| Objective:To investigate the diagnostic value of diffusion-weighted imaging (DWI) in uterine cervical cancer, and its application value in the differentiation of metastatic lymph nodes from non-metastatic lymph nodes in uterine cervical cancer; to evaluate the feasibility of DWI in predicting and monitoring the early response of uterine cervical cancer to concurrent chemoradiation and choose the optimal time that can detect the tumor response to therapy as early as possible;to analyze the application value of the maximum ADCdifferenee value in evaluating the pathologic grade of uterine cervical cancer. Materials and Methods:One hundred and five patients with uterine cervical cancer were enrolled in this study from October 2006 to January 2010. Uterine cervical cancer was confirmed in all patients,either by surgery (forty-six cases) or by biopsy (fifty-nine cases).The control group was composed of 15 female patients with uterine leiomyomas.All of them received total hysterectomy, normal cervix was confirmed by both negative cytology of cervical mucosa and surgery. Conventional MRI and DWI examinations were performed using a 1.5 T (Twin Excite, GE Healthcare) clinical scanner with diffusion factor of 0 and 1000s/mm2. Fifty-six of them received DWI examinations with diffusion factor of 0 and 100s/mm2,0 and 3000s/mm2.Thirty-three patients who received concurrent chemoradiation underwent conventional MRI and DWI follow up examinations at three days, seven days, fourteen days, one month and two months after the therapy initiated. Ninety-four cases of these histopathologically proven uterine cervical cancers were analyzed for tumor cellularity density. Fifty-six histological specimens were stained with CD34,VEGF. By using an optical microscope, tumor's MVD count and VEGF expression were measured.Parameters evaluated consisted of①ADC values of uterine cervical cancer and normal cervix;②short-axis diameter, long-axis diameter, L/S ratio, S/L ratio, mean ADC value, minimum ADC value and relative ADC (rADC) value of both metastatic and non-metastatic lymph nodes;③the maximum ADCdifference value of tumor on gray-scale ADCdifference map;④the longest tumor diameter and ADC value of tumor before and after therapy. The following data were analyzed:①Independent-Samples T Test was used to compare mean ADC value between uterine cervical cancer and normal cervix, also between squamous carcinoma and adenocarcinoma of the uterine cervix.Receiver operating characteristics (ROC) analysis was performed in order to evaluate the diagnostic performance of ADC value in differentiating uterine cervical cancer from normal cervix and squamous carcinoma from adenocarcinoma as well.From the ROC curves, the optimal cutoff value was extracted;One-Way ANOVA (analysis of variance) was used to compare mean ADC value and tumor cellularity density among three different pathologic grades of uterine cervical cancer. Pearson correlation was performed to determine whether ADC value of uterine cervical cancer was significantly related to tumor cellularity density. In addition, Spearman rank correlation coefficient was calculated to evaluate the correlation between ADC value of uterine cervical cancer and the pathologic grade of tumor;③Independent-Samples T Test and Wilcoxon rank sum test was used to compare each of the size-based criteria (i.e.short-axis diameter, long-axis diameter, L/S ratio and S/L ratio), the signal intensity on T2WI, mean ADC value,minimum ADC value, mean rADC value and minimum rADC value between metastatic lymph nodes and non-metastatic lymph nodes.To evaluate the correlation between the short-and long-axis diameters and ADC value in metastatic and non-metastatic lymph nodes, Pearson and Spearman rank correlation analysis was performed;④A ROC analysis was used to compare the diagnostic performance of each ADC-based criteria on DWI and each of the conventional size-based criteria on T2WI for the differentiation of metastatic lymph nodes from non-metastatic lymph nodes.Thereafter, the area under the curve was evaluated.From the ROC analysis, Youden's index was used in determining the optimal threshold value for differentiating metastatic lymph nodes from non-metastatic lymph nodes;⑤One-Way ANOVA (analysis of variance) was used to compare the differences of the maximum ADCdifference value and MVD count among three different pathologic grades of uterine cervical cancer. Chi-square test was performed to compare the