Objective Comparison Of Lesion Detectability In Low And Medium-Energy Collimator Iodine-123 Mibg Images Using A Channelized Hotelling Observer

PHYSICS IN MEDICINE AND BIOLOGY(2017)

引用 5|浏览19
暂无评分
摘要
Iodine-123 mIBG imaging is widely regarded as a gold standard for diagnostic studies of neuroblastoma and adult neuroendocrine cancer although the optimal collimator for tumour imaging remains undetermined. Low-energy (LE) high-resolution (HR) collimators provide superior spatial resolution. However due to septal penetration of high-energy photons these provide poorer contrast than medium-energy (ME) general-purpose (GP) collimators. LEGP collimators improve count sensitivity. The aim of this study was to objectively compare the lesion detection efficiency of each collimator to determine the optimal collimator for diagnostic imaging.The septal penetration and sensitivity of each collimator was assessed. Planar images of the patient abdomen were simulated with static scans of a Liqui-Phil T anthropomorphic phantom with lesion-shaped inserts, acquired with LE and ME collimators on 3 different manufacturers' gamma camera systems (Skylight (Philips), Intevo (Siemens) and Discovery (GE)). Two-hundred normal and 200 single-lesion abnormal images were created for each collimator. A channelized Hotelling observer (CHO) was developed and validated to score the images for the likelihood of an abnormality. The areas under receiver-operator characteristic (ROC) curves, Az, created from the scores were used to quantify lesion detectability. The CHO ROC curves for the LEHR collimators were inferior to the GP curves for all cameras. The LEHR collimators resulted in statistically significantly smaller Azs (p < 0.05), of on average 0.891 +/- 0.004, than for the MEGP collimators, 0.933 +/- 0.004. In conclusion, the reduced background provided by MEGP collimators improved I-123 mIBG image lesion detectability over LEHR collimators that provided better spatial resolution.
更多
查看译文
关键词
iodine-123,neuroblastoma,collimator,channelized Hotelling observer,lesion detectability
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要