Abstract
Background/Aim: Exctracapsular extension (ECE) in prostate cancer has a high impact on treatment decision. MRI might predict presence of ECE non-invasively. Patients and Methods: Triplanar T2w-sequences, DWI (diffusion weighted imaging) and DCE (dynamic contrast-enhanced imaging) of 34 patients with PCa were analyzed to prior prostatectomy. Sensitivity (SS) and specificity (SP) of T2w, apparent diffusion coefficient (ADC), plasma flow (PF) and mean transit time (MTT) normalized by PCa/normal tissue ratio for prediction of CI (capsular infiltration)/ECE were determined by area-under-the-receiver-operating-characteristics analysis. Results: SS/SP for detecting ECE was 29/85. AUC (area under the curve) of ECE cases was 0.98/0.92/0.69 (cut-off-ratios 3.2/0.51/0.46), SS 93/100/86% and SP 95/80/50% for PF-/MTT-/ADC-ratios, respectively. PF- and MTT-ratios between CI and without CI/ECE differed significantly (PF, p<0.0001; MTT, p=0.0134) with SS/SP 84/89% for PF and SS/SP 52/100% for MTT-ratios. No significant differences regarding ADC-ratios were identified. Conclusion: ECE/CI can be assessed by quantitative DCE analysis with great diagnostic confidence and higher specificity than ADC.
- Magnetic resonance imaging
- prostate cancer
- dynamic contrast-enhanced MRI
- diffusion weighted MRI
- extracapsular extension
Prostate cancer (PCa) remains a major health concern among men in both developing and developed countries. In 2013, approximately 3 million men were living with a history of PCa in the United States, and an additional 233,000 cases were diagnosed in 2014 (1). Nerve-sparing radical prostatectomy (RP) is the treatment standard for clinically localized PCa, especially in low- and intermediate-risk patients (2, 3). In high-risk patients, treatment options also include beside surgery RT (radiation therapy), hormonal therapy, or a combination (4). The choice of the treatment strategy depends on cancer stage and grade, as well as patient comorbidity and age. The presence of extracapsular extension (ECE) and seminal-vesicle-infiltration (SVI) also has a significant impact, as patients presenting with ECE or SVI are not eligible for RP (5). Moreover, ECE, which is poorly detected by standard clinical tests, can appear in early stages of PCa with an incidence of up to 60% in patients with clinical stage T1/T2 PCa (6).
Due to its excellent intrinsic soft-tissue-contrast and ability to provide functional and morphologic information within one examination, Magnetic Resonance Imaging (MRI) has evolved into an established tool for the diagnostic evaluation of PCa. T2w-sequences are most commonly used for the assessment of ECE. However, studies have found significant variation in the accuracy of T2w-sequences for the detection of ECE and SVI with reported values ranging between 50% and 92% depending upon reader experience (7, 8).
In this context, functional MRI techniques such as dynamic-contrast-enhanced (DCE) and diffusion-weighted (DWI) MRI are increasingly utilized to improve the accuracy of ECE detection. DCE-MRI allows a non-invasive evaluation of tumor perfusion and vascular permeability measuring the changes in gadolinium concentration in blood and tissue over time. The differences in signal-intensity-time courses observed in benign and malignant tissue can either be expressed semi-quantitatively by parameters derived from the gadolinium-concentration-time curves or quantitatively as Plasma Flow (PF) and Mean Transit Time (MTT) derived from pharmacokinetic models (9, 10).
Scan parameters of the T2 weighted, DCE MRI and DWI MRI sequences.
DWI is used to measure the Brownian motion of water molecules in biological tissue. Due to the complexity of biologic tissues, actual diffusion coefficients of water protons cannot be measured directly. Thus, diffusion is measured indirectly by calculating a diffusion coefficient (the apparent-diffusion-coefficient, ADC) (11, 12). Although preliminary results have suggested a benefit to ADC and perfusion measurements for detection of ECE, no cut-off values have been established for quantitative discrimination between prostate carcinoma with and without ECE.
The purpose of this study was to evaluate the diagnostic performance of quantitative functional MR-parameters, such as ADC, PF and MTT for the assessment of ECE in PCa patients versus T2w morphologic imaging.
