Elastography: Theory into Clinical Practice

Abstract

To the Editor:Ultrasonography has increasingly been consideredas a useful tool for the diagnosis and management ofbreast lesions (1). With the introduction of highfre-quency digital transducers ranging from 7 to 14 mHz,it is now possible to assess the morphology of theselesions and to categorize them according to the criteria suggested by the most recent BIRADS lexicon (2).However, the study of undetermined masses remains agreat challenge as these lesions present benign sonographic features but no additional data that allowthem to be classified as benign (3). Trials are beingconducted using complementary diagnostic methodsthat allow the categorization of theses masses asbenign, probably benign or suspicious for malignancyand in this context elastography is a promising devel-opment (47).The ultrasound elastography study represents anextension of the most ancient tool used in medicine,i.e., palpation, where the physician assesses the shapeand rigidity of the target organ (8). When externalpressure is applied (stress) onto a target area, thispressure causes tissue deformation (strain). Using specific softwares, tissue deformation is expressed ascolor variation on the ultrasound and may be n usefultool, particularly for the study of breast lesions basedon the assumption that malignant lesions tend to be510 times more rigid than benign (47).This technique was first described by JonathanOphir et al. (4) during the 1990s. Currently, there aretwo main research lines to determine ultrasound elas-tography clinical applicability. The first line ofresearch is based on the assessment of the mass sizebefore and after compressing the target areas with theuse of a software that makes soft lesions appearlighter and rigid lesions appear darker, and in whichmalignant lesions tend to appear more evident thanthe benign lesions (5,6). The other research line isbased on the use of a software that applies a differentcolor spectrum to tissues according to their rigidity,ranging from red to soft tissues, green to intermediatetissues, and dark blue to rigid tissues (Figs. 1 and 2)(7). There is no consensus as to which is the best technique or a classification defining its clinical application, which main limitations are the interobservervariability described in previous studies and how compression should be performed for the study.In a recent study conducted at Santa Casa deMiserico´rdia de Sa˜o Paulo and Centro de TomografiaComputadorizada (CTCGeˆnese) using the ultrasoundsystem Sonix SP (Ultrasonix Medical Corporation,Vancouver, BC, Canada) with multifrequency linearprobes of 514 MHz, and a special software designedfor the device, 256 lesions were evaluated based onthe color spectrum variability. Lesions were classified,using a 4point scoring system developed by theauthors, in two phases: during compression and afterdecompression (almost static compressions). From 256lesions, 215 were benign and 41 malignant; the positive predictive value was 81.8%, the negative predictive value was 97.17%, sensibility was 85.71%,specificity was 96.26%, and the diagnostic accuracywas 94.53%. When compared with conventionalultrasound, a significant improvement in the ability ofobservers to distinguish benign from malignant lesionswas observed (Fig. 3). Through the comparison ofimages during compression and after decompression,it is possible to create parameters to compare thebehavior of target lesions with the adjacent breast tis-sue. The variation in lesion size seen during the studywas not considered a good diagnostic criterion as themargins were not well defined during the dynamicstudy and this was believed to have contributed forthe interobserver variability described in previousseries.The main application of elastography is in the complementary study of undetermined masses, usuallyclassified as BIRADS 3, which mainly comprise complicated cysts, fibroadenomas, and carcinomas withatypical appearance. These lesions are often source ofanxiety, stress, and fear for patients who frequentlyend up choosing to have a biopsy to histologically