Synthetic Patient Data Generation and Evaluation in Disease Prediction Using Small and Imbalanced Datasets.

The increasing prevalence of chronic non-communicable diseases makes it a priority to develop tools for enhancing their management. On this matter, Artificial Intelligence algorithms have proven to be successful in early diagnosis, prediction and analysis in the medical field. Nonetheless, two main issues arise when dealing with medical data: lack of high-fidelity datasets and maintenance of patient's privacy. To face these problems, different techniques of synthetic data generation have emerged as a possible solution. In this work, a framework based on synthetic data generation algorithms was developed. Eight medical datasets containing tabular data were used to test this framework. Three different statistical metrics were used to analyze the preservation of synthetic data integrity and six different synthetic data generation sizes were tested. Besides, the generated synthetic datasets were used to train four different supervised Machine Learning classifiers alone, and also combined with the real data. F1-score was used to evaluate classification performance. The main goal of this work is to assess the feasibility of the use of synthetic data generation in medical data in two ways: preservation of data integrity and maintenance of classification performance.

Multiple vertebral fractures after suspension of denosumab. A series of 56 cases.

BACKGROUND: Denosumab is a monoclonal antibody approved for the treatment of postmenopausal osteoporosis. The withdrawal of denosumab produces an abrupt loss of bone mineral density and may cause multiple vertebral fractures (MVF). OBJECTIVE: The objective of this study is to study the clinical, biochemical, and densitometric characteristics in a large series of postmenopausal women who suffered MVF after denosumab withdrawal. Likewise, we try to identify those factors related to the presence of a greater number of vertebral fractures (VF). PATIENTS AND METHODS: Fifty-six patients (54 women) who suffered MVF after receiving denosumab at least for three consecutive years and abruptly suspended it. A clinical examination was carried out. Biochemical bone remodelling markers (BBRM) and bone densitometry at the lumbar spine and proximal femur were measured. VF were diagnosed by magnetic resonance imaging MRI, X-ray, or both at dorsal and lumbar spine. RESULTS: Fifty-six patients presented a total of 192 VF. 41 patients (73.2%) had not previously suffered VF. After discontinuation of the drug, a statistically significant increase in the BBRM was observed. In the multivariate analysis, only the time that denosumab was previously received was associated with the presence of a greater number of VF (P = .04). CONCLUSIONS: We present the series with the largest number of patients collected to date. 56 patients accumulated 192 new VF. After the suspension of denosumab and the production of MVF, there was an increase in the serum values of the BBRM. The time of denosumab use was the only parameter associated with a greater number of fractures.