The effect of baseline morphology and its change during treatment on the accuracy of Response Evaluation Criteria in Solid Tumours in assessment of liver metastases

PurposeTumour response assessment to therapy is crucial in oncology. We analysed the morphology of liver metastases (LM) in gastrointestinal stromal tumour (GIST) patients to determine whether uni-dimensional measurement of lesions by Response Evaluation Criteria in Solid Tumours (RECIST), accurately reflects lesion volume.Materials and methodsThe volumes of LM (n = 139) from a GIST patient cohort were measured using computed tomography (CT) at baseline, 3, 6 and 12 months after commencement of imatinib therapy. Baseline measurements were obtained by two independent investigators and inter-observer agreement assessed using Bland–Altman plots. Actual lesion volumes (VACTUAL) were measured and compared with volumes based on the RECIST measure (VRECIST), and with volumes based on three orthogonal measures (VELLIPSOID) at several time-points.ResultsAt baseline, the inter-observer bias for VACTUAL was just 1.8%. VRECIST and VELLIPSOID overestimated VACTUAL by a mean of 35% and only 9% respectively (P < 0.0001 for both). At baseline, 44% (61/139) of LM were classified as spheroidal and 56% (78/139) as ellipsoidal. During treatment, only 42% of LM retained their original morphology. The remainder demonstrated significant changes in morphology (from spheroidal to ellipsoidal and vice versa) over time, while the RECIST measure did not reflect such changes.ConclusionsThe morphology of LM in GIST is rarely spherical (an underlying assumption for RECIST) and can change considerably during imatinib therapy. In this setting, measurements using RECIST do not reflect changes in size and morphology. Additionally, whilst VELLIPSOID is a more suitable surrogate for volume estimation, it is still somewhat limited by the morphology and orientation of such lesions. Studies are warranted to further explore the clinical impact of these findings.