Atrial fibrillation risk prediction from the 12-lead electrocardiogram using digital biomarkers and deep representation learning.

Aims: This study aims to assess whether information derived from the raw 12-lead electrocardiogram (ECG) combined with clinical information is predictive of atrial fibrillation (AF) development. Methods and results: We use a subset of the Telehealth Network of Minas Gerais (TNMG) database consisting of patients that had repeated 12-lead ECG measurements between 2010 and 2017 that is 1 130 404 recordings from 415 389 unique patients. Median and interquartile of age for the recordings were 58 (46-69) and 38% of the patients were males. Recordings were assigned to train-validation and test sets in an 80:20% split which was stratified by class, age and gender. A random forest classifier was trained to predict, for a given recording, the risk of AF development within 5 years. We use features obtained from different modalities, namely demographics, clinical information, engineered features, and features from deep representation learning. The best model performance on the test set was obtained for the model combining features from all modalities with an area under the receiver operating characteristic curve (AUROC) = 0.909 against the best single modality model which had an AUROC = 0.839. Conclusion: Our study has important clinical implications for AF management. It is the first study integrating feature engineering, deep learning, and Electronic medical record system (EMR) metadata to create a risk prediction tool for the management of patients at risk of AF. The best model that includes features from all modalities demonstrates that human knowledge in electrophysiology combined with deep learning outperforms any single modality approach. The high performance obtained suggest that structural changes in the 12-lead ECG are associated with existing or impending AF.

Near infrared light examination as part of the management of sporadic pancreatic head insulinoma: Case report.

INTRODUCTION: We report the case of a 77-year-old female patient with the diagnosis of pancreatic head insulinoma, in whom we used near infrared light (NIR) to detect synchronous pancreatic tumors and potential secondary lymph node or liver involvement. The patient presented with hypoglycemia manifesting by lipothymia. With the diagnosis of secretory neuroendocrine tumor (insulinoma) of the pancreatic head, cephalic pancreatoduodenectomy with the preservation of the pylorus was performed after NIR visualization of the pancreatic tumor mass. At 6, 12, 18 months postoperatively, the patient no longer had hypoglycemia and her general state was good. CONCLUSION: NIR with indocyanine green (ICG) evidences pancreatic neuroendocrine tumors, as well as possible synchronous tumors and secondary lymph node or liver involvement.

Detection of epithelial specific cell adhesion molecules in colon cancer and the correlation with clinical and pathological characteristics EpCAM expression in colon cancer.

AIM: The aim of this study was to evaluate the correlation between EpCAM expression in colon cancer tissue and the clinico-pathological characteristics of the patients. MATERIAL AND METHODS: This is a prospective, longitudinal, observational study on 80 patients undergoing for colon cancer between January - December 2017. EpCAM expression at tumoral level was analyzed in relation with clinical and pathological variables of the patients using anti-EpCAM specific antibody. RESULTS: EpCAM expression was predominant in tumoral tissue compared to normal colonic mucosa and most of the cases (58.7%) showed increased EpCAM expression. Although increased EpCAM expression was observed in advanced stages and in patients with advanced locoregional disease, there was no statistically significant correlation with the clinical and pathological characteristics of the patients. DISCUSSION: The majority of the analyzed samples showed increased EpCAM expression in tumoral tissue suggesting its involvement in the carcinogenesis process. Numerous studies have identified EpCAM overexpression in colon cancer as a negative prognostic factor, being associated with advanced stage of the disease and a poor prognosis of the patient but results are inconsistent. Nevertheless, assessing a possible correlation between EpCAM expression at tumoral level and clinico- pathological characteristics is dependent on the type of antibody used to identify the molecule of interest. CONCLUSIONS: EpCAM detection in colon cancer using anti-human CD326/EpCAM clone VU-1D9 does not allow the correlation between its expression and the clinico-pathological characteristics of the patients and it should only be used for EpCAM identification in colon cancer tissues. KEY WORDS: Cancer, Colon, EpCAM, Immunohistochemistry.

Predictors of 1-year postoperative mortality in radical colon cancer surgery.

AIM: The aim of this study was to determine the impact of patient, tumor and surgery-related parameters on 1-year postoperative mortality in a cohort of patients operated in a single tertiary center. MATERIALS AND METHODS: The study included 605 patients diagnosed with colon cancer between January 2013 and December 2015 that underwent radical surgery in a tertiary center. Patient demographics, comorbidities, preoperative biological parameters alongside with tumor and surgery-related factors were prospectively recorded and then analyzed in relation 1-year postoperative mortality. RESULTS: One-year mortality rate in the study group was 10.9%. Independent risk factors in relation to 1-year mortality were advanced TNM stage (OR 3.10, 1.10 - 8.75 95% CI ), emergency surgery (OR 1.91, 1.11 - 3.74 95% CI ), location of the tumor in the ascending colon (OR 2.17, 1.32 - 3.57 95% CI ), multiorgan resections (OR 2.07, 1.15 - 3.74 95% CI), age over 63 years (OR 2.05, 1.16 - 3.62 95% CI) and the history of alcohol consumption (OR 2.058, 1.17 - 3.61 95% CI ). DISCUSSION: Postoperative complications are still being reported in colon cancer surgery, despite technological progress and constant research in the field. So far, factors that influence postoperative mortality have been mostly studied up to 30 days postoperatively. According to some recent papers, reporting 30-day mortality data can underestimate accurate communication of postoperative adverse events. Thus, 1-year mortality in colon cancer surgery could be a better indicator of the impact on surgery on postoperative period of this patients and factors that influence it should be well known. KEY WORDS: Surveillance, Colon cancer, 1-year mortality.