- square test, where appropriate. Sensitivities, specificities, positive predictive values (PPV), negative predictive val- ues (NPV), and likelihood ratios (LR) were calculated for tumor sizes at 2-cm intervals. Receiver operating characteristic (ROC) curves were plotted based on these data. The ROC curve plots sensitivity versus 1-specificity (true positives versus false-negatives) at each tumor size threshold and can be used as a method of assessing the accuracy of a test. Additionally, the ROC curve graphically demonstrates the tradeoff between sensitivity and specificity for the different threshold val- ues of a diagnostic test. The ROC curve may also high- light the appropriate threshold value for calling a diag- nostic test positive.
We defined sensitivity as the percentage of ACC that would have been correctly identified preoperatively if the tumor was greater than or equal to a particular size threshold. Specificity was defined as the percentage of benign cortical tumors that would have been correctly identified preoperatively if the tumor was less than a particular size threshold. The LR for each tumor size threshold was calculated by determining the ratio of the probability of that tumor size occurring in a patient with
| Variables | SEER malignant (n=457)* | UCSF benign (n=47)+ |
|---|---|---|
| Gender, n(%) | ||
| Male | 196 (43) | 15 (32) |
| Female | 261 (57) | 32 (68) |
| Race, n (%) | ||
| Caucasian | 398 (87) | 32 (68) |
| African American | 29 (6) | 2 (4) |
| Other | 30 (7) | 13 (28) |
| Age at presentation, y | ||
| Range | 0-92 | 5-77 |
| Mean ± SD | 51.3 ± 18.6 | 49.9 ± 15.5 |
*Data from SEER database for ACC cases from 1988 to 2000.
+Data from UCSF for benign adrenal cortical neoplasms from 1993 to 2003.
ACC (true positives) to the probability of that tumor size in a patient without cancer (false positives). A test of no predictive value yields a likelihood ratio of 1. The higher the likelihood ratio, the greater the patient’s probability of having disease when the test result is pos- itive. In this case, a higher likelihood ratio indicates a higher probability that the tumor size threshold will be exceeded in a patient with an ACC. Data are presented as mean ± standard deviation unless otherwise noted.
RESULTS
We identified 457 patients with ACC in the SEER da- tabase between 1988 and 2000. From the UCSF surgical database, we identified 47 patients with benign adrenal cortical neoplasms (excluding aldosteronomas) who un- derwent adrenalectomy between January 1, 1993, and July 1, 2003; 44 of 47 patients had tumor size data available.
The demographic characteristics of the patients in the two groups are summarized in Table 2. The gender dis- tribution was similar between the two groups (p = 0.15). The distribution of races between the two groups was different in the proportion of Caucasian versus non- Caucasian patients (p = 0.02); 32% of the patients with an adenoma were non-Caucasian and only 13% of the patients with ACC were non-Caucasian. There was no statistically significant difference in age between the two groups (p = 0.74).
The exact surgical approach is not recorded in the SEER database. Of the 457 patients with ACC, 123 were classified as having radical surgery; 48 had total resection; 157 had partial or simple resection;, 54 had no
| SEER malignant n=376* | UCSF benign n=44+ | |
|---|---|---|
| Tumor size, cm Mean + SD | 12.0 ± 5.6 | 4.2 ± 1.9* |
| Tumor size stratification (cm), n (%) | ||
| <2 | 5 (1) | 1 (2) |
| 2.0-3.9 | 6 (2) | 22 (50) |
| 4.0-5.9 | 24 (6) | 12 (27) |
| 6.0-7.9 | 45 (12) | 7 (16) |
| 8.0-9.9 | 63 (17) | 1 (2) |
| ≥ 10 | 233 (62) | 1 (2) |
*Tumor size data were available for 376 of 457 malignant tumors from the SEER database.
*Tumor size data were available for 44 of 47 benign tumors for data from UCSF.
*Mean tumor size differed significantly between the two groups (Mann- Whitney rank sum test, p < 0.05).
SEER, Surveillance, Epidemiology, and End Results; UCSF, University of California, San Francisco.
surgery; 46 had exploration or biopsy or both; 12 had excisional biopsy; 3 underwent local tumor destruction; and 1 underwent debulking. The extent of operation was unknown for 13.
