Purpose: Intraindividual comparison of quantitative malignant liver tumor response analysis using computed tomography. The RECIST and WHO evaluation results were compared to the volumetry results.
Materials and method: Consecutive CT follow-up investigations (portal-venous phase, collimation 3 mm, increment 2 mm) of 82 patients were analyzed retrospectively. The median interval was 56 (30 - 455) days. The patients showed a total of 198 (median 3, range 1 - 5) malignant liver lesions. The evaluation was performed by 2 radiologists using the OncoTREAT software (Mevis) in consensus. The results were classified according to RECIST (Response Evaluation Criteria in Solid Tumors, stable disease: - 30 % to + 20 %) and WHO (stable: - 50 % to + 25 %) and compared to the volumetric analysis (stable: - 65 % to + 44 %). Both the continual follow-up changes and the classified results (complete and partial remission, no change and progression) were analyzed.
Results: The classified RECIST and WHO results agreed with the volumetric analysis in 71/82 (87 %) of cases kappa (RECIST) = 0,699, kappa (WHO) = 0,741). This included different patients thus showing the agreement between the RECIST and WHO evaluations in 68/82 (83 %) cases (kappa = 0,656). The estimation of the relative tumor development was clearly different in all procedures. Relative tumor changes are not directly comparable because of underlying one-, two- and three-dimensional structures.
Conclusion: If the tumor size is regarded as a suitable parameter of tumor development, quantifying procedures according to RECIST and WHO are of limited use. Relatively reliable tumor response evaluation can only take place by means of thin section investigation and volumetric analysis on the basis of digital image data.