Summary Background Whether some benign ovarian cysts can develop into cancerous cysts is not known. If a large proportion of ovarian cancers arose in this way, it might be possible to remove the benign cysts in a screening programme before they became malignant. We used follow-up data from a cohort of 5479 self-referred women without symptoms, who participated in a ultrasonographic-screening trial for early ovarian cancer between June, 1981, and August, 1987. We assessed whether the removal of persistent ovarian cysts from these women was associated with a reduction in the expected number of deaths from ovarian cancer in the cohort as a whole. Methods The expected number of deaths from all causes, all cancers, and ovarian, breast, and colorectal cancers were calculated for the study cohort by the standard life-table method. The actual number of deaths and each cause were obtained and the proportional mortality ratio was calculated for each cause of death. Findings 5135 (95%) of the participants in the original trial were traced. During the screening, five of these women were found to have stage I epithelial ovarian cancer and 88 had benign epithelial ovarian tumours. The number of reported deaths from all causes (387 [50% of expected]), all cancers (221 [71%]), and ovarian cancer (22 [90%]) was lower than expected because of the “healthy-volunteer effect”. Proportional mortality ratios were 100% (by definition) for all cancers, 141% for breast cancer, 128% for ovarian cancer (95% CI 87·7–187·6, p=0·19), 84% for colorectal cancer, and 48% for lung cancer. Interpretation The removal of persistent ovarian cysts was not associated with a decrease in the proportion of expected deaths from ovarian cancer relative to other cancers during follow-up. For population-based screening of healthy women without a family history of ovarian cancer, a screening test is required that is specific and sensitive to early malignant disease, and inexpensive. Lancet 2000; 355: 1060–63 See Commentary page ??? Introduction The lifetime risk of epithelial ovarian cancer for women living in western industrial countries is about 1·4% (ie, one woman in 70 will develop the disease). The overall 5-year survival rate is about 39%, but this rate varies substantially according to histological features of the tumour, and how far the disease has progressed at the time of diagnosis and treatment. 1 There is, however, evidence that the overall survival rate might be about 90% if primary tumours could be detected and removed when the transformed cells are apparently confined within the ovaries—ie, at stage 1. 2 By contrast, the survival rate is only 23% for patients with stage III or IV disease. Consequently, attempts are being made to develop effective screening procedures for morphological, vascular, or immunological signs of early ovarian cancer in symptom-free women. One research strategy is to use ultrasonography at each screening; another is to measure the concentration of serum tumour antigens and scan only those women with a positive result; and a third is to scan only those women with evidence of an inherited, predisposing gene for ovarian cancer. Whichever method is chosen, pelvic ultrasonography is the principal technique for the final assessment of whether a patient who has been screened should be referred for surgical investigations, largely because ultrasonography is thought to be sensitive to the presence of early-stage disease. 3 A common aim of each programme is to develop screening procedures with a high detection rate and a low false-positive rate for malignant disease in situ. There is the distinct possibility, however, that some benign ovarian cysts may have the potential to become malignant disease at a later date. 4 If this disease course were proven, it would have a major effect on the viability of population-based screening programmes. Very little is known about the genetic, biochemical, and morphological changes associated with the transformation of epithelial ovarian cells and the components of metastasis. The primary cancers can be divided into two subtypes: those cancers that arise de novo from the surface epithelium or its inclusion cysts; and those that are present within an existing lesion—either non-neoplastic or neoplastic. 5 Examples of early de novo epithelial cancers have been described in terms of size, location, and metastatic potential. 6.7 Furthermore, a substantial increase has been found in the proportion of epithelial pseudo- stratification, epithelial invagination, inclusion cysts, papillomatosis, and stromal activity in ovaries from women with a history of familial or hereditary ovarian cancer. 8 Whether these findings represent premalignant change is not certain. 9 Circumstantial evidence for the existence of premalignant ovarian abnormalities has been collated. 4,10,11 In particular, there is a time delay of about 10–15 years between the peak incidence of benign ovarian tumours and the corresponding peak for malignant tumours. 4 Furthermore, the number of benign tumours detected by ARTICLES 1060 THE LANCET • Vol 355 • March 25, 2000 Benign ovarian cysts and ovarian cancer: a cohort study with implications for screening Timothy J B Crayford, Stuart Campbell, Thomas H Bourne, Helen J Rawson, William P Collins Department of Epidemiology and Public Health, Guy’s, King’s, and St Thomas’ School of Medicine, King’s College Hospital, London SE5 9PJ (T J B Crayford MFPHM, H Rawson MSc); Department of Obstetrics and Gynaecology, St George’s Hospital Medical School, London SW17 0RE (Prof S Campbell DSc, T H Bourne PhD); and Academic Department of Obstetrics and Gynaecology, Guy’s, King’s, and St Thomas’ School of Medicine, King’s College Hospital, London SE5 9PJ, UK (Prof W P Collins Dsc) Correspondence to: Dr Timothy J B Crayford (e-mail: tim@crayford.net)