For the 44 benign adrenal tumors in the UCSF data- base, 42 were removed by the lateral transabdominal laparoscopic approach, and 2 were removed by the pos- terior laparoscopic approach. No operations were con- verted to open procedures.
The size of the resected tumor was available for 376 of 457 ACC. Mean tumor size differed between the malig- nant and benign groups (12.0 ± 5.6 versus 4.2 ± 1.9 cm; p < 0.05). The tumor size distribution is listed in Table 3.
Sensitivities, specificities, PPV, and NPV were calcu- lated based on incremental tumor size thresholds and are listed in Table 4. Likelihood ratios for each of the tumor size thresholds were used to draw the ROC curve for tumor size as a predictor of malignancy for all adrenal cortical neoplasms (Fig. 2).
Stratification of SEER data by stage at presentation revealed that of the 457 patients with ACC, 192 had local disease, 94 had regional invasion, 145 had distant disease, and in 26, the stage was unknown. In subgroup analysis of the 192 ACC patients presenting with local disease only, the mean tumor size was 11.2 ± 5.4 cm. Tumor size distribution is listed in Table 5. Sensitivities, specificities, PPV, and NPV are listed in Table 6. The LR
| Threshold tumor size, cm | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Likelihood ratio |
|---|---|---|---|---|---|
| ≥2 | 99 | 2 | 90 | 17 | 1.01 |
| ≥ 4 | 97 | 52 | 95 | 68 | 2.03 |
| ≥6 | 91 | 80 | 97 | 50 | 4.43 |
| ≥ 8 | 79 | 95 | 99 | 34 | 17.32 |
| ≥ 10 | 62 | 98 | 99 | 23 | 27.27 |
*Sensitivities, specificities, positive predictive values (PPV), negative predictive values (NPV), and likelihood ratios (LR) were calculated for size as a predictor of malignancy for tumor size thresholds in 2-cm increments.
for each tumor size threshold was used to draw the ROC curve (Fig. 3) for this subgroup analysis.
To make a judgment about the predictive value of a diagnostic test to detect a disease, the prevalence (or pretest probability) of disease in the population to which the test is being applied must be known. The probability of disease when a diagnostic test result is positive is de- pendent on the probability of disease in the population to which the test is being applied. In this study, sensitiv- ities, specificities, positive and negative predictive val- ues, and likelihood ratios were calculated from data that included 173 (80%) malignant and 44 benign cortical adenomas. This is dramatically different from what is found in the true population of adrenal cortical neo- plasms, in which only up to 5% (of incidentalomas) are believed to be malignant.3-5
To illustrate the effect of pretest probability on the posttest probability of disease, and to estimate the pre- dictive value of size for malignancy, probabilities were calculated and are displayed in Table 7 using pretest probabilities of 80% and 5%, respectively. The likeli- hood ratio is derived from the sensitivity and specificity of a test, which, in theory, does not change regardless of the patient population to which the test is applied, so is
1.0
0.9
sensitivity
0.8
0.7
0.6
0.5
0.0
0.2
0.4
0.6
0.8
1.0
1-specificity
the same in both examples. In the first example in Table 7, for tumors larger than 8 cm, there appears to be a high posttest probability of cancer (99%) when a pretest probability of 80% is used. In the second part of Table 7, we set the pretest probability of malignancy at 5% be- cause only approximately 5% of incidentally discovered adrenal adenomas are found to be malignant. If the like- lihood ratios derived from this study are valid for the larger population of all adrenal cortical neoplasms, as- suming a cancer prevalence of 5%, the posttest proba- bility of a tumor 8 cm or larger being malignant is 47%.
DISCUSSION
Without evidence of regional invasion or metastatic dis- ease on imaging study, no preoperative diagnostic test can reliably differentiate between malignant and benign adrenal cortical neoplasms. Because ACCs are rare, few centers have accumulated enough patients to produce data reflective of the general population with ACC. So
| SEER malignant n=173 (%)* | UCSF benign n=44 (%)+ | |
|---|---|---|
| Tumor size, cm Mean + SD | 11.2 ± 5.4 | 4.2 ± 1.9* |
| Tumor size stratification (cm), n (%) | ||
| <2 | 2 (1) | 1 (2) |
| 2.0-3.9 | 5 (3) | 22 (50) |
| 4.0-5.9 | 10 (6) | 12 (27) |
| 6.0-7.9 | 23 (13) | 7 (16) |
| 8.0-9.9 | 37 (21) | 1 (2) |
| ≥ 10 | 96 (55) | 1 (2) |
*Tumor size data were available for 173 of 192 local malignant tumors from the SEER database.
