Texture analysis of medical images

The analysis of texture parameters is a useful way of increasing the information obtainable from medical images. It is an ongoing field of research, with applications ranging from the segmentation of specific anatomical structures and the detection of lesions, to differentiation between pathological and healthy tissue in different organs. Texture analysis uses radiological images obtained in routine diagnostic practice, but involves an ensemble of mathematical computations performed with the data contained within the images. In this article we clarify the principles of texture analysis and give examples of its applications, reviewing studies of the technique.     

Differentiation of hepatocellular carcinoma and hepatic metastasis from cysts and hemangiomas with calculated T2 relaxation times and the T1/T2 relaxation times ratio

PURPOSE: To determine the diagnostic capability of the T1 and T2 relaxation times and the T1/T2 relaxation times ratio generated with the mixed turbo spin echo (mixed-TSE) pulse sequence, in order to discriminate between hepatocellular carcinoma (HCC)/metastases and hemangiomas/cysts. MATERIALS AND METHODS: A retrospective review of 36 MR examinations implementing the mixed-TSE pulse sequence demonstrated 70 focal hepatic lesions. Quantitative MR algorithms were used to generate T1 and T2 relaxation times, and the T1/T2 relaxation times ratio for each lesion. A two-sample t-test compared mean T1 and T2 relaxation times, and the T1/T2 relaxation times ratio, by lesion type: carcinoma/metastases and hemangiomas/cysts. Sensitivity and specificity for discriminating carcinoma/metastases from hemangiomas/cysts with T2 relaxation time thresholds of 112 and 125 msec, as well as a ratio of T1/T2 relaxation times of 5.8, were calculated. RESULTS: Using a T2 relaxation time threshold of 112 msec, 92% sensitivity and 100% specificity discriminating cysts/hemangiomas from HCC/liver metastasis was demonstrated. With a threshold of 125 msec, 96% sensitivity and 98% specificity was demonstrated. There was no correlation between calculated T1 relaxation times and type of lesion. Using a T1/T2 relaxation times ratio of 5.8, 100% sensitivity and specificity were demonstrated. CONCLUSION: Although there is high sensitivity and specificity associated with the use of T2 relaxation times alone to discriminate carcinoma/metastases from hemangiomas/cysts, using the T1/T2 relaxation times ratio threshold of 5.8 allowed proper classification of all lesions.     

Gadolinium-enhanced T1-weighted renal and abdominal MR imaging: quantitative discrepancy between clinical and in vitro findings

RATIONALE AND OBJECTIVES: This study was conducted to compare the magnetic resonance (MR) contrast medium enhancement of abdominal organs in vivo with the signal intensity (SI) values of known in vitro gadolinium solutions. MATERIALS AND METHODS: A phantom was imaged with the MR contrast medium gadodiamide (Omniscan; Nycomed, Princeton, NJ) of solutions at full-strength (0.5 mmol/mL), one-third, 1/10, and 1/100 concentrations. A fat-suppressed fast spoiled gradient-echo pulse sequence with flip angles ranging from 10 degrees to 170 degrees (at 20 degrees increments) was performed with a 1.5-T magnet. In 12 subjects, the SIs of abdominal organs were determined with identical imaging parameters, before and after administration of gadodiamide injection at 0.1 mmol/kg. RESULTS: As anticipated, the plot of SI in relation to gadodiamide concentration is nonlinear, with a decrease in SI due to T2 effects at concentrations above 0.05 mmol/mL. The kidney showed the highest SI after gadodiamide enhancement (125.2 +/- 11.6 [standard error] at 2.5 minutes), followed by the liver (76.5 +/- 11.5 at 1 minute) and spleen (57.26 +/- 9.35 at 30 seconds). The SI of the renal medulla (114.2 +/- 9.8 at 4.5 minutes) was approximately one-third that in phantom observations. CONCLUSION: The authors observed a marked discrepancy between empirical contrast medium performance in abdominal organs and SI values for comparable gadodiamide concentrations in vitro. One possible reason is the intracellular compartmentalization of water molecules in vivo. These results suggest a need for a better understanding of MR contrast medium performance in vivo.     

