This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Karlan, B. Y. Articles by McIntosh, M. Search for Related Content PubMed PubMed Citation Articles by Karlan, B. Y. Articles by McIntosh, M.

The Quest for Ovarian Cancer's Holy Grail: Can CA-125 Still Be the Chalice of Early Detection?

Women's Cancer Research Institute; Division of Gynecologic Oncology, Cedars-Sinai Medical Center; Geffen School of Medicine at University of California, Los Angeles, Los Angeles, CA

Martin McIntosh

Molecular Diagnostics Program, Fred Hutchinson Cancer Research Center, Seattle, WA

Since its discovery more than 25 years ago, the CA-125 tumor antigen has become the standard for monitoring ovarian cancer patients' response to therapy and for surveillance for recurrent disease. Its usefulness for screening, early detection, and risk prediction, however, has been limited by both false-positive and false-negative results. Thus, the quest for an effective ovarian cancer screening test has become a goal intensely sought after by investigators, both in the clinic and in the laboratory. Two approaches to improve early detection include evaluating ever more sophisticated uses of CA-125 alone, or looking toward the discovery of novel ovarian cancer biomarkers to augment or replace CA-125. We believe that, despite the work by Hermsen et al¹ published in this issue of the Journal of Clinical Oncology, the latter approach, involving new biomarker discovery and multiplexing combinations of markers, holds greater promise of successfully achieving this goal.

Hermsen et al¹ studied a cohort of 424 high-risk women attending their Family Cancer Clinic and report on the performance of CA-125 in finding early cancers and precancers in this population. These women were between the ages of 21 and 70 years, and nearly 80% were premenopausal; approximately one third were known to carry BRCA mutations, and one third had a prior history of breast cancer. The study's findings remind us once again of the many physiologic and benign conditions that can influence an individual's CA-125 value, such as age, menopausal status, fibroids, and pregnancy²—all factors that contribute to the false-positive and false-negative results that hamper CA-125's usefulness in screening.³

The authors claim that the change in serial CA-125 values, at least the difference between the last value taken 6 to 13 months before surgery and the one drawn at the time of surgery, was strongly predictive of ovarian cancer. This change-from-previous measure is one form of a longitudinal algorithm that has been hypothesized as a potential boon to early detection. There are others as well, including the Risk of Ovarian Cancer (ROC) algorithm of Skates et al, which uses a highly complex algorithm to assess specific growth shapes predictive of cancer.⁴ Another algorithm used in screening a high-risk population of women (similar to the one reported here by Hermsen et al) is the Parametric Empiric Bayes (PEB) algorithm.^5,6 The PEB algorithm is intended for CA-125 results or novel markers and is, perhaps, more appropriate for the study population than either the ROC or simple change-score that the authors applied. In the PEB model, each woman's complete screening history, not only her most recent index value, acts as her own control, and the slightest deviation from that person-specific norm is used to detect elevated values at a fixed specificity. This algorithm is vastly superior to any single-threshold cutoff for CA-125 and takes into account the individual's underlying nonovarian conditions, such as fibroids or her menopausal status.

It is appealing to think of CA-125 as a dynamic marker of ovarian cancer growth in this way. Its behavior on the "down slope" predicts ovarian cancer cell death and response to therapy, and its rise on the "up slope" is often the first indication of recurrent disease.⁷ De novo ovarian cancer growth has been more difficult to define; however, each of these "dynamic" approaches should certainly be able to improve screening compared to simple threshold-cutoff approaches. But the field needs to discuss how often they matter in a clinical sense: how many stage I ovarian cancers can be detected using CA-125 alone? Adding to other evidence against CA-125 alone as being sufficient to be clinically relevant is found in Table 4 of the article by Hermsen et al. The CA-125 (and ultrasound) results described therein are really quite discouraging! The serial CA-125 values preceding the diagnosis of ovarian carcinoma in these six cases do not show any trend that would suggest a precancer screening window to allow for early intervention. That is, even the most complex and most simple biomarker rules would have led to the same screening decisions as did the CA-125 values in these index cases. The prediagnostic CA-125 values reported do not demonstrate any trend, up or down, and do not suggest any increased risk of ovarian cancer beyond these women's known high risk because of their family histories or BRCA gene mutations.

Thus, despite the great attention to novel uses of CA-125, all attempts are falling short of demonstrating CA-125's clinical usefulness as an early diagnostic biomarker in this high-risk group of women. One is left still searching for new markers or a composite of markers that would point to early diagnosis.

With the vast majority of women still presenting with advanced metastatic disease, and the unlikely event that monitoring CA-125 alone will make much of a difference, there is an urgent need for effective means to detect ovarian cancer while it is still confined to the ovary and imminently curable. This need is, perhaps, best recognized in those women at high risk for developing ovarian cancer because of their family histories.⁸ Women who have inherited a deleterious germline mutation in BRCA1 or BRCA2 are acutely aware of their greatly elevated risk. Many have seen family members succumb to the disease. They often comment that they feel like a "ticking time bomb" in light of the recognized shortcomings of our current ovarian cancer screening modalities, transvaginal sonography, and CA-125 testing. Furthermore, these women are vulnerable to the hype surrounding novel screening tests that promise to deliver high sensitivity and specificity for early detection.^9,10 It is our responsibility to safeguard these women by critically evaluating candidate biomarkers and assuring that appropriate validation studies are performed.

