Associate Professor
Creighton University
University of Arizona College of Medicine
Gynecologic Oncologist, Arizona Oncology

The diagnostic performance of CA125 for the detection of ovarian and non-ovarian cancer in primary care: A population-based cohort study
PLoS Med 17(10): e1003295. https://doi.org/10.1371/journal.pmed.1003295

KOL Insight and Perspective

Background: The large majority of ovarian cancers present with Stage III or IV disease (1).   The inability to identify a screening test for ovarian cancer has led to an ongoing disappointing 5-year survival rate for this disease (2).  CA125 is a serum test that is associated with the presence of ovarian and other cancers.  Currently, the test is used either to 1) triage patients to a gynecologic oncology surgeon (in the presence of a pelvic mass) or 2) to monitor treatment response in patients with known ovarian cancer (3).  CA125 has not yet proven to be an effective screening test, with or without transvaginal ultrasound, in a population of asymptomatic women (2).  Despite disappointment as a screening test, questions remain as to the performance of CA125 in a population of women with symptoms that could be linked to the presence of ovarian cancer or whom are presenting to primary care practices.

Study Design: Funston et al studied women who had CA125 testing in a large database of primary care practices in the United Kingdom over a roughly three-and-a-half-year period.   The authors calculated measurements of CA125 test diagnostic accuracy in this population of women.  In addition, cancer probability with both age and CA125 levels was also calculated.

Results: The authors identified 50,780 women who had CA125 testing during the study period.  During the study period 0.9% were diagnosed with ovarian cancer and 2.6% with non-ovarian cancer.  With a cut-off of Ca125 ³ 35 U/mL, 15.2% had ovarian cancer in women ³ 50 years old versus 3.4% in women less than 50.  The majority of ovarian cancers were advanced stage (48.5%) but, of note, 172 (37.7%) were diagnosed at Stage I or II.

Of those women who had recorded symptoms (47.8% of study population), the most common symptom was abdominal pain (49.2%) followed by distension/bloating, changes in bowel habits, fatigue and pelvic pain.  In the general study population, a CA125 cut-off value of ³35 U/mL led to a PPV of 10.1% with a NPV of 99.8% and a sensitivity and specificity of 77% and 93.8%. respectively, for ovarian cancer.  The PPV improved when limiting to women >50 or women including all cancers but was slightly lower when only including invasive ovarian cancers.

In the UK National Institute for Health and Care Excellence, a CA124 level equating to 3% estimated cancer probability, is the threshold at which urgent referral to cancer specialist is recommended.  The authors estimated CA125 levels for various groups where this 3% cutoff is noted.  In the general population a CA125 level of 53 U/ml met these criteria for a diagnosis of ovarian cancer, a level of 68 for invasive ovarian cancers and level of 18 for all cancers.  For women <50 years old, the threshold was met at a CA125 level of 89 with 104 for a 40-year-old woman and 32 for a 70-year-old woman.

Conclusions: The authors conclude that CA125 in this population of women attending primary care clinic and having a test performed demonstrates a higher PPV than a screening population but a lower PPV than a population with a pelvic mass.  CA125 levels that direct further workup may vary for differing age groups.  In this study, a noteworthy limitation is that the reason for ordering a CA125 test was unknown although some women, roughly half, had documented symptoms.

Clinical Implications: The CA125 test could be useful in a general primary care population especially when taking into consideration variations in the test for various age groups and recognizing the possible association with non-ovarian cancers.

Bottom Line: When ordering a CA125 test in a primary care population rather than a screening population, the performance of the test may improve.  However, clinicians may want to take into consideration varying degrees of abnormal CA125 levels in different age groups.  If ovarian cancer is not identified, perhaps attention should be paid to ruling out other non-ovarian cancers rather than aborting the workup altogether.