Diagnostic value of full-field digital mammography for breast carcinoma / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To evaluate the diagnostic value of full-field digital mammography for breast cancer.</p><p><b>METHODS</b>The clinical data and mammograms of 230 patients with breast diseases between January 2008 and July 2008 were collected and reviewed. Craniocaudal (CC) and mediolateral oblique (MLO) view mammograms were performed in all patients before surgery. Three experienced radiologists in breast imaging assessment analyzed and classified all the mammograms according to breast imaging reporting and data system (BI-RADS). The sensitivity, specificity and accuracy were evaluated according to their pathological diagnosis. The reasons resulting in false-negative and false-positive diagnosis were also analyzed.</p><p><b>RESULTS</b>Of the 238 samples, 130 had a malignant breast tumors and 108 cases of benign breast lesions. One hundred and nine of the 130 malignant breast tumors were invasive ductal carcinoma. Fifty-seven of the 108 benign breast lesions were breast adenosis. Masses or masses with microcalcification were the most frequently seen signs of the malignant tumors, accounting for 40.8% and 20.8%, respectively. The sensitivity, specificity and accuracy of FFDM in detecting breast carcinoma were 90.8%, 87.0% and 89.1%, respectively. The false-negative signs including negative X-ray finding (5 cases) and focal asymmetric densities (4 cases). The false-positive signs were masses with spiculate, indistinctive or lobulated margin leading to misdiagnosing the lesions as malignant tumors.</p><p><b>CONCLUSION</b>Full-field digital mammography (FFDM) is helpful in detection of breast cancers in women, with a higher sensitivity, specificity and accuracy, and has an important clinical application value.</p>

MR imaging features of different renal cell carcinoma subtypes / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To compare and analyze the MRI features of different renal cell carcinoma (RCC) subtypes.</p><p><b>METHODS</b>The MR images of 81 surgically and pathologically confirmed renal cell carcinomas from 79 patients were reviewed retrospectively. The MR imaging features of lesions in plain scan, the degree and patterns of lesion enhancement (homogeneous, heterogeneous, peripheral), and tumor spreading patterns were analyzed. In order to evaluate the diagnostic validity of differentiating RCC subtypes using signal enhancement, receiver operating characteristic curves (ROC) were generated. The cutoff value of post-contrast signal intensity to noise ratios (SNR) of the tumor parenchyma were also generated in order to differentiate clear cell RCC from other subtypes.</p><p><b>RESULTS</b>Of the 81 lesions, 58 were clear cell carcinomas, 10 chromophobe cell carcinomas, 8 papillary cell carcinomas, and 5 unclassified RCC. All the chromophobe cell subtype tumors showed a homogeneous density (P < 0.05). The clear cell subtype tumors were likely heterogenous, and also showed heterogenous enhancement with mixed signal than other subtypes (P < 0.05). The cutoff value of SNR, which was used to differentiate clear cell subtype from the other subtypes, were 616 (corticomedullary phase), 579 (nephrographic phase) and 278 (excretory phase), retrospectively. The nephrographic phase is the most appropriate for differentiation, with a sensitivity of 62.1%, specificity of 91.3%, positive predictive value of 94.7%, negative predictive value of 48.8% and an accuracy value of 70.3%. No significant difference was found in tumor spreading patterns among all subtypes of RCC.</p><p><b>CONCLUSION</b>MR imaging features, particularly tumor heterogeneity and degree of enhancement are useful in differentiation of the renal cell carcinoma subtypes, and in choosing an individualized therapy.</p>