Computed tomographic-pathologic correlation of gross tumor volume and clinical target volume in non-small cell lung cancer - A pilot experience

Context.-Computed tomographic (CT) scan data are used regularly in radiatio n treatment planning for patients with lung cancer. To our knowledge, the r elationship of the CT images of tumors and their corresponding microscopic extent has not yet been studied in detail. Objective.-To correlate tumor sizes on CT with tumor sizes measured microsc opically (ie, the gross tumor volume [GTV]-clinical target volume margin) i n non-small cell lung cancers. Design.-Prospective pilot study. Setting.-Single institution. Patients.-Patients with operable non-small cell lung cancer were identified preoperatively. Interventions.-Once the surgical specimen was available, it was oriented wi th the surgeon and the pathologist. Seven whole-mount, cross-sectional hist ologic glass slides were made from 5 tumors using formalin fixation and hem atoxylin-eosin staining. The pathologist then outlined the cancer-containin g area under the microscope (Micro-GTV) and the area of surrounding inflamm atory response (Micro-GTV + inflammation). Preoperative CT scans were used for outlining tumor on the corresponding slice (CT-GTV). Main Outcome Measures.-Correlation of the areas of Micro-GTV, Micro-GTV + i nflammation, and CT-GTV was performed. Results.-There was an obvious trend that the CT-GTV was bigger than the Mic ro-GTV, except in specimen 1, in which the 2 areas were about equal. Howeve r, on comparing the values for the CT-GTV and the Micro-GTV + inflammation, the difference between the 2 areas became smaller. Conclusions.-Modern CT scans might overestimate the GTV in non-small cell l ung cancer. The GTV-clinical target volume margin could actually be zero or even a negative value. The findings in this small study are interesting an d provoking. Further study with a larger number of patients and more rigid quality control is warranted to confirm our findings.