In Chronic Spontaneous Urticaria, Complete Response to Antihistamine Treatment Is Linked to Low Disease Activity.

INTRODUCTION: The use of predictors of response to a specific treatment in patients with chronic spontaneous urticaria (CSU) can improve disease management, help prevent unnecessary healthcare costs, and save time. In this study, we aimed to identify predictors of complete response to standard-dosed and higher than standard-dosed antihistamine treatments in patients with CSU. METHODS: Medical records of 475 CSU patients, 120 of them <18 years old, from 3 different centers were analyzed. We used 15 machine learning (ML) models as well as traditional statistical methods to predict complete response to standard-dosed and higher than standard-dosed antihistamine treatment based on 17 clinical parameters. RESULTS: CSU disease activity, which was assessed by urticaria activity score (UAS), was the only clinical parameter that predicted complete response to standard-dosed and higher than standard-dosed antihistamine treatment, with ML models and traditional statistics, for all age groups. Based on ROC analyses, optimal cut-off values of disease activity to predict complete response were UAS <3 and UAS <4 for standard-dosed (area under the ROC curve [AUC] = 0.69; p = 0.001) and higher than standard-dosed (AUC = 0.79; p = 0.001) antihistamine treatments, respectively. Also, ML models identified lower total IgE (<150 IU/mL) as a predictor of complete response to a standard-dosed antihistamine and lower CRP (<3.4 mg/mL) as a predictor of complete response to higher than standard-dose antihistamine treatment. DISCUSSION: In this study, we showed that patients with UAS <3 are highly likely to have complete response to standard-dosed AH and those with a UAS <4 are highly likely to have complete response to higher than standard-dosed AH treatment. Low CSU disease activity is the only universal predictor of complete response to AH treatment with both ML models and traditional statistics for all age groups.

On the identification of thyroid nodules using semi-supervised deep learning.

Detecting malign cases from thyroid nodule examinations is crucial in healthcare particularly to improve the early detection of such cases. However, malign thyroid nodules can be extremely rare and is hard to find using the traditional rule based or expert-based methods. For this reason, the solutions backed by Machine Learning (ML) algorithms are key to improve the detection rates of such rare cases. In this paper, we investigate the application of ML in the healthcare domain for the detection of rare thyroid nodules. The utilized dataset is collected from 636 distinct patients in 99 unique days in Turkey. In addition to the texture feature data of the Ultrasound (US), we have also included the scores of different assessment methods created by different health institutions (e.g., Korean, American and European thyroid societies) as additional features. For detection of extremely rare malign cases, we use auto-encoder based neural network model. Through numerical results, it is shown that the auto-encoder based model can result in an average Recall score of 0.98 and a Sensitivity score of 1.00 for detecting malign and non-malign cases from the healthcare dataset outperforming the traditional classification algorithms that are trained after Synthetic Minority Oversampling Technique (SMOTE) oversampling.