Portal de Periódicos da CAPES

Agreement between Competing Imaging Measures of Response of Hepatocellular Carcinoma to Yttrium-90 Radioembolization

2010; Elsevier BV; Volume: 21; Issue: 4 Linguagem: Inglês

10.1016/j.jvir.2009.11.020

1535-7732

Eugene Duke, Jie Deng, Saad Ibrahim, Robert J. Lewandowski, Robert K. Ryu, Kent T. Sato, Frank H. Miller, Laura Kulik, Mary F. Mulcahy, Andrew C. Larson, Riad Salem, Reed A. Omary,

Radiomics and Machine Learning in Medical Imaging

Purpose There are currently at least six major competing criteria used to determine response to yttrium-90 and other liver-directed therapies, including: (i) Response Evaluation Criteria in Solid Tumors (RECIST); (ii) World Health Organization (WHO), (iii) volumetric, (iv) two-dimensional (2D) European Association for the Study of the Liver (EASL), and (v) three-dimensional (3D) EASL criteria; and (vi) functional diffusion-weighted (DW) magnetic resonance (MR) imaging. This study evaluated agreement among these competing tumor response classification schemes based on quantitative measurements of tumor size, necrosis, and changes in water mobility. Materials and Methods In this retrospective study, 20 patients with hepatocellular carcinoma (HCC) underwent 90Y radioembolization. The patients' tumor burden before and 3–6 months after treatment was assessed with MR imaging. The percent change in size of tumors was used to classify patients into response categories. κ and agreement statistics were used to compare concordance among the different criteria. Results Conventional size criteria (RECIST, WHO, and volumetric) all had a substantial level of agreement (κ = 0.76–0.78) when classifying patients into response categories. However, the conventional size criteria in relation to 2D or 3D EASL had only slight to moderate concurrence, with κ statistics as low as 0.06. Two-dimensional EASL criteria and functional DW MR imaging resulted in the highest response rates, 55% (n = 11) and 75% (n = 15), respectively, whereas conventional size criteria produced lower response rates. Conclusions Classification of HCC response to 90Y radioembolization is related to which of the competing criteria are used. It is recommended that anatomic imaging criteria be used as the primary method to determine response and functional imaging criteria be used as a complementary secondary method. There are currently at least six major competing criteria used to determine response to yttrium-90 and other liver-directed therapies, including: (i) Response Evaluation Criteria in Solid Tumors (RECIST); (ii) World Health Organization (WHO), (iii) volumetric, (iv) two-dimensional (2D) European Association for the Study of the Liver (EASL), and (v) three-dimensional (3D) EASL criteria; and (vi) functional diffusion-weighted (DW) magnetic resonance (MR) imaging. This study evaluated agreement among these competing tumor response classification schemes based on quantitative measurements of tumor size, necrosis, and changes in water mobility. In this retrospective study, 20 patients with hepatocellular carcinoma (HCC) underwent 90Y radioembolization. The patients' tumor burden before and 3–6 months after treatment was assessed with MR imaging. The percent change in size of tumors was used to classify patients into response categories. κ and agreement statistics were used to compare concordance among the different criteria. Conventional size criteria (RECIST, WHO, and volumetric) all had a substantial level of agreement (κ = 0.76–0.78) when classifying patients into response categories. However, the conventional size criteria in relation to 2D or 3D EASL had only slight to moderate concurrence, with κ statistics as low as 0.06. Two-dimensional EASL criteria and functional DW MR imaging resulted in the highest response rates, 55% (n = 11) and 75% (n = 15), respectively, whereas conventional size criteria produced lower response rates. Classification of HCC response to 90Y radioembolization is related to which of the competing criteria are used. It is recommended that anatomic imaging criteria be used as the primary method to determine response and functional imaging criteria be used as a complementary secondary method.