Total body screening: predicting actionable findings.

Abstract

RATIONALE AND OBJECTIVES: Total body screening, despite its popularity, has not been evaluated in clinical trials. Even the appropriate target for screening has not been addressed. In this study, we determined the variables from a subject's demographic and medical and family history that are predictive of actionable findings on total body screening. MATERIALS AND METHODS: Over a 3-year period, 982 self-referred subjects underwent total body screening with multislice computed tomography and completed a demographic and medical history questionnaire. The study sample was divided into training and testing samples. Univariate and multiple-variable statistical methods were used on the training sample to derive models that predict actionable lung findings, actionable heart findings, actionable abdomen/pelvis findings, and any actionable findings on total body screening. The training models were then applied and evaluated on the test sample. RESULTS: A subject's age at the time of screening was the single most important predictor and often the only significant predictor of actionable findings. Among subjects younger than 40 years of age, 22.5% had actionable findings; this number nearly doubled, to 43.5%, for subjects between 40 and 49, and increased to 80% for subjects 80 years and older. Overall, every increase of 10 years in age brings an increase of 1.6 in the likelihood of an actionable finding. CONCLUSIONS: Total body screening targeted at older subjects has the highest yield of actionable findings. The efficacy and cost-effectiveness of total body screening for older subjects is unknown and needs further assessment.