Application Of3.0T MR In Aggressiveness Of Endometrial Cancer And Prostate Cancer

Part1. Can Magnetic Resonance Spectroscopy Differentiate Endometrial cancer?Objectives:To investigate whether the choline-containing compounds (Cho) obtained from three-dimension1H magnetic resonance (MR) spectroscopy can differentiate endometrial cancer (ECa) from benign lesions in endometria or in submucosa (BLs-ESm) and is associated with the aggressiveness of ECa.Methods:Thirty-eight patients with ECa and19patients with benign lesions in endometria or in submucosa (BLs-ESm) confirmed by hysterectomy or lesions resection were included. They underwent preoperative multi-voxel1H MR spectroscopy (MRS) with a3.0-T system. The MRS was performed with3D CSI techniques based on PRESS with sufficient lipid. The average number of acquisition was6for the water suppression spectra and2for the unsuppressed water spectra. The voxels of interest with size of7×7×7mm3were selected in the lesions of solid components, avoiding cystic or necrotic areas. All the metabolite quantization was performed within jMRUI v.4.0software. The ratio of sum of Cho to sum of water (Cho/water) was the statistical unit.Cho/water=(∑Choi)/(∑wateri), where, n was the total number of included voxels for a patient.If the lesion was ECa, the tumor stage, grade, type (type Ⅰ:estrogen-dependent; type Ⅱ:estrogen-independent) and size (maximum diameter) were determined by a pathologist. The coefficient of variation (CV) was used to describe the variability of Cho/water in one lesion. where, SD(Cho/waterall vox) was the Standard Deviation (SD) of Cho/water of all voxels in one lesion; Mean(Cho/waterall vox) was the Mean of Cho/water of all voxels in one lesion.The mean Cho/water was compared using independent sample t-test between ECa and BLs-ESm, as well as between type Ⅰ ECa and type Ⅱ ECa. The receiver operating characteristic (ROC) curve analyses was used to determine an optimal threshold to distinguish between ECa and BLs-ESm, as well as between type Ⅰ ECa and type Ⅱ ECa. Comparations of mean Cho/water among different stages ECa were performed using one-way analysis of variance (ANOVA). The Tukey test was used in post-hoc multiple comparisons. The relationship between stage and mean Cho/water was analyzed with Spearman correlation analysis. The different grades tumors were analyzed using the same statistical methods with different stages tumors. The relationship between mean Cho/water and tumor size was analyzed with Pearson correlation analysis. The same statistical analyses were performed for CV. P<0.05was considered to have a significant difference. Statistical analysis was performed with SPSS for Windows, version17.0.Results:The mean Cho/water (±SD) for ECa [(3.02±1.43) x10-3] was significantly higher than that for BLs-ESm [(1.68±0.33)×10-3](P<0.001). The area under the ROC curve (AUC) to differentiate ECa from BLs-ESm was0.817; and the Yonden index was0.711. When the Cho/water threshold to differentiate ECa from BLs-ESm was2.18x10-3, the sensitivity and specificity were0.711and1, respectively.The mean Cho/water for type Ⅱ ECa (estrogen-independent)[(4.42±1.53)×10-3] was significantly higher than that for type Ⅰ ECa (estrogen-dependent)[(2.65±1.17)×10-3](P=0.001). The AUC to differentiate type Ⅰ from type Ⅱ ECa was0.833; and the Yonden index was0.517. When the Cho/water threshold to differentiate type Ⅰ from type Ⅱ ECa was3.41×10-3, the sensitivity and specificity were0.75and0.767, respectively.The mean Cho/water was (2.57±1.37)×10"3for FIGO Ia ECa;(3.02±0.87)×10-3for FIGO Ib ECa;(4.31±2.13) x10-3for FIGO Ⅱ ECa;(3.52±1.36)×10-3for FIGO Ⅲ ECa. There were no significant differences among different stages tumors (P=0.107), but the Cho/water was positively correlated with the tumor stage (the Spearman coefficient r=0.386, P=0.017). However, the mean Cho/water of local invasive lesions (FIGO Ⅱ and Ⅲ,(3.81±1.62) x10-3) was significantly higher than localized lesions (FIGO Ⅰ,(2.70±1.24) x10-3)(P=0.029).The mean Cho/water was (2.43±1.52)×10-3for G1ECa;(3.46±1.35)×10-3for G2ECa;(2.85±1.29)×10-3for G3ECa. There were no significant differences among different grades tumors (P=0.142); and there was no significant correlation between Cho/water and tumor grade (the Spearman coefficient r=0.235, P=0.156). The Cho/water was positively correlated with tumor size (Pearson coefficient r=0.333, P=0.041).The mean CV (±SD) for ECa [(33.98±13.02)%] was significantly higher than that for BLs-ESm [(26.65±11.83)%](P=0.039). The mean CV for type Ⅱ ECa [(39.39±10.11)%] was higher than that for type Ⅰ ECa [(32.54±13.46)%](P=0.19), but the difference was not significantly. The mean CV was (27.14±12.33)%for la ECa;(40.31±8.61)%for Ib ECa;(41.75±10.47)%for Ⅱ ECa;(40.90±12.23)%for Ⅲ ECa (P=0.009); there were significant differences only between la and Ib ECa (P=0.048), as well as between la and Ⅲ ECa (P=0.049). The CV was positively correlated with tumor stage (Spearman coefficient r=0.537, P=0.001). The mean CV was (27.96±13.24)%for G1ECa;(35.40±11.47)%for G2ECa;(40.46±14.17)%for G3ECa; there were no significant differences among different grades tumors (P=0.102). There was no significant correlation between the CV and tumor grade (Spearman coefficient r=0.29, P=0.077). The CV was positively correlated with tumor size (Pearson coefficient r=0.34, P=0.037).Conclusions:the Cho/water obtained from MRS can differentiate ECa from BLs-ESm and differentiate type Ⅱ ECa from