Hosch Waldemar, Junghanss Thomas, Stojkovic Marija, Brunetti Enrico, Heye Tobias, Kauffmann Günter W, Hull William E
Department of Radiodiagnostics, University Hospital, Heidelberg, Germany.
NMR Biomed. 2008 Aug;21(7):734-54. doi: 10.1002/nbm.1252.
Cystic echinococcosis is a worldwide disease caused by larval stages of the parasite Echinococcus granulosus (canine tapeworm). In clinical practice, staging of cyst development by ultrasonography (US) has allowed treatment options to be tailored to individual patient needs. However, the empirical correlation between cyst morphology and parasite viability is not always dependable and has, until now, required confirmation by invasive assessment of cyst content by light microscopy (LM), for example. Alternatively, high-field 1H MRS may be used to examine cyst fluid ex vivo and prepare detailed quantitative metabolite profiles, enabling a multivariate metabolomics approach to cyst staging. One-dimensional and two-dimensional 1H and 1H/13C MRS at 600 MHz (14.1 T) was used to analyze 50 cyst aspirates of various US and LM classes. MR parameters and concentrations relative to internal valine were determined for 44 metabolites and four substance classes. The high concentrations of succinate, fumarate, malate, acetate, alanine, and lactate found in earlier studies of viable cysts were confirmed, and additional metabolites such as myo-inositol, sorbitol, 1,5-anhydro-D-glucitol, betaine, and 2-hydroxyisovalerate were identified. Data analysis and cyst classification were performed using univariate (succinate), bivariate (succinate vs fumarate), and multivariate partial least squares discriminant analysis (PSL-DA) methods (with up to 48 metabolite variables). Metabolic classification of 23 viable and 18 nonviable cysts on the basis of succinate alone agreed with LM results. However, for seven samples, LM and MRS gave opposing results. Reclassification of these samples and two unclassified samples by PLS-DA prediction techniques led to a set of 50 samples that could be completely separated into viable and nonviable MRS classes with no overlap, using as few as nine variables: succinate, formate, malate, 2-hydroxyisovalerate, acetate, total protein content, 1,5-anhydro-D-glucitol, alanine, and betaine. Thus, future noninvasive in vivo applications of MRS would appear promising.
囊型包虫病是一种由寄生虫细粒棘球绦虫（犬绦虫）的幼虫阶段引起的全球性疾病。在临床实践中，通过超声检查（US）对囊肿发育进行分期，使得治疗方案能够根据个体患者的需求进行调整。然而，囊肿形态与寄生虫活力之间的经验性关联并不总是可靠的，到目前为止，例如需要通过光学显微镜（LM）对囊肿内容物进行侵入性评估来加以证实。或者，高场1H磁共振波谱（MRS）可用于离体检查囊肿液，并制备详细的定量代谢物谱，从而实现对囊肿分期的多变量代谢组学方法。使用600 MHz（14.1 T）的一维和二维1H以及1H/13C MRS来分析50份不同US和LM类别的囊肿抽吸物。测定了44种代谢物和四类物质相对于内部缬氨酸的MR参数和浓度。早期对存活囊肿研究中发现的琥珀酸盐、富马酸盐、苹果酸盐、乙酸盐、丙氨酸和乳酸盐的高浓度得到了证实，并且还鉴定出了其他代谢物，如肌醇、山梨醇、1,5-脱水-D-葡萄糖醇、甜菜碱和2-羟基异戊酸。使用单变量（琥珀酸盐）、双变量（琥珀酸盐与富马酸盐）和多变量偏最小二乘判别分析（PSL-DA）方法（最多使用48个代谢物变量）进行数据分析和囊肿分类。仅基于琥珀酸盐对23个存活囊肿和18个非存活囊肿进行的代谢分类与LM结果一致。然而，对于7个样本，LM和MRS给出了相反的结果。通过PLS-DA预测技术对这些样本和2个未分类样本进行重新分类，得到了一组50个样本，这些样本可以使用低至9个变量：琥珀酸盐、甲酸盐、苹果酸盐、2-羟基异戊酸、乙酸盐、总蛋白含量、1,5-脱水-D-葡萄糖醇、丙氨酸和甜菜碱，完全分离为无重叠的存活和非存活MRS类别。因此，未来MRS的非侵入性体内应用似乎很有前景。
