Search

Hydrocephalus is the leading indication for pediatric neurosurgical care worldwide. Identification of postinfectious hydrocephalus (PIH) verses non-postinfectious hydrocephalus (NPIH), as well as the pathogen involved in PIH is crucial for developing an appropriate treatment plan. Accurate identification requires clinical diagnosis by neuroscientists and microbiological analysis, which are time-consuming and expensive. In this study, we develop a domain enriched AI method for CT-based infection diagnosis in hydrocephalic imagery. State-of-the-art (SOTA) convolutional neural network (CNN) approaches form an attractive neural engineering solution for addressing this problem as pathogen-specific features need discovery. Yet black-box deep networks often need unrealistic abundant training data and are not easily interpreted. In this paper, a novel brain attention regularizer (BAR) is proposed, which encourages the CNN to put more focus inside brain regions in its feature extraction and decision making. Our approach is then extended to a hybrid 2D/3D network that mines inter-slice information. A new strategy of regularization is also designed for enabling collaboration between 2D and 3D branches. Our proposed method achieves SOTA results on a CURE Children's Hospital of Uganda dataset with an accuracy of 95.8% in hydrocephalus classification and 84% in pathogen classification. Statistical analysis is performed to demonstrate that our proposed methods obtain significant improvements over the existing SOTA alternatives. . Such attention regularized learning has particularly pronounced benefits in regimes where training data may be limited, thereby enhancing generalizability. To the best of our knowledge, our findings are unique among early efforts in interpretable AI-based models for classification of hydrocehpalus and underlying pathogen using CT scans.

Background The etiology of neonatal sepsis is often not identified. Molecular methods can identify pathogens that culture-based methods miss. Most cases of neonatal sepsis globally are treated empirically per WHO guidelines with intravenous ampicillin and gentamicin, which may not be the best regimen for all pathogens. Methods We prospectively enrolled 800 neonates presenting with signs of sepsis to two Ugandan hospitals. Blood and cerebrospinal fluid were subjected to 16S rRNA sequencing, which identified Paenibacillus thiaminolyticus in 33/800 (4%) neonates. We confirmed the presence of P. thiaminolyticus by quantitative polymerase chain reaction (PCR). We describe neonatal and birth characteristics, presenting signs, and 12-month developmental outcomes for neonates with paenibacillosis. We performed antibiotic susceptibility testing and genomic analyses on three clinical isolates successfully grown in the laboratory. Results Neonates presented at a median age of 3 (1, 7) days. Fever (86%), irritability (78%) and seizures (52%) were common presenting signs (Figure). Most neonates were born vaginally (73%) at a medical facility (79%). Twelve (36%) had an adverse outcome: 5 (15%) neonates died; 4 (14%) survivors developed postinfectious hydrocephalus and three (9%) additional survivors had neurodevelopmental impairment. All three isolates were resistant to vancomycin, two were resistant to penicillin and ampicillin and one was unlikely to be sensitive to ceftriaxone; all were susceptible to gentamicin and meropenem. The genomes of all three strains contained multiple beta-lactamase genes and a cluster of genes that encodes a type IV pilus. Clinical signs at presentation for neonates with good and poor outcomes followng paenibacillosis Conclusion Molecular methods such as 16S rRNA sequencing and PCR can be used to improve the identification of pathogens causing neonatal sepsis. Paenibacillosis is an important emerging cause of neonatal sepsis in Uganda and is likely an underrecognized cause of postinfectious hydrocephalus in the region and possibly elsewhere. Antibiotics commonly used for neonatal sepsis may be inadequate for the treatment of paenibacillosis. Additional studies to understand the pathophysiology and optimal treatment of this novel infection are urgently needed to prevent neonatal mortality and morbidity including postinfectious hydrocephalus. Disclosures Jessica E. Ericson, MD, MPH, Abbvie: Advisor/Consultant Abhaya Kulkarni, MD, MSc, PhD, Medtronic: Advisor/Consultant.

Paenibacillus thiaminolyticusmay be an underdiagnosed cause of neonatal sepsis. We prospectively enrolled a cohort of 800 neonates presenting with a clinical diagnosis of sepsis at two Ugandan hospitals. Quantitative polymerase chain reaction specific toP. thiaminolyticusand to thePaenibacillusgenus were performed on the blood and cerebrospinal fluid (CSF) of 631 neonates who had both specimen types available. Neonates with virus detected in either specimen type were considered to potentially have paenibacilliosis, (37/631, 6%). We described antenatal, perinatal, and neonatal characteristics, presenting signs, and 12-month developmental outcomes for neonates with paenibacillosis vs. clinical sepsis. Median age at presentation was 3 (interquartile range 1, 7) days. Fever (92%), irritability (84%) and seizures (51%) were common. Eleven (30%) had an adverse outcome: 5 (14%) neonates died during the first year of life; 5 of 32 (16%) survivors developed postinfectious hydrocephalus and one (3%) additional survivor had neurodevelopmental impairment without hydrocephalus. These results highlight the need to consider local pathogen prevalence and the possibility of unusual pathogens when determining antibiotic choice for neonatal sepsis.