Simon Mpooya

Background: Schistosomiasis-related periportal fibrosis (PPF) diagnoses rely on point-of-care ultrasound imaging. Yet, the extent of agreement between sonographers, limitations of existing World Health Organization (WHO) protocols, and the applicability of ultrasound video remain poorly understood. Methods: In 2024, six sonographers performed ultrasound scans for 3302 participants aged 5–92 years within SchistoTrack, a community-based cohort in rural Uganda. We adapted and standardised the WHO Niamey protocol for the collection of curated ultrasound images of liver fibrosis patterns related to Schistosoma mansoni (ranging A–F) and liver ultrasound video sweeps. PPF was defined as patterns C–F. Images and videos were read for patterns of fibrosis by second sonographers who were blinded to field diagnoses. Second sonographers recorded their clinical confidence, the probe position and any ultrasound interference. Outcomes considered were simple agreement for individual patterns and agreement on whether any PPF pattern was present, using Cohen’s κ . The influence of participant characteristics and image quality factors was evaluated against agreement outcomes. Results: Abnormal liver fibrosis patterns (B–F) were identified in 32.8% (1082/3302) of participants, and among these 2074 images were read by a second sonographer from 1040 participants. The percentage agreement was 73.0% (1515/2074) for broad pattern categories. Severe fibrosis patterns had the least agreement; E and F patterns had κ = 0.430 and 0.149, respectively. The D pattern had the most agreement ( κ = 0.717). 94.0% of second sonographer readings were within one step of severity of the field diagnosis. Participant characteristics were not consistently associated with disagreement, though agreement clustered by the field sonographer. Clinical confidence, breathing/movement, fat interference and probe position were positively associated with disagreement. There was no agreement among the 1054 videos that were read by a second sonographer: κ = 0.001 for PPF agreement and ranged 0–0.062 for individual patterns. Conclusion: We have presented the largest diagnostic study of inter-sonographer agreement for schistosomal liver fibrosis patterns. Disagreement between adjacent patterns of liver fibrosis may indicate that the WHO Niamey protocol should be updated to capture a continuous progression of liver fibrosis rather than distinct patterns. The greatest disagreement was among severe patterns, which supports routine second reading to ensure life-threatening cases are prioritised and appropriately triaged.

Schistosomiasis causes a complex, difficult to diagnose form of liver fibrosis with high rates of life-threatening morbidity in resource-poor settings where there are often no trained sonographers. Protocols for diagnosis of schistosomiasis-related liver fibrosis have focused on difficult-to-acquire and subjective ultrasound images dependent on extensive expertise. Here we present SchistoTrackVideoNet, the first deep learning-based video model trained on easy-to-acquire standardised ultrasound video sweeps for classification of schistosomiasis-related liver fibrosis. This video-based classification model was trained and evaluated on video sweeps from 2140 participants aged 5--87 years from three districts in rural Uganda. We tested the model at a clinically-relevant sensitivity threshold ($\geq$90\%) and achieved positive predictive values of 0.0968--0.5556 for diverse presentations of liver fibrosis. Our findings show potential for the use of easy-to-acquire video sweeps for diagnosis of schistosomiasis-related liver fibrosis and our model provides a proof-of-concept for deep learning applied to liver ultrasound video for diagnosis of schistosomiasis-related liver morbidity.

Liver fibrosis is a major cause of death in low- and middle-income country contexts. In rural, poor areas of sub-Saharan Africa, schistosomiasis is an underestimated cause of liver fibrosis. Despite the need for increased diagnostic capacity for schistosomiasis-related liver fibrosis, there are no automated, clinically-validated tools to diagnose schistosomiasis-related liver fibrosis. We present SchistoTrackNet which is, to our knowledge, the first deep learning-based model for distinguishing distinct presentations of schistosomiasis-related liver fibrosis of varying severity. Ultrasound images from 1533 participants aged 5–84 years from three districts in rural Uganda were used to train and evaluate the presented models. The models were evaluated by assessing failure cases and by comparing results with re-readings performed by sonographers experienced in diagnosis of schistosomiasis morbidity. Our models show potential to enable automated reading of ultrasound images for schistosomiasis-related liver fibrosis to allow large-scale surveillance of schistosomiasis morbidity and contribute towards the World Health Organization target to eliminate schistosomiasis as a public health problem.

