No abstract

Artificial Intelligence is creating a paradigm shift in health care, with phenotyping patients through clustering techniques being one of the areas of interest. Objective: To develop a predictive model to classify heart failure (HF) patients according to their left ventricular ejection fraction (LVEF), by using available data from Electronic Health Records (EHR). Subjects and methods: 2854 subjects over 25 years old with a diagnosis of HF and LVEF, measured by echocardiography, were selected to develop an algorithm to predict patients with reduced EF using supervised analysis. The performance of the developed algorithm was tested in heart failure patients from Primary Care. To select the most influentual variables, the LASSO algorithm setting was used, and to tackle the issue of one class exceeding the other one by a large amount, we used the Synthetic Minority Oversampling Technique (SMOTE). Finally, Random Forest (RF) and XGBoost models were constructed. Results: The full XGBoost model obtained the maximum accuracy, a high negative predictive value, and the highest positive predictive value. Gender, age, unstable angina, atrial fibrillation and acute myocardial infarct are the variables that most influence EF value. Applied in the EHR dataset, with a total of 25,594 patients with an ICD-code of HF and no regular follow-up in cardiology clinics, 6170 (21.1%) were identified as pertaining to the reduced EF group. Conclusion: The obtained algorithm was able to identify a number of HF patients with reduced ejection fraction, who could benefit from a protocol with a strong possibility of success. Furthermore, the methodology can be used for studies using data extracted from the Electronic Health Records.

We report the histopathological, immunohistochemical (IHC), and molecular findings in 3 patients with adult pancreatoblastoma, including 2 with autopsy features. The tumors were located in the tail and body of the pancreas, and the 2 autopsy examinations revealed liver and lung metastases. Histopathologically the neoplasms were composed of solid epithelial elements with nested or trabecular growth patterns, fibrous stroma, and squamoid clusters. Keratin 19 was positive mainly in squamoid corpuscles, and trypsin or chymotrypsin was positive in the acinar component. Neuroendocrine differentiation was observed in all tumors, and nuclear β-catenin expression in 2 tumors. Despite nuclear β-catenin expression, CTNNB1 mutation was found only in tumor 2. APC mutation was detected in tumor 1, and SMAD4 as well as MEN1 mutations in tumor 3. This last tumor also revealed chromosomal instability with many chromosomal losses and gains. The follow-up showed regional or distant metastases in all patients. Two patients died of disease after 3 and 26 months of follow-up and 1 patient is alive with no evidence of disease 6 years and 2 months after surgery. Adult pancreatoblastoma can display genetic heterogeneity, diverse histological appearance, and overlapping IHC findings. As a result, the differential diagnosis with other adult pancreatic tumors, such as acinar cell carcinoma, neuroendocrine neoplasm, solid pseudopapillary neoplasm, and mixed tumors may be challenging, especially when dealing with limited tumor tissue. The identification of squamoid corpuscles is essential for diagnosis. Although molecular findings might provide useful information, the integration of clinical, radiological, and histopathological findings is essential in pancreatoblastoma diagnosis.

Objective:To assess the impact of BP changes on the risk of cardiovascular events and mortality in type 2 diabetes.Design and method:The sample was selected using the HER of the Valencian Community. Type 2 diabetics were selected through ICD codes and retrospectively evaluated from January 2012 to December 2016. To evaluate BP changes, the follow-up time was divided in six-month blocks and the average of BP for each interval was considered. SBP was categorized: < 120 mmHg; 120–129 mmHg (reference); 130–139 mmHg; 140–149 mmHg; > 150 mmHg. To evaluate the influence of DBP, the population was stratified according to categories of average DBP: < 60 mmHg; 60–69 mmHg; 70–79 mmHg; > 80 mmHg. Information about cardiovascular events, including Stroke and Acute Coronary Syndrome (ACS) was extracted from the ICD codes. Total mortality was determined by matching records and death certificates. Time-varying Cox regression for SBP stratified by DBP categories was used to assess the risk associated with changes of SBP. The models were adjusted by age, sex, HBA1c, KDIGO, previous cardiovascular events and use of cardiovascular drugs.Results:156 363 type 2 diabetics patients were included (mean age 69.5(11.5), 48.7% females, mean glycated Hb 7.05 (1.35%) 21% in secondary prevention, 21.4% under insulin treatment). The average number of BP measurements was 14. During an average follow-up of 4.18 y, there were 13399 deaths, 15100 strokes and 6295 ACS. In the fully adjusted time-varying Cox regression, having a SBP< 120 mmHg was a significant risk factor for death across all categories of DBP, whereas SBP> 130 mmHg conferred protection. For the case of ACS and stroke, a J-curve phenomenon was observed with significantly higher risk for those with SBP> 150 mmHg or < 120 mmHg across categories with DBP> 60 mmHg. This J-curve phenomenon was also observed graphically using restricted cubic splines (Figure).Conclusions:Although difficult to achieve in clinical practice, lower boundaries for BP goals in diabetics have to be taking into account.