Post-mortem Lodox® as a potential substitute for histologic confirmation of pneumonia in sudden infant death investigation

Cook, Tracy Leanne

Post-mortem Lodox® as a potential substitute for histologic confirmation of pneumonia in sudden infant death investigation

Thesis / Dissertation

2024

Publisher

University of Cape Town

Department

Department of Pathology

Faculty

Faculty of Health Sciences

Abstract

Background: Lodox®, low dose digital x-rays, may be used as a cost-effective screening tool in developing countries for the post-mortem diagnosis of pneumonia in the medicolegal investigation of sudden unexpected infant death, avoiding the need to perform an internal examination. However, interpretation of the Lodox® is not standardised, and this modality has not yet been validated or tested against the recognised best standard in autopsy practice, histologic examination. Objective: This study aimed to evaluate whether interpretation of a post-mortem Lodox® scan is sufficient to diagnose pneumonia in alleged sudden infant deaths, without the need for confirmatory histology. Methods: A retrospective cross-sectional study was performed. Lodox® scans and lung histology from 100 infant decedents less than 1 year of age, admitted to Salt River Mortuary between 2011-2021, were independently interpreted according to standardised guideline by a senior forensic pathology registrar (100 cases), radiologist and anatomical pathologist (25 cases each). The data were captured in a Microsoft Excel® spreadsheet and subjected to statistical analysis. Results: Fifty-seven male infants (57%) and forty-three female infants (43%) were included in the study. The majority of infants were 1 to 5 months of age (67%) with an average age of 3.25 months and the most frequent age encountered being 2 months old. More than half of the Lodox® scans and lung histology were categorised as minor/non-specific changes. Lodox® categorisation was mostly inconsistent with histology categorisation (weighted-k: 0,14; CI: - 0,02-0,29) with a negligible relationship between the two modalities (Pearson co-efficient: 0,14 (p:0,16) and Spearman co-efficient: 0,15 (p:0,13)). A strongly positive relationship between registrar and anatomical pathologist histology categorisations was observed (Pearson co- efficient: 0,4 (p:0,05) and Spearman co-efficient: 0,42 (p:0,04)). Fair inter-rater agreement was noted between the registrar and radiologist categorisations; however, these findings displayed broad confidence intervals. (Lodox® weighted-k: 0,27; CI: - 0,09-0,59; histology weighted-k: 0,29; CI: -0,03-0,49). Moderate correlation noted between registrar and radiologist Lodox® categorisations was statistically insignificant (Pearson co-efficient: 0,35 (p:0,09) and Spearman co-efficient: 0,31 (p:0,11)). The Lodox® displayed a low sensitivity of 32% (17%, 51%) with a low positive predictive value of 42% (22%, 63%). Conversely, a high specificity of 80% (68%, 88%; p:0,21) in addition to a high negative predictive value of 72% (61%, 82%; p:0,21) was observed. Conclusion: The potential utilisation of the Lodox® scan as a substitute for histologic diagnosis of pneumonia in sudden infant death would make a vast difference in forensic practice in lower income countries, where caseloads are high and access to sophisticated imaging is limited. The small sample size in this study (n:100) did not yield statistically significant information to conclude that a relationship exists between Lodox® and histology categorisation. Nonetheless, results suggest that the diagnosis of pneumonia on Lodox® may be unreliable, resulting in failure to identify unnatural causes of death that may not be apparent on history, Lodox® and external examination, as well as compromise public health interventions and cause of death data. Further studies are required to properly evaluate post mortem Lodox® as a potential substitute for internal examination and histologic diagnosis of pneumonia in sudden infant death investigation.