Hermsen et al also highlight the importance of recognizing fallopian tube carcinoma as part of the BRCA mutation phenotype and the need for careful pathologic examination of the full length of the fallopian tubes in these cases.^11,12 They discovered twice the number of individual fallopian tube dysplasia compared with ovarian epithelial proliferative lesions in their prophylactic surgery population. Although it remains unclear whether these proliferative epithelial lesions of the tubal epithelium and ovarian surface are indeed the premalignant lesion, the frequency of these findings in the high-risk cohort is noteworthy. Other investigators have noted that up to 50% of fallopian tubes from BRCA mutation carriers demonstrate tubal hyperplasia or dysplasia.^13,14 Although there may be significant observer bias in these numbers, there is little doubt that the fallopian tubes are at a significant risk of malignant transformation and that tubal carcinomas are much more common in BRCA mutation carriers.

A marker capable of predicting adnexal dysplasia would be a monumental advance for this disease. Interventions could be provided before the development of cancer. Young women could delay prophylactic surgery until dysplasia was predicted. The authors claim that an absolute value of CA-125 can fit this bill. In reality, however, their results do not appear to have any substantial clinical utility. Although the median CA-125 values between the dysplastic and normal epithelium cohorts reached statistical significance (14 v 10 U/mL), the ranges of the CA-125 values overlapped almost entirely, and this finding has no real clinical significance.

So, the quest for an effective screening test for early ovarian cancer must continue. It's heartening to think that there may be a reliable precancerous lesion that could be detected preoperatively, and this may open up a new avenue for chemoprophylaxis and prevention. In the meantime, careful validation of candidate biomarkers followed by prospective clinical trials testing the markers' performance will be required. Algorithms such as the PEB that account for longitudinal biomarker changes within an individual woman, and allow for novel or composite markers' behavior to be examined as well, should be considered in future trials. It seems likely that CA-125 will remain a part of future ovarian cancer screening algorithms, but its best partners to achieve the goal of early detection remain elusive.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The authors indicated no potential conflicts of interest.

AUTHOR CONTRIBUTIONS

Conception and design: Beth Y. Karlan

Administrative support: Beth Y. Karlan

Collection and assembly of data: Beth Y. Karlan

Data analysis and interpretation: Beth Y. Karlan, Martin McIntosh

Manuscript writing: Beth Y. Karlan, Martin McIntosh

Final approval of manuscript: Beth Y. Karlan

REFERENCES

1. Hermsen BBJ, von Mensdorff-Pouilly S, Berkhof J, et al: Serum CA-125 in relation to adnexal dysplasia and cancer in women at hereditary high risk of ovarian cancer. J Clin Oncol 25:1383-1389, 2007[Abstract/Free Full Text]

2. Pauler DK, Menon U, McIntosh M, et al: Factors influencing serum CA125II levels in healthy postmenopausal women. Cancer Epidemiol Biomarkers Prev 10:489-493, 2001[Abstract/Free Full Text]

3. Urban N, McIntosh MW, Andersen M, et al: Ovarian cancer screening. Hematol Oncol Clin North Am 17:989-1005, 2003[CrossRef][Medline]

4. Skates SJ, Xu FJ, Yu YH, et al: Toward an optimal algorithm for ovarian cancer screening with longitudinal tumor markers. Cancer 76:2004-2010, 1995[CrossRef][Medline]

5. McIntosh MW, Urban N, Karlan B: Generating longitudinal screening algorithms using novel biomarkers for disease. Cancer Epidemiol Biomarkers Prev 11:159-166, 2002[Abstract/Free Full Text]

6. Crump C, McIntosh MW, Urban N, et al: Ovarian cancer tumor marker behavior in asymptomatic healthy women: Implications for screening. Cancer Epidemiol Biomarkers Prev 9:1107-1111, 2000[Abstract/Free Full Text]

7. Santillan A, Garg R, Zahurak ML, et al: Risk of epithelial ovarian cancer recurrence in patients with rising serum CA125 levels within the normal range. J Clin Oncol 23:9338-9343, 2005[Abstract/Free Full Text]

8. Friedman LC, Kramer RM: Reproductive issues for women with BRCA mutations. J Natl Cancer Inst Monogr 34:83-86, 2005[Abstract/Free Full Text]

9. Petricoin EF, Ardekani AM, Hitt BA, et al: Use of proteomic patterns in serum to identify ovarian cancer. Lancet 359:572-577, 2002[CrossRef][Medline]

10. Ransohoff DF: Lessons from controversy: Ovarian cancer screening and serum proteomics. J Natl Cancer Inst 97:315-319, 2005[Abstract/Free Full Text]

11. Carcangiu ML, Radice P, Manoukian S, et al: Atypical epithelial proliferation in fallopian tubes in prophylactic salpingo-oophorectomy specimens from BRCA1 and BRCA2 germline mutation carriers. Int J Gynecol Pathol 23:35-40, 2004[CrossRef][Medline]

12. Colgan TJ, Murphy J, Cole DEC, et al: Occult carcinoma in prophylactic oophorectomy specimens. Am J Surg Pathol 25:1283-1289, 2001[CrossRef][Medline]

13. Hermsen BBJ, van Diest PJ, Berkhof J, et al: Low prevalence of (pre) malignant lesions in the breast and high prevalence in the ovary and fallopian tube in women at hereditary high risk of breast and ovarian cancer. Int J Cancer 119:1412-1418, 2006[CrossRef][Medline]

14. Piek JMJ, van Diest PJ, Zweemer RP, et al: Dysplastic changes in prophylactically removed fallopian tubes of women predisposed to developing ovariancancer. J Pathol 195:451-456, 2001[CrossRef][Medline]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Karlan, B. Y. Articles by McIntosh, M. Search for Related Content PubMed PubMed Citation Articles by Karlan, B. Y. Articles by McIntosh, M.