Victor Anguajibi

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: Schistosomiasis causes substantial chronic morbidity in sub-Saharan Africa, yet case definitions, clinical management guidance, and health worker training for schistosomiasis-related morbidity remain limited. Methodology: We conducted a qualitative needs assessment for schistosomiasis morbidity management. Workshops were held over one day in each of Pakwach, Buliisa, and Mayuge Districts in Uganda in October 2024. 105 government health workers participated including clinicians, nurses, laboratory technicians, sonographers, and district health managers from health facilities at different levels of care. The workshops comprised six structured sessions: presentations on schistosomiasis burden in Uganda and the SchistoTrack cohort, a clinical case report by an expert clinician, an interactive session on patient case studies from the SchistoTrack cohort, mapping of patient pathways, anonymous participation and feedback, and demonstrations of schistosomiasis diagnosis. Workshop discussions were documented through notes taken in English and analysed using qualitative thematic analysis as per Braun and Clarke. Findings: Health workers demonstrated substantial gaps in understanding schistosomiasis case definitions, particularly in distinguishing current infection from chronic morbidity and in grading disease severity. Patient pathways for schistosomiasis morbidity management were fragmented and inconsistent, with weak triage, unclear referral and feedback mechanisms, and limited follow-up across facility levels. Health facilities lacked essential capacity and resources, including routine access to praziquantel outside mass drug administration, diagnostic reagents, functional ultrasound equipment, trained sonographers, and standardized training and reference tools. Collectively, these gaps contributed to inconsistent clinical decision-making and under-recognition of severe schistosomiasis-related morbidity. Conclusions: Integrating case management into routine health services through standardized case definitions, clearer patient pathways, and targeted practical training for health workers is essential to complement preventive chemotherapy and reduce preventable morbidity. The engagement framework and patient case studies used here can support needs-based assessments in other endemic settings to inform the development of context-appropriate clinical guidance and training programmes. Author summary: Schistosomiasis is a disease common in sub-Saharan Africa that can cause serious, lifelong health problems, even after the infection itself has been treated. While mass drug administration programmes have reduced infections, we found that health facilities are often not equipped to recognize or manage the chronic illness that remains after mass drug administration. In this study, we worked with government health workers in three rural districts of Uganda to understand their experiences, challenges, and training needs related to schistosomiasis morbidity. Through interactive workshops, we learned that many health workers struggle to distinguish between current infection and long-term organ damage, especially liver disease. Patient care pathways were often unclear, referrals were poorly coordinated, and essential tools, such as diagnostic supplies, ultrasound access, and routine availability of treatment, were frequently missing. These gaps make it difficult to identify severe cases early and provide consistent care. We show that strengthening routine health services through clearer case definitions, better training, and more coordinated patient pathways is essential to reduce preventable suffering from schistosomiasis. Our approach offers a practical way to support health systems in other endemic settings facing similar challenges.

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.

One in 25 deaths worldwide is related to liver disease, and often with multiple hepatosplenic conditions. Yet, little is understood of the risk factors for hepatosplenic multimorbidity, especially in the context of chronic infections. We present a novel Bayesian multitask learning framework to jointly model 45 hepatosplenic conditions assessed using point-of-care B-mode ultrasound for 3155 individuals aged 5-91 years within the SchistoTrack cohort across rural Uganda where chronic intestinal schistosomiasis is endemic. We identified distinct and shared biomedical, socioeconomic, and spatial risk factors for individual conditions and hepatosplenic multimorbidity, and introduced methods for measuring condition dependencies as risk factors. Notably, for gastro-oesophageal varices, we discovered key risk factors of older age, lower hemoglobin concentration, and severe schistosomal liver fibrosis. Our findings provide a compendium of risk factors to inform surveillance, triage, and follow-up, while our model enables improved prediction of hepatosplenic multimorbidity, and if validated on other systems, general multimorbidity.