Use of PET/CT with Cosyntropin Stimulation to Identify and Localize Adrenal Rest Tissue following Adrenalectomy in a Woman with Congenital Adrenal Hyperplasia

Abstract

Context:

Adrenalectomy is an experimental treatment option for select patients with congenital adrenal hyperplasia who have failed medical therapy. After adrenalectomy, adrenal rest tissue can remain in extraadrenal locations, cause recurrent hyperandrogenism, and be difficult to localize.

Objective:

The aim of the study was to investigate the usefulness of positron emission tomography/computerized tomography (PET/CT) in identifying adrenal rest tissue.

Subject:

A female with salt-wasting 21-hydroxylase deficiency who had bilateral adrenalectomy at age 17 yr presented with hyperandrogenism at age 32 yr. Pelvic magnetic resonance imaging and ultrasound imaging were nondiagnostic for the source of androgen production.

Methods and Results:

A baseline F-18 labeled fluoro-2-deoxy-d-glucose (18F-FDG) PET/CT scan showed no active uptake; however, a second scan preceded by a 250-μg cosyntropin injection identified three areas of active uptake near both ovaries. Subsequent ovarian venous sampling showed elevations in 17-hydroxyprogesterone, androstenedione, and 21-deoxycortisol in both ovarian veins compared to a peripheral vein at baseline and more so after cosyntropin administration. At laparoscopy, three well-circumscribed nodules (2.4 × 0.9 × 1.3 cm, 1.2 × 1.5 × 1.5 cm, and 2 × 1.5 × 1 cm) lying lateral to the fallopian tubes adjacent to the broad ligaments were removed. The paraovarian nodules and previously removed adrenal glands had similar histology and immunohistochemistry. Postoperatively, androgen concentrations were undetectable, with no response to cosyntropin stimulation.

Conclusions:

Patients with CAH after an adrenalectomy may experience recurrent hyperandrogenism due to adrenal rest tissue. 18F-FDG PET/CT with cosyntropin stimulation accurately identified adrenal rest tissue not visualized with conventional imaging, allowing for successful surgical resection.

Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency leads to insufficient cortisol and excess androgen production. Treatment with glucocorticoid suppresses ACTH and lowers androgen levels. Undertreatment leads to virilization, whereas overtreatment may cause iatrogenic Cushing syndrome; often the balance is difficult to achieve (1).

A complicating factor in the management of CAH is the presence of adrenal rest tissue, extraadrenal tissue often found around the gonads that retains adrenocortical biosynthetic function. The adrenal cortex and gonads both arise from the urogenital ridge. One hypothesis suggests that adrenocortical cells can migrate with the gonads prenatally as they descend from the upper abdomen to their final location (2). Testicular adrenal rest is a common finding in classic CAH (3). In contrast, detection of adrenal rest in other locations is difficult and not well described. Adrenal rest cells may become functional and clinically important in conditions with excess ACTH, such as CAH. In particular, after bilateral adrenalectomy, which has been used in rare cases to treat CAH (4, 5), remnant adrenal rest tissue outside the adrenal bed can produce androgens.

F-18 labeled fluoro-2-deoxy-d-glucose (18F-FDG) positron emission tomography/computerized tomography (PET/CT) is used to assess hypermetabolic foci with increased glycolysis. Although more commonly used to detect malignant deposits, hypermetabolism may also be encountered in benign processes.

Here, we report a case study of the use of PET/CT with cosyntropin stimulation to localize metabolically active adrenal rest tissue not visualized by conventional imaging in an adrenalectomized patient with CAH, its correlations with ovarian venous sampling, and histological evaluation after surgical removal of the tissue.

Subject and Methods

A 32-yr-old female with CAH due to salt-wasting 21-hydroxylase deficiency [homozygous exon 6 cluster: p.I236N, p.V237E, p.M239K (c.710T>A, c.713T>A, c.719T>A)], whose history through age 24 was previously reported (6, 7), presented at age 16 with virilization and primary amenorrhea. Androgen production from hyperplastic adrenal glands was suspected, and venous sampling of the adrenal, ovarian, and renal veins did not indicate other androgen sources. Iodocholesterol scan showed increased uptake in both adrenals. Because of concern for hyperandrogenemia despite appropriate medical treatment, laparoscopic bilateral adrenalectomy was performed.

After adrenalectomy, her biochemical hyperandrogenemia resolved, but she presented at age 24 (7 yr postoperatively) with a return of hirsutism. A brain magnetic resonance imaging (MRI) showed a 4-mm pituitary lesion. Increased glucocorticoid dosing resulted in androgen suppression but led to weight gain. Ectopic adrenal rest tissue was suspected but was not seen on repeated imaging over several years, including ultrasound, CT, and MRI.

At age 32, the patient re-presented with hyperandrogenism and was enrolled in a protocol (www.ClinicalTrials.gov, Identifier no. NCT00250159) approved by The Eunice Kennedy Shriver National Institute of Child Health and Human Development's Institutional Review Board (Bethesda, MD). Consent was obtained after a full explanation of all procedures.