Segmentation of small structures in MR images: Semiautomated tissue hydration measurement

Segmentation of small anatomic structures In noisy magnetic resonance (MR) images is inherently challenging because the edge Information is contained in the same high-frequency image component as the noise. The authors overcame this obstacle in the analysis of the sural nerve in the ankle by processing images to reduce noise and extracting edges with an edge detection algorithm less sensitive to noise. Anatomic accuracy of the segmentation was confirmed by a neuroradiologist. A nerve hydration coefficient was determined from the signal intensity of the nerve in these segmented images. These semiautomated measurements of hydration agreed closely with those obtained with a previously described manual method (n=44, P=.76). Each image in the study was analyzed identically, with no modification of the computer algorithm parameters. The data suggest that this robust method may be useful in a multicenter evaluation of diabetes treatment protocols.     

A texture analysis method for MR images of airway dilator muscles: a feasibility study

Objectives:

Airway dilator muscles play an important role in the analysis of breathing-related symptoms, such as obstructive sleep apnoea. Texture analysis (TA) provides a new non-invasive method for analysing airway dilator muscles. In this study, we propose a TA methodology for airway dilator muscles and prove the robustness of this method.

Methods:

15 orthognathic surgery patients underwent 3-T MRI. Computerized TA was performed on 20 regions of interest (ROIs) in the patients'' airway dilator muscles. 53 texture parameters were calculated for all ROIs. The robustness of the TA method was analysed by altering the locations, sizes and shapes of the ROIs.

Results:

Our study shows that there is significant difference in TA results as the size or shape of ROI changes. The change of location of the ROI inside the studied muscle does not affect the TA results.

Conclusions:

The TA method is valid for airway dilator muscles. We propose a methodology in which the number of co-occurrence parameters is reduced by using mean values from four different directions (0°, 45°, 90° and 135°) with pixel spacing of 1 pixel.     

STIR vs. T1-weighted fat-suppressed gadolinium-enhanced MRI of bone marrow edema of the knee: computer-assisted quantitative comparison and influence of injected contrast media volume and acquisition parameters

PURPOSE: To compare short tau inversion recovery (STIR) and T1-weighted (T1w) gadolinium (Gd)-enhanced fat-suppressed MRI of bone marrow edema (BME) of the knee, and investigate the influence of injected contrast media volume and variation of major acquisition parameters on apparent BME volume and signal contrast. MATERIALS AND METHODS: STIR and T1w Gd-enhanced fat-suppressed images were obtained from 30 patients with BME of the knee. Two groups of patients were examined with different MR scanners, acquisition parameters, and contrast media volumes. For both sequences, BME volume and signal contrast were assessed by computer-assisted quantification, and were compared through their arithmetic means and correlation coefficients (r(2)). The injected contrast media volume was also correlated with BME volume and signal contrast differences between sequences. RESULTS: A strong correlation between the STIR and Gd-enhanced T1w images was found for BME volume (r(2) = 0.96-0.99) and BME signal contrast (r(2) = 0.86-0.94). Despite the differences in MR acquisition parameters and injected contrast media volume, both sequences depicted an almost identical BME volume in both groups. Contrast media volume showed a moderate correlation (r(2) = 0.40) with BME volume differences. CONCLUSION: STIR is the optimum method for determining the size and signal contrast of BME. The injected contrast media volume appears to have only a limited influence on apparent BME volume.     

Cardiac amyloidosis: MR imaging findings and T1 quantification, comparison with control subjects

In cardiac amyloidosis an interstitial deposition of amyloid fibrils causes concentric thickening of the atrial and ventricular walls. We describe the results of tissue characterization of the myocardium by T1 quantification and MRI findings in a patient with cardiac amyloidosis. The T1 time of the myocardium was elevated compared to that in individuals without amyloidosis. The T1 time of the myocardium was 1387 +/- 63 msec (mean value obtained from four measurements +/- standard deviation [SD]) in the patient with cardiac amyloidosis, while the reference value obtained from the myocardium of 10 individuals without known myocardial disease was 1083 +/- 33 msec (mean value +/- SD). In combination with other MR findings suggestive of amyloidosis, such as homogeneous thickening of the ventricular and atrial walls, thickening of the valve leaflets, restrictive filling pattern, and reduction of systolic function, T1 quantification may increase diagnostic confidence.     

Paradoxical signal pattern of mediastinal cysts on T2-weighted MR imaging: phantom and clinical study

Purpose

To evaluate the intracystic MRI (magnetic resonance imaging) signal intensity of mediastinal cystic masses on T2-weighted images.

