Abstract
Objective
To compare dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) and T2-weighted imaging (T2WI) at 3.0T for detection of prostate cancer.
Methods
Patients with elevated prostate-specific antigen underwent T2WI and DCE-MRI prior to prostate needle biopsy. The sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of T2WI and DCE-MRI to diagnose prostate cancer were evaluated. The relationship between Gleason score and prostate cancer detection by DCE-MRI was evaluated.
Results
Prostate adenocarcinoma was histopathologically confirmed in 44/75 patients. DCE-MRI had significantly higher sensitivity, accuracy and NPV than T2WI. The detection rate of prostate cancer by DCE-MRI was significantly better for tumours with Gleason score 7–9 than for those Gleason score 4–6.
Conclusion
DCE-MRI at 3.0T can significantly improve prostate cancer detection using simple visual diagnostic criteria, compared with T2WI.
Introduction
Prostate cancer is the most common malignant tumour and the second most deadly cancer in men, in the developed world. 1 Diagnosis of prostate cancer is based on digital rectal examination, serum concentration of prostate-specific antigen (PSA), transrectal ultrasound-guided (TRUS) biopsy and magnetic resonance imaging (MRI). MRI has been widely used to aid prostate cancer detection and tumour staging. T2-weighted imaging (T2WI), in which tumour tissue appears hypointense relative to the normal peripheral tissue, 2 has been used for morphological prostate tumour detection and localization, but the specificity of T2WI is low because benign prostatic hyperplasia, prostatitis, fibrosis and postbiopsy haemorrhage also cause T2 hypointensities.3–5 In addition, some prostate tumours appear normal on T2WI, leading to low sensitivity of this method.6,7 Functional MR techniques (such as dynamic contrast-enhanced MRI [DCE-MRI], diffusion-weighted imaging [DWI], and magnetic resonance spectroscopy [MRS]) have been used to increase the diagnostic accuracy of MR in prostate cancer.8–10
Dynamic contrast-enhanced MRI is useful in imaging tumour vascularization, vascular permeability and perfusion. 11 The method involves intravascular injection of contrast agents and imaging of their concentrations in blood and tissue over time. 11 Analysing contrast agent uptake in tissues usually involves the generation of a semiquantitative signal intensity/time curve or more complicated quantitative approaches using pharmacokinetic models. The highly specialized and time-consuming nature of these analyses likely prevents widespread clinical implementation of these methods,12,13 and, since they are nonstandardized, often generate unclear results.14–17 Studies have focused on the analysis of raw DCE T1-weighted images (T1WI), which can be more easily implemented in daily clinical practice, compared with other imaging modalities.18–23 These studies had low temporal resolution, minimal dynamic series and an acquisition time that was too brief to allow prostate cancer detection or comprehensive measurement of tumour haemodynamics.
The aims of the present study, therefore, were to evaluate the clinical value of 3.0T DCE-MRI in detecting prostate cancer, and to compare imaging results with TRUS-guided biopsy findings.
Patients and methods
Study population
The study recruited consecutive male patients with elevated PSA (>4.0 ng/ml) and/or prostate nodule detected during digital rectal examination who attended the Department of Radiology, Zhujiang Hospital of Southern Medical University, Guangzhou, Guangdong Province, China, for diagnosis between January 2012 and July 2013. Patients were required to have undergone both MRI and a subsequent transrectal prostate biopsy, and to not have received any treatment for prostate cancer.
The study was approved by the ethics committee of the Zhujiang Hospital of Southern Medical University, Guangzhou, China. Written informed consent was obtained from all patients prior to enrolment.
MRI
All MRI examinations were performed on a 3.0T whole-body multitransmit scanner system (Achieva TX, Philips Healthcare, Best, The Netherlands) using a 16-channel SENSE XL torso coil. T2-weighed, turbo spin-echo images with spectrally selective attenuated inversion recovery were obtained in the axial and coronal planes (repetition time [TR] 1483 ms; echo time [TE] 70 ms; slice thickness 5 mm; interslice gap 1 mm; number of slices 20; field of view [FOV] 240 × 240 mm; matrix size 256 × 256 pixels). DCE-MRI was performed using a three-dimensional (3D) T1-fast field-echo (FFE) sequence in the axial plane (TR 5.5 ms; TE 1.7 ms, slice thickness 6 mm [reconstructed to 3 mm]; interslice gap 0 mm; FOV 230 × 230 mm; flip angle 15°; matrix size 256 × 256 pixels). DCE-MRI images were scanned from the apex to the base of the prostate and a total of 20 slices were obtained. A 20-slice volume was obtained every 2.9 s and imaging comprised eight precontrast volumes and 96 postcontrast volumes. Postcontrast imaging was initiated immediately after administering 0.1 mmol/kg body weight gadopentetate dimeglumine (Magnevist®, Bayer Schering Pharma, Germany) at 2.5 ml/s via the cephalic vein. Contrast agent injections were followed by a 15-ml saline flush. The DCE-MRI examination time was 5 min and 6 s.