Results

PET/CT scans

A PET/CT scan from the lower head to mid thigh was performed 60 min after an injection of approximately 15.3 mCi of 18F-FDG. A low-dose spiral CT was integrated with the PET data for attenuation correction and anatomical localization. Emission and attenuation-corrected three-dimensional cine, transverse, coronal, and sagittal images were obtained. Sixty hours later, a second PET/CT was obtained 30 min after a cosyntropin bolus was given.

The PET/CT scan after cosyntropin identified two areas of active uptake near the left ovary (maximum standard uptake value, 19.9) and a possible area near the right ovary; these areas were not visualized on the baseline scan (Fig. 1, A–F). In light of these findings, nodules corresponding to the areas of PET/CT active uptake were identified on a prior MRI, but these could not have been distinguished from other nonspecific findings without the PET/CT localization.

Fig. 1.

18F-FDG PET/CT before and after cosyntropin stimulation and laparoscopic surgery findings. Baseline scans (A and C) show no areas of abnormal uptake. Scans after cosyntropin stimulation show three areas of increased uptake (B), including two definitive foci seen near the left ovary (D and E) and one possible foci near the right ovary (F), all denoted by arrows. An additional potential area of uptake on the left (B) was determined to be intraluminal within the bowel, without corresponding CT abnormality, probably a physiological variant of uptake. Three well-circumscribed nodules, corresponding to the three areas of uptake on PET/CT and lying lateral to the fallopian tubes, were isolated and surgically removed as depicted in the surgical drawing, top down view (G).

Ovarian venous sampling

The patient underwent ovarian venous sampling to evaluate the biochemical activity of the areas noted on PET/CT. Baseline samples taken in duplicate from the right and left ovarian veins (ROV, LOV) and the peripheral vein (PV) were assayed for 17-hydroxyprogesterone (17-OHP), androstenedione (A4), and 21-deoxycortisol (21-deoxy-F), a hormone produced exclusively in the adrenal glands (8). All samples were analyzed at the Mayo Medical Labs by HPLC/tandem mass spectrometry. 17-OHP (ROV, 417 ng/dl; LOV, 1250 ng/dl; PV, 76 ng/dl), A4 (ROV, 1038 ng/dl; LOV, 3780 ng/dl; PV, 36 ng/dl), and 21-deoxy-F (ROV, 72 ng/dl; LOV, 176 ng/dl; PV, 15 ng/dl) were 4- to 32-fold higher in the ROV and 9- to 124-fold higher in the LOV compared with the PV.

Cosyntropin was administered through a separate peripheral vein, and blood was collected 15, 30, and 45 min later. 17-OHP (peak values: ROV, 43,500 ng/dl; LOV, 132,000 ng/dl; PV, 311 ng/dl) and 21-deoxy-F (peak values: ROV, 6800 ng/dl; LOV, 18,000 ng/dl; PV, 68 ng/dl) increased approximately 100-fold from baseline in the ovarian veins but only 4-fold in the PV. A4 (peak values: ROV, 10,100 ng/dl; LOV, 17,700 ng/dl; PV, 70 ng/dl) increased 10-fold in the ROV, 5-fold in the LOV, and 2-fold in the PV.

Surgery

Three well-circumscribed nodules lying lateral to the fallopian tubes in the outer leaves of the broad ligaments were removed laparoscopically (two left nodules, 2.4 × 0.9 × 1.3 cm, and 1.2 × 1.5 × 1.5 cm; and one right nodule, 2 × 1.5 × 1 cm) (Fig. 1G). All three rests were ovoid and reddish-tan in color.

Histology

The three paraovarian nodules did not demonstrate ovarian parenchyma or typical adrenal zonation/medulla histologically. Tissues were composed of irregular aggregates of compact, lipid-depleted eosinophilic cells in a rich fibrotic background with frequent lymphoid aggregates with germinal centers. Cells were frequently enlarged, with nuclear atypia/hyperchromasia and inclusions. Lipofuscin-like pigment was occasionally identified (Fig. 2, A and B). The adrenals (Fig. 2, C and D), which were previously excised (7), revealed some adrenal medulla, but the cortex was hyperplastic, filled with compact, lipid-depleted eosinophilic cells. Occasional lymphoid aggregates and minimal fibrosis were present. The adrenal glands and paraovarian nodules shared histological similarities, in keeping with the diagnosis of hyperplastic adrenal rest tissue. However, the paraovarian nodules appeared to show more fibrosis, cytomegaly with nuclear atypia/inclusions, and lymphoid infiltrates than the adrenal tissue.

Fig. 2.

Histology and immunohistochemistry profile of paraovarian rests and removed adrenals. Hematoxylin and eosin staining of the paraovarian specimens (magnification, A, ×4; B, ×10) shows irregular nests of compact, lipid-depleted, enlarged eosinophilic cells in a rich fibrotic background with frequent lymphoid aggregates with germinal centers. In the patient's previously removed adrenal glands, the cortex was compact, hyperplastic with lipid-depleted eosinophilic cells, occasional lymphoid aggregates, and minimal fibrosis (C, ×4; D, ×10). Immunohistochemistry profiles for both paraovarian tissue (E–H) and the adrenals (I–L) were similar: cells are positive for CD56 (E and I), melan A (F and J), inhibin (G and K), and synaptophysin (H and L).

