Despite advancements in detection and treatment, tuberculosis (TB), an infectious illness caused by the Mycobacterium TB bacteria, continues to pose a serious threat to world health. The TB diagnosis phase includes a patient’s medical history, physical examination, chest X-rays, and laboratory procedures, such as molecular testing and sputum culture. In artificial intelligence (AI), machine learning (ML) is an advanced study of statistical algorithms that can learn from historical data and generalize the results to unseen data. There are not many studies done on the ML algorithm that enables the prediction of treatment success for patients with pulmonary TB (PTB). The objective of this study is to identify an effective and predictive ML algorithm to evaluate the detection of treatment success in PTB patients and to compare the predictive performance of the ML models. In this retrospective study, a total of 1236 PTB patients who were given treatment under a randomized controlled clinical trial at the ICMR-National Institute for Research in Tuberculosis, Chennai, India were considered for data analysis. The multiple ML models were developed and tested to identify the best algorithm to predict the sputum culture conversion of TB patients during the treatment period. In this study, decision tree (DT), random forest (RF), support vector machine (SVM) and naïve bayes (NB) models were validated with high performance by achieving an area under the curve (AUC) of receiver operating characteristic (ROC) greater than 80%. The salient finding of the study is that the DT model was produced as a better algorithm with the highest accuracy (92.72%), an AUC (0.909), precision (95.90%), recall (95.60%) and F1-score (95.75%) among the ML models. This methodology may be used to study the precise ML model classification for predicting the treatment success of TB patients during the treatment period.
Background: In the medical field, various deep learning (DL) algorithms have been effectively used to extract valuable information from unstructured clinical text data, potentially leading to more effective outcomes. This study utilized clinical text data to classify clinical case reports into tuberculosis (TB) and non-tuberculosis (non-TB) groups using natural language processing (NLP), a pre-processing technique, and DL models. Methods: This study used 1743 open-source respiratory disease clinical text data, labeled via fuzzy matching with ICD-10 codes to create a labeled dataset. Two tokenization methods preprocessed the clinical text data, and three models were evaluated: the existing Text-CNN, the proposed Text-CNN with t-test, and Bio_ClinicalBERT. Performance was assessed using multiple metrics and validated on 228 baseline screening clinical case text data collected from ICMR–NIRT to demonstrate effective TB classification. Results: The proposed model achieved the best results in both the test and validation datasets. On the test dataset, it attained a precision of 88.19%, a recall of 90.71%, an F1-score of 89.44%, and an AUC of 0.91. Similarly, on the validation dataset, it achieved 100% precision, 98.85% recall, 99.42% F1-score, and an AUC of 0.982, demonstrating its effectiveness in TB classification. Conclusions: This study highlights the effectiveness of DL models in classifying TB cases from clinical notes. The proposed model outperformed the other two models. The TF-IDF and t-test showed statistically significant feature selection and enhanced model interpretability and efficiency, demonstrating the potential of NLP and DL in automating TB diagnosis in clinical decision settings.
Background: Multisystem inflammatory syndrome in children (MIS-C), a sequela of severe acute respiratory syndrome coronavirus-2 infection (SARS-CoV2), has been progressively reported worldwide, with cardiac involvement being a frequent presentation. Although the clinical and immunological characteristics of MIS-C with and without cardiac involvement have been described, the immunological differences between cardiac and non-cardiac MIS-C are not well understood. Methods: The levels of type 1, type 2, type 17, other proinflammatory cytokines and CC chemokines and CXC chemokines were measured using the Magpix multiplex cytokine assay system in MIS-C children with MIS-C cardiac (MIS-C (C) (n = 88)) and MIS-C non-cardiac (MIS-C (NC) (n = 64)) abnormalities. Results: MIS-C children with cardiac manifestations presented with significantly increased levels of cytokines such as IFN-γ, IL-2, TNFα, IL-5, IL-1α, IL-1β, IL-6, IL-10 and IL-12p70 and chemokines such as CCL2, CCL3, CCL11 and CXCL10 in comparison to MIS-C children without cardiac manifestations. Clustering analysis revealed that cytokines and chemokines could clearly distinguish MIS-C children with and without cardiac manifestations. In addition, these responses significantly diminished and normalized 9 months after treatment. Conclusions: This is one of the first studies characterizing and differentiating systemic inflammation in MIS-C with and without cardiac involvement from a low- and middle-income country (LMIC). Our study contributes to the existing body of evidence and advances our knowledge of the immunopathogenesis of MIS-C in children.
Background: Globally, neonatal fungal sepsis (NFS) is a leading cause of neonatal mortality, particularly among vulnerable populations in neonatal intensive care units (NICU). The use of spatial frailty models with a Bayesian approach to identify hotspots and risk factors for neonatal deaths due to fungal sepsis has not been explored before. Methods: A cohort of 80 neonates admitted to the NICU at a Government Hospital in Tamil Nadu, India and diagnosed with fungal sepsis through blood cultures between 2018–2020 was considered for this study. Bayesian spatial frailty models using parametric distributions, such as Log-logistic, Log-normal, and Weibull proportional hazard (PH) models, were employed to identify associated risk factors for NFS deaths and hotspot areas using the R version 4.1.3 software and QGIS version 3.26 (Quantum Geographic Information System). Results: The spatial parametric frailty models were found to be good models for analyzing NFS data. Abnormal levels of activated thromboplastin carried a significantly higher risk of death in neonates across all PH models (Log-logistic, Hazard Ratio (HR), 95% Credible Interval (CI): 22.12, (5.40, 208.08); Log-normal: 20.87, (5.29, 123.23); Weibull: 18.49, (5.60, 93.41). The presence of hemorrhage also carried a risk of death for the Log-normal (1.65, (1.05, 2.75)) and Weibull models (1.75, (1.07, 3.12)). Villivakkam, Tiruvallur, and Poonamallee blocks were identified as high-risk areas. Conclusions: The spatial parametric frailty models proved their effectiveness in identifying these risk factors and quantifying their association with mortality. The findings from this study underline the importance of the early detection and management of risk factors to improve survival outcomes in neonates with fungal sepsis.
Background: Globally, Neonatal fungal sepsis (NFS) is a leading cause of neonatal mortality, particularly among vulnerable populations in neonatal intensive care units (NICU). The use of spatial frailty models with a Bayesian approach to identify hotspots and risk factors for neonatal deaths due to fungal sepsis has not been explored before. Methods: A cohort of 80 neonates admitted to the NICU and diagnosed with fungal sepsis through blood culture at a Government Hospital in Tamil Nadu, India, during 2018-2020 was considered for this study. The Bayesian spatial frailty models using parametric distributions, including Log-logistic, Log-normal, and Weibull proportional hazard (PH) models were employed to identify associated risk factors for deaths of NFS and hotspot areas using R software. Results: The spatial parametric frailty models were found to be good models for ana-lysing NFS data. Abnormal levels of activated thromboplastin had a significantly higher risk of death across all PH models (Log-logistic, Hazard Ratio (HR), 95% Credible In-terval (CI): 22.12, (5.40,208.08); Log-normal: 20.87, (5.29,123.23); Weibull: 18.49, (5.60,93.41). Haemorrhage had a higher risk of death for the Log-normal (1.65, (1.05,2.75)) and Weibull models (1.75, (1.07,3.12)). Villivakkam, Tiruvallur, and Poo-namallee blocks were identified as high-risk areas. Conclusions: The spatial parametric frailty models proved effective in identifying these risk factors and quantifying their association with mortality. The findings from this study underline the importance of early detection and management of risk factors to improve survival outcomes in neonates with fungal sepsis.