Srisairam Achuthan, Ph.D.
Senior Scientific Programmer, Research Informatics Division, Department of Information Sciences, City of Hope, CA, USA
Applying Machine learning for Automated Identification of Patient Cohorts
Srisairam Achuthan, Mike Chang, Ajay Shah, Joyce Niland
Patient cohorts for a clinical study are typically identified based on specific selection criteria. In most cases considerable time and effort are spent in finding the most relevant criteria that could potentially lead to a successful study. For complex diseases, this process can be more difficult and error prone since relevant features may not be easily identifiable. Additionally, the information captured in clinical notes is in non-coded text format. Our goal is to discover patterns within the coded and non-coded fields and thereby reveal complex relationships between clinical characteristics across different patients that would be difficult to accomplish manually. Towards this, we have applied machine learning techniques such as artificial neural networks and decision trees to determine patients sharing similar characteristics from available medical records. For this proof of concept study, we used coded and non-coded (i.e., clinical notes) patient data from a clinical database. Coded clinical information such as diagnoses, labs, medications and demographics recorded within the database were pooled together with non-coded information from clinical notes including, smoking status, life style (active / inactive) status derived from clinical notes. The non-coded textual information was identified and interpreted using a Natural Language Processing (NLP) tool I2E from Linguamatics.
![Delegate Talk: A Mobile Phone Application for Daily Physical Activity Monitoring in Chronic Obstructive Pulmonary Disease @ Amriteshwari Hall | Vallikavu | Kerala | India](http://www.amritabioquest.org/conference/2013/wp-content/uploads/sites/2/2015/02/biomed.jpg)
H S M Kort, J-W J Lammers, S N W Vorrink, T Troosters
Introduction
Chronic Obstructive Pulmonary Disease (COPD) is a disabling airway disease with variable extrapulmonary effects that may contribute to disease severity in individual patients (Rabe et al. 2007). The world health organization predicts that COPD will become the third leading cause of death worldwide by 2030. Patients with COPD demonstrate reduced levels of spontaneous daily physical activity (DPA) compared with healthy controls (Pitta et al. 2005). This results in a higher risk of hospital admission and shorter survival (Pitta et al. 2006). Pulmonary rehabilitation can help to improve the DPA level, however, obtained benefits decline after 1–2 years (Foglio et al. 2007).
Purpose
In order to maintain DPA in COPD patients after rehabilitation, we developed a mobile phone application. This application measures DPA as steps per day, measured by the accelerometer of the smartphone, and shows the information to the patient via the display of the mobile phone. A physiotherapist can monitor the patient via a secure website where DPA measurements are visible for all patients. Here, DPA goals can be adjusted and text messages sent.
Method
Three pilot studies were performed with healthy students and COPD patients to test the application for usability, user friendliness and reliability with questionnaires and focus groups. Subjects also wore a validated accelerometer. For the Randomized Controlled Trial (RCT) 140 COPD patients will be recruited in Dutch physiotherapy practises. They will be randomised in an intervention group that receives the smartphone for 6 months and a control group. Measurements include lungfunction, dyspnea, and exercise capacity and are held at 0, 3, 6 and 12 months.
Results and Discussion
The application was found to be useful, easy to learn and use. Subjects had no problems with health care professionals seeing information on their physical activity performance. They do find it important to be able to determine who can see the information. Correlations between the accelerometer and the measurements on DPA of the smartphone for steps per hour were 0.69 and 0.70 for pilot studies 1 (students) and 2 (COPD patients) respectively. The version of the application in pilot study 3 contained an error, which made correlations with the accelerometer unusable. The RCT study is now being executed.
![Delegate Talk: Inefficient NETosis: Cause for Predisposition to Recurrent Infections in Type 2 Diabetes @ Acharya Hall | Vallikavu | Kerala | India](http://www.amritabioquest.org/conference/2013/wp-content/uploads/sites/2/2015/02/gtm.jpg)
Manjunath Joshi, Apoorva Lad, Bharat Prasad Alevoor, Aswath Balakrishnan, Lingadakai Ramachandra and Kapaettu Satyamoorthy
Pathological conditions during Type 2 Diabetes (T2D) are associated with elevated risk for common community acquired infections due to poor glycemic control. Multiple studies have indicated specific defects in innate and adaptive immune function in diabetic subjects. Neutrophils play an important role in eliminating pathogens as an active constituent of innate immune system. Apart from canonically known phagocytosis mechanism, neutrophils are endowed with a unique ability to produce extracellular traps (NETs) to kill pathogens by expelling DNA coated with bactericidal proteins and histone. NETosis is stimulated by diverse bacteria and their products, fungi, protozoans, cytokines, phorbol esters and by activated platelets. Considering deregulation of metabolic and immune response pathways during pathological state of diabetes and NETosis as a potential mechanism for killing bacteria, we therefore, investigated whether hyperglycemic conditions modulate formation of neutrophil NETs and attempted to identify underlying immunoregulatory mechanisms. Freshly isolated neutrophils from normal individuals were cultured in absence or presence of high glucose (different concentrations) for 24 hours and activated with either LPS (2 mg/ml) or PMA (20 ng/ml) or IL-6 (20 ng/ml) for 3 hours. NETs were visualized and quantified by addition of DNA binding dye SYTOX green using fluorescence microscope and fluorimetry. NETs were quantified in Normal and diabetic subjects. Serum IL-6 levels were measured using ELISA technique. NETs bound elasatse were quantified in normal and diabetic subjects in presence or absence of DNase. Bacterial killing assays were performed upon infecting E.coli with activated neutrophils from normal and diabetic subjects. Microscopy and fluorimetry analysis suggested dramatic impairment in NETs formation under high glucose conditions. Extracellular DNA lattices formed in hyperglycemic conditions were short lived and unstable leading to rapid disintegration. Subsequent, time course experiments showed that NETs production was delayed in hyperglycemic conditions. To validate our findings more closely to clinical conditions, we investigated the neutrophil activation and NETs formation in diabetic patients. Upon stimulation with LPS for three hours, neutrophils from diabetic subjects responded weakly to LPS and lesser NETs were formed; whereas, neutrophils from normal individuals showed robust release of NETs. In few patients we found short and imperfect NETs in basal conditions suggesting constitutive activation of neutrophils in diabetic subjects. Interestingly, NETs bound elastase activity was reduced in diabetes subjects when compared to non-diabetic individuals, indicating a dysfunction of one of the important protein component of NETs during diabetes. Neutrophils from diabetic subjects released higher levels of IL-6 without any stimulation suggesting an existence of constitutively activated pro-inflammatory state. IL-6 induced NETs formation and was abrogated by high glucose. Weobserved that glycolysis inhibitor 2-DG resensitize the high glucose attenuated LPS and IL-6 induced NETs. a) NETs are influenced by glucose homeostasis, b) IL-6 as potent inducer of energy dependent NETs formation and c) hyperglycemia mimics a state of constitutively active pro-inflammatory condition in neutrophils leading to reduced response to external stimuli making diabetic subjects susceptible for infections.