Prediction of Pathological and Radiological Nature of Glioma by Mass Spectrometry Combined With Machine Learning
Prediction of Pathological and Radiological Nature of Glioma by Mass Spectrometry Combined With Machine Learning
复制标题
质谱结合机器学习预测胶质瘤的病理和放射学性质
DOI：
10.1093/neuopn/okaa026
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Kinouchi Hiroyuki
中科院分区：
文献类型：
--
作者：
Suzuki Keiko;Yoshimura Kentaro;Kawataki Tomoyuki;Hanihara Mitsuto;Takeda Sen;Kinouchi Hiroyuki
背景我们之前开发了一种采用质谱和机器学习的医疗诊断流程。它不注释癌症特有的分子标志物，而是使用完整的质谱来预测胶质瘤的特性。目的通过简单的样品制备程序验证我们的诊断方法在预测胶质瘤的病理和放射学特性方面的能力。方法共有10例接受手术切除的胶质瘤患者和4例非胶质瘤患者入组。从 88 个样本中总共获取了 1020 个质谱。为了检验我们开发的诊断流程的预测能力，我们对病理和放射学结果进行了10倍交叉验证，并计算了与病理诊断（世界卫生组织[WHO]分级、MIB-1标记指数[LI]、异柠檬酸脱氢酶[IDH]-1基因突变、5-氨基乙酰丙酸[5-ALA]荧光阳性）和放射学信息（钆）等常规方法的一致率。液体衰减反转恢复[FLAIR]成像的[Gd]增强区和高强度区）。结果WHO恶性分级的预测准确率为91.37%。 MIB-1 LI≥10%和IDH-1突变阳性者分别为82.84%和87.75%。我们的方法对 5-ALA 阳性病变的准确预测率为 95.00%。本方法对 FLAIR 高信号区域的预测准确率为 82.36%，对 Gd 增强区域的预测准确率为 81.27%。结论我们的方法在病理学和放射学方面实现了更高的胶质瘤预测率。正在进行研究开发一个验证队列来验证神经胶质瘤标本的生物学特征。
BACKGROUNDWe have previously developed a medical diagnostic pipeline that employs mass spectrometry and machine learning. It does not annotate molecular markers that are specific to cancer but uses entire mass spectra for predicting the properties of glioma.OBJECTIVETo validate the power of our diagnostic method in predicting the pathological and radiological properties of glioma with a simple sample preparation procedure.METHODSA total of 10 patients with glioma and 4 nonglioma patients who went through surgical resection were enrolled in our hospital. A total of 1020 mass spectra were acquired from 88 specimens. In order to examine the prediction power of the diagnostic pipeline that we have developed, we performed 10-fold cross-validation for pathological and radiological findings and calculated agreement rates with the conventional methods such as pathological diagnosis (World Health Organization [WHO] grading, MIB-1 labeling index [LI], mutations in the isocitrate dehydrogenase [IDH]-1 gene, and positive 5-aminolevulinic acid [5-ALA] fluorescence) and radiological information (gadolinium [Gd]-enhanced area and high-intensity area on fluid-attenuated inversion recovery [FLAIR] imaging).RESULTSPrediction accuracy for WHO malignant grade was 91.37%. Those for MIB-1 LI≥ 10% and IDH-1 mutation-positive were 82.84% and 87.75%, respectively. Our method achieved an accurate prediction of 95.00% for the 5-ALA-positive lesion. The present method displayed an accuracy of 82.36% in predicting the area of FLAIR hyperintensity and 81.27% for the Gd-enhanced area.CONCLUSIONOur methodology achieved a higher rate of prediction of glioma in terms of pathology and radiology. Research is ongoing to develop a validation cohort to verify the biological profiles of glioma specimens.