The investigation of prognosis prediction in T4 nasopharyngeal cancer with the semiquantitative measurements of FDG PET images

碩士 === 長庚大學 === 醫學物理暨影像科學研究所 === 97 === The aim of this study mainly lay in retrospectively investigating pre-treatment defined parameters for prognosis prediction in 71 T4NxM0 nasopharyngeal cancer(NPC) patients, who received treatment in Linkou Chang Gung Memorial Hospital between December 2001 an...

Full description

Bibliographic Details
Main Authors: Ming Jui Chung, 鍾銘芮
Other Authors: T. C. Yen
Format: Others
Published: 2009
Online Access:http://ndltd.ncl.edu.tw/handle/44329178569104137822
Description
Summary:碩士 === 長庚大學 === 醫學物理暨影像科學研究所 === 97 === The aim of this study mainly lay in retrospectively investigating pre-treatment defined parameters for prognosis prediction in 71 T4NxM0 nasopharyngeal cancer(NPC) patients, who received treatment in Linkou Chang Gung Memorial Hospital between December 2001 and March 2006. Those parameters include GTV and the semiquantitative parameters, such as SUVmax, SUVmean, and total lesion glycolysis(TLG), measured from the initial scanning of 18F-FDG PET before treatment. After 41.6±20 months (2~79 mo) mean duration of follow-up after treatment, we combined those parameters with other clinical risk factors to analyze the overall survival rate(OS), relapse free survival rate(RFS), local-regional relapse free survival rate(LRRFS), local failure free survival rate(LFFS), regional failure free survival rate(RFFS), distant relapse free survival rate(DRFS), disease specific survival rate(DSS). Survival was estimated by the Kaplan-Meier method with log-rank test as the univariate analysis for each parameters with the optimal cutoff respectively derived from receiver operating characteristic(ROC) curve of each survival endpoint. The multivariate analysis using the Cox’s proportional hazards models was made to identify the independent prognosticator with the statistical significance. The result discovered that 5 year OS is 71.8% (the optimal cutoff is SUVmean=4.26 g/ml; HR=4.11, p=0.012), 5 year RFS is 66.2% (TLG=348.32 SUV.ml; HR=3.26, p=0.004), 5 year LRRFS is 76.1% (GTV=59.27 ml; HR=6.20, p=0.001), 5 year LFFS is 78.9% (TLG=348.32 SUV.ml; HR=5.46, p=0.002), 5 year RFFS is 90.1% (GTV=62.28 ml; HR=10.78, p=0.028), 5 year DRFS is 81.7% (only TLG=598.29 SUV.ml has the slightly statistical significance; HR=3.17, p=0.055), 5 year DSS is 85.9% (TLG=348.32 SUV.ml; HR=21.82, p=0.003). TLG provided superior prognostic information compared to traditional risk factors in predicting DSS, RFS, and LFFS rates, which demonstrated that TLG may be more valuable than SUV for predicting the prognosis outcome of the patient before they treated by a certain treatment. To improve patient's life quality and survival rate, it truly still need to further study the performance of TLG in other cancer types in the future, to assist the oncologists work out more appropriate treatment planning.