Background: The effect of incidental pulmonary embolism (PE) on long-term prognosis in cancer patients is unclear. This study assessed the characteristics of cancer and venous thromboembolism (VTE) and the effect of incidental PE identified by oncologists on long-term survival of patients with cancer.Methods and Results:This single-center, retrospective, cohort study used hospital-based cancer registry data from the Osaka International Cancer Institute linked with electronic medical records and administrative data from Japan's Diagnosis Procedure Combination Per-diem Payment System. Overall, 15,689 cancer patients underwent contrast-enhanced thoracic computed tomography during 2010-2018. After excluding patients with missing data, symptomatic patients, or patients with suspected PE, 174 with incidental PE (PE+ group) and 13,197 with no PE (PE- group) were identified. The total incidence of incidental PE was 1.3%. No deaths from thrombotic events were identified in the PE+ group. Both groups were adjusted for cancer- and VTE-related characteristics using inverse probability weighting. After adjusting for immortal time bias in the PE+ group, Kaplan-Meier analysis revealed that all-cause mortality was higher in the PE+ group (hazard ratio, 2.26; 95% confidence interval, 1.53-3.33). A Cox proportional hazard model revealed that metastatic cancer and a history of curative treatment were significant prognostic factors, whereas central PE and residual proximal deep vein thrombosis were not.
Conclusions: Incidental PE in cancer patients indicates poorer prognosis. Cancer-related but not thrombosis-related factors determine prognosis.
Keywords: Immortal time bias; Prognostic factor; Pulmonary embolism; Venous thromboembolism.