Materials and methods

A phantom study was performed to evaluate the signal intensity of a mediastinal cystic mass phantom (rubber balloon containing water) adjacent to a cardiac phantom pulsing at the rate of 60/min. T2-weighted images (sequence, fast spin echo [FSE] and single shot fast spin echo [SSFSE]) were acquired for the mediastinal cystic mass phantom. Further, a clinical study was performed in 33 patients (16 men, 17 women; age range, 19-85 years; mean, 65years) with thymic cysts or pericardial cysts. In all patients, T2-weighted images (FSE and SSFSE) were acquired. The signal intensity of cystic lesion was evaluated and was compared with that of muscle. A region of interest (ROI) was positioned on the standard MR console, and signal intensity of the cystic mass (cSI), that of the muscle (mSI), and the rate of absolute value of cSI–mSI to standard deviation (SD) of background noise (|cSI–mSI|/SD = CNR [contrast-to-noise ratio]) were measured.

Results

The phantom study demonstrated that the rate phantom-ROI/saline-ROI was higher in SSFSE (0.36) than in FSE (0.19). In clinical cases, the degree of the signal intensity was higher in SSFSE than in FSE. The CNR was significantly higher in SSFSE (mean ± standard deviation, 111.0 ± 47.6) than in FSE (72.8 ± 36.6) (p < 0.001, Wilcoxon signed-rank test).

Conclusions

Anterior mediastinal cysts often show lower signal intensity than the original signal intensity of water on T2-weighted images. SSFSE sequence reduces this paradoxical signal pattern on T2-weighted images, which may otherwise cause misinterpretation when assessing cystic lesions.     

Preoperative evaluation of gallbladder carcinoma: efficacy of combined use of MR imaging,MR cholangiography,and contrast-enhanced dual-phase three-dimensional MR angiography

PURPOSE: To determine the efficacy of the combined use of magnetic resonance (MR) imaging, MR cholangiography (MRC), and MR angiography (MRA) in the preoperative evaluation of gallbladder carcinoma. MATERIALS AND METHODS: During a 20-month period, 41 patients with proven gallbladder carcinomas were referred for MR examination, including MR imaging, MRC, and gadolinium-enhanced dual-phase MRA to determine the operability of their gallbladder carcinoma. Eighteen patients who underwent surgery within six days of the MR examination were included in this study. All MR images were analyzed in order to assess bile duct invasion, vascular invasion, hepatic invasion or metastasis, lymph node metastasis, and invasion into adjacent organs. RESULTS: Surgical and histopathologic findings revealed hepatic invasion in nine patients, bile duct invasion in nine, vascular invasion in three, and lymph node metastasis in 10. The sensitivity and specificity of MR examination were, respectively, 100% and 89% for bile duct invasion, 100% and 87% for vascular invasion, 67% and 89% for hepatic invasion, and 56% and 89% for lymph node metastasis. CONCLUSION: The "all-in-one" MR protocol, including MR imaging, MRC, and MRA, can be an effective diagnostic method in the preoperative work-up for gallbladder carcinoma.     

Differentiation between cavernous hemangiomas and untreated malignant neoplasms of the liver with free-breathing diffusion-weighted MR imaging: Comparison with T2-weighted fast spin-echo MR imaging

Objective

To test interobserver variability of ADC measurements and compare the diagnostic performances of free-breathing diffusion-weighted (FBDW) with that of T2-weighted FSE (T2WFSE) MR imaging for differentiating between cavernous hemangiomas and untreated malignant hepatic neoplasms.

Materials and methods

Thirty-five patients with cavernous hemangiomas and 35 with untreated hepatic malignant neoplasms had FBDW and T2WFSE MR imaging. Hepatic lesions were characterized with ADC measurement and visual evaluation. Interobserver agreement for ADC measurement was calculated. Association between ADC value and lesion type was assessed using univariate analysis. Sensitivity, specificity and accuracy of ADC values and visual evaluation of MR images for the diagnosis of untreated malignant hepatic neoplasm were compared.

Results

ADC measurements showed excellent interobserver correlation (intraclass correlation coefficient = 0.980). Malignant neoplasms had lower ADC values than hemangiomas for the two observers (1.11 × 10⁻³ mm²/s ± .21 × 10⁻³vs. 1.77 × 10⁻³ mm²/s ± .29 × 10⁻³ for observer 1 and 1.11 × 10⁻³ mm²/s ± .19 × 10⁻³vs. 1.79 × 10⁻³ mm²/s ± .32 × 10⁻³ for observer 2) and univariate analysis found significant correlations between lesion type and ADC values. Depending on ADC threshold value, accuracy for the diagnosis of malignant neoplasm varied from 82.9% to 94.3%. Using visual evaluation, FBDW showed better specificity and accuracy than T2WFSE MR images for the diagnosis of malignant neoplasm (97.1% vs. 77.1% and 94.3% vs. 62.9%, respectively).