difference of VEGF expression among different pathologic grades.Pearson correlation was used to determine whether the maximum ADCdifference value was significantly related to MVD count. In addition, Spearman rank correlation coefficient was calculated to evaluate the correlation of the maximum ADCdifference value and the pathologic grade of tumor and VEGF expression, and also the correlation between MVD count and the pathologic grade of tumor;⑥Independent-Samples T Test was used to compare pretreatment ADC value, pretreatment tumor diameter, ADC value and percentage ADC change and also the longest tumor diameter and percentage diameter change during therapy between CR and PR group;⑦Pearson correlation was performed to evaluate the correlation between pretreatment ADC value and pretreatment tumor diameter, pretreatment ADC value and percentage size reduction of tumor after two months of therapy, pretreatment tumor diameter and percentage size reduction of tumor after two months of therapy, and between those ADC parameters during therapy and therapeutic efficacy as well;⑧With aggregate analysis of these parameters, an optimal time to detect the tumor response to therapy as early as possible was chosen. A ROC analysis was performed to determine the optimal threshold value.Results:①The mean ADC value of the uterine cervical cancer was lower than that of the normal cervix, and the difference between the two groups was significant (p=0.000).The area under the ROC curve for differentiating uterine cervical cancer from normal cervix was 0.991. From the ROC curve,an ADC of 1.285×10-3mm2/s was determined as a cutoff value for differentiating uterine cervical cancer from normal cervix.With this threshold from ROC curve, the sensitivity and specificity was 97.1% and 100%,respectively. The mean ADC value of squamous carcinoma was significantly lower than that of adenocacinoma (p=0.003).The area under the ROC curve for differentiating squamous carcinoma from adenocarcinoma was 0.809.From the ROC curve, an ADC of 0.966×10-3mm2/s was determined as a cutoff value for differentiating squamous carcinoma from adenocarcinoma. With this threshold from ROC curve, the sensitivity and specificity was 84.1% and 73.3%,respectively;②Tumor cellularity density of different pathologic grades varied significantly (p=0.000), and statistical difference was observed between either two groups (p=0.000, p=0.000, p=0.000).The ADC value of different pathologic grades varied significantly as well (p=0.000),and statistical difference was observed between either two groups (p=0.001,p=0.000, p=0.030).There was a significant positive linear correlation between tumor cellularity density and the pathologic grade of tumor (p=0.000).The ADC value of uterine cervical cancer correlated negatively with cellularity density (p=0.000) and the pathologic grade of tumor (p=0.000);③There were significant differences between metastatic and non-metastatic lymph nodes in the short-axis and long-axis diameter (p=0.000,p=0.000).However, there were no significant differences between metastatic and non-metastatic lymph nodes in the L/S ratio or S/L ratio (p=0.261, p=0.157).The signal intensity of lymph nodes on T2WI was not significantly different between metastatic and non-metastatic lymph nodes (p=0.166), although the former was slightly higher than the latter one.The mean ADC,minimum ADC,mean rADC and minimum rADC value of metastatic lymph nodes was significantly lower than that of non-metastatic lymph nodes (p=0.000, p=0.000,p=0.000,p=0.000).The correlations between short-and long-axis diameters and mean ADC,minimum ADC, mean rADC and minimum rADC both in metastatic and non-metastatic lymph nodes were found to be insignificant (p>0.05);④The short-axis and long-axis diameters, mean ADC, minimum ADC, mean rADC and minimum rADC values were all contributory to differentiate metastatic lymph nodes from non-metastatic lymph nodes.The area under the ROC curve was greater for minimum ADC than that for short-axis diameter, long-axis diameter and other ADC criteria. On the basis of ROC analysis, a lymph node with a minimum ADC value equal or less than 0.983×10-3mm2/s was considered to be a metastatic lymph node. Using the minimum ADC value criteria, the sensitivity and specificity for differentiating metastatic from non-metastatic lymph nodes were 94.6% and 91.8%,respectively;⑤The maximum ADCdifference value of grade one, grade two and grade three was (0.712±0.195)×10-3mm2/s,(1.273±0.162)×10-3mm2/s and (1.748±0.292)×10-3mm2/s, respectively. There was statistical difference among them (p=0.000), and statistical difference was observed between either two groups (p=0.000,p=0.000,p=0.000).The MVD count of different pathologic grades varied significantly as well (p=0.000), and it was (30.83±2.98) for well differentiated, (43.86±3.92) for moderately differentiated and (54.19±5.36) for poorly differentiated uterine cervical cancer. Statistical difference was observed between either two groups (p=0.000,p=0.000,p=0.000). Positive correlation was found between the maximum ADCdifference value and the pathologic grade of tumor and also between MVD count and the pathologic grade of tumor (p= 0.000,p=0.000).No significant difference was obtained between the level of VEGF expression and pathologic grade of tumor (p=0.222).The maximum ADCdifference value correlated positively with both MVD count and the level of VEGF expression (p=0.000,p=0.000);⑥The pretreatment ADC value of CR was significantly lower than that of PR (p=0.007).Although the pretreatment tumor diameter of PR group was larger than that of CR group, there was no significant difference between them (p=0.661).The ADC value for both CR and PR group increased to different extents after the initiation of concurrent chemoradiation. There was no significant difference of ADC value between CR and PR group at either follow up time during therapy (p>0.05).The percentage ADC change of CR at either follow up time was greater than that of PR group, and the differences were significant at the time of fourteen days (p=0.007) and one month (p=0.001).The tumor diameter decreased gradually for both CR and PR group, but there was no significant difference of tumor diameter or percentage diameter change between CR and PR group at either follow up time during therapy (p>0.05);⑦No significant correlation was found between pretreatment ADC and pretreatment tumor diameter (p=0.072). There was a significant and moderate negative linear correlation between the percentage size reduction of tumor after two months of treatment and pretreatment ADC value (p=0.045),while the correlation between the percentage size reduction of tumor after two months of treatment and pretreatment tumor diameter was found to be insignificant (p=0.508).At the time of three days,seven days, fourteen days and one month after the therapy initiated, there was positive correlation between the percentage ADC change and the percentage size reduction of tumor after two months of therapy (p=0.014,p=0.026,p=0.000,p=0.001),especially at the day of fourteen. No significant correlation was found between ADC value at either follow up time and the percentage size reduction of tumor after two months of therapy (p>0.05);⑧The day of fourteen after the therapy initiated can be considered as the optimal time for monitoring the early treatment response of uterine cervical cancer to concurrent chemoradiation. With the percentage ADC change of 35.4% at the time of fourteen days after the therapy initiated, the sensitivity and specificity for determining its treatment response group were 97.3% and 66.4%, respectively. Conclusion:①DWI has a potential ability to differentiate between normal and cancerous tissue in the uterine cervix, and it can indicate the histologic type of uterine cervical cancer as well. The ADC value of uterine cervical cancer represents tumor cellularity density,thus providing a new method for evaluating the pathologic grade of tumor;②The minimum ADC shows superior diagnostic performance and it outperforms all size-based criteria and other ADC criteria,since the Az value of minimum ADC was greater than that of short-axis diameter, long-axis diameter, mean ADC,mean rADC and minimum rADC;③The maximum ADCdifferenee value can represent the grades of differentiation and may provide valuable information of tumor microcirculation and perfusion, thus allowing a promising new method for non-invasively assessing the pathologic grade of tumor, and it can serve as a substitution for assessing tumor angiogenesis;④It is feasible to use DWI to predict and monitor the treatment response in patients with uterine cervical cancer that undergoing concurrent chemoradiation, and the optimal time for monitoring the early treatment response of uterine cervical cancer to concurrent chemoradiation is the day of fourteen after the therapy initiated.
|
PDF Full Text Request