Case Report

A 30 year old was referred to the Department of Reproductive Endocrinology for ovarectomy because of persistently elevated estradiol levels in the range of 114–118 pg/mL (enhanced estradiol assay; ADVIA Centaur Immunoassay, Siemens, Malvern, PA). She reported amenorrhea and hypoestrogenic symptoms, including severe hot flashes and vaginal dryness. She had stage 4A breast carcinoma and was on ovarian suppressive therapy with leuprolide acetate (Lupron; AbbVie Inc, North Chicago, IL) and fulvestrant (Faslodex; Cayman Chemical Co, Ann Arbor, MI), an estrogen receptor (ER) antagonist. Despite undergoing a radical mastectomy, radiation, and chemotherapy, she had progressive metastatic disease.

Our clinical evaluation focused on the assessment of estrogen bioactivity. A pelvic examination revealed a pale pink vaginal mucosa with decreased rugation. Ultrasound revealed an endometrial lining thickness of 2.8 mm and quiescent ovaries. Repeat immunoassays for follicle-stimulating hormone and estradiol (Estradiol 33540; Beckman Coulter Access Immunoassay System, Brea, CA) in our in vitro fertilization laboratory demonstrated an estradiol level of 12 pg/mL and follicle-stimulating hormone of 4.64 mIU/mL. These findings were consistent with pituitary suppression and low estradiol production.

Comment

Reduction of estrogen is key in treating ER-positive breast cancer. Fulvestrant, a selective ER antagonist, is a structural analog of 17β-estradiol and is indicated for use in metastatic breast cancer. In contrast to gonadotropin-releasing hormone agonists, fulvestrant is not known to inhibit ovarian function directly.

For this patient, the measurement of serum estradiol was utilized as a surrogate marker for estrogen activity by the oncologist. The measurement of low estradiol levels is challenging because of the lack of standardization at these levels and the variation in sensitivity/specificity among different immunoassays. The lower range of estradiol can be less than 1 pg/mL in breast cancer patients treated with aromatase inhibitors.

The gold standard assay for measurement of estradiol is extraction and purification with gas/liquid chromatography-tandem mass spectrometry. This purification/separation step improves specificity and sensitivity by removing estrogen metabolites such as estrone sulfate that can cross-react with antibodies in the immunoassay. Many new automated assays lack this critical step.

Addition of the purification step in commercial immunoassay kits results in approximately 70% reduction in artifact in postmenopausal women using an aromatase inhibitor, resulting in more accurate estradiol levels. The cross-reactivity of the antibody with metabolites indicates decreased specificity of binding to the epitope, leading to an overestimation of estradiol. Antibody cross-reactivity with other estrogen-like molecules including fulvestrant may also lead to errors because of the similarity in the steroid structure. The presence of heterophile antibodies can also result in the discrepancy of estradiol measurement. Unfortunately, the serum samples were not serially diluted in the initial assay to determine whether heterophile antibodies were present.

In this case, there was a clear discrepancy between the clinical evaluation of estrogen activity and the reported estradiol levels. Both estradiol measurements were performed using immunoassay kits that used a competitive assay format without purification. The sensitivity ranges of the first and second assays were similar (12-3000 pg/mL and 20-4800 pg/mL). The difference between the assays was the specificity of the antibody to the epitope of the substrate measured, with the first assay using an antiestradiol sheep monoclonal antibody and the second assay using an antiestradiol rabbit polyclonal antibody. Theoretically, the monoclonal-derived antibody targets a single epitope of the estradiol moiety, whereas the polyclonal-derived antibody targets many epitopes.

Importantly, the second assay had increased analytical specificity of the antibody with less than 5% cross-reactivity to substances similar in structure to estradiol. Thus, the discrepancy of the estradiol measurement was most likely caused by antibody cross-reactivity with metabolites in the assay.

Clinicians often base their management decisions on laboratory tests, without taking into account the entire clinical scenario. In this case, the clinical bioassay for estrogen activity did not agree with the laboratory results. When this discrepancy occurs, ovarian function as measured by follicular growth and endometrial lining thickness can be quickly evaluated by ultrasound to provide immediate feedback on estrogen production. It is important for gynecologists to recognize the limitations and variations in estradiol measurement at low levels. Accurate interpretation can avoid erroneous decisions in management of patients who are on estrogen suppressive therapy including unnecessary surgical intervention.