Nonlinear Heart Rate Variability based artificial intelligence in lung cancer prediction

Lung cancer is uncontrolled growth of cells that occurs due to exposure to smoke, radiation and chemicals, which causes chronic stress and associated with impaired autonomic nervous system. Nonlinear heart rate variability (HRV) analysis has been suggested to uncover the performance status of lung cancer subjects and distinguish them from healthy controls. The present work obtained tachogram from recorded electrocardiogram of 104 lung cancer subjects and 30 healthy controls to extract HRV indices. The obtained results suggested lowered HRV (altered autonomic nervous system tone) values from Eastern Cooperative Oncology Group (ECOG) 1 to ECOG4. Subject males had higher HRV measures than their female counterparts. The HRV parameters decreased from ECOG PS of 1 to 4. Control females had higher HRV measures than control males. There was no association between age and HRV measures. Statistically, nonlinear HRV features were observed significant. ANN exhibited ECOG1 83.3%, ECOG2 50%, ECOG3 90%, ECOG4 95% and Controls 86.7%. The prediction analysis using artificial neural network (ANN) and support vector machine (SVM) scoring an accuracy of 93.09% and 100% with nonlinear HRV indices as input thus has been suggested to be a tool of prognostic importance.