Scientific Paper

Measurement of tumor volumes improves RECIST-based response assessments in advanced lung cancer: Objective: This study was designed to (1) estimate the reproducibility of measurement for tumor volumes their longest tumor diameters (LDs); (2) estimate the potential impact of using changes in tumor volumes instead of LDs as the basis for response assessments; and (3) create a data set to help qualify image analysis tools.
Methods: We studied patients with advanced lung cancer who had been followed longitudinally with x-ray computed tomography (CT) in a multi-national trial. A total of 71 time-points from 10 patients with 13 morphologically complex target lesions were analyzed.  A total of 6,461 volume measurements and their corresponding LDs were made by 7 independent teams using their own work flows and image analysis tools.  Inter-team agreement and overall inter-rater concurrence were characterized. 
Results: Inter-team agreement between volume measurements was better than between LD measurements (Iota = 0.945 v. 0.734, p = 0.005).  Variability in determining the nadir was lower for volumes than for LDs (p = 0.005).  Use of standard thresholds for the RECIST-based method and use of experimentally determined cutoffs for categorizing responses showed that volume measurements had a significantly greater sensitivity for detecting partial responses and disease progression. Earlier detection of progression would have led to earlier changes in patient management in a majority of cases.
Conclusions: Our findings indicate that measurement of changes in tumor volumes is adequately reproducible.  Using tumor volumes as the basis for response assessments could have a positive impact on both patient management and clinical trials. The value of volume measurements as criteria for lung tumor response assessments should be further explored.
(Published in Translational Oncology, Volume 5, Issue 1, pp. 19-25, Feb 2012)