Novel Applications and Refinements of Ultrasound Techniques in Perinatal and Infant Death Investigation

The decline in parental consent for perinatal autopsies has led to the development of less invasive autopsy techniques, primarily using imaging and in some cases acquiring tissue samples through laparoscopic techniques via small incisions. Whilst post-mortem MRI (PMMR) has been extensively tested, and shown to be a suitable modality for determining the cause of death/demise in the perinatal cohort, it is not widely available. This thesis explores the utility of a poorly explored, yet more accessible, cheaper, dynamic imaging modality widely used in ‘live’ paediatric imaging for the purposes of a less invasive autopsy – the perinatal post-mortem ultrasound (PMUS). I present a systematic review of the limited literature of PMUS diagnostic accuracy for perinatal death investigation encompassing 4 publications with 455 cases in total. This reveals an overall pooled whole body sensitivity rate of 73.3% [95% CI 59.9, 83.5] and specificity rate of 96.6% [95% CI 92.6, 98.4] . A comprehensive imaging protocol for whole body PMUS is also presented based on my own experience in scanning 272 perinatal deaths. An analysis in 130 of these cases with autopsy as a reference standard, showed that brain and abdominal diagnoses yielded the highest sensitivity rates (90.9% and 92.3% respectively), with spinal, cardiac and thoracic diagnoses yielding the lowest sensitivity rates (50%, 50% and 57.1% respectively). Imaging of the brain and heart however were the least likely to be of diagnostic quality (76.8% and 78.3% diagnostic cases respectively), particularly in macerated fetuses. In a subset of cases where PMUS and 1.5T PMMR were performed, there was no significant difference in whole body diagnostic accuracy rates (concordance rates for PMUS versus PMMR of 86.4% [95%CI 77.7, 92.0] versus 88.6% [95% CI 80.3, 93.7]), although PMMR yielded fewer non-diagnostic brain and cardiac examinations (2.9% and 2.9% non-diagnostic brain and cardiac PMMR cases versus 22.8% and 14.7% non-diagnostic PMUS cases). In the second part of my thesis, I describe the development of an ‘incisionless’ ultrasound guided biopsy method using a single entry site for the biopsy needle – the umbilical vein. This ‘INTACT’ biopsy method allowed for a ‘non-invasive’ autopsy with tissue sampling, with a biopsy success rate of 76.1% overall for all organs, with highest individual organ success rates >90% for heart and lungs. I conclude by discussing how best to incorporate PMUS into clinical practice and suggest areas for future researc