Image: a). This model can accurately identify AD lesions in the training data. b). LASSO (REC and AAG) and LR (REC and AAG) models showed good classification performed on the test dataset. c). All genes showed significant time-dependent downregulation in his LASSO and LR models with dupilumab treatment. d). LASSO(REC) and LR(AAG) model scores were positively correlated with SCORD. e). Correlations between LASSO (REC) and LR (AAG) model scores and immune cell infiltration.
opinion more
Credits: author
Atopic dermatitis (AD) is the most common chronic inflammatory disease in children and adults. This condition severely impacts the patient’s quality of life. Worldwide prevalence ranges from 7% to 30% in children and 1% to 10% in adults. Clearly, early diagnosis and individualized treatment of AD require reliable and accurate assessment methods.
The use of machine learning (ML) in the medical field has grown significantly in recent years. ML can help detect and classify diseases, improve predictions, and personalize treatments. Although ML is primarily used in dermatology to identify skin lesions and histopathological images such as vitiligo and psoriasis, few models have been reported at the molecular level for Alzheimer’s disease. Until recently, his Songjiang Wu, a doctoral candidate in dermatology at the Third University, was reported. Xiangya Hospital, Central South University, China and their co-authors have established several relatively stable and reliable diagnostic and efficacy assessment predictive models for AD based on ML algorithms.
Researchers used publicly available RNA transcriptome data from AD lesional and nonlesional lesions with three different ML algorithms to develop six predictive models of AD. Lasso, Linear Regression (LR), Random Forest (RF). The model showed excellent performance in discriminating between AD lesions and non-lesions (AUC > 0.8).
The samples that received biotherapy showed a positive correlation between the model score and SCORAD (“Scoring for Atopic Dermatitis”) and a negative correlation with duration of treatment, showing a trend of improvement.
“These results demonstrate the potential of the model, especially for assessing the therapeutic efficacy of biologics and small molecule drugs. However, due to the small sample size and lack of sample quality, The correlation coefficient between the two models and SCORAD was not high enough,” explained Wu.
The team published their findings in the KeAi journal basic research.
According to corresponding authors Qinghai Zeng and Jing Chen, ML-based models show good predictive performance in the diagnosis and treatment effects of Alzheimer’s disease, suggesting new options for early diagnosis and intervention.
Going forward, the team plans to collect patient samples for validation and evaluation of model stability.
###
Contact author: Qinghai Zeng, zengqinghai@csu.edu.cn
KeAi Publishing was established by Elsevier and China Science Publishing & Media Ltd to bring high quality research to the world. In 2013, our focus shifted to open access publishing. Today, we proudly publish over 100 world-class open access English-language journals across all scientific disciplines. Many of these titles are published by us in partnership with prestigious societies and academic institutions such as the National Natural Science Foundation of China (NSFC).
journal
basic research
research method
Computational simulation/modeling
Research theme
human tissue sample
article title
A Machine Learning-Based Predictive Model for Atopic Dermatitis Diagnosis and Evaluation
Conflict of Interest Statement
The authors declare no conflicts of interest regarding this work
Disclaimer: AAAS and EurekAlert! We are not responsible for the accuracy of news releases posted on EurekAlert!. Use of information by contributing institutions or via the EurekAlert system.