Background: Schistosoma mansoni is a leading cause of hepatosplenic disease in sub-Saharan Africa. Yet, associations with current Schistosoma mansoni infections and hepatosplenic organometry remain unclear in the context of widespread mass drug administration and co-endemic infections. Methods: From January to February 2024, we conducted a community-based, cross-sectional study nested within the SchistoTrack cohort in three districts of Uganda. Liver and spleen dimensions were assessed via point-of-care B-mode ultrasound for 3121 individuals. Organ dimensions were classified using the standard deviations from height-standardized internal reference values derived from an infection-free population. Multinomial logistic regressions were run for children (5–17 years) and adults (18+ years) separately. Population attributable fractions (PAFs) were used to estimate the proportion of abnormal organometry statistically attributable to each infection. Key exposures were S. mansoni , malaria, human immunodeficiency virus (HIV), and hepatitis B virus (HBV) alongside a comprehensive set of social, biomedical, and other covariates controlled for. Results: Moderate-to-severe splenic enlargement was observed in 29.1% (438/1507) of children and 23.3% (376/1614) of adults. Among adults, 20.9% (337/1614) had left liver lobe enlargement and 18.8% (303/1614) had right liver lobe shrinkage. In children, severe splenic enlargement was statistically attributable to malaria (PAF 46.7%; Relative Risk Ratio (RRR) 3.96, 95% CI 2.64–5.92) and S. mansoni infection intensity (PAF 23.6%; RRR 1.12, 95% CI 1.04–1.20). In adults, S. mansoni intensity was associated with moderate left liver lobe enlargement (PAF 12.4%; RRR 1.11, 95% CI 1.04–1.18). In adults, HIV was associated with severe left liver lobe shrinkage (RRR 4.50, 95% CI 1.19–17.00) and severe splenomegaly (RRR 3.62, 95% CI 1.58–8.33), while HBV was associated only with severe left liver lobe shrinkage (RRR 2.54, 95% CI 1.07–6.03). Praziquantel treatment in the past year showed inconsistent associations and no clear protective pattern. Conclusion: Current S. mansoni infection intensity remains associated with splenomegaly in children despite controlling for concurrent malaria positivity, and with hepatomegaly in adults despite HIV and HBV associations.

Background: The global burden of multimorbidity is increasing yet poorly understood, owing to insufficient methods available for modelling complex systems of conditions. In particular, hepatosplenic multimorbidity has been inadequately investigated. Methods: From 17 January to 16 February 2023, we examined 3186 individuals aged 5-92 years from 52 villages across Uganda within the SchistoTrack Cohort. Point-of-care B-mode ultrasound was used to assess 45 hepatosplenic conditions. Three graph learning methods for representing hepatosplenic multimorbidity were compared including graphical lasso (GL), signed distance correlations (SDC), and co-occurrence. Graph kernels were used to identify thresholds of relevant condition inter-dependencies (edges). Graph neural networks were applied to validate the quality of the graphs by assessing their predictive performance. Clinical utility was assessed through medical expert review. Findings: Multimorbidity was observed in 54·65% (1741/3186) of study participants, who exhibited two or more hepatosplenic conditions. Conditions of mildly fibrosed vessels were most frequently observed (>14% of individuals). Percentage thresholds were found to be 50·16% and 64·46% for GL and SDC, respectively, but could not be inferred for co-occurrence. Thresholded GL and SDC graphs had densities of 0·11 and 0·17, respectively. Both thresholded graphs were similar in predictive utility, although GL produced marginally higher AUCs under certain experiments. Both GL and SDC had significantly higher AUCs than co-occurrence. Numerous conditions were predicted with perfect sensitivity using both GL and SDC with graph convolutional network with five input conditions. Interpretation: The most common method for multimorbidity (co-occurrence) provided an uninformative representation of hepatosplenic conditions with respect to sparsity and predictive performance. More clinically useful graphs were computed when algorithms consisted of statistical assumptions, such as graphical lasso. Future work could apply the pipeline developed here for clinically relevant multimorbidity representations. Funding: NDPH Pump Priming Fund, John Fell Fund, Robertson Foundation, UKRI EPSRC (EP/X021793/1).

