by Gaia Schiavon, Alessandro Ruggiero, Patrick Schöffski, Bronno van der Holt, Dave J. Bekers, Karel Eechoute, Vincent Vandecaveye, Gabriel P. Krestin, Jaap Verweij, Stefan Sleijfer, Ron H. J. Mathijssen
Assessment of tumor size changes is crucial in clinical trials and patient care. We compared imatinib-induced volume changes of liver metastases (LM) from gastro-intestinal stromal tumors (GIST) to RECIST and Choi criteria and their association with overall survival (OS). Methods
LM from 84 GIST patients (training and validation set) were evaluated using manual and semi-automated Computed Tomography measurements at baseline, after 3, 6 and 12 months of imatinib. The ability of uni-dimensional (1D) and three-dimensional (3D) measurements to detect size changes (increase/decrease) =20% was evaluated. Volumetric response cut-offs were derived from minimally relevant changes (+20/-30%) by RECIST, considering lesions as spherical or ellipsoidal. Results
3D measurements detected size changes =20% more frequently than 1D at every time-point (P=0.008). 3D and Choi criteria registered more responses than RECIST at 3 and 6 months for 3D-spheres (P=0.03) and at all time-points for 3D-ellipsoids and Choi criteria (P<0.001). Progressive disease by 3D criteria seems to better correlate to OS at late time-points than other criteria. Conclusion
Volume criteria (especially ellipsoids) classify a higher number of patients as imatinib-responders than RECIST. Volume discriminates size changes better than diameter in GIST and constitutes a feasible and robust method to evaluate response and predict patient benefit.