Blood steroid levels predict survival in endometrial cancer and reflect tumor estrogen signaling
Peer reviewed, Journal article
MetadataVis full innførsel
OriginalversjonGynecologic Oncology. 2020, 156 (2), 400-406. https://doi.org/10.1016/j.ygyno.2019.11.123
Objective Blood-based biomarkers are attractive due to ease of sampling and standardized measurement technology, reducing obstacles to clinical implementation. The objective of this study was to evaluate a clinically available method of steroid hormone measurement for its prognostic potential in endometrial cancer. Methods We quantified seven steroid hormones by liquid chromatography-tandem mass spectrometry in 100 endometrial cancer patients from a prospective cohort. Abdominal fat distribution was assessed from abdominal computed tomography (CT) scans. Steroid hormone levels were compared to clinical characteristics, fat distribution and gene expression in primary tumor samples. Results Low levels of 17OH-progesterone, 11-deoxycortisol and androstenedione were associated with aggressive tumor characteristics and poor disease specific survival (p = .003, p = .001 and p = .02 respectively). Adjusting for preoperative risk based on histological type and grade, low 17OH-progesterone and 11-deoxycortisol independently predicted poor outcome with hazard ratios of 2.69 (p = .033, 95%CI: 1.09–6.68) and 3.40 (p = .020, 1.21–9.51), respectively. Tumors from patients with low steroid level displayed increased expression of genes related to mitosis and cell cycle progression, whereas high steroid level was associated with upregulated estrogen signaling and genes associated with inflammation. Estrone and estradiol correlated to abdominal fat volume in all compartments (total, visceral, subcutaneous, p < .001 for all), but not to the visceral fat proportion. Patients with higher levels of circulating estrogens had increased expression of estrogen signaling related genes. Conclusion Low levels of certain endogenous steroids are associated with aggressive tumor traits and poor survival and may provide preoperative information independent of histological biomarkers already in use.