Liver diseases are a leading cause of death worldwide, with an estimated 2 million deaths each year. Causes of liver disease are difficult to ascertain, especially in sub-Saharan Africa where there is a high prevalence of infectious diseases such as hepatitis B and schistosomiasis, along with alcohol use. Point-of-care ultrasound often is used in low-resource settings for diagnosis of liver disease due to its portability and low cost. For classification models that can automatically stage liver disease from ultrasound video, the region of interest is liver tissue. A fully-automated pipeline for liver tissue identification in ultrasound video is presented. Ultrasound video data was collected using a low-cost, portable ultrasound machine in rural areas of Uganda. The pipeline first detects the diaphragm in each ultrasound video frame, then segments the diaphragm to ultimately use this segmentation to infer the position of liver tissue in each frame. This pipeline outperforms directly segmenting liver tissue with an intersection over union of 0.83 compared to 0.62. This pipeline also shows improved results with respect to the ease of clinical interpretation and anticipated clinical utility.

Background Vaccination is one of the most successful public health interventions for preventing infectious diseases. A successful vaccination program depends on high coverage, and health care workers (HCWs) play a pivotal role in ensuring high uptake of vaccines in the population. COVID-19 vaccines have been proven to be efficacious, and vaccination campaigns have been ongoing, however there is a perceived high vaccine hesitancy even among health care workers in Uganda. This study aimed at describing the facilitators, barriers to and level of uptake of COVID 19 vaccines among healthcare workers in primary healthcare facilities in an urban setting in Uganda. Methods We conducted an online cross-sectional survey among healthcare workers in private and public healthcare facilities in Entebbe municipality between July 2021 and August 2021. Data was collected using an online questionnaire. Uptake of the vaccines among healthcare workers was analysed as proportions, and logistic regression was used to analyse barriers and facilitators to uptake of COVID 19 vaccines. Results The study enrolled 360 participants, with 61.7% (n=222) females. A total of 236 (65.6%) healthcare workers had received at least one dose of COVID 19 vaccine with higher uptake among females 64%(n=151). Age above 40 years (OR 4.29), participating in COVID 19 vaccine related activities (OR 4.18) and having had a negative SARS-COV-2 test result (OR 1.79) increased the odds of having been vaccinated. Working in either a private for profit (OR 0.23) or a private not for profit (OR 0.19) reduced the odds of having been vaccinated. History of having cared for a COVID 19 patient and having a positive SARS-COV-2 test result did not influence the uptake of the vaccines in the study population. Conclusion Vaccine uptake among healthcare workers was close to the World Health Organisation (WHO) recommended uptake of 70% by mid-2022.

