See commentary on pages 2854–5, this issue.
CA 125 and the detection of recurrent ovarian cancer†
A reasonably accurate biomarker for a difficult disease
Article first published online: 7 APR 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 12, pages 2850–2853, 15 June 2010
How to Cite
Bast, R. C. (2010), CA 125 and the detection of recurrent ovarian cancer. Cancer, 116: 2850–2853. doi: 10.1002/cncr.25203
- Issue published online: 11 JUN 2010
- Article first published online: 7 APR 2010
- Manuscript Revised: 23 NOV 2009
- Manuscript Accepted: 23 NOV 2009
- Manuscript Received: 16 NOV 2009
Since the first description of the assay by our group more than 25 years ago,1 the measurement of serum cancer antigen 125 (CA 125) has contributed to the care of patients with ovarian cancer in several different ways. The original application of CA 125 was in monitoring the response of ovarian cancer during chemotherapy and in detecting persistent disease after primary treatment.2 In patients whose tumors shed sufficient quantities of CA 125 to be elevated in peripheral blood, biomarker levels have tracked tumor volume with up to 90% accuracy.3 After cytoreductive surgery and combination chemotherapy, persistent elevation of CA 125 levels has been reported to be correlated with persistent disease in >90% of cases.2, 3 However, CA 125 is not optimally sensitive, and up to 50% of patients with normal levels of CA 125 after chemotherapy were found to have small volumes of persistent disease at second-look surgery.2, 3
At the time of diagnosis, elevated CA 125 has been combined with ultrasonography4 and other biomarkers5-7 to identify those patients with pelvic masses who are most likely to have ovarian cancer and who would benefit from referral to a gynecologic oncologist for primary surgery. Perhaps the most promising application of serial CA 125 determinations is to identify a small fraction of healthy postmenopausal women who would benefit from transvaginal sonography to detect ovarian cancer at an early stage.8 Early results from the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) trial involving 200,000 women in the United Kingdom suggested that rising CA 125 followed by transvaginal sonography can nearly double the fraction of tumors detected in early stage and requires no more than 3 surgeries for each case of ovarian cancer detected.8 A National Cancer Institute Specialized Program of Research Excellence (SPORE)- sponsored trial in the United States has confirmed these observations, albeit on a much smaller scale.9 The UKCTOCS trial is adequately powered to detect a survival advantage and results will be available within the next 3 to 4 years.
CA 125 has also been used over the years to detect recurrence of ovarian cancer after primary therapy. Greater than one-half of the women treated for advanced disease with cytoreductive surgery and combination chemotherapy will experience a complete clinical response with normalization of CA 125 levels and without evidence of macroscopic disease on imaging studies. If “second-look” surgeries are performed, greater than one-half of patients in “clinical remission” will have macroscopic or microscopic metastases that fall beneath the limits of resolution for imaging and the sensitivity of CA 125. Even when second-look procedures are negative, the majority of patients will experience recurrence of disease within months to years. Increasing levels of CA 125 precede the signs and symptoms of disease recurrence by 3 months to 5 months in as many as 70% of cases.10, 11 The practice in the United States has been to monitor CA 125 every 3 months during the first years after primary treatment, on the assumption that the detection of disease recurrence would translate into more effective treatment of small-volume disease.