Conclusion

FBDW imaging provides reproducible quantitative information and surpasses the value of T2WFSE MR imaging for differentiating between cavernous hemangiomas and untreated malignant hepatic neoplasms.     

A computerized global MR image feature analysis scheme to assist diagnosis of breast cancer: a preliminary assessment

Objectives

To develop a new computer-aided detection scheme to compute a global kinetic image feature from the dynamic contrast enhanced breast magnetic resonance imaging (DCE-MRI) and test the feasibility of using the computerized results for assisting classification between the DCE-MRI examinations associated with malignant and benign tumors.

Materials and Methods

The scheme registers sequential images acquired from each DCE-MRI examination, segments breast areas on all images, searches for a fraction of voxels that have higher contrast enhancement values and computes an average contrast enhancement value of selected voxels. Combination of the maximum contrast enhancement values computed from two post-contrast series in one of two breasts is applied to predict the likelihood of the examination being positive for breast cancer. The scheme performance was evaluated when applying to a retrospectively collected database including 80 malignant and 50 benign cases.

Results

In each of 91% of malignant cases and 66% of benign cases, the average contrast enhancement value computed from the top 0.43% of voxels is higher in the breast depicted suspicious lesions as compared to another negative (lesion-free) breast. In classifying between malignant and benign cases, using the computed image feature achieved an area under a receiver operating characteristic curve of 0.839 with 95% confidence interval of [0.762, 0.898].

Conclusions

We demonstrated that the global contrast enhancement feature of DCE-MRI can be relatively easily and robustly computed without accurate breast tumor detection and segmentation. This global feature provides supplementary information and a higher discriminatory power in assisting diagnosis of breast cancer.     

Usefulness of magnetic resonance imaging for evaluation of cardiovascular invasion: evaluation of sliding motion between thoracic mass and adjacent structures on cine MR images

PURPOSE: To determine the feasibility and usefulness of magnetic resonance imaging (MRI) for evaluating cardiovascular invasion of a thoracic mass by demonstrating the sliding motion between the mass and adjacent structures. MATERIALS AND METHODS: Twenty-six patients (17 males and nine females, mean age = 49 years) were included in this study. They all had thoracic masses with equivocal cardiovascular invasion on chest CT scan and/or MRI that were surgically confirmed. The pathologic diagnoses were teratoma/thymic tumor (N = 12), lung cancer (N = 9), and other thoracic tumor (N = 5). Conventional T1/T2, contrast-enhanced, and breathheld ECG-gated cine MRI using a steady-state free precession (SSFP) technique was performed. The results were compared to the surgical reports. RESULTS: The cine MR images showed the presence of sliding motion in 39 structures in 20 patients, which were surgically confirmed as not being invaded, and 15 structures in six patients with the absence of sliding motion noted as tumor invasion. Therefore, the accuracy of the cine MR images was 94.4% (51/54) for evaluating cardiovascular invasion of a thoracic mass. CONCLUSION: MRI can provide additional information and improve the accuracy of preoperative staging for predicting cardiovascular invasion of a thoracic mass by evaluating the sliding motion.     

Machine learning study of several classifiers trained with texture analysis features to differentiate benign from malignant soft‐tissue tumors in T1‐MRI images

Purpose:

To study, from a machine learning perspective, the performance of several machine learning classifiers that use texture analysis features extracted from soft‐tissue tumors in nonenhanced T1‐MRI images to discriminate between malignant and benign tumors.

Materials and Methods:

Texture analysis features were extracted from the tumor regions from T1‐MRI images of clinically proven cases of 49 malignant and 86 benign soft‐tissue tumors. Three conventional machine learning classifiers were trained and tested. The best classifier was compared to the radiologists by means of the McNemar's statistical test.

Results:

The SVM classifier performs better than the neural network and the C4.5 decision tree based on the analysis of their receiver operating curves (ROC) and cost curves. The classification accuracy of the SVM, which was 93% (91% specificity; 94% sensitivity), was better than the radiologist classification accuracy of 90% (92% specificity; 81% sensitivity).

Conclusion:

Machine learning classifiers trained with texture analysis features are potentially valuable for detecting malignant tumors in T1‐MRI images. Analysis of the learning curves of the classifiers showed that a training data size smaller than 100 T1‐MRI images is sufficient to train a machine learning classifier that performs as well as expert radiologists. J. Magn. Reson. Imaging 2010;31:680–689. © 2010 Wiley‐Liss, Inc.     

