Part 1 Consistency evaluation of two regions of interest methods in analyzing blood oxygenation level dependent functional MRI of nasopharyngeal carcinomaObjective:To investigate the consistency evaluation of two regions of interest(ROI)methods in analyzing BOLD-fMRI of nasopharyngeal carcinoma(NPC).Material and Methods:forty-one cases with NPC confirmed by pathological biopsy underwent MRI examination with both conventional sequence and BOLD-fMRI before any treatment.The parameter map(T2*map)was generated automatically upon the BOLD-fMRI imaging complete.The T2*value of tumor was measured using two ROI place methods,including the contour method(T2*(contour))and the random method(T2*(random)),respectively.The difference between T2*(contour) and T2*(random) was evaluated using a rank sum test.The correlation between them was evaluated using Spearman correlation,being strongly correlated when 0.8≤|r|.The consistency evaluation between them was assessed using Bland-Altman plot and Passing&Bablok regression.Results:There was not a significant difference between T2*(contour) and T2*(random)(z=-0.026,P=0.919).T2*(contour) showed a linear correlation with T2*(random)(r=0.933,P<0.001).There was not a systematic difference,proportion difference,and random difference between T2*(contour) and T2*(random).Their consistency between two ROI methods was at a good level.Conclusion:The consistency of the two ROI methods for analyzing BOLD-fMRI was at a good level.There were two alternative ROI methods,the contour method or random method.Part 2 Value of blood oxygenation level dependent functional MRI for the evaluation of short-term chemoradiotherapy efficacy in nasopharyngeal carcinomaObjective:To investigate the feasibility of blood oxygenation level dependent functional MRI(BOLD-fMRI)for evaluating short-term therapeutic effect of chemoradiotherapy in nasopharyngeal carcinoma(NPC).Materials and Methods:41 cases with NPC confirmed by pathological biopsy exposed to the chemoradiotherapy being the induction chemotherapy(IC)followed by the concurrent chemoradiotherapy(CCRT)and adjuvant chemotherapy(AC)therapy.The conventional MRI was performed pre-IC and after CCRT+AC.They were used to measure the maximal length of the primary and metastatic tumors,and the minimum length of the pathological lymph node,and then the regression rate of lesions was calculated for each patient.BOLD-fMRI was performed pre-and post-IC,then the corresponding T2*value of tumor was measured(T2*(baseline) and T2*(IC)),respectively.All date was measured by two observers,and the consistency evaluation between them was assessed using intra-class correlation coefficient,being good consistency when ICC>0.75.According to the response evaluation criteria in solid tumors(Version 1.1),the patients were divided into the complete response(CR group)and not(non-CR)groups.The consistency evaluation between two observers was assessed using intra-class correlation coefficient,being good consistency when 0.75≤ICC.The T stage,T2*(baseline),and T2*(IC) were compared between the CR and non-CR groups.The correlation of the regression rate of tumor with the T stage,T2*(baseline),and T2*(IC) was tested,respectively.The receiver operating characteristic curve(ROC)was used to compare the capacity of T stage,T2*(baseline),and T2*(IC) in predicting complete response to the chemoradiotherapy for the cases with NPC.Results:T stage,T2*(baseline),and T2*(IC) were statistically different between CR and non-CR groups(P<0.05),and they showed a linear correlation with the regression rate of tumor(r=-0.481,0.748,and 0.617;all P<0.05).For predicting the complete response of the cases with NPC post-chemoradiotherapy,the area under ROC was 0.778,0.903,and 0.763 for T stage,T2*(baseline),and T2*(IC) respectively(all P<0.05).T2*(baseline) was the best one that can predict the complete response for the cases with NPC while it is larger than 37.5 ms(P<0.05).Conclusion:The BOLD-fMRI based T2*(baseline) can predict the short-term chemoradiotherapy efficacy in the patients with NPC,playing a supplementary role to the conventional T stage method. |