Introduction/Background*Ovarian cancer is the leading cause of mortality from gynaecological malignancy. Survival improves with early diagnosis, however, early detection in primary care is challenging. The current blood test, cancer antigen 125 (CA125), has limited sensitivity and specificity for early disease. Human Epididymis 4 (HE4) is a promising diagnostic biomarker. We aimed to investigate the diagnostic accuracy and clinical utility of serum HE4 in a symptomatic primary care population.

Methodology We conducted a prospective observational study testing HE4 on primary care serum CA125 samples from women with suspected ovarian cancer in Manchester, UK, between April 2018 and April 2019. Serum HE4 was measured using chemiluminescent enzyme immunoassays following routine CA125 testing for clinical care. HE4 thresholds of 77pmol/L and 150pmol/L were used. The primary outcome was final diagnosis within 12 months of testing. Clinical outcomes were collected from hospital electronic patient records. Receiver operator characteristic (ROC) curves with area under the curve (AUC), sensitivity and specificity were calculated for CA125 and HE4 both alone and in combination. Age adjusted HE4 thresholds were calculated with linear regression models.

Result(s)*1247 patients were included, with a mean age of 50 years (SD 15.7). 100 women had epithelial ovarian cancer; including 82 invasive and 18 borderline ovarian tumours. There was little difference in overall performance of CA125 and HE4 (AUC 0.932 vs 0.914 respectively). At a threshold of 77pmol/L, HE4 alone had a better sensitivity than CA125 [89% (95%CI 81.2-94.4) vs 81% (95% CI 71.9-88.2)] but a worse specificity [75.6% (95%CI 73-78) vs 92.2% (95%CI 90.4-93.6)]. HE4 and CA125 combined had improved sensitivity compared with CA125 alone (93%, 95%CI 86.1-97.1), but at a significant cost to specificity (70%, 95%CI 67.3-72.6). Serum HE4 levels were correlated with increasing age (p <0.001) and worsening eGFR (p<0.001). Age adjusted HE4 cut-offs marginally improved the specificity of CA125, however the numbers were small per age category and require validation in larger cohorts.

Conclusion*HE4 adds little to current diagnostic pathways in primary care. Age-adjusted thresholds may improve accuracy, but not sufficiently to recommend routine use at present.