Refining ultrasound screening in ovarian cancer

• Main Author: Sharma, A.
• Published: University College London (University of London) 2014
• Subjects: 616.99
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.626807

Ovarian cancer (OC) is the leading cause of gynaecological cancer mortality in developed countries. There is concern that ultrasound (first/second line test in all screening strategies) might be better at detecting Type I OC than the aggressive Type II OC. If ultrasound is to be part of population screening, there is a need for quality assurance (QA), given the subjective nature of transvaginal scan (TVS). In the United Kingdom Collaborative Trial of Ovarian Cancer Screening, 48230 women had initial scan. Prospective cohort studies of women with ultrasound detected inclusion cysts (IC) and abnormal adnexal morphology (unilocular, multilocular, unilocular solid, multilocular solid cysts and solid masses) were undertaken. On median follow up of 6.13 years, no increased OC risk was associated with IC (Relative Risk 2.32, 95%CI 0.86-6.28). The absolute risk (AR) of overall OC associated with abnormal adnexal morphology at three years was 1.08%, 0.73% for Type I OC and 0.34% for Type II OC. At one year from first scan, sensitivity for Type II OC was highest for multilocular solid cyst at 27.8% with a PPV of 1.5%. TVS visualisation of normal ovary was the chosen metric to assess impact of the QA programme. 43867 scans were analysed to assess the effect of non-subjective factors on ovarian visualisation. Previous hysterectomy, tubal ligation, increasing age, unilateral oophorectomy and rising BMI decreased visualisation whereas increasing age at menopause and infertility increased visualisation. These factors were included in a Generalised Estimating Equation model and comparison undertaken of observed versus adjusted visualisation rate (VR) between individual sonographers/trial centres showing that while VR should be adjusted for non-subjective factors, the magnitude of differences between observed and adjusted VR was small. An accreditation programme for scanning postmenopausal ovaries was developed. Sonographers were assessed on knowledge of ultrasound protocol, three monthly VR, practical scanning technique, central training day attendance and submitted scan images. A 48-item questionnaire study (150 women/centre) on scan experience (including pain/discomfort) was conducted to assess the acceptability/ability to deliver ultrasound screening. TVS was well accepted with only 3.5% women reporting moderate/severe pain. This thesis provides accurate risk estimates for overall OC, Type I and Type II OC associated with ultrasound detected adnexal morphology in an asymptomatic postmenopausal population. In addition, it describes the development of QA, training/accreditation of sonographers and delivery of large scale ultrasound screening.