*Tumor size data were available for 44 of 47 benign tumors for data from UCSF.
Mean tumor size differed significantly between the two groups (p < 0.05).
| Threshold tumor size, cm | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Likelihood ratio |
|---|---|---|---|---|---|
| ≥2 | 99 | 2 | 80 | 33 | 1.01 |
| ≥ 4 | 96 | 52 | 89 | 77 | 2.01 |
| ≥6 | 90 | 80 | 95 | 67 | 4.40 |
| ≥ 8 | 77 | 95 | 99 | 51 | 16.91 |
| ≥ 10 | 55 | 98 | 99 | 36 | 24.42 |
*Sensitivities, specificities, positive predictive values (PPV), negative predictive values (NPV), and likelihood ratios (LR) were calculated for size as a predictor of malignancy for tumor size thresholds in 2-cm increments.
we reviewed a comprehensive national cancer registry to identify a large cohort with ACC. From this database we were able to identify 457 patients with ACC treated between 1988 and 2000. We reviewed our experience at UCSF and found 47 patients who had adrenalectomies for benign adrenal cortical adenomas between 1993 and 2003. We then compared the demographic and clinical profiles of our patients with those in the SEER database with the aim of identifying preoperative predictors for differentiating between malignant and benign adrenal cortical neoplasms. Because the purpose of this study was to identify differences between malignant and be- nign neoplasms that cannot otherwise be distinguished by standard imaging modalities, we also performed sub- group analysis of patients with ACC presenting with local disease only.
As expected, the mean ACC size was significantly larger than that of the benign cortical adenomas. We stratified the tumors by size, which revealed that in- creased tumor size correlated with a higher likelihood of malignancy. It has previously been observed that larger adrenal tumor size is associated with an increased risk of malignancy,2,9 and our study confirms this observation and provides a more precise estimate of this risk.
Although the SEER database contains valuable demo-
1.0
0.9
sensitivity
0.8
0.7
0.6
0.5
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1-specificity
graphic profiles and survival data, it is limited in several areas, including information on diagnostic and imaging tests, specific mode of surgical treatment (eg, laparo- scopic versus open), recurrence, and followup. Only if databases prospectively record these details of how adre- nal cortical neoplasms are treated will we have more useful information in the future. In addition, compari- son of a national cancer database with an institutional database is not optimal for drawing accurate conclusions about the disease in the larger population, especially when the proportion of malignant disease is dramati- cally overrepresented. So we determined the posttest probability of disease assuming a pretest probability of malignancy more reflective of the general popula- tion (5% incidence of cancer among adrenal incidentalomas).
Our data on adrenal cortical lesions presenting with only localized disease suggest that the likelihood of ma- lignancy doubles for tumors larger than or equal to 4 cm (to 10%) and increases more than ninefold for tumors larger than or equal to 8 cm (47%). A tumor size thresh- old of 4 cm has a sensitivity of 96% and a specificity of 52% for malignancy. These findings compare favorably with those of Terzolo and associates,3 who found that a tumor size of 5 cm had a sensitivity of 93% and speci- ficity of 64% for malignancy. Our results also reveal that a tumor size threshold of 8 cm has a sensitivity of 77% and a specificity of 95% for malignancy.