Patients and Methods
Patients. A total of 34 patients (mean age 67 years, range=52-78 years) with pathologically proven PCa (Gleason score [GS] median 7, range=6-9) were included in this retrospective analysis. The patients were referred for prostate MRI prior to RP by the department of urology between 12/2009 and 12/2012. Institutional review board (IRB) approval was obtained (2013-824R-MA) and written informed consent waived for this retrospective study. Every MRI exam was indicated clinically and no additional examination or therapy was performed.
A total of 4 patients in this series were part of a different study, previously published in Investigative Radiology (13), which focused on distinguishing PCa from prostatitis using DCE-MRI at 1.5 T vs. 3 T. Furthermore, all patients were part of a study published in European Journal of Radiology, comparing sonoelastography to MRI (14).
MRI protocol. MRI was performed on a 3T MR-system (Magnetom TimTrio, Siemens Healthcare Sector, Erlangen, Germany) utilizing a body as well as endorectal coil (Medrad Inc., Pittsburgh, PA, USA). The endorectal coil (ERC) was inserted after digital rectal examination and was inflated with 40-80 ml of air. If not contraindicated, 40 mg of N-butylscopolaminium (Buscopan®, Boehringer Ingelheim GmbH, Ingelheim, Germany) was injected intravenously to suppress bowel motion. The prostate MRI protocol consisted of a high-resolution T2-weighted triplane turbo-spin-echo (TSE) sequence, an axial, fat-suppressed, single-shot, echo-planar diffusion-weighted imaging (EPI-DWI) sequence, a 3D chemical-shift imaging MR spectroscopy (MRS), and an axial 2D T1-weighted (T1w) dynamic contrast-enhanced (DCE) scan using spoiled gradient echo saturation recovery (Turbo FLASH).
The EPI-DWI sequence was acquired prior to gadolinium application and ADC maps were calculated automatically. The DCE-MRI examination was performed after bolus injection of 0.1 mmol/kg (2.5 ml/sec) body weight of gadolinium chelate (Dotarem®, Guerbet, Roissy CdG Cedex, France). Imaging parameters of T2w, DCE, and DW sequences are detailed in Table I. Within 1 to 3 weeks after the MRI, all patients underwent RP. The prostate was histopathologically assessed by a pathologist with 7 years experience. The pathologist identified areas of PCa in a standardized manner with the prostate being divided into 16 parts and using the haematoxylin and eosin stain (H&E stain). A histological capsular penetration was recognized if neoplastic glands were outside the prostatic capsular in the periprostatic tissue, according an ECE (14).
Quantitative analysis of DCE MRI (PF, MTT) and DWI MRI (ADC) for ECE, CI and non ECE/CI cases.
Post-processing and Data-analysis. Post-processing of the DCE-MRI data was performed by a radiologist (M.S.), with 4 years experience in prostate MRI, using an open source software tool for quantitative MRI perfusion analysis (UMM Perfusion, OsiriX DICOM viewer, Geneva, Switzerland) (15). The contrast enhancement curves were analyzed via a model-free deconvolution analysis as described previously (16, 17). The arterial input function (AIF) was measured in a plane wherein the common femoral artery was clearly opacified with the arrival of contrast agent and quantitative color-coded maps of PF and MTT were calculated (16, 17).
On the quantitative color-coded PF and MTT as well as ADC maps, a single region of interest (ROI) was drawn in areas with histologically proven PCa in concordance with the histopathological analysis scheme (see Figure 1). A second ROI was placed in a contralateral area without pathological findings.
For all ROIs PF, MTT and ADC values were measured as the mean of pixel values. A ROI volume of 0.3×0.3 cm2 was used for all examinations. For the normalization and standardized comparison of PF, MTT, and ADC values among the different exams, ratios of these parameters in PCa and normal tissue were calculated in every patient.
Two other radiologists with 10 (D.J.D.) and 15 (S.O.S) years experience in prostate MRI separately evaluated the high-resolution triplane T2w-images for ECE. Both radiologists were aware of the diagnosis of PCa but were otherwise blinded to any clinical and pathological information. By definition, ECE was diagnosed on the triplane T2w-images if at least one of the following criteria were available: irregular capsular bulge, disruption of the prostatic capsule, extension into the periprostatic fat, obliteration of the recto prostatic angle, or asymmetry/involvement of the neurovascular bundle (18, 19).