A recent study, presented as a plenary paper by Rustin and van der Burg at the 2009 American Society of Clinical Oncology (ASCO) Annual Meeting, has questioned the value of monitoring patients for disease recurrence with CA 125.12 In this trial conducted previously in the United Kingdom, patients presumed to be in complete remission after primary therapy underwent CA 125 determinations every 3 months, but were blinded to the results. When CA 125 values doubled outside the normal range, patients were randomized to either have their physicians informed of the rising value or not. Some 265 patients in the “early” group were treated at the discretion of the participating physicians with second-line and occasionally third-line chemotherapy. Another 264 patients in the “delayed” group were treated with second-line chemotherapy when their recurrent disease became symptomatic or clinically obvious, some 4.8 months later than the “early” group, in whom treatment had been based on CA 125 levels. The study accrued slowly and required more than 9 years to complete, but in the final analysis no difference was observed with regard to overall survival, nor was any improvement found in quality of life as a result of the earlier detection of recurrent disease. The quality of life deteriorated in both groups, but this occurred 2.6 months sooner in the group treated “early,” and was correlated primarily with the side effects of chemotherapy, particularly fatigue. CA 125 had accurately predicted disease recurrence, but earlier treatment had not had a significant impact on clinical outcome and had slightly, but significantly, hastened a decline in quality of life.
Rustin and van der Burg12 should be congratulated for their persistence and organizational skills in performing a study in multiple institutions over nearly a decade. Their trial addresses an important problem and challenges the status quo. Physicians in the United Kingdom and the United States have been appropriately concerned by the anxiety surrounding each CA 125 determination in a fraction of patients.13 Small increases in CA 125 can also prompt negative imaging studies, with their associated inconvenience and expense. In addition, CA 125 can rise persistently in the absence of abnormalities on imaging or physical examination, posing the therapeutic dilemma of whether and how to treat a rising CA 125 level. On the basis of the recent ASCO report, the UK investigators have argued that there is no value in the routine measurement of CA 125 in the follow-up of ovarian cancer patients who attain a complete response after first-line treatment and that the practice should change.
However, since the initiation of the UK trial of Rustin et al in 1996,12 both the use of CA 125 and the standards for chemotherapy for recurrent disease have evolved. Before we change clinical practice and abandon monitoring for recurrence of ovarian cancer based on a single negative study, it will be important to consider the limitations of the trial, as well as the rationale for treating recurrent ovarian cancer at an early stage.
Although the “early” and “delayed” arms of the trial were well-balanced with regard to many relevant prognostic variables, there are several technical problems with the trial design that may have led inadvertently to an imbalance of the arms. Patients were not stratified based on the degree of primary cytoreduction or tumor grade. Trial participants were restaged after primary surgery and chemotherapy with CA 125 determination and imaging, but the modalities and criteria for imaging were not standardized. Even with computed tomography imaging, at least one-half of those patients with a normal CA 125 level will be found to have macroscopic residual disease at second-look surgery and this is more likely in the context of suboptimal initial cytoreduction. Randomization of more patients with suboptimal cytoreduction, high-grade tumors, and macroscopic residual disease after treatment to the “early” arm could well have nullified a modest improvement in overall survival.
Of greater concern is that the CA 125 level had to double outside the normal range before physicians were informed of potential disease recurrence. At the time the study was planned, this was a reasonable benchmark and produced nearly 5 months of lead time. We now know that increases in CA 125 within the normal range of 35 U/mL can precede disease recurrence with an even greater lead time.14, 15
In addition, physicians were simply informed that the patient's CA 125 level was rising and each physician was free to choose whether to treat, when to treat, and what drugs to use. Randomization was completed within 1 month after the CA 125 determination. Although 75% of patients had chemotherapy initiated within a little more than a month after randomization, 3 months were required for 90% of patients to have second-line chemotherapy initiated and 4% of patients on the “early” arm of the trial never received treatment. With only 4.8 months of lead time, treatment for approximately one-quarter of the patients was unacceptably delayed.
Of critical importance is the finding that only approximately one-third of patients received a combination of carboplatin and a taxane; approximately two-thirds received single-agent therapy or a combination that lacked a taxane. The ICON 4/AGO trial was completed in the United Kingdom during the same years as the study of Rustin et al12 and demonstrated that a combination of carboplatin and paclitaxel produced significantly longer progression-free and overall survivals than did treatment with carboplatin alone as a single agent.16 Consequently, approximately one-quarter of patients on the “early” arm were treated late and approximately two-thirds received suboptimal treatment by today's standards. If these factors are independently assorted, only 25% of patients would have been treated promptly with optimal chemotherapy and could have benefited from the early detection of disease recurrence.
