OBJECTIVES: To assess diagnostic strategies for distinguishing benign from malignant adnexal masses. DATA SOURCES: MEDLINE(R) and reference lists of recent reviews; discharge data from the Nationwide Inpatient Sample. REVIEW METHODS: The major diagnostic methods evaluated were bimanual pelvic examination, ultrasound (morphology and Doppler velocimetry), MRI, CT, FDG-PET, CA-125, and scoring systems that incorporated multiple clinical, laboratory, and radiologic findings. Meta-analysis using a random-effects model was used to estimate pooled sensitivity and specificity for discriminating benign from malignant. We reviewed evidence for followup strategies for masses considered benign, and for adverse outcomes of diagnostic surgery. We also reviewed published models of the natural history of ovarian cancer and compared the impact of assumptions about natural history on outcomes. RESULTS: The majority of studies did not describe whether patients presented with asymptomatic masses detected through screening or with symptoms. Prevalence of malignant masses in a U.S. postmenopausal screening population was approximately 0.1 percent, while benign masses were found in 0.8 to 1.8 percent of women. Pooled (a) sensitivity and (b) specificity were: bimanual exam (a) 0.45, (b) 0.90; ultrasound morphology scores (a) 0.86 to 0.91, (b) 0.68 to 0.83; Doppler resistive index (a) 0.72, (b) 0.90; pulsatility index (a) 0.80, (b) 0.73; maximum systolic velocity (a) 0.74, (b) 0.81; presence of vessels (a) 0.88, (b) 0.78; combined morphology and Doppler (a) 0.86, (b) 0.91; MRI (a) 0.91, (b) 0.88; CT (a) 0.90, (b) 0.75; FDG-PET (a) 0.67, (b) 0.79; and CA-125 (a) 0.78, (b) 0.78. Both sensitivity and specificity of CA-125 were better in postmenopausal than in premenopausal women. In modeled outcomes, combinations of imaging and CA-125 were both more sensitive and more specific than either alone. Performance of scoring systems in validation studies was consistently worse than in development studies; the highest demonstrated specificity observed was 0.91, with a concurrent sensitivity of 0.74. Evidence on followup strategies was sparse, although one large study provided good evidence for safely following unilocular cysts less than 10 cm in diameter. Overall complication rates in studies of surgically managed adnexal masses were low, but important clinical information was not reported. CONCLUSIONS: All diagnostic modalities showed trade-offs between sensitivity and specificity, but the available literature does not provide sufficient detail on relevant characteristics of study populations to allow confident estimation of the results of alternative diagnostic strategies. Although modeling studies may prove useful in evaluating diagnostic algorithms, further work is needed to explore the implications of uncertainty about the natural history of ovarian cancer.
|Original language||English (US)|
|Number of pages||145|
|Journal||Evidence report/technology assessment|
|State||Published - Feb 1 2006|