Venous thromboembolism (VTE) is a term referring to blood clots in the veins. Lung cancer compared to other cancer types carry a higher risk. The symptoms of venous thromboembolism are pain when breathing, shortness of breath, swelling, pain in the legs, typically in the thigh, enlarged arm or leg, skin warming, and skin stripes of red. It can strike anyone at any age and result in serious disease, a disability, and occasionally even death. Chemicals released by tumors increase the cancer patient’s body’s propensity to clot the blood. It occurs in almost one in five persons with lung cancer. Currently, to find out the early stage of blood clots, several feature extraction approaches utilized in medicine use CT datasets with increased accuracy. Despite the implementation of many techniques, data interference and poor consistency still exist. Expeditious detection has been achieved using the novel approach. This paper proposed the fuzzy rough set-based feature selection (FRST-FS) algorithm has been utilized for huge datasets with the rule of membership function to avoid overlap between the data and to accelerate the binary processes; harmony search algorithm (HSA) operates with excellent reliability for feature extraction. The results indicate that FRS-FS-HSA outperforms conventional feature selection strategies with a shorter processing time, achieving significant dimensionality reduction and higher classification accuracy.