Routine immunostaining on the paraovarian and previously obtained adrenal specimens revealed cells positive for inhibin, synaptophysin, melan A, and CD56 (Fig. 2, E–L), a profile consistent with adrenal cortex. Chromogranin staining was minimal, and HMB45 was negative (data not shown).

Electron microscopy (data not shown) from adrenal and paraovarian rests showed cells with overlapping ultrastructural features characteristic of adrenal cortex morphology: abundant cytoplasm and small round smooth nuclei, many mitochondria, lipid droplets, lipofuscin granules, and stacks of well-developed rough endoplasmic reticulum. Although the ultrastructure could suggest Leydig cells, the lack of Reinke's crystals, clinical history, anatomic location, morphology, and immunohistochemistry support an adrenal cortex origin.

Biochemical follow-up

On postoperative day 6, the patient's baseline 17-OHP and 21-deoxy-F were undetectable; testosterone measured 9 ng/dl, and A4 was 24 ng/dl. There was no rise in hormone concentrations after cosyntropin stimulation.

Discussion

This case demonstrates the possibility of identifying adrenal rest tissue by 18F-FDG PET/CT in cases where traditional imaging, such as MRI, CT, and ultrasound, is unrevealing. The difficulty in detecting adrenal rest tissue in this female patient is in direct contrast to males with testicular adrenal rest tumors (TART), which are easily and commonly detected by both ultrasound and MRI (9–13).

Ovarian adrenal rest tumors (OART) are thought to be less common than TART (9, 12); however, the inability to identify adrenal rest tissue using traditional imaging in women may mask the true prevalence. Van Wyk and Ritzen (5) reported follow-up of 18 CAH patients who had bilateral adrenalectomy. Interestingly, eight patients had elevated steroid precursors postoperatively while on reduced hydrocortisone dose, suggesting adrenal rest activation.

Even among symptomatic patients, OART has been difficult to detect on imaging, most likely because rest tissue is concealed by other abdominal anatomy, in contrast to the relatively unobscured TART. In a recent review of seven symptomatic cases of OART in CAH, only two were found by imaging, whereas the others were detected surgically or on autopsy (14). Additionally, three patients with Cushing disease treated with adrenalectomy subsequently developed symptoms of adrenal hormone excess but were only found to have OART by invasive techniques (14, 15). The present case is unusual because the tissue was not located in the ovaries, but was instead several centimeters from the ovaries. Interestingly, our patient's adrenal rest tissue demonstrated similar histology to her previously excised adrenal glands, but with greater fibrosis and lymphocytic infiltrates. Similarly, progressive fibrosis has been observed in TART (16).

Two recent case reports also used 18F-FDG PET/CT to localize adrenal rest tissue in females (14, 17). PET/CT revealed uptake in a 4.7 × 2.4-cm adnexal mass that was easily visualized by ultrasound in an 18-yr-old hyperandrogenic adrenalectomized CAH female (14). Surgery and subsequent pathology confirmed that the mass was adrenal rest. Additionally, a 27-yr-old female with Cushing disease who had persistent Cushingoid features after bilateral adrenalectomy was found to have FDG hypermetabolism in the bilateral suprarenal regions both at baseline and after cosyntropin; however, surgery was not performed, and thus, pathological confirmation was lacking (17). In both of these cases, the rest tissue was visible on baseline PET/CT; however, our patient required cosyntropin stimulation to increase the metabolic activity in the rest tissue to detectable levels. Perhaps our patient's endogenous ACTH production was blunted by her glucocorticoid treatment, whereas the other CAH patient may have been on lower dose therapy, and the Cushing disease patient had an untreated ACTH-producing pituitary adenoma leading to up-regulation of adrenal activity.

Our case also adds biochemical confirmation of androgen production from the adrenal rest tissue, both through venous sampling preoperatively and cosyntropin stimulation postoperatively. Interestingly, our patient did not have elevated androgen production near her ovaries when ovarian venous sampling was performed as a teenager. Although we suspect the adrenal rest was present at that time, it likely was not hypertrophied enough to produce high levels of androgens until years later.

In our case, the size of each adrenal rest nodule was at least 1.5 cm in the largest dimension. It is unclear whether smaller adrenal rest tissue could be detected by PET/CT because the sensitivity of PET/CT for detection of masses less than 1 cm is limited (18), which may prevent successful utilization of PET/CT to find adrenal rest tissue. Additionally, to our knowledge, this technique has only been tested in three patients; more extensive experience is required to understand other limitations.

In summary, elusive adrenal rest tissue often complicates the management of CAH, especially in females. Adrenalectomy as a treatment for CAH, while not routinely recommended, may be considered in select cases (19). However, patients are at risk for adrenal rest activation. 18F-FDG PET/CT offers a promising tool for the detection of ectopic adrenal rest tissue. Further studies are needed to delineate the use of PET/CT in the management of patients with CAH.