False positive readings of HE4

Posted 12 years ago

A couple of weeks ago a friend and gynaecologist complained about HE4 and in particular about false positive readings. Traditionally we use biomarkers such as HE4 and CA125 to triage patients with pelvic masses into low and high risk of malignancy. If there is a high risk of malignancy, gynaecologists traditionally refer the patient for a gynaecological oncology opinion. In that case, we all would like to see that this referral was warranted. My colleague thought hat HE4 raises the alarm of a malignancy too often.

When HE4 was introduced not quite two years ago we thought that HE4 could help us all, GPs, general gynaecologists and cancer surgeons with our clinical decision-making.

First, I thought that HE4 would be helpful to reduce anxiety in young women diagnosed with pelvic masses and raised "cancer marker" CA125. The CA125 in these women can be scaringly high (as high as several hundred U/ml) as many of us would know. By contrast, the majority of these women will be diagnosed with an endometrioma. However, until they actually get to the operating theatre for surgical exploration and until the final histopathology results are available a few days later they worry to death about the possibility of ovarian cancer.

The second main indication that I saw was for elderly women who we could spare major surgery if the HE4 was negative. Often these women suffer from a number of medical co-morbidities, look after sick family members, etc. and surgery is the last thing they want.

Sometimes we also see women who have a normal CA125 but in fact their pelvic mass is cancerous. When we introduced HE4 to Queensland I did hope that HE4 would be elevated in those cases so that an appropriate referral to a gynaecological oncologist can be made. While that expectation was met a few times, there were more cases where the HE4 was falsely elevated.

What is the cause of false positive HE4?

First, HE4 increases with age. In a publication that we currently prepare for submission, HE4 is also high in patients with hypertension and diabetes.

A recent paper from Spain analysed 575 samples of healthy men and woman and analysed the causes of abnormally high levels fro CA125 and HE4. False positive readings for CA125 were found in 37% of patients and false positive readings for HE4 were found in only 12% of patients.

The most common causes for false positive CA125 values were benign gynaecological conditions (endometrioisis, adenomyosis, fibroids). CA125 can also be elevated in patients with arthritis and diverticular disease.

For HE4, renal failure was the most common cause of false positive readings. Patients with renal failure also had higher HE4 levels than patients with benign gynaecological conditions. Patients with effusions (not specified), with liver or lung disease also had false positive HE4 readings.

The authors noted that HE4’s diagnostic specificity in patients with benign gynaecological conditions was higher than CA125. In patients with benign gynecologic conditions, concentrations were above the upper limit of the reference interval in 1.3% of patients for HE4 compared to 33.2% for CA 125. Comparing CA125 and HE4 to correctly predict malignancy, HE4 was consistently between 10% and 20% more accurate.

It seems that HE4 is consistently superior to CA125 and I always like to combine the two markers. I suspect no biomarker ever will be 100% accurate. However, sometimes I quite like to review the reasons why we misinterpret data and come to wrong conclusions.

First, HE4 is obviously elevated in a number of conditions that are linked with renal impairment. I saw a patient last week with a pelvic mass that looked not suspicious on ultrasound. She had marked renal impairment and indeed, her HE4 was around the 200 mark. The more severe the renal impairment, the higher the HE4 serum level. The HE4 companies who market the HE4 kits should provide us with more research on this association so that we are able to make better decisions in the future.

The ROMA algorithm is too simplistic and not good enough. To distinguish between ovarian cancer and benign pelvic masses, the ROMA cut-off values depend on whether our patient is pre- or postmenopausal. By contrast, HE4 continuously and steadily increases with age. In addition, the algorithm has been developed using data from patients with mostly advanced ovarian cancer. In these patients we don’t need tumour markers because a diagnosis of ovarian cancer can be made clinically (omental cake, ascites, etc). We need comparative data from patients with ovarian cancer confined to the ovary (stage 1 ovarian cancer) versus benign controls.

Overall, HE4 seems to do a good job but the current algorithms that recommend clinical decisions are out-dated and require readjustment. In the meantime, we need to adjust for renal function when we interpret HE4 results.