Distinguishing hemangiomas from malignant solid hepatic lesions: a comparison of heavily T2-weighted images obtained before and after administration of gadoxetic acid

Purpose:

To compare the use of heavily T2‐weighted images obtained before and after administration of gadoxetic acid in differentiating hemangiomas from malignant solid hepatic lesions.

Materials and Methods:

Heavily T2‐weighted images (TE = 150 msec) were obtained for 70 patients (42 men and 28 women) with 74 focal hepatic lesions (25 hepatocellular carcinomas [HCC], 22 metastases, and 27 hemangiomas) ≤3 cm in diameter before and after gadoxetic acid‐enhanced dynamic magnetic resonance imaging (MRI). Quantitative analysis was performed using receiver operating characteristic (ROC) curves with lesion‐to‐liver signal intensity difference‐to‐noise ratio (SDNR) on precontrast and postcontrast images. Qualitative analysis was also performed by two blinded reviewers.

Results:

The SDNR of the solid lesions was significantly higher on the postcontrast (1.66 ± 1.18) than on the precontrast (1.38 ± 1.07) images (P = 0.0012), while the SDNR of hemangiomas was comparable for pre‐ and postcontrast images (P = 0.8164). The best SDNR cutoff values for distinguishing solid lesions from hemangiomas were ≤1.85 (Az = 0.948) for precontrast and ≤2.58 (Az = 0.901) for postcontrast images (P = 0.057). Reader performances for distinguishing hemangiomas from solid lesions were comparable between the precontrast (Az = 0.975 and 0.970 for readers 1 and 2) and postcontrast (Az = 0.977 and 0.972) images (P = 0.899 and 0.946).

Conclusion:

Heavily T2‐weighted images obtained after administration of gadoxetic acid have a diagnostic capability comparable to precontrast images for differentiating between small hemangiomas and malignant solid lesions of the liver. J. Magn. Reson. Imaging 2011;. © 2011 Wiley‐Liss, Inc.     

Magnetic resonance imaging in patients with pancreatitis: evaluation of signal intensity and enhancement changes

PURPOSE: To evaluate the utility of unenhanced and enhanced T1-weighted fat-suppressed (T1-FS) magnetic resonance imaging (MRI) in detecting pancreatitis. MATERIALS AND METHODS: 1.5-T MRI was performed in 25 patients with acute and 23 patients with chronic pancreatitis and in 20 control subjects without known pancreatic disease. T1-FS spin-echo and contrast-enhanced arterial-predominant (DYN1) and portal-predominant (DYN2) fast multiplanar spoiled gradient-echo (FMPSPGR) sequences were evaluated. These three sets of images were evaluated both subjectively for decreased or heterogeneous signal intensity (rating scale, 0-3) and objectively (region of interest (ROI)) in the head, body, and tail of the pancreas, in each patient. RESULTS: Good correlation between subjective assessment and objective data was demonstrated. The T1-FS sequence showed an abnormality with greater frequency (T1-FS > DYN1, 81/144 scores; T1-FS = DYN1, 63/144 scores; T1-FS < DYN1, 0/144 scores) and magnitude (average subjective score, 2.48 vs. 1.74; P < 0.0003) than that of the contrast-enhanced FMPSPGR (decreased or heterogeneous enhancement). The overall sensitivity and specificity of MRI was 92% and 50%, respectively. On the basis of signal intensity and enhancement, MRI was not able to differentiate acute from chronic pancreatitis. CONCLUSION: MRI was highly sensitive for disease detection, particularly using the T1-FS sequence, but using the sequences described, was not able to differentiate acute from chronic pancreatitis.     

