A high-resolution ultrasound device has shown great accuracy in detecting suspected meningitis in newborns and infants, potentially offering a non-invasive alternative to lumbar puncture, the traditional diagnostic method. This is the main conclusion of an international study led by the Barcelona Institute for Global Health (ISGlobal), a centre supported by the ”la Caixa” Foundation, in collaboration with hospitals in Spain, Mozambique, and Morocco. The results have been published in the journal Pediatric Research.
Meningitis is an inflammation of the membranes surrounding the brain and spinal cord. When caused by bacteria or fungi, it can be fatal if not diagnosed and treated early. Even in cases where the disease is overcome, it can leave serious after-effects, such as neurological damage or cognitive disorders. Despite medical advances in recent decades, meningitis remains a major threat to child health, especially in low- and middle-income countries, where limited access to early diagnosis exacerbates its impact.
Current diagnosis: invasive and impractical
At present, diagnosing meningitis requires a lumbar puncture to collect cerebrospinal fluid, which is then analysed in the lab for signs of inflammation, such as elevated white blood cell counts.
This is an invasive technique, with associated risks and significant practical limitations. In high-income countries, it is performed routinely even in cases of low suspicion, resulting in a large number of “normal” procedures and therefore a low diagnostic yield. Conversely, in low-income countries, the lack of resources means the test is rarely performed, leading to significant underdiagnosis or, in many cases, empirical and often inaccurate diagnoses.
The alternative: an ultrasound device
The aim of the study was to validate the NEOSONICS device, which uses high-frequency ultrasound applied through the baby’s open fontanelle —a membranous gap between the bones of the skull, which has not yet closed— to visualise and analyse cerebrospinal fluid. A deep learning algorithm interprets the images, identifies and counts the cells, and determines whether there are inflammatory signs consistent with meningitis.
The study, funded by the Bill and Melinda Gates Foundation, was conducted between 2020 and 2023 and included over 200 newborns aged up to 24 months from the Spanish hospitals Sant Joan de Déu, La Paz and Quironsalud, as well as Maputo Central Hospital (Mozambique) and the Hôpital d’Enfants de Rabat-Ibn Sina (Morocco). “The device was able to correctly classify 17 out of 18 meningitis cases and 55 out of 58 controls without meningitis,” explains Sara Ajanovic, researcher at ISGlobal and lead author of the study. “Specifically, it detected high white blood cell levels in cerebrospinal fluid with approximately 94% sensitivity and 95% specificity.”
The new device — cost-effective, portable and easy to use — could not only reduce the number of lumbar punctures, but also be used in clinically unstable patients where lumbar puncture is contraindicated. “Introducing a non-invasive tool could reduce unnecessary antibiotic use, prevent complications associated with lumbar puncture, and improve both early diagnosis and non-invasive monitoring of treatment response,” explains Quique Bassat, ISGlobal’s director general, ICREA researcher and senior author of the study.
Artificial Intelligence and the future of diagnostics
The validation of NEOSONICS marks a first step towards its future incorporation into clinical practice. At the same time, other studies coordinated by ISGlobal are exploring the potential of integrating Artificial Intelligence (AI) with ultrasound technology to enhance interpretation of results. Using advanced algorithms, AI enables image metrics to be analysed and texture patterns associated with inflammatory cells in cerebrospinal fluid to be detected, thus improving the system’s diagnostic capacity in cases of infant meningitis.
References
Ajanovic S, Jobst B, Jiménez J, Quesada R, Santos F, Carandell F, Lopez-Azorín M, Valverde E, Ybarra M, Bravo MC, Petrone P, Sial H, Muñoz D, Agut T, Salas B, Carreras N, Alarcón A, Iriondo M, Luaces C, Sidat M, Zandamela M, Rodrigues P, Graça D, Ngovene S, Bramugy J, Cossa A, Mucasse C, Buck WC, Arias S, El Abbass C, Tligi H, Barkat A, Ibáñez A, Parrilla M, Elvira L, Calvo C, Pellicer A, Cabañas F, Bassat Q; UNITED study group. Non-Invasive Meningitis Screening in Neonates and Infants: Multicentre International Study. Pediatr Res. 2025 Jul 23. https://doi.org/10.1038/s41390-025-04179-7
Sial, H., Carandell, F., Ajanovic, S., Jiménez, J., Quesada, R., Santos, F., Buck, W. C., Sidat, M., UNITED Study Consortium, Bassat, Q., Jobst, B., & Petrone, P. (2024). Novel AI-driven infant meningitis screening from high resolution ultrasound imaging. In medRxiv. https://doi.org/10.1101/2024.08.29.24312709
Journal
Pediatric Research
Method of Research
Observational study
Subject of Research
People
COI Statement
B.J., J.J., R.Q., F.S., and F.C. are employed by Kriba, the company developing the noninvasive device. S.A. received honoraries from Kriba for the scientific coordination of the multi-site study at ISGlobal. The Biomedical Data Science team at ISGlobal received training fees and consulting fees from Kriba to assist in the development of deep learning methods.