Postimaging prostate biopsy
Within 2 weeks of MRI, all patients underwent an extended 12- to 18-core TRUS-guided biopsy. All samples were obtained from the peripheral zone bilaterally, including the base, mid-gland, apex, and the bilateral transitional zone. Each sample was histologically analysed by the same pathologist with 15 years’ genitourinary experience, and was determined to be cancerous or noncancerous according to the pathology.
MRI analyses
The T2WI and DCE-MRI data were prospectively analysed by two radiologists who were blinded to the clinical data (X-H.Z. and X-Y.Q., with 10 and 13 years’ experience reading prostate MR images, respectively). On T2WI, prostate cancer was defined as hypointense nodules in the peripheral zone, or areas of homogeneous hypointensity with ill-defined margins and no visible capsule in the transitional zone of the prostate gland. DCE-MRI evaluation of unmodified T1 images, before and after administration of the contrast agent, defined prostate cancer as nodular foci that showed early and strong enhancement, and rapid washout relative to the background in the peripheral zone and transitional zone. Nodular foci showing early and strong enhancement followed by a plateau phase were not considered to be prostate cancer, and neither were nodular foci showing persistent enhancement. The prostate was divided into eight regions on MRI, corresponding to TRUS-guided biopsy (the bilateral peripheral zone including the base, mid-gland, apex and the bilateral transitional zone). MR regions were classified as cancerous or noncancerous according to the pathological results of corresponding biopsy cores.
Statistical analyses
Patients with at least one region with positive MR and biopsy findings were considered true positives (TP). Patients with negative MR and biopsy findings in all regions were considered true negatives (TN). Sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) were calculated according to the following formulae: Sensitivity = TP/number of patients with cancer at biopsy; Specificity = TN/number of patients without cancer at biopsy; Accuracy = TP + TN/number of patients; PPV = TP/number of patients with positive MRI; NPV = TN/number of patients with negative MRI.
Data were presented as mean ± SD or n (%). The sensitivity, specificity and accuracy of each technique for the diagnosis of prostate cancer were compared using McNemar test. Positive predictive value (PPV) and negative predictive value (NPV) were compared using χ2-test; χ2-test was also used to correlate Gleason scores and prostate cancer detection by DCE-MRI. All statistical analyses were performed using SPSS® version 13.0 (SPSS Inc., Chicago, IL, USA) for Windows®. P-values <0.05 were considered statistically significant.
Results
The study included 75 patients (mean age 69 ± 8 years; range 50–83 years; median PSA concentration 14.19 ng/ml; range 4.59–470.3 ng/ml). Prostate adenocarcinoma was histopathologically confirmed in 44 patients (58.7%). MR images of two representative patients with prostate adenocarcinoma are shown in Figures 1 and 2. Biopsy findings of the remaining 31 patients without cancer included benign prostatic hyperplasia (n = 26) and prostatitis (n = 5).
Magnetic resonance (MR) images from a 72-year-old male with prostate cancer (Gleason score = 4 + 3). A: T-weighted image showed abnormal hypointensities in the left peripheral zone (arrow). B: Dynamic contrast-enhanced (DCE)-MRI soon after contrast injection showing early and strong enhancement of lesions in the left peripheral zone (arrow). C: In later DCE-MRI acquisitions, lesions exhibited washout (arrow). The colour version of this figure is available at: http://imr.sagepub.com. Magnetic resonance (MR) images from a 69-year-old male with prostate cancer (Gleason score = 3 + 4). A: T2-weighted image showing abnormal hypointense area in the left transitional zone (arrow). B: Dynamic contrast-enhanced (DCE)-MRI soon after contrast injection showing early and strong enhancement of lesions in the left transitional zone (arrow). C: In later DCE-MRI acquisitions, lesions exhibited washout (arrow). The colour version of this figure is available at: http://imr.sagepub.com.
Comparison of dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) and T2-weighted imaging (T2WI) in the diagnosis of prostate adenocarcinoma (n = 75 patients, n = 44 histopathologically confirmed tumours).
Data presented as n (%).
aMcNemar test.
bχ2-test.
NS, not statistically significant (P ≥ 0.05).
