Symptom Screening Indexes Do Not Detect Early Ovarian Cancer

Roxanne Nelson

January 20, 2012

July 20, 2012 — In an attempt to diagnose ovarian cancer at an earlier and more curative stage, the use of symptom indexes is increasing. However, these indexes might need to be revisited, according to a study published in the January 18 issue of the Journal of the National Cancer Institute.

In the study, the investigators developed 2 new indexes, but found that their performance was similar to the most established measure, the Goff index.

They also found that sensitivity decreased in all indexes when the symptoms detected in the 3 months before a diagnosis of ovarian cancer were excluded.

"At best, a symptom index might advance diagnosis of ovarian cancer by 3 months or more in two thirds of women," write the authors, led by Peter Sasieni, PhD, from the Wolfson Institute of Preventive Medicine, Queen Mary University of London, United Kingdom.

Thus, there is room for improvement. Diagnosing ovarian cancer when the patient is already symptomatic is unlikely to make a major difference in outcome, write Patricia Hartge, ScD, from the National Cancer Institute, and James L. Speyer, MD, from the Langone Cancer Center, New York University, New York City, in an accompanying editorial. They wonder if efforts should be focused on "the time between when a tumor is large enough to cause mild symptoms and when it is large enough to cause symptoms that prompt women to call a physician."

Goff Index Is the Standard

The value of a symptom index is largely dependent on how it is used and how index-positive women are managed, the study authors explain. Therefore, the method of symptom assessment can influence index performance.

Some evidence suggests that patient-reported symptoms can be used as a screening tool and might be helpful in detecting the disease early. However, the authors note that one of the primary challenges to this approach is the low specificity of ovarian cancer symptoms. This has the potential to cause patient anxiety, unnecessary interventions, and an increased workload for healthcare systems.

Currently, the most widely evaluated tool is the Goff index. It was developed using questionnaire data from case patients with primary ovarian cancer and control subjects who were undergoing ultrasound or who were at high risk for ovarian cancer and were enrolled in a screening study. When used in women 50 years and older, the Goff index had a sensitivity of 66.7% (in fewer than 75 women with cancer) and a specificity of 90.0% (in fewer than 245 control subjects) for symptoms occurring in the year before diagnosis in a confirmatory group of case patients (Cancer. 2007;109:221-227).

Other groups have used the Goff index in various settings and obtained comparable estimates of sensitivity and specificity, the authors note.

The hope, they write, is that a symptom index can be used as an adjunct to current strategies for women at low to moderate risk for ovarian cancer. But for such an approach to effective, "symptoms must be present sufficiently before diagnosis to allow for time to screen, evaluate, and intervene," the authors report.

The method of symptom assessment, however, has been known to affect the nature and threshold of symptoms reported, and thus affect the performance of the index. The methods of symptom assessment include self-completed questionnaires vs patient interviews and symptom checklists vs open-ended questions.

Study Details

In this study, Dr. Sasieni and colleagues sought to determine the effectiveness of the symptom data and estimate the potential for expediting a diagnosis using this tool. Their study objectives included developing and evaluating new symptom indexes, evaluating the Goff index, and assessing the impact of the data source on index performance. They also wanted to quantify the loss of sensitivity if symptoms that developed within a few months of diagnosis were excluded.

The cohort consisted of 194 women 50 to 79 years of age who had recently been diagnosed with ovarian cancer and 268 control subjects who underwent ovarian cancer screenings. The symptoms and the date of onset were assessed from 3 sources: questionnaires, telephone interviews, and general practitioner notes. Data from questionnaires and practitioner notes were used to develop 2 new symptom indexes. The sensitivity and specificity of the new indexes and the Goff index were calculated for 2 periods: 0 to 11 months and 3 to 14 months prior to diagnosis.

Benefits Might Be Overemphasized

The authors found that the results for the 2 new indexes were similar to those observed in previous reports on the Goff index, both qualitatively (symptoms included) and quantitatively (sensitivity and specificity). They also noted that the sensitivity of the symptoms was more pronounced in late-stage disease than in early-stage disease.

When symptoms that began in the 3 months before a diagnosis were excluded, sensitivity decreased for all indexes and all data sources.

The specificity of the indexes was lower for telephone interviews than for questionnaires and practitioner notes. For example, during the 3 to 14 months before diagnosis, the specificity for telephone interviews vs questionnaires with index 1 was 19.2% vs 10.4% (difference, 8.8%), with index 2 was 14.4% vs 6.7% (difference, 7.7%), and with the Goff index was 7.2% vs 1.5% (difference, 5.7%).

The authors conclude that their 2 new indexes were qualitatively and quantitatively similar to the Goff index, and the "small differences between the 3 indexes indicate that there is little to gain from deriving new symptom indexes."

They suggest that even though a symptom index could advance the diagnosis of ovarian cancer, the benefits of these indexes might be overemphasized, given that most symptoms of the disease emerge in the 3 months before diagnosis. "For a more specific index, the sensitivity would be approximately one third," they write.

In Which Direction?

In their editorial, Drs. Hartge and Speyer point out that it has been difficult to develop a sensitive and specific screening method to detect ovarian cancer early. The reasons have to do with the biology of ovarian cancer, the arithmetic of screening, and the fact that "the clinical characteristics of the disease and its treatment collude" to make it difficult to detect the disease at an early enough stage to matter.

They ask whether researchers should be putting more effort into understanding the etiology of the disease, finding methods of prevention, and developing improved therapies for women with biologically aggressive disease.

"These other research directions remain critical as the search continues for better ways to find ovarian cancer early," the editorialists conclude.

The study was funded by Cancer Research UK and the Eve Appeal/Oak Foundation. Coauthor Ian Jacobs, from the University of Manchester, United Kingdom, reports being a consultant for Becton, Dickinson and Company and a nonexecutive director and board member of Abcodia Ltd. Coauthor Usha Menon, MD, from the University College London, United Kingdom, reports a financial interest in Abcodia Ltd.

J Natl Cancer Inst. 2012;104:82-83, 114-124. Editorial, Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: