Relationship Among Retention Index With Clinical Factors And Lymph Nodes Metastasis For Dual-phase¹⁸F-FDG Imaging In NSCLC

Objective： To investigate the relationship among primary lesion RetentionIndex（RI）with tumor size（long diameter，short diameter），pathologic type，degree ofdifferentiation，hilar and/or mediastinal lymph nodes metastasis for dual-phase¹⁸F-FDGimaging in Non-Small Cell Lung Cancer（NSCLC）.Materials and methods：Retrospective analysis from Jun2010to Sep2012,78patients underwent dual-phase PET/CT in the nuclear medicine department of The FirstAffiliated Hospital of Dalian Medical University were suspected lung cancer beforeexamination，without any anti-tumor therapy or nvasive examination，then wereconfirmed NSCLC by surgical resection，a total of78primary lung lesions were found，all patients underwent lung cancer resection and regional lymph nodes dissection，58cases of adenocarcinoma，15cases of squamous cancer and5cases of adenosquamouscarcinoma were pathologically confirmed；There were21cases of well-differentiated，6cases of high-moderately differentiated，12cases of moderately differentiated，9casesof moderate-poorly differentiated and10cases of poorly differentiated in58cases ofadenocarcinoma；13cases with lymph nodes metastasis and45cases without metastasisin adenocarcinoma；8cases with lymph nodes metastasis and7cases without metastasisin squamous cell carcinoma. All patients were taken early whole body PET/CTscanning after injection of¹⁸F-FDG60min，delay chest scanning and routine chestbreath-hold CT scanning in120-150min. Sketched3D region of interest （ROI）according original lesion location in Siemens MMWP workstation.Early imagingmaximum standardized uptake value （SUVmaxE） and Delay imaging maximumstandard uptake value （SUVmaxD） of primary tumor were calculated by computerautomaticly，RI=（SUVmaxD-SUVmaxE）/SUVmaxE×100%；Long diameter andshort diameter were measured on maximum cross-section of primary tumor in breath-hold CT scanning of lung window. Long diameter,short diameter and RI werenonnormal distribution， all represented by median and four percentile of M（Q1-Q3）.Correlation between RI with long diameter and short diameter of primarylesion，and RI with cell differentiation degree of adenocarcinoma used spearmancorrelation analysis；used rank-sum test of non parametric test to analysis difference ofprimary lesion RI in adenocarcinoma,squamous cell carcinoma and adenosquamouscarcinoma，and differences of primary lesion RI with or without lymph nodes metastasisgroups in adenocarcinoma and squamous cell carcinoma either entirety or separately.Results：1.Primary lesion RI was25.00（15.38-31.09）%，long diameter was2.39（1.70-3.25）cm，and short diameter was1.82（1.30-2.65） cm.Correlation between RI and longdiameter was （r=0.21，P=0.06），correlation between RI and short diameter was （r=0.29，P=0.07），there were no statistically significant.2.Primary lesion RI of adenocarcinoma， squamous cell carcinoma andadenosquamous carcinoma were25.48（15.11-31.05），23.24（15.33-38.71） and31（23.50-47.50） respectively，there was no statistically significant difference among thegroups（X2=7.57，P>0.05）.3.Primary lesion RI of adenocarcinoma well-differentiated,high-moderatelydifferentiated，moderately differentiated，moderate-poorly differentiated and poorlydifferentiated were13.00（5.50-18.00），25.00（22.00-31.25），25.50（21.00-35.50），28.56（25.53-37.02） and33.96（28.43-48.68） respectively，RI and cell differentiation wasnegatively correlated （r=-0.69，P <0.05）.4.When adenocarcinoma and squamous cell carcinoma were entirety analysis，primary lesion RI was31（19.50-35.50） in lymph nodes metastasis group，which washigher than23（13.50-28.00） of non-metastasis group （Z=-2.16，P<0.05）；Primary lesionRI of adenocarcinoma with hilar and/or mediastinal lymph nodes metastasis group washigher than that without metastasis group （Z=-3.17，P<0.05）；RI of squamous cellcarcinoma with lymph nodes metastasis group was24.00（-7.00-31.00），and withoutlymph nodes metastasis group was23.00（20.25-39.00）， the difference was notstatistically significant between two groups （Z=-0.58，P>0.05）.Conclusion：1.No obvious correlation was found between the primary lesion and lesion size isRI.in other words，tumor size had little effect on RI.2.Primary lesion RI was not significant difference among different types of pathology（adenocarcinoma，squamous cell carcinoma，adenosquamous carcinoma） inNSCLC.in other words，pathology types had little effect on RI.3.Higher RI of primary lesion，less differentiated in adenocarcinoma，lower RI ofprimary lesion，higher differentiated in adenocarcinoma.4.NSCLC （adenocarcinoma and squamous cell carcinoma） primary lesions RI wascorrelate with hilar and/or mediastinal lymph nodes metastasis.Higher RI of primarylesion，higher occurrence of hilar and/or mediastinal lymph nodes metastasis；Higher RIof primary lesion，higher occurrence of hilar and/or mediastinal lymph nodes metastasisin adenocarcinoma；But primary lesion RI was not correlate with metastasis insquamous cell carcinoma.