MR imaging and proton spectroscopy of the breast: how to select the images useful to convey the diagnostic message
MR imaging and proton spectroscopy of the breast: how to select the images useful to convey the diagnostic message
复制标题
DOI：
10.1007/s11547-007-0193-x
复制
发表时间：
2007-10-01
影响因子：
8.9
通讯作者：
Sardanelli, F.
中科院分区：
文献类型：
--
作者：
Fausto, A.;Magaldi, A.;Sardanelli, F.
目的。本研究的目的是提出一种简短的方法来总结乳房磁共振（MR）检查，包括预对比和增强对比的动态研究和质子谱（H-1-MRS），以传达诊断信息。材料和方法。在米兰大学圣多纳托Policlinico放射科，乳房MR通常在1.5 T下进行：36层轴向2D短时间反转恢复（STIR）序列；快速静脉注射0.1 mmol/kg Gd-DOTA前后（1个造影前和4个造影后）128分区三维梯度回声冠状序列（1 mm(3)各向同性体素）。后处理包括时间减法（对比后减去对比前），最大强度投影（MIPs），小区域的百分比增强-时间曲线，以及轴向和/或矢状面多平面重建。获得单体素(HMRS)-H-1来表征局灶性病变。应用此协议，每次检查生成1200多张图像。我们只选择了四个早期减动态期的mip：一个轴向类似于颅侧x线乳房x线片，一个冠状和两个侧位类似于侧位90度x线乳房x线片。对于报告中描述的每个病变，我们选择了五个项目，包括三张图像，一张图表和一张表格：STIR图像，预对比和减对比后图像（形态学），增强百分比-时间曲线和生成曲线的原始数据表（动力学）。如果已经进行了H-1-MRS，我们添加其他五个项目：两个后处理光谱（代谢）和三个定位感兴趣体积的图像。只有选定的项目才会印在胶片上并附在报告上。选择的项目通常从4个（未检测到病变）到14个（一个病变，同时研究H-1-MRS），到34个（5个病变，其中一个也研究H-1-MRS）。与生成的图像总数相比，报告中出现的项目百分比分别为0.33%（4/ 1200）、1.17%（14/ 1200）和2.83%（34/ 1200）。乳腺MR成像和H-1-MRS可以有效地总结，仅呈现所有生成图像的最小部分。
Purpose. The purpose of this study was to propose a short way to summarise a breast magnetic resonance (MR) examination including a precontrast and contrast-enhanced dynamic study and proton spectroscopy (H-1-MRS) in order to convey the diagnostic message.Materials and methods. At the Department of Radiology of the Policlinico San Donato (University of Milan), breast MR is routinely performed at 1.5 T as follows: 36-slice axial 2D short-time inversion-recovery (STIR) sequence; 128-partition 3D gradient-echo coronal sequence (1-mm(3) isotropic voxel) before and after rapid automatic intravenous injection of 0.1 mmol/kg of Gd-DOTA (one precontrast and four postcontrast phases). Postprocessing includes temporal subtraction (postcontrast minus precontrast), maximum intensity projections (MIPs), percent enhancement-to-time curves for small regions of interest, and axial and/or sagittal multiplanar reconstructions. Single-voxel (HMRS)-H-1 is acquired to characterise focal lesions. Applying this protocol, more than 1,200 images are generated for each examination. We select only four MIPs of an early subtracted dynamic phase: one axial similar to craniocaudal X-ray mammographic views, one coronal, and two lateral similar to lateral 90 degrees X-ray mammographic views. For each lesion described in the report, we select five items, including three images, one graph, and one table: STIR image, precontrast and subtracted postcontrast images (morphology), percent enhancement-to-time curves and a table of raw data generating the curves (dynamics). If H-1-MRS has been performed, we add other five items: two postprocessed spectra (metabolism) and three images localising the volume of interest. Only the selected items are printed on films and attached to the report.Results. The selected items range usually from four (no detected lesion) to 14 (one lesion, studied also with H-1-MRS), to 34 (five lesions, one of them studied also with H-1-MRS). The percentage of items presented with the report if compared with the total number of generated images is equal to 0.33% (4/1,200), 1.17% (14/1,200), and 2.83% (34/1,200), respectively.Conclusions. Breast MR imaging and H-1-MRS can be effectively summarised presenting only a minimal fraction of all generated images.