Lung cancer remains the primary cause of cancer-related deaths, and early detection is essential to reduce its mortality rate. In India, where the availability of advanced medical facilities is limited, an X-ray-based pulmonary nodule detection model will be a cost-effective and accessible solution for the early detection of lung cancer. In daily clinical practice, it has been observed that on the thoracic imaging of different pulmonary diseases, the incidence rate of pulmonary nodules is quite low. Hence, these lesions on chest X-ray can be considered as the presence of an anomaly. In order to enhance the accuracy of pulmonary nodule detection methodology using digital X-rays, this paper introduces a novel, cost-effective methodology for areas with limited healthcare facilities existed. The study utilizes the “Swash Thoracic Imaging dataset,” a private dataset created in collaboration between Peerless Hospital and the University of Calcutta, which features an imbalanced distribution of cases and includes various pulmonary abnormalities that mimic respiratory disease symptoms. Notably, few datasets have incorporated multiple types of pulmonary conditions in prediction models, making this dataset uniquely comprehensive. In order to address the data imbalance, pulmonary nodules are treated as anomalies. Considering Principal Component Analysis (PCA), the dataset’s dimensionality is reduced, and the Support Vector Data Descriptor (SVDD) is integrated with Convolutional Neural Networks (CNNs) to enhance anomaly detection. The proposed method achieves a sensitivity of 94.78% and specificity of 93.44%, outperforming current nodule detection techniques on chest X-rays in terms of accuracy.