Longitudinal analysis of MRI T2 knee cartilage laminar organization in a subset of patients from the osteoarthritis initiative: A texture approach

Cartilage magnetic resonance imaging T₂ relaxation time is sensitive to hydration, collagen content, and tissue anisotropy, and a potential imaging‐based biomarker for knee osteoarthritis. This longitudinal pilot study presents an improved cartilage flattening technique that facilitates texture analysis using gray‐level co‐occurrence matrices parallel and perpendicular to the cartilage layers, and the application of this technique to the knee cartilage of 13 subjects of the osteoarthritis initiative at baseline, 1‐year follow‐up, and 2‐year follow‐up. Cartilage flattening showed minimum distortion (～ 0.5 ms) of mean T₂ values between nonflattened and flattened T₂ maps. Gray‐level co‐occurrence matrices texture analysis of flattened T₂ maps detected a cartilage laminar organization at baseline, 1‐year follow‐up, and 2‐year follow‐up by yielding significant (P < 0.05) differences between texture parameters perpendicular and parallel to the cartilage layers. Tendencies showed higher contrast, dissimilarity, angular second moment, and energy perpendicular to the cartilage layers; and higher homogeneity, entropy, variance, and correlation parallel to them. Significant (P < 0.05) longitudinal texture changes were also detected reflecting subtle signs of a laminar disruption. Tendencies showed decreasing contrast, dissimilarity, and entropy; and increasing homogeneity, energy, and correlation. Results of this study warrant further investigation to complete the assessment of the usefulness of the presented methodology in the study of knee osteoarthritis. Magn Reson Med, 2010. © 2010 Wiley‐Liss, Inc.     

Gadolinium-enhanced 3D MR angiography of renal artery stenosis: a pilot comparison of maximum intensity projection, multiplanar reformatting, and 3D volume-rendering postprocessing algorithms

RATIONALE AND OBJECTIVES: The authors compared diagnostic accuracy of maximum intensity projection (MIP), multiplanar reformatting (MPR), and three-dimensional (3D) volume rendering (VR) in the evaluation of gadolinium-enhanced 3D magnetic resonance (MR) angiography of the renal arteries. They hypothesized that VR is as accurate as or more accurate than MIP and MPR at depicting renal artery stenosis. MATERIALS AND METHODS: The study group comprised 28 consecutive patients who underwent gadolinium-enhanced 3D MR angiography of the renal arteries. Studies were postprocessed to display images in MIP, MPR, and VR formats. Digital subtraction angiography (DSA), when performed (nine of 28 patients), was the standard for comparison. For each main renal artery, an estimate of percentage stenosis was made for any stenoses detected by three independent radiologists. For calculation of sensitivity, specificity, and accuracy, MR angiographic stenosis estimates were categorized as mild (0%-39%), moderate (40%-69%), or severe (> or = 70%). DSA stenosis estimates of 70% or greater were considered hemodynamically significant. RESULTS: Analysis of variance demonstrated MIP estimates of stenosis were statistically greater than VR estimates in two readers and greater than MPR estimates in all readers for all patients. MIP images also showed the largest mean difference from DSA stenosis estimates for all three readers. For both VR and MPR, mean differences between MR angiographic stenoses estimates and DSA estimates reached significance for only one reader, whereas, for MIP versus DSA, mean differences reached significance for all three readers. Although not statistically significant compared with DSA, accuracies of VR (87%) and MPR (89%) were greater than that of MIP (81%). CONCLUSION: In this pilot study, MIP was the least accurate of the three image display algorithms tested. VR and MPR yielded similar values for each method of comparison.     

DWI、T2加权像及综合应用对前列腺癌诊断价值的研究

目的:通过对分别应用T2WI及DWI两种成像方法及其综合应用对前列腺癌进行定性诊断的比较,探讨两种方法各自及综合应用对前列腺癌的诊断价值。方法:应用1.5T MR成像仪对42例前列腺疾病患者进行前列腺磁共振高分辨成像,其中前列腺癌15例,良性病变包括良性前列腺增生及慢性前列腺炎症等27例。在MRI常规成像基础上行DWI扫描,扩散敏感系数(b)值分别为0、1000s/mm2。采用盲法阅片,按照前列腺6分区法将两种MR技术分别所得图像数据分区评估测量并采用5分制评分,根据两种技术的5分制评分进行T2加权像综合DWI方法评分,将所得结果与病理结果对照。利用SPSS 11.5分别做ROC曲线分析比较,计算各组方法诊断的敏感度、特异度及准确度等。结果:各组诊断方法的敏感度、特异度及ROC曲线下面积(Az)如下:T2WI:88.2%,67.2%和0.848±0.030,最佳诊断界值为3;ADC诊断:82.4%,81.6%和0.860±0.033,最佳界值为4;ADC结合T2WI诊断:78.4%,92.5%和0.922±0.021,最佳界值为4.结论:通过独立评估T2WI及DWI两种方法均可以有效诊断前列腺癌;两种成像方法综合应用诊断前列腺癌的准确度显著高于两种成像技术独立诊断,其诊断结果与病理分析一致性较好。