The need for surgical resection of a nonfunctioning adrenal tumor is determined by tumor size or imaging features that are suspicious for malignancy. Based on the current limited information, we recommend that pa- tients who have adrenal tumors with imaging findings of local invasion, regional lymphadenopathy, or metastases should have an open adrenalectomy. Ultrasound or CT- guided fine-needle aspiration (FNA) biopsy is advocated by some experts10,11 to distinguish malignant from be-
| Pretest probability, % | Pretest odds | Threshold tumor size, cm | Likelihood ratio | Posttest odds | Posttest probability, % |
|---|---|---|---|---|---|
| 80 | 4.00 | ≥2 | 1.01 | 4.04 | 80 |
| ≥4 | 2.01 | 8.04 | 89 | ||
| ≥6 | 4.40 | 17.60 | 95 | ||
| ≥ 8 | 16.91 | 67.64 | 99 | ||
| ≥ 10 | 24.42 | 97.68 | 99 | ||
| 5 | 0.05 | ≥2 | 1.01 | 0.05 | 5 |
| ≥ 4 | 2.01 | 0.11 | 10 | ||
| ≥6 | 4.40 | 0.23 | 19 | ||
| ≥8 | 16.91 | 0.89 | 47 | ||
| ≥ 10 | 24.42 | 1.29 | 56 |
*These data were derived from subgroup analysis of only those malignant adrenal cortical cancers presenting with localized disease. In reality, only about 5% of incidentally found adrenal cortical neoplasms are malignant, not 80%. So in this example, we have calculated the posttest probability of disease based on 80% prevalence and compared it with 5% prevalence. The pretest probability can be expressed as an odds ratio. The pretest odds are multiplied by the likelihood ratio to yield the posttest odds. The odds are then converted back to probabilities. A test of no value does not change the pretest probability, but a valuable test substantially changes the magnitude of the posttest probability of disease.
nign adrenal tumors, but we believe it is contraindicated unless trying to rule out metastatic disease. A disadvan- tage of FNA is the high false-negative rate, so it often provides little helpful information. There is also a poten- tial for complications, including seeding of the needle tract, hemorrhage, or fatal hemodynamic instability if the tumor is a pheochromocytoma.12 For these reasons, we do not recommend or use FNA. In essence, cytologic findings by FNA rarely alter clinical decision making about the need for adrenalectomy.
Different size criteria for removal of nonfunctioning adrenal tumors have been advocated, including equal to or larger than 3,13,14 4,4,15-20 5,3,13,21,22 or 69,23 cm. Our approach for the past several years has been to resect adrenal incidentalomas larger than 3 cm in young pa- tients with low operative risk, and those 5 cm or larger in older patients with significant comorbidities.13,24 Le- sions initially too small to merit resection are removed if growth is observed during the followup period.24 The 44 benign cortical adenomas with known tumor size in this series were all removed laparoscopically. If we had strictly followed a 6-cm threshold for performing open adrenal- ectomy, 20% of these benign tumors would have been removed by open adrenalectomy. An 8-cm threshold would have resulted in a 4.5% open resection rate for these benign tumors.
Intraoperative findings, rather than a strict reliance on tumor size, should determine whether a patient under- goes laparoscopic versus open adrenalectomy for adrenal cortical tumors. Our current operative strategy for pa- tients with adrenal cortical neoplasms with apparently
local disease (even large tumors) is to begin with a diag- nostic laparoscopy. Even if the lesion is thought to be benign, we routinely resect the periadrenal fatty connec- tive tissue in continuity with the adrenal gland. If there is intraoperative suspicion of malignancy, such as difficulty in dissection because of tumor adhesions, tumor inva- sion into adjacent structures, or technical difficulty be- cause of tumor size, we do not hesitate to convert to either open or hand-assisted adrenalectomy. In this series of benign adrenal cortical neoplasms, there were no con- versions from laparoscopic to open adrenalectomy.
We believe that the majority of adrenal cortical neo- plasms can be safely resected laparoscopically. Based on the results of this study, we recommend that all adrenal cortical tumors larger than or equal to 4 cm be removed. There should be a high suspicion for ACC when an adrenal lesion is 8 cm or larger or there are imaging findings suspicious for malignancy. Surgeons should not hesitate to convert to the open or hand-assisted tech- nique when dissection is difficult because of large tumor size, adhesions, or invasion that might risk capsular frac- ture or tumor spillage.
Author Contributions
Study conception and design: Sturgeon, Kebebew Acquisition of data: Sturgeon, Shen, Duh, Kebebew Analysis and interpretation of data: Sturgeon, Kebebew Drafting of manuscript: Sturgeon
Critical revision: Duh, Clark, Kebebew Statistical expertise: Sturgeon, Kebebew
Obtaining funding: Kebebew Supervision: Kebebew
REFERENCES
1. Kebebew E, Siperstein AE, Clark OH, Duh QY. Results of laparoscopic adrenalectomy for suspected and unsuspected ma- lignant adrenal neoplasms. Arch Surg 2002;137:948-951; dis- cussion 952-953.
