RESUMO
Background: Datasets on rare diseases, like pediatric acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL), have small sample sizes that hinder machine learning (ML). The objective was to develop an interpretable ML framework to elucidate actionable insights from small tabular rare disease datasets. Methods: The comprehensive framework employed optimized data imputation and sampling, supervised and unsupervised learning, and literature-based discovery (LBD). The framework was deployed to assess treatment-related infection in pediatric AML and ALL. Results: An interpretable decision tree classified the risk of infection as either "high risk" or "low risk" in pediatric ALL (n = 580) and AML (n = 132) with accuracy of â¼79%. Interpretable regression models predicted the discrete number of developed infections with a mean absolute error (MAE) of 2.26 for bacterial infections and an MAE of 1.29 for viral infections. Features that best explained the development of infection were the chemotherapy regimen, cancer cells in the central nervous system at initial diagnosis, chemotherapy course, leukemia type, Down syndrome, race, and National Cancer Institute risk classification. Finally, SemNet 2.0, an open-source LBD software that links relationships from 33+ million PubMed articles, identified additional features for the prediction of infection, like glucose, iron, neutropenia-reducing growth factors, and systemic lupus erythematosus (SLE). Conclusions: The developed ML framework enabled state-of-the-art, interpretable predictions using rare disease tabular datasets. ML model performance baselines were successfully produced to predict infection in pediatric AML and ALL.
RESUMO
INTRODUCTION: Fungal infections in patients with COVID-19 was one of the most debated topics during the pandemic. OBJECTIVES: To analyze the clinical characteristics and evolution of people living with HIV/AIDS and coinfection with cryptococcus and COVID-19 (group A) or without it (group B). MATERIALS AND METHODS: This is an analytical and retrospective study. We reviewed medical records of patients with meningeal cryptococcosis between April 2020 and May 2021. RESULTS: We studied 65 people living with HIV/AIDS and with cryptococcosis infection diagnosed from April 2020 to May 2021. Fifteen patients with HIV/AIDS suffered from cryptococcosis and COVID-19, and out of these, 14 presented meningitis (group A), while 28 suffered from meningeal cryptococcosis, but did not have COVID-19 (group B). CONCLUSIONS: No statistically significant differences were observed between the two groups (A and B) considering: intracranial hypertension, presence of Cryptococcus antigens in cerebrospinal fluid, sensorium deterioration or mortality. The detection of Cryptococcus antigens in serum by lateral flow assay was highly effective to rapidly diagnose cryptococcosis in patients with HIV/AIDS who also developed COVID-19. Patients of both groups consulted for cryptoccocosis sometime after, in comparison with the pre-pandemic cases related to this infection.
Introducción: Las infecciones fúngicas en pacientes con COVID-19 fue uno de los temas más debatidos durante la pandemia. Objetivo: Analizar las características clínicas y la evolución de personas con VIH/SIDA que presentaron la asociación de criptococosis meníngea y COVID-19 (grupo A), y compararlas con aquellas personas con VIH/SIDA que padecieron criptococosis meníngea, pero sin infección de COVID-19 (grupo B). Materiales y métodos: Se realizó un estudio analítico y retrospectivo en el que se revisaron las historias clínicas de pacientes que padecieron criptococosis meníngea entre abril de 2020 y mayo de 2021. Resultados: Se estudiaron 65 pacientes con HIV/SIDA y con criptococosis, diagnosticados entre abril de 2020 y mayo de 2021 (63 habían desarrollado sida y 2 eran negativos para VIH). De estos, 15 de los pacientes con sida padecían criptococosis y COVID-19, y 14 presentaban meningitis (grupo A), mientras que 28 pacientes padecieron criptococosis meníngea, pero no tuvieron COVID-19 (grupo B). Conclusiones: No se observaron diferencias estadísticamente significativas, entre los dos grupos, respecto a la hipertensión intracraneal, la presencia de antígenos de criptoccoco en líquido cefalorraquídeo, el deterioro del sensorio o la mortalidad. La detección de antígenos de Cryptococcus en suero por ensayo de flujo lateral fue efectiva para diagnosticar rápidamente criptococosis en personas con VIH/sida y con infección de COVID-19. Se observó que los pacientes de ambos grupos consultaron tarde por criptococosis en comparación con los casos prepandémicos de esta infección.