Background: Intestinal schistosome infections are known to cause periportal fibrosis (PPF). Yet, the epidemiology of PPF remains poorly understood, especially in settings endemic with Schistosoma mansoni . Methods: We randomly sampled 1442 households from 38 villages in Mayuge, Buliisa, and Pakwach Districts of Uganda within the SchistoTrack Cohort to examine 2834 individuals aged 5-90 years. PPF was diagnosed using ultrasound and image patterns C-F from the Niamey Protocol. S. mansoni infection status/intensity was diagnosed by Kato-Katz microscopy and point-of-care circulating cathodic antigens (POC-CCA). Schistosome infection, coinfections, and comorbidities were examined as exposures for PPF. Logistic regressions were run with standard errors clustered by household. Findings: PPF prevalence was 12·10% (343/2834), varying from 5·00-19·46% across districts. S. mansoni prevalence by Kato-Katz, and POC-CCA trace negative and positive was 43·37% (1229/2834), 40·86% (1158/2834), and 65·73% (1863/2834) respectively. Individual schistosome infection status/intensity was not correlated with the likelihood of PPF. Living in a village where adults had 2-fold increased PPF likelihood. Interpretation: Current individual schistosome infections alone are uninformative for PPF. History of HIV+ and underlying chronic hepatitis/early cirrhosis-like disease were risk factors and could be investigated for PPF surveillance and management. Funding: Nuffield Department of Population Health Pump Priming Fund, Wellcome Trust Institutional Strategic Support Fund (204826/Z/16/Z), John Fell Fund, Robertson Foundation Fellowship, and UKRI EPSRC Award (EP/X021793/1). Research in context: Evidence before this study: Morbidity due to parasitic infection is a complex interplay of current and past exposures. World Health Organization (WHO) guidelines for elimination of schistosomiasis as a public health problem assume current infection is a reliable proxy indicator of prevalent morbidity. Community infection thresholds are defined in guidelines and, when met, the assumption is there is no schistosomiasis-related morbidity. There is a lack of evidence for the association of infection with prevalent morbidity in the context of repeated treatment from routine mass drug administration. To evaluate WHO guidelines, there is a need for large-scale population- based, cross-sectional studies in endemic areas where current infection is compared with current morbidity at the same timepoint. A cross-sectional design also enables the investigation of a wide range of risk factors to assess the relative importance of current schistosome infection. Periportal fibrosis is a schistosomiasis-associated severe morbidity with clinical consequences such as portal hypertension, upper gastrointestinal tract bleeding, and ultimately premature death. Yet, little is known about the distribution of this disease; no information is available on its most basic epidemiology including age and gender-specific likelihoods. It is schistosomiasis-specific or attributable to schistosome pathology unlike more subtle conditions with complex aetiologies (e.g. anaemia). Hence, investigating periportal fibrosis serves as a first line, conservative approach to evaluating World Health Organization guidelines for elimination of schistosomiasis-related morbidity as a public health problem. A systematic literature search as part of an ongoing metanalysis was prospectively registered on PROSPERO (CRD42022333919). Databases were searched on 18 ^th May 2022 and included the Cochrane Central Register of Controlled Trials, Embase, Global Health, Global Index Medicus, and Medline. The following general terms were used: “Schistosoma” AND “fibrosis” AND “intensity” AND “infection” AND “periportal OR liver”. Studies were included that considered Schistosoma mansoni , S. japonicum, or S. mekongi species. Only original research articles in English were considered. No restriction on date of publication, age, gender, or region was applied. Infection was required to be diagnosed as opposed to self-reported. Periportal fibrosis was defined by the study authors. Added value of this study: No population-wide or adjusted analyses were found to characterize the age-specific likelihood of periportal fibrosis. Studies focused on unadjusted associations in nonrandomly sampled populations of narrow age groups. In a single case (Weigand et al 2021) where adjusted analyses were completed (with age and gender considered), national programmatic data was used with sparsely nonrandomly sampled schoolchildren and limited frequency of the outcome of periportal fibrosis. Ultrasound data collection protocols and validation across studies were poorly reported. There was a lack of investigations on coinfections and comorbidities with only eight studies initiated (or published) from 2003 onwards after the start of mass drug administration in sub-Saharan Africa. Implications of all available evidence: To our knowledge, this study is the first to characterize the epidemiology of periportal fibrosis with respect to the most common intestinal schistosome pathogen ( S. mansoni ). We conducted a comprehensive, population-based study of all ages (5+ years) eligible for mass drug administration in an area that has received at least 13 rounds of treatment. Here we provide clear evidence for the lack of association of current S. mansoni infection status and intensity with periportal fibrosis irrespective of the diagnostic for schistosome infection. No support was found for current WHO elimination guidelines despite using arguably the most biologically specific and severe morbidity associated with schistosomiasis. We also characterized the age-specific likelihood of periportal fibrosis, identifying a transitional age as young as 25 years. We identified future avenues for research into coinfections such as HIV and hepatitis B that appear to influence periportal fibrosis status even after controlling for a wide range of biosocial determinants of schistosome infection, treatment, and unrelated liver fibrosis. Future work is needed to understand if/how coinfections alter the pathogenesis of periportal fibrosis. Importantly, World Health Organization guidelines should be differentiated for schistosomiasis morbidities to discourage the use of infection status/intensity/prevalence as a proxy indicator for monitoring the elimination of periportal fibrosis as a public health problem.