Statistical analysis. Statistical analysis was performed with a statistical software package (SAS System, release 9.2, SAS Institute Inc., Cary, NC, USA). DWI- and DCE-MRI parameters are presented by their mean values and standard deviations; for ordinal scaled parameters median values are given. Kappa coefficients have been calculated to assess the strength of agreement between two readers. In order to compare the mean values of two groups, 2 sample t-tests have been used. For the comparison of GS, Cochran Armitage trend test has been applied. Spearman correlation coefficients have been calculated to find a relationship between GS and quantitative parameters. For diagnostic evaluations, logistic regression analyses have been performed. Areas under the curve (AUCs) have been compared using a test according to De Long et al. (20).
Results
All examinations were successfully completed without any adverse events. Pathologically, ECE was present in 14 patients, whereas capsular infiltration (CI) was found in 11 patients. The median GS was 7 in all patients.
On the triplanar T2w-sequences, ECE was correctly diagnosed in only 4 out of 14 patients by each of the two experienced readers resulting in sensitivity (SS) of 29% and specificity (SP) of 85% for both readers. Hence, there were 10 false negative cases and 3 false positive cases for ECE based upon the evaluation of T2w-sequences alone. The interobserver agreement was poor with κ=0.1. CI was detected by reader 1 (D.J.D.) with a SS/SP of 96/44% and by reader 2 (S.O.S.) with a SS/SP 68/44%.
A total of 34 areas with PCa and 34 control areas were included in the ROI analysis. The mean values of PF, MTT and ADC in PCa and contralateral normal prostate tissue for all three groups (ECE, CI, no ECE/CI) are summarized in Table II.
Statistically significantly higher PF-ratios (p<0.0001), lower MTT-ratios (p<0.0001) and lower ADC-ratios (p=0.02) were found in areas of PCa with ECE versus areas of PCa without. Furthermore, we found a positive correlation between PF-ratios and GS (r=0.4517, p=0.0073). However, differences in MTT- and ADC-ratios with GS were not observed.
Receiver operating characteristic (ROC) analysis identified maximum areas-under-the-curve (AUC) for cut-off-ratios of 3.2 with PF, 0.51 with MTT and 0.46 with ADC (Figure 2). Using these ratios ECE could be diagnosed with SS/SP of 93/95% (PF), 100/80% (MTT) and 86/50% (ADC), respectively. The AUC of ECE cases was 0.98, 0.92, and 0.69 for PF, MTT and ADC-ratios, respectively.
Furthermore, there were statistically significant differences between tumors with and without CI regarding PF-ratios (p<0.0001) and MTT-ratios (p=0.0134), and in contrast, no significant difference regarding ADC-ratios could be identified (p=0.1203). Logistic regression analysis revealed AUCs of 0.89 (PF-ratio), 0.80 (MTT-ratio) and 0.69 (ADC-ratio). Optimal cut-off-ratios were 2.56 for PF-ratio (SS/SP, 84/89%) and 0.42 for MTT-ratio (SS/SP, 52/100%).
69-year old man with PCa and ECE, Gleason score 7 (4+3), (a) axial histopathologic slice, pT3a (b) low signal intensity on the axial T2w image without morphological signs of ECE and (c) restricted diffusion on the ADC map (823.5 sec/mm2, ratio* 0.40), DCE MRI demonstrates increased PF (d) and decreased MTT (e) values/ratios (PF, 14.87 ml/100 ml/min, ratio* 6.44; MTT 91.27 sec, ratio* 0.33). *PCa/normal tissue.
ROC (receiver operating characteristics) analysis. ROC analysis for detection a) of ECE (extracapsular extension) and b) of CI (capsular infiltration) with PF (plasma flow) ratios, MTT (mean transit time) ratios, and ADC (apparent diffusion coefficient) ratios. The AUC (area under the curve) of ECE/CI cases was 0.98/0.89, 0.92/0.80 and 0.69/0.60 for PF, MTT, and ADC ratios, respectively.
