Outcomes from ovarian cancer screening in the PLCO trial: Histologic heterogeneity impacts detection, overdiagnosis and survival

Abstract

Aim: A mortality benefit from screening for ovarian cancer has never been demonstrated. The aim of this study was to evaluate the screening outcomes for different histologic subtypes of ovarian cancers.

Methods: Women in the screening arm of the Prostate, Lung, Colorectal and Ovarian Screening Trial underwent CA-125 and transvaginal ultrasound annually for 3-5 years. We compared screening test characteristics (including overdiagnosis) and outcomes by tumour type (type II versus other) and study arm (screening versus usual care).

Results: Of 78,215 women randomised, 496 women were diagnosed with ovarian cancer. Of the tumours that were characterised (n = 413; 83%), 74% (n = 305) were type II versus 26% other (n = 108). Among screened patients, 70% of tumours were type II compared to 78% in usual care (p = 0.09). Within the screening arm, 29% of type II tumours were screen detected compared to 54% of the others (p < 0.01). The sensitivity of screening was 65% for type II tumours versus 86% for other types (p = 0.02). 15% of type II screen-detected tumours were stage I/II, compared to 81% of other tumours (p < 0.01). The overdiagnosis rate was lower for type II compared to other tumours (28.2% versus 72.2%; p < 0.01). Ovarian cancer-specific survival was worse for type II tumours compared to others (p < 0.01). Survival was similar for type II (p = 0.74) or other types (p = 0.32) regardless of study arm.

Conclusions: Test characteristics of screening for ovarian cancer differed for type II tumours compared to other ovarian tumours. Type II tumours were less likely to be screen diagnosed, early stage at diagnosis or overdiagnosed.

Keywords: CA-125; Cancer screening; Ovarian cancer; Overdiagnosis.