type Ⅰ ECa. Cho/water can distinguish between localized ECa (FIGO Ⅰ) and local invasive ECa (FIGO Ⅱ+Ⅲ). There were no significant differences among different stages ECa, as well as among different grades ECa. However, Cho/water of ECa increased with the increase of tumor stage and tumor size. The inhomogeneity of Cho/water in ECa was greater than that in BLs-ESm. The Cho/water inhomogeneity of Ib and Ⅲ ECa was significantly greater than that of la ECa. The Cho/water inhomogeneity of ECa increased with the increase of tumor stage and size. There is no relationship between the Cho/water inhomogeneity and tumor grades. Part2. Diffusion-Weighted Imaging of Prostate Cancer on3.0T MR: Relationship between Apparent Diffusion Coefficient Values and Ki-67ExpressionObjectives:To investigate the relationship between apparent diffusion coefficient (ADC) values and the Ki-67staining index (Ki-67SI), a tumor proliferation marker, in prostate cancer (PCa).Methods:Forty-three patients of PCa and thirty-six patients of benign prostatic hyperplasia (BPH) confirmed by transrectal ultrasound (TRUS)-guided biopsy and/or radical prostatectomy (RP) underwent MRI. The median age of patients was70years (range,51-83years). The PSA ranged from1.2to100ng/ml (median,42.63ng/ml). The Gleason score ranged from5to9. Axial free-breathing DW imaging was performed with single-shot echo-planar imaging technique with the following parameters:b values,0and800s/mm2; TR/TE,6200/93ms; matrix,160×120. The center of slice groups, section thickness and FOV were same to those for axial T2W images. Two radiologists determined the PCa suspicious regions in consensus using the combined information of multi-parametric MRI. The average number of biopsy was approximate13(range,10-16) in this study. To achieve a good spatial consistency between TRUS scans and MR sections, two criteria were followed for the biopsy. First, ultrasound axial and sagittal scans should generate the gland morphology similar to that obtained with the axial and sagittal T2W images, with some anatomic structures (i.e. seminal vesicles, veromontanum, urethra, and some hyperplasia-nodules) as landmarks. Second, the biopsy cores were labeled to specify the location of the biopsy and marked on the T2W images at the corresponding sites. Pathological analysis was performed for all samples. Each patient of PCa was assigned a Gleason score by a pathologist.The IHC staining was done by a pathologist with Ki-67mouse monoclonal antibody (MIB-1). Briefly, Paraffin sections (4μm) were placed in the oven (60℃) overnight. All sections were deparaffinized in xylene and rehydrated in graded alcohol. Endogenous peroxidase was blocked. Antigen retrieval was performed. Add primary antibody and secondary antibody. Diaminobenzidine was used as the final chromogen, and the slides were then counterstained with hematoxylin. Ki-67was considered positive when the nuclei of cells stained brown.A lesion contained more than one ROI when a tumor or a BPH nodule was bigger. The ROIs of normal prostate tissue proven by pathology were drawn in PZ of BPH patients. The mean ADC value of a lesion was calculated by averaged the ADC values of all voxels in all ROIs of a lesion, mean ADC=(∑ADCiSi)/∑Si, where, n was the number of ROIs for a lesion; ADCi was the mean ADC of the ith ROI; Si was the area of the ith ROI. Because the areas of ROIs in a lesion often differed greatly, the area Si was as the weight of the ADCi; the equation1can calculate the mean ADC accurately.The ADC values of PCa were dichotomized for the analyses as either Ki-67SI≤3.5%versus>3.5%or Ki-67SI≤7.1%versus>7.1%. The reasons were as follows: using the3.5%cut-point Ki-67SI was independently prognostic for5-year risk of distant metastases; the7.1%cut-point was associated with distant metastases and disease-specific survival. The ADC values of PCa with different level of Ki-67SI, as well as the Ki-67SI of PCa and BPH, were compared with independent-sample t-test. The ADC values of PCa, BPH and normal prostate tissue were compared with one-way ANOVA. The Turkey test was used in post hoc multiple comparisons. The Pearson correlation test was used to correlate between any two variables of the ADC values, Gleason scores and Ki-67SI. For all statistical analyses, P<0.05was considered to have a significant difference. Statistical analysis was performed with SPSS for Windows, version13.0.Results:The Ki-67SI of PCa (7.23±5.29)%was significantly higher than that of BPH (2.11±1.90)%(P<0.001). The mean (±standard deviation [SD]) Ki-67SI of PCa (7.23±5.29)%was higher than that of BPH (2.11±1.90)%(P<0.001). The mean (±SD) ADC value (10-3mm2/s) of PCa (0.850±0.155) was lower than that of BPH (1.173±0.245)(P<0.001). The ADC values of PCa were negatively correlated with the Ki-67SI (r=-0.459, P=0.002). The mean ADC values of PCa with Ki-67>3.5%and≤3.5%were (0.803±0.094) and (0.936±0.208), respectively. The former was significantly lower than the latter (P=0.031). The ADC values of PCa with Ki-67>7.1%and≤7.1%were (0.779±0.081) and (0.906±0.178), respectively. The former was significantly lower than the latter (P=0.004).Conclusions:The ADC values can reflect the proliferative activity of PCa and BPH, and can differentiate the two kinds of lesions. The ADC values were negatively correlated with the Ki-67SI, as well as the Gleason score, of PCa. Therefore, the DW imaging may be used to predict the proliferative activity and differentiated degree of PCa.