Two large prospective randomised trials have evaluated the impact of ovarian cancer screening in a low-risk general population on mortality. The PLCO trial (78126 women) used an absolute Ca125 value and TVUS as the screening intervention. The UKCTOCS study randomised 202638 women (1:1:2) to multimodal (using the risk-of-ovarian-cancer-algorithm (ROCA)) or TVUS based screening and controls. The use of a longitudinal-Ca125 ROCA analysis, strict call recall system as well as tight protocol-driven centralised co-ordination and management were unique to UKCTOCS. The PLCO study found no mortality benefit and a high 15% serious complication rate from surgical evaluation. The UKCTOCS study reported a 84%-85.9% sensitivity, 99.8% specificity and an acceptable PPV of 2.7 surgeries/cancer diagnosis. The complication rate was 3%. The authors reported a stage shift of 14% (40% vs 26%) ≤Stage-3a disease which is extremely noteworthy. The primary Cox regression analysis indicated a 15% non-significant mortality benefit with multimodal and 11% with US screening. However, interestingly the post-hoc analysis found a potential statistically significant delayed mortality benefit of 23% at >7 years post-randomisation. Unfortunately, the pre-specified UKCTOCS analysis plan did not account for any statistical adjustment for a potential delayed effect on mortality. Further follow-up of the cohort is being undertaken to confirm this. This is essential to confirm whether a stage-shift with lower volume disease and higher R’0’ cyto-reduction rates can translate into a mortality benefit.
Unlike high risk women (e.g. Lynch Syndrome), screening for endometrial cancer is not currently recommended in the low risk population as most women are symptomatic and present with early stage disease. Given increasing obesity and rising population incidence, this becoming an area of increasing concern. A nested case-control analysis within the UKCTOCS TVUS cohort of 37078 women shows that. The optimum cut-off for endometrial thickness (ET)=5·15 mm, with a RR=25·2(CI 16·5,38·5). An ET=10mm, gives a sensitivity=54.1%(CI: 45.3,62.8) and leads to 17 diagnostic interventions/case diagnosed. An ET=5mm increases sensitivity(80.5%; CI:72.7,86.8) but requires 56 procedures/case diagnosed. Screening only the top quartile of the population (responsible for 40% cancers) yields a sensitivity=84.3%(CI:71.4,93.0) and specificity=89.9%(CI:89.3,90.5) for endometrial cancer using a cut off 6.75mm. Better identification of a higher risk category/sub-population may improve screening performance in asymptomatic women.