Radiographic screening for cancer. Proposed paradigm for requisite research.

Computed tomography (CT) imaging as an excellent approach to the detection and characterization of small solitary pulmonary nodules (SSPN) raises three questions: (1) How often does CT imaging lead to detection of SSPN? (2) How often is such an SSPN malignant? (3) If malignant, how curable is it? The first question pertains to decisions about screening use of CT (clinical or mass screening), the second to decisions about screening for SSPN and diagnosis of malignancy given SSPN, and the third--in the context of known curability at ordinary clinical diagnosis--to decisions about screening for SSPN, diagnosis given SSPN and intervention given malignant SSPN. We present a three component study design that addresses these questions. The first is directed primarily to the first question. Some 1000 persons at high risk for lung cancer will be screened for SSPN using screening-type CT. The primary aim is to determine the prevalence of CT-detectable SSPN as a joint function of risk-relevant aspects of the person. The second component addresses the prevalence of malignancy among the detected cases of SSPN. To develop the prevalence function, a larger series of CT-detected SSPN will be obtained by developing a multi-center SSPN "registry." A subsequent, third component will focus on the registered cases of malignant SSPN screening incidentally detected and address their curability on the basis of long-term follow-up. This design, in lieu of a randomized trial, may represent a new paradigm for applied research on radiologic technologies in cancer screening, given its advantages in terms of research efficiency and implications to decisions about diagnostic workup and therapeutic intervention.