2. Ross NS, Aron DC. Hormonal evaluation of the patient with an incidentally discovered adrenal mass. N Engl J Med 1990;323: 1401-1405.
3. Terzolo M, Ali A, Osella G, Mazza E. Prevalence of adrenal carcinoma among incidentally discovered adrenal masses. A ret- rospective study from 1989 to 1994. Gruppo Piemontese Inci- dentalomi Surrenalici. Arch Surg 1997;132:914-919.
4. Thompson GB, Young WF Jr. Adrenal incidentaloma. Curr Opin Oncol 2003;15:84-90.
5. Young WF Jr. Management approaches to adrenal incidentalo- mas. A view from Rochester, Minnesota. Endocrinol Metab Clin North Am 2000;29:159-185.
6. Vassilopoulou-Sellin R, Schultz PN. Adrenocortical carcinoma. Clinical outcome at the end of the 20th century. Cancer 2001; 92:1113-1121.
7. Pommier RF, Brennan MF. An eleven-year experience with ad- renocortical carcinoma. Surgery 1992;112:963-970; discussion 970-971.
8. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Public-Use Data (1973-2001). National Cancer Institute, DCCPS, Surveillance Research Program, Can- cer Statistics Branch; 2004.
9. Copeland PM. The incidentally discovered adrenal mass. Ann Intern Med 1983;98:940-945.
10. Lumachi F, Borsato S, Tregnaghi A, et al. CT-scan, MRI and image-guided FNA cytology of incidental adrenal masses. Eur J Surg Oncol 2003;29:689-692.
11. Lumachi F, Borsato S, Brandes AA, et al. Fine-needle aspiration
cytology of adrenal masses in noncancer patients: clinicoradiologic and histologic correlations in functioning and nonfunctioning tumors. Cancer 2001;93:323-329.
12. McCorkell SJ, Niles NL. Fine-needle aspiration of catecholamine- producing adrenal masses: a possibly fatal mistake. AJR Am J Roentgenol 1985;145:113-114.
13. Duh QY. Adrenal incidentalomas. Br J Surg 2002;89:1347-1349.
14. Staren ED, Prinz RA. Selection of patients with adrenal inciden- talomas for operation. Surg Clin North Am 1995;75:499-509.
15. Dackiw AP, Lee JE, Gagel RF, Evans DB. Adrenal cortical car- cinoma. World J Surg 2001;25:914-926.
16. Proye C, Pattou F. Adrenocortical carcinoma: Nonfunction- ing and functioning. In: Clark OH, Duh QY, eds. Textbook of endocrine surgery. Philadelphia: WB Saunders; 1997: 490-496.
17. Farley D, van Heerden JA. Incidental adrenal masses. In: Doherty GM, Skogseid B, eds. Surgical endocrinology. Phila- delphia: Lippincott Williams & Wilkins; 2001:213-219.
18. Miccoli P, Bernini GP. Adrenocortical carcinoma. In: Doherty GM, Skogseid B, eds. Surgical endocrinology. Philadelphia: Lip- pincott Williams & Wilkins; 2001:263-271.
19. Herrera MF, Grant CS, van Heerden JA, et al. Incidentally dis- covered adrenal tumors: an institutional perspective. Surgery 1991;110:1014-1021.
20. Graham DJ, McHenry CR. The adrenal incidentaloma: guide- lines for evaluation and recommendations for management. Surg Oncol Clin North Am 1998;7:749-764.
21. Al-Fehaily M, Duh QY. Adrenocortical carcinoma. In: Clark OH, Duh QY, Perrier N, Jahan T, M., eds. Endocrine tumors. Hamilton, ON: BC Decker; 2003:123-130.
22. Al-Fehaily M, Duh QY. Malignant adrenal neoplasms. Probl Gen Surg 2003;20:92-102.
23. NIH State-of-the-Science Statement on management of the clinically inapparent adrenal mass (“incidentaloma”). NIH Consens State Sci Statements 2002;19:1-23.
24. Lal G, Duh QY. Laparoscopic adrenalectomy-indications and technique. Surg Oncol 2003;12:105-123.