• 微信公众号:机器学习炼丹术
  • 笔记:陈亦新
  • 参考论文:Dual-time-point F-18 FDG PET/CT imaging for differentiating the lymph nodes between malignant lymphoma and benign lesions

introduction

Malignent Lymphoma (ML)恶性淋巴瘤是最常见的血液系统恶性肿瘤之一。

PET/CT offers advantages over converntional imaging techniques, because it can demonstrate abnormal netabolic activity and extension of malignant diseases.

PET相比传统的技术更有优势,因为它可以显示异常的代谢活动和恶性疾病的扩展

However, many inflammatory lesions such as pneumonia, sarcoidosis, rheumatoid arthritis, etc., also have elevated F-18 FDG uptake in PET, leading to false-positive results.

然而许多炎症病变,如肺炎、结节病、类风湿性关节炎等,在PET中的f-18 FDG的摄取也会升高,产生假阳性的结果。

本研究的目的就是评估,dual-time-point F-18 FDG PET/CT scan在恶性淋巴瘤初期诊断中的有效性,并分析早期50分钟扫描和100分钟延迟扫描的SUVmax的值。

Patient population

这个入组人群的描述挺好的。我觉得可以学习一下:

Dual-time-point识别恶性淋巴瘤_ML

F-18 FDG PET/CT scans

这个描述机器和扫描的,也可以学习一下:

Dual-time-point识别恶性淋巴瘤_算法_02

Dual-time-point识别恶性淋巴瘤_sed_03

image analysis

这里文章很老是2012年的,采用的分类方式也是four-score visual scale。

这里值得一提的是SUV的计算方式写的很标准:

Dual-time-point识别恶性淋巴瘤_ML_04

除了SUV,还用到了一个指标,RI-SUV, we tried to evaluate an usefulness of difference between early SUV max and delayed SUVmax in this study as a more simple index:

Dual-time-point识别恶性淋巴瘤_sed_05

statistical analysis

A paired 2-tailed t test was used to compare differences bwtween ML and BLN in those indices.

关于统计的部分先留着坑,之后填。