Ovarian venous sampling to guide localization and treatment of female hyperandrogenism

Abstract

Virilization with plasma testosterone levels greater than 150 ng/dl raises the possibility of an ovarian or adrenal tumor. While testosterone-producing adrenal tumors are almost always seen on imaging, ovarian tumors may remain occult. In that case, ovarian venous sampling may identify and localize the source of testosterone excess. We report the results of this technique in four patients. Case reports from a tertiary care center. Ovarian ultrasound (US) and MRI were obtained and used to calculate right (R) and left (L) ovarian volumes (OVs). For ovarian venous sampling, catheters were placed and blood was drawn sequentially and simultaneously from each ovarian vein and a peripheral vein before and every 10 minutes after administration of hCG (2000 or 5000 IU IM). The ratios of ovarian:peripheral testosterone concentrations (O:P Testo) were calculated and the right (R) and left (L) ratios compared. A maximal (max) O:P Testo step-up of 2 or more was considered to localize pathology if the contralateral value was near 1. All four patients presented with virilization and hyperandrogenism (all Testo yr> upper normal range). Plasma inhibin A and/or B levels were measured in 3 patients, and were normal or low. All estradiol levels were normal. There were no ovarian masses on US or MRI. 12-year old Patient 1 presented with precocious puberty but no menarche (plasma testosterone 158 ng/dl). ROV was 24 cm3 and LOV was 23 cm3. The R O:P Testo was 4 and the L O:P Testo was 1. The right ovary was removed and found to be hyperthecotic. A subsequent plasma testosterone level was 46 ng/dl. 18-year old Patient 2 (plasma testosterone 219 ng/dl) had a ROV of 14.1 cm3 and LOV of 9.1 cm3. The O:P Testo was 2.7 on the R at 20 minutes, but anomalous blood flow prevented definitive lateralization. (Left O:P Testo = 1.2.) Retrograde flow from the L kidney descended through the ovarian vein and then crossed via pelvic collateral veins to the R ovarian vein where it ascended normally to join the upper inferior vena cava. Thus, elevated R O:P Testo could represent testosterone secretion by either ovary. Given these considerations, surgery was not recommended, and the patient continues to be followed. 62-year old Patient 3 (plasma testosterone 876 ng/dl) had ROV of 2.6 cm3 and LOV 2.8 cm3. Ten minutes after hCG stimulation the R O:P Testo was 1.5 and the L O:P Testo was 1.8. Histologic examination after bilateral oophorectomy revealed a Sertoli-Leydig tumor on the left ovary and cortical stromal hyperplasia of the right. The post-operative plasma testosterone level was 20 ng/dl. 71-year old Patient 4 (plasma testosterone 216 ng/dl) had ROV of 1.7 cm3 and LOV of 1.5 cm3. The max R O:P Testo was 3.6 at 10 minutes post-hCG and the max L O:P Testo was 0.9 at time 0. After bilateral oophorectomy (based on surgical findings) histologic examination showed bilateral hyperthecosis. The plasma testosterone level was 52 ng/dl after surgery. Ovarian venous sampling contributed to the successful identification and removal of an ovarian testosterone source in three patients with unilateral or bilateral hyperthecosis, or with a Sertoli-Leydig tumor and hyperplasia. A fourth patient did not have surgery. Thus, ovarian venous sampling may be a helpful guide to treatment decisions in cases of female hyperandrogenism.