In this trial, CA 125 accurately predicted disease recurrence. An apparent failure to impact on survival was correlated with the inadequacy of therapy for recurrent disease. In the future, how are we to improve treatment for recurrent and primary ovarian cancer? On average, women with ovarian cancer survive only 12 to 18 months after clinically apparent disease recurrence, but there is a small fraction of women who survive up to a decade after responding to multiple drugs, individually and in combination. Currently, there are at least 7 conventional drugs available to treat recurrent ovarian cancer, producing an objective response rate. In the absence of an effective predictive test, oncologists generally prescribe single drugs or 2-drug combinations sequentially, requiring 2 to 3 months to determine the response to each regimen. Waiting for recurrent disease to grow to a point at which it causes symptoms or can be readily palpated will shorten the interval available to test these conventional agents and to provide patients an opportunity to benefit.
Just as important, the CA 125 trial underscores the critical need to improve the treatment of patients with recurrent ovarian cancer. At the current time, there are more than 400 new drugs being developed to treat different forms of cancer. Almost certainly, combinations of these agents will be required for optimal benefit. Currently, only 4% of Americans with cancer participate in clinical trials. For ovarian cancer, this fraction may be even lower in that a smaller number of patients meet Response Evaluation Criteria In Solid Tumors (RECIST) criteria. By unnecessarily delaying the detection of disease recurrence, we are likely to further decrease participation in clinical trials, because women will have fewer months with adequate performance status.
The trial by Rustin and van der Burg does mandate that oncologists discuss with each patient whether she wishes to have her disease monitored for recurrence with CA 125.12 If a patient did not want to be treated with multiple or novel agents at the time of disease recurrence, she could be reassured that the early detection of recurrence would not translate into longer life. If, however, a patient would wish to receive multiple conventional drugs or to consider participating in clinical trials, monitoring her CA 125 level would provide additional months for treatment.
In a cost-conscious healthcare environment, 1 possible outcome of the study reported by Rustin and van der Burg at ASCO12 would be for Medicare and other third-party payers to deny coverage for monitoring disease recurrence with CA 125 or other biomarkers, regardless of the wishes of patients or physicians. This would be unfortunate, because it would be based on a single limited study, would ignore the progress made in monitoring and therapy since 1996, and would assume that there will be no further improvements over the next several years.
CONFLICT OF INTEREST DISCLOSURES
Dr. Bast receives royalties for CA 125 and has served on the scientific advisory boards of Fujirebio Diagnostics, Inc and Vermillion. This commentary was supported, in part, by the National Cancer Institute Ovarian SPORE 5P50 CA83639 and by the Zarrow Foundation.
- 6A panel of 7 biomarkers improves specificity in discriminating malignant from benign pelvic masses. Proc Int Gynecol Cancer Soc. 2006. Abstract 0288., , , et al.
- 7Validation of a panel of 7 proteomic biomarkers in an independent and blinded sample set. Proc Int Gynecol Cancer Soc. 2006. Abstract 0255., , , et al.
- 8Sensitivity and specificity of multimodal and ultrasound screening for ovarian cancer, and stage distribution of detected cancers: results of the prevalence screen of the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). Lancet Oncol. 2009; 10: 327-340., , , et al.
- 9Optimizing a 2-stage strategy for early detection of ovarian cancer [abstract]. Proceedings of the National Cancer Institute Translational Science Meeting, Location, Inclusive Dates, 2008. Abstract 292., , , et al.
- 12A randomized trial in ovarian cancer (OC) of early treatment of relapse based on CA125 level alone versus delayed treatment based on conventional clinical indicators (MRC OV05/EORTC 55955 trials) [abstract]. J Clin Oncol. 2009; 27( 18 suppl): 5s., .