A total of nine patients had Gleason scores ≤6 (4, n = 1; 5, n = 2; 6, n = 6); the remaining 35 tumours were scored ≥ 7 (7, n = 21; 8, n = 10; 9, n = 4). The DCE-MRI prostate cancer detection rate was significantly higher for tumours with Gleason scores of 7–9 than those with scores of 4–6 (97 vs 67%; P = 0.029).
Discussion
The utility of T2WI in differentiating between prostate cancer and other abnormalities is limited. To improve prostate cancer detection, conventional T2WI is complemented by functional MR techniques such as DCE-MRI, DWI, and MRS.8–10 The present study investigated the use of T2WI and DCE-MRI at 3.0T to differentiate cancerous prostate tissue from noncancerous tissue, and found that DCE-MRI was significantly more sensitive, accurate and had better NPV for cancer detection than T2WI. Specificity and PPV were not significantly different between DCE-MRI and T2WI, however.
Prostate cancer generally develops as multiple lesions, and the sensitivity and specificity of DCE-MRI therefore vary between reports.20–23 Using a per-sector analysis to measure sensitivity and specificity, a per-patient analysis was found to have higher sensitivity but decreased specificity. 21 Others found a dramatic difference in specificity between per-sector and per-patient analyses for DCE-MRI that was thought to be due to the very high proportion of prostate regions without cancer, which could lead to an overestimation of specificity in per-sector analysis. 20 In addition, a single false-positive sector was sufficient for a patient to be counted as a false-positive result on per-patient analysis. 20 Other studies have suggested that multifocal prostate cancer usually comprises a dominant, potentially aggressive tumour that would warrant therapy, but other smaller foci are clinically insignificant and should only be monitored.24,25 The dominant tumour is usually considered to be the largest tumour body. The present study, therefore, estimated sensitivity, specificity, accuracy, PPV and NPV of DCE-MRI for prostate cancer using a per-patient analysis.
Most prostate cancers have higher vascularity and permeability than background prostate tissue, showing early and strong enhancement, and rapid washout relative to background prostate tissue. 26 This is confirmed by the findings of the present study, where our per-patient analysis resulted in greater sensitivity, specificity and accuracy than those of others.20,21 This may be due to several factors. First, the use of 3.0T in the current study allows for higher signal-to-noise ratio than the 1.5T used by others.20,21 Secondly, temporal resolution was 2.9 s in our study, but 15–20 s in the other studies.20,21 Early enhancement is more difficult to identify using longer acquisition times (e.g., >15 s). 15 Finally, the time period between bolus injection and peak intensity has been shown to be 70–180 s in cancerous tissue (mean 103 s) compared with 200–300 s (mean 250 s) in noncancerous tissue. 16 The delay time in the present study was nearly 300 s, considerably longer than the 120–180 s used by others.20,21 This extended delay time is sufficient to enable accurate visualization of the initial increase in signal enhancement and gradual changes in signal intensity, in cancerous and noncancerous tissue.
It is unclear whether the detection rate of prostate cancer is correlated with Gleason score.19,27–30 Some studies have found no significant correlation;27,28 and others have reported that detection rates were significantly higher for high-grade than low-grade tumours.19,29,30 Consistent with these findings, the results of the present study indicate that the detection rate for tumours with Gleason score 7–9 was significantly higher than that for tumours with score 4–6.19,29,30
There were several limitations to this study. First, we did not use an endorectal coil, which gives a high signal-to-noise ratio but has drawbacks regarding patient discomfort, time and cost. Secondly, our reference standard was prostate biopsies, and it is possible that some patients with negative biopsy findings could have had cancer. Moreover, MRI findings do not correspond directly to biopsy results because the entire prostate was not evaluated during biopsy. 18 Comparing MR results with histological specimens of the whole prostate following radical prostatectomy will be required in future studies. Finally, needle biopsy often underestimates Gleason score compared with prostatectomy, but this would not change our findings of the relationship between Gleason score and prostate cancer detection by DCE-MRI.
In conclusion, DCE-MRI at 3.0T can significantly improve prostate cancer detection using simple visual diagnostic criteria compared with T2WI. In addition, DCE-MRI has the potential to detect tumours with high Gleason scores.
Footnotes
Declaration of conflicting interest
The authors declare that there are no conflicts of interest.
Funding
This research received no specific grant from any funding agency in the public, commercial, or nonprofit sectors.
Acknowledgements
We thank Dr Yu Wang, Department of Pathology of Zhujiang Hospital, Southern Medical University, Guangzhou, China, for her help with pathological diagnosis for this study.