Computer-aided diagnosis (CAD) systems have enormous potential in medical imaging and diagnostic radiology, assisting radiologists in acquiring, managing, storing, and reporting digital medical images from various imaging techniques. In lymphadenopathy, the sole criterion to determine an abnormal lymph node is enlarged size; yet CT cannot display abnormal architecture in a normalsized node, which is the most significant shortcoming of the technique and a source of most false-negative results from CT examinations. We employed the deep convolutional neural network ResNet-34 to classify lymph node lesions in CT images from Abdominal Lymphadenopathy patients and Healthy Controls. We created a single database containing 1400 source CT images for Abdominal Lymphadenopathy patients (n=700) and Healthy Controls (n=700). Images were resized, normalized, and arranged in m batches of 16 images before supervised training, testing, and cross-validation of the ResNet-34, to identify and label lesions with automatic volume delineation of target areas. The ResNet-34 had high diagnostic accuracy, with an AUC of 0.9957 for Abdominal Lymphadenopathy and 1.00 for Healthy Control. Thus, the two groups had identically high sensitivity and specificity values of 99.57% and 100%. The added effect of ResNet-34 is a success rate of 99.57% and 100% for classifying random CT images, with an overall accuracy of 99.79% in the testing subset for classifying lymph node lesions. We believe the final layer of ResNet-34's output activation map is a powerful tool for diagnosing lymph node lesions of lymphadenopathy from CT images because of its high classification precision.