Screening for Ovarian Cancer: Could it Cause More Harm than Good?
Abstract & Commentary
By Robert L. Coleman, MD, Professor, University of Texas; M.D. Anderson Cancer Center, Houston, is Associate Editor for OB/GYN Clinical Alert.
Dr. Coleman reports no financial relationship relevant to this field of study.
Source: Buys SS, et al. Effect of screening on ovarian cancer mortality: The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial. JAMA 2011;305:2295-2303.
Synopsis: Annual screening with CA125 and transvaginal ultrasound among U.S. women aged 55-74 did not improve disease-specific ovarian cancer mortality or alter stage at diagnosis among detected cases compared with usual care. Further, diagnostic evaluation in response to false-positive screens was associated with adverse complications.
Due to the frequent identification of metastatic disease at diagnosis and its high associated mortality, ovarian cancer has been the persistent focus of many screening efforts. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer screening trial is a randomized controlled trial of 78,216 women aged 55 to 74 years assigned to undergo either annual screening (n = 39,105) or usual care (n = 39,111) at 10 screening centers across the United States.1 Annual screening consisted of CA125 for 6 consecutive years and transvaginal ultrasound (TVU) for 4 consecutive years. Women randomized to "usual care" were not offered annual screening with CA-125 or TVU, but instead received well-woman care. Participants were enrolled between November 1993 and July 2001 and followed for a maximum of 13 years (median [range], 12.4 years [10.9-13.0 years]) for cancer diagnoses and death. The primary objective was to evaluate the impact of this screening algorithm on disease-specific mortality. Secondary objectives were: ovarian cancer incidence, stage at diagnosis, overall mortality, and complications associated with screening examinations and diagnostic procedures. The diagnosis of ovarian cancer was made in 212 women (5.7 per 10,000 person-years) in the screened cohort and 176 (4.7 per 10,000 person-years) in the usual care group (rate ratio [RR], 1.21; 95% confidence interval [CI] 0.99-1.48). There were 118 deaths caused by ovarian cancer (3.1 per 10,000 person-years) in the intervention group and 100 deaths (2.6 per 10,000 person-years) in the usual care group (mortality RR, 1.18; 95% CI 0.82-1.71). Of 3285 women with false-positive results, 1080 underwent surgical follow-up; of whom, 163 women experienced at least 1 serious complication (15%). There was no difference in deaths due to other causes (excluding ovarian, colorectal, and lung cancer) between the two groups (RR, 1.01; 95% CI 0.96-1.06). The authors conclude that among this cohort of women undergoing annual CA125 and TVU, screening did not result in a reduction in ovarian cancer mortality, and diagnostic evaluation following a false-positive screening test result was associated with complications.
Just over 2 years ago, I discussed in OB/GYN Clinical Alert the recently published data of the PLCO study, but the data included only four rounds of screening. At that time, compliance with the planned intervention had stabilized and there appeared to be a serial reduction in the ratio of false-positive screens for each ovarian cancer identified. However, this ratio (20 to 1) was considered unacceptable for general implementation of the strategy because the primary endpoint (reduction in disease-specific mortality) was not met. The current and final outcomes from this large screening effort have essentially discounted screening with annual CA125 and TVU in the general population. In fact, it appears indiscriminant use of annual screening with these tools may cause harm. However, there are several interesting and informative observations, which bear some discussion. First, one of the principle tenets of screening is that an abnormal test should (preferably) identify pre-invasive disease, where intervention could positively affect outcome. In this case, the screening tests (CA125 and TVU) detected a higher rate of ovarian cancer in the screened women compared to the usual care group. An important consideration is there were no differences in the stage of cancer at diagnosis or overall mortality rates between the two groups, therefore suggesting that these additional cases were of no clinical significance. This potential concern that screening for ovarian cancer would not show reduced mortality was raised recently, since mathematical models show evidence of the diseases' heterogeneity.2 These analyses suggested that "screenable" cancers may have a different biologic behavior (e.g., low-grade serous cancer), and as such, may not measurably impact overall survival, since these cancers are rare in the screened population and have a favorable prognosis.
A second important observation is that screening did not alter the detection of early- and late-stage cases. It is well appreciated that screening cannot reliably detect a pre-invasive state. However, mortality could still be reduced in a screening program if the "positive" screens, on the whole, were associated with an earlier stage than that seen in the general population. The condition for this "stage migration" could be met if the outcome of early-stage cancer is better than late-stage cancer. For instance, stage I disease is associated with an 85% or higher 5-year survival, and is significantly better than stage III/IV (about 40% 5-year survival); in addition, there are more "cures" among early-stage patients. Unfortunately, the PLCO trial clearly demonstrates that annual screening with CA125 and TVU is neither sufficiently sensitive nor specific to impact stage at diagnosis. However, this hypothesis is a potential promise for other algorithms, such as the risk of ovarian cancer analysis (ROCA), which is being explored in the yet to be reported U.K. Collaborative Trial of Ovarian Cancer Screening (UKCTOCS).3 In their prevalence report, authors appeared to identify a higher proportion of early-stage cases when serial measures of CA125 were evaluated using a mathematical prediction model and TVU triage. Mortality data from this trial is expected in 2014.
Finally, this trial serves to remind us that responding to a patient's concern, such as ovarian cancer, by utilizing unproven testing can in fact backfire and cause harm. The overall false-positive rate per round of screening was about 5%. This led to operative complications — about one for every five operations performed. The majority of operations were indicated by abnormalities on TVU. It is not known whether different criteria for morphological aberration of the ovaries could improve the positive-predictive value. However, similar results have been reported in the initial reports from the UKCTOCS trial and from a randomized controlled trial in Japanese women.4 Ultimately, serial blood-based biomarkers will need to be developed to identify high-risk women (higher prevalence) or women with early changes associated with disease onset.
- Buys SS, et al. Effect of screening on ovarian cancer mortality: The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial. JAMA 2011;305:2295-2303.
- Havrilesky LJ, et al. Development of an ovarian cancer screening decision model that incorporates disease heterogeneity: Implications for potential mortality reduction. Cancer2011;117:545-553.
- Menon U, et al. Sensitivity 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.
- Kobayashi H, et al. A randomized study of screening for ovarian cancer: A multicenter study in Japan. Int J Gynecol Cancer 2008;18:414-420.