A Blood Test for Tomorrow: AI Unlocks Early Detection of Paediatric Long COVID

Commentary by Dr. Donald Greig

This commentary builds on my newsletter from last week, which examined the prevalence of chronic fatigue syndrome and its connection to long COVID.

In an exciting leap forward for paediatric healthcare, a groundbreaking study has unveiled a revolutionary method for diagnosing Long COVID in children—one that hinges on the power of Artificial Intelligence (AI). Conducted by a team from Università Cattolica del Sacro Cuore in Rome, in collaboration with Fondazione Policlinico Universitario Agostino Gemelli IRCCS and Ospedale Paediatrico Bambino Gesù IRCCS, this research paves the way for accurate, early diagnosis through a simple blood test.

The Evolution of Long COVID Diagnosis

Imagine a world where diagnosing Long COVID in children could be as straightforward as a routine blood draw. This possibility is becoming a reality, thanks to a unique molecular signature identified in the blood plasma of paediatric patients. The study, published on January 24, 2025 in Paediatric Research, highlights how an AI-powered tool was able to analyze this data, achieving an impressive diagnosis accuracy rate of 93%.

Dr. Danilo Buonsenso, a leading paediatric researcher at the Catholic University’s Faculty of Medicine and Surgery, spearheaded this research alongside a distinguished team including Dr. Nicola Cotugno from Bambino Gesù IRCCS. Their collaboration signifies a monumental step forward in understanding and diagnosing this condition in younger populations.

Understanding the Impact of Long COVID in Children

Long COVID affects approximately 0.5% of paediatric patients exposed to SARS-CoV-2, manifesting as lingering symptoms that can endure for 8-12 weeks or longer. These symptoms—often debilitating—can profoundly impact a child's daily life. Current evidence suggests that children over the age of 10 are particularly susceptible, regardless of the severity of their initial viral infection.

While previous studies identified Long COVID markers in adults, similar findings in children were conspicuously absent until now.

Key Findings from the Study

The research involved analyzing blood samples from 112 young individuals aged 0-19 years. Among them, 34 had received a clinical diagnosis of Long COVID, while others included patients with active infections and healthy controls.

Using advanced proteomic analysis, the researchers discovered a distinct protein signature in the blood of those diagnosed with Long COVID. This signature included elevated levels of specific pro-inflammatory and pro-angiogenic chemokines, which AI models successfully utilized to differentiate Long COVID cases with remarkable accuracy—0.93 for diagnosis, 0.86 for specificity, and 0.97 for sensitivity.

Implications for Future Diagnosis and Treatment

The findings not only reveal a unique inflammatory profile associated with paediatric Long COVID but also hold promise for developing a straightforward diagnostic test based on blood samples. This could facilitate timely and comprehensive care for affected children.

Dr. Cotugno emphasizes that the immunological data from the study could lead to targeted therapeutic interventions, while Dr. Buonsenso calls for increased funding to explore effective treatment approaches for this immune-mediated condition.

A New Dawn in Paediatric Health

The integration of AI in medical diagnostics signifies a transformative era in healthcare, particularly for children suffering from Long COVID. This research not only enhances our understanding of the condition but also sets the stage for practical applications that could significantly improve patient outcomes.

As we move forward, the implications of this study could reshape how we approach paediatric health, making early and accurate diagnosis a reality for countless children facing the challenges of Long COVID.

Source

Article by UNIVERSITA CATTOLICA DEL SACRO CUORE

Revolutionary AI Blood Test Detects Long COVID in Children With 93% Accuracy

A groundbreaking study has revealed that Long Covid in children can be detected through specific blood markers, analyzed with AI, paving the way for accurate and early diagnosis.

Paediatric Long Covid Diagnosis

One day, diagnosing Long Covid in children could be as simple as a blood test, thanks to Artificial Intelligence (AI). A recent study conducted by researchers from Università Cattolica del Sacro Cuore in Rome—along with Fondazione Policlinico Universitario Agostino Gemelli IRCCS and Ospedale Paediatrico Bambino Gesù IRCCS—has identified a unique molecular signature of Long Covid in children’s blood plasma. Using this data, an AI-powered tool was able to diagnose the condition with an impressive 93% accuracy.

The study was published today (January 24) in Paediatric Research, a journal by the Nature Group, and was led by Dr. Danilo Buonsenso, a Paediatric researcher at the Catholic University’s Faculty of Medicine and Surgery and a paediatrician at the Paediatric Infectious Diseases Unit of Fondazione Policlinico Gemelli IRCCS. Dr. Nicola Cotugno, from the Clinical Immunology and Vaccinology Unit at Bambino Gesù IRCCS Paediatric Hospital, also contributed to the research.

The study involved, among others, Dr. Piero Valentini, a researcher in General and Specialist Paediatrics at the Faculty of Medicine and Surgery of the Catholic University and director of the Paediatric Infectious Diseases Operative Unit of the Fondazione Policlinico Gemelli IRCCS, and Dr. Paolo Palma, head of the Complex Operative Unit of Clinical Immunology and Vaccinology of the Bambino Gesù Children’s Hospital IRCCS.

Understanding Long Covid in Children

Long Covid affects on average 0.5% of Paediatric patients exposed to SARS-CoV-2. This condition, also known as Post COVID or Post-Acute Sequelae of SARS-CoV-2, is characterized by the persistence of signs and symptoms for at least 8-12 weeks, symptoms not present before the viral infection, with a negative impact on daily life. Long COVID affects patients of almost all age groups, and among Paediatric patients those older than 10 years seem to be most affected, regardless of the severity of the initial infection.

In adults, the Long COVID ‘sign’ was found in the blood of adults, but a similar finding in the Paediatric population was lacking.

Key Insights from the Study

The experts analyzed the blood of 112 young people aged 0-19 years, of whom 34 had a clinical diagnosis of Long COVID, 32 had the active infection at the time of the study, 27 had Multisystem Inflammatory Syndrome (MIS-C, a severe hyper-inflammatory reaction that almost always requires intensive care) and 19 healthy control peers.

The experts performed an analysis of the blood protein component (proteomics) and saw that, compared to the control groups, the Paediatric Long COVID was characterized by a higher presence in plasma of the pro-inflammatory and pro-angiogenic chemokine sets CXCL11, CXCL1, CXCL5, CXCL6, CXCL8, TNFSF11, OSM, STAMBP1a. An artificial intelligence model based on proteomic profiling was able to identify long Covid with an accuracy of 0.93, a specificity of 0.86, and a sensitivity of 0.97.

Potential Implications for Future Diagnosis

Long Covid in Paediatric patients also presents a distinct protein signature in plasma, characterized by increased inflammation in general and at the level of blood vessel walls (endothelium), as already seen in adults. The discovery could lead to the development of a simple routine diagnostic test based on a blood sample, allowing timely and complete care of the Paediatric patient with Long Covid.

“The immunological data produced in this study provide the evidence needed to identify therapeutic targets to be tested in efficacy and safety studies in Paediatric Long Covid,” Dr. Cotugno of the Bambino Gesù Children’s Hospital explains.

“This work demonstrates incontrovertibly that Long Covid, also in Paediatric age, is an organic immune-mediated disease, for which new funding is needed to study the best therapeutic approaches,” Dr. Buonsenso concludes.