In Vivo and Ex Vivo Elasticity Measurements of Human Breast Cancers

Abstract

Purpose: To compare the quantitative and qualitative in vivo elasticity features of human breast cancers on shear-wave elastography (SWE) with ex vivo tumor elasticity and histopathologic features. Materials and Methods: This prospective study was conducted with institutional review board approval, and written informed consent was obtained from all patients. The elasticity (Youngs moduli, in kilopascals [kPa]) of 30 invasive breast cancers was measured in vivo using SWE and ex vivo using a customized indentation system at similar levels of precompression. Quantitative SWE values (maximum [Emax], mean, minimum, and standard deviation) and qualitative SWE features (maximum visual color stiffness, homogeneity of elasticity, and the presence of signal void areas) of the breast cancers were analyzed. Indentation tests were performed for a representative 5-mm-thick slab of each surgical specimen, which corresponded to the ultrasound image section. Histopathologic evaluation was performed to assess the relative amount of fibrous components (1: low, 2: moderate, 3: high) and the presence of necrosis in the tumor. The correlation between the quantitative in vivo and ex vivo elasticity values was analyzed using Pearsons correlation coefficient. The quantitative elasticity values were compared between the subgroups according to the histopathologic variables using the Kruskal-Wallis test. Imaging-pathologic correlation was performed for the qualitative in vivo elasticity features of the SWE images. Results: Histologic type of the 30 invasive breast cancers (mean size, 22.8 mm

range, 11-43 mm) was ductal carcinoma, not otherwise specified type. Breast cancers with higher fibrous component scores were stiffer on ex vivo elasticity measurements (median of ex vivo Emax, 20.5 kPa for score 1, 118.0 kPa for score 2, and 200.3 kPa for score 3) (P=.001). The quantitative in vivo elasticity value exhibited a moderate level of correlation with the ex vivo elasticity of the breast cancers (Emax at no precompression, r=0.552, P=.002

Emax at minimal precompression, r=0.530, P=.003). The qualitative SWE feature of signal void area was present in 43% (13 of 30) of the breast cancers. Of the 13 cancers with signal void areas on SWE, 2 cancers had a gross necrosis in the area of signal void (median of ex vivo Emax, 62.6 kPa). The other 11 cancers without necrosis had dense collagen deposition in the tumor and showed high ex vivo tumor elasticity (median of ex vivo Emax, 279.3 kPa). Seventeen cancers without signal void area had no gross necrosis and showed low ex vivo tumor elasticity (median of ex vivo Emax, 45.3 kPa) (P<.001). Conclusion: Quantitative in vivo elasticity measurement with SWE showed a moderate level of correlation with ex vivo tumor elasticity in the human breast cancers. Qualitative SWE feature of signal void area was related to intratumoral necrosis or substantial stiffness and might potentially be used as an indicative finding of malignancy.