University of Rochester Medical Center researchers have created a risk model that can predict with 98 percent certainty which cancer patients will not get a blood clot.

Blood clots are common in cancer patients and on the rise. Alok Korana, M.D. and colleagues discovered five variables that contribute to understanding a patient's risk of clots: site of cancer (pancreas, stomach, brain and lung are worst); body mass index above 35; and three measures within a blood sample including platelet, hemoglobin and white blood cell count. Patients at higher risk are candidates to receive blood thinners.

The researchers also discovered that a key to preventing clots lies with a biomarker called tissue factor (TF). TF is overexpressed by solid tumors. They hope to develop a blood test that could be run at the time of the tumor biopsy that would analyze TF and predict the likelihood of a dangerous clot forming. They have already shown the potential benefit of a test as pancreatic patients with a high TF expression had a venous thromboembolism rate of 26.3 percent compared with 4.5 percent of patients with low TF expression.