Discussion
This study demonstrates the successful identification of ECE and CI of PCa via use of quantitative DCE-MR parameters such as PF and MTT. In addition, ECE of PCa has been detected by quantitative ADC measurements. Both PF- and MTT-ratios successfully differentiated cases with and without ECE and CI. PF-ratios correlated with GS significantly; although, the relationship between MTT- or ADC-ratios and GS was not statistically significant. In distinction, the diagnostic performance of morphologic T2w-images was poor.
The reported SS and SP of T2w-sequences for the detection of ECE in PCa vary widely, ranging from 14% to 97%, and this variation is a well-known limitation of the technique (7, 8). Reader experience is thought to play an important role in assessing ECE and SVI with MRI; however, Fütterer et al. demonstrated only incremental gains in SS/SP from 50 to 59%/93 to 96% for ECE in a study of 120 patients; SS for detection of SVI raised from 43 to 71% without change in SP (21). Similar to those findings, both readers in our study had more than 10 years experience in prostate MRI, however the SS and SP for ECE detection with high resolution, triplane T2w-images remained poor with no interobserver agreement. Moreover, cases of early ECE were frequently misclassified as CI. This mischaracterization was often related to known pitfalls in prostate MRI leading to misinterpretation (22), such as artifacts caused by the interposition of the faces of the ERC and the air filled rectum.
Numerous studies have demonstrated the added benefit of DWI-MRI for both localization as well as local staging of PCa (23, 24). These studies showed that ADC values, calculated from DWI measurements, correlate with both GS and local tumor aggressiveness. Moreover, low ADC values are a biomarker for the risk of recurrence following prostatectomy (24). In a recent study by Hara T et al. including 132 PCa patients, ECE was diagnosed with respective accuracy and SP of 72% and 62% on ADC maps. However, even when all hypointense foci with capsular contact present on ADC maps were defined as representing ECE, the SS did not exceed 58% (25). In our study, we also found statistically significant differences in ADC-ratios between cases with and without ECE, but not in cases with CI. Giganti et al. showed significant different ADC values and ADC-ratios in patients with and without ECE (26). The SS and SP for detection of ECE were moderate (SS/SP 85/50%). The reason for this could be the magnetic susceptibility artifacts arising from the interface of the air-filled ERC and dorsal prostate tissue may have limited the robustness of ADC calculations for nearby lesions and decreasing diagnostic accuracy.
In the last decade several studies have demonstrated the benefit of DCE-MRI for diagnosis and localization of PCa, particularly in the detection of initial ECE in intermediate and high-risk patients (27). However, to the best of our knowledge, no study thus far has attempted a quantitative assessment of ECE with DCE-MRI. In a study by Bloch et al. with 32 PCa patients, a SS/SP of 91/86% was obtained utilizing a 1.5 T MRI-system and color-coded scheme of the DCE images. In that work, the authors defined the presence of extracapsular bright red pixel clusters larger than 3 mm in diameter as being positive for ECE (19). In a preliminary study by Fütterer et al. using a 3 T MRI-system, the criteria for ECE on DCE-MRI were: presence of asymmetric high peak enhancement, washout, and shorter onset time or an increased time-to-peak parameter. The fusion of parametric DCE and T2w-images resulted in a SS/SP of 65/95% versus 59/96% for the T2w-images alone (21). In a recent study by Min et al. the SS, SP, and accuracy of DCE-MRI for detecting ECE at 3 Tesla were, respectively, 27%, 96%, and 65%. The relatively poor sensitivity reported in Min et al. may be related to the presence of post-biopsy hemorrhage. This was present in more than 90% of patients, since the time between biopsy and MR was 13.9±5.9 days. However, SS, SP, PPV, and NPV (65%, 87.5%, 76.5%, 80%) did increase utilizing a combined evaluation including T1w, T2w, DCE, and DWI (28).
The key element of our study was the quantitative evaluation of the DCE-MRI parameters calculated by a deconvolution analysis (9, 15, 16) which allowed differentiation between CI, ECE and non-ECE. As reported by Schlemmer et al. (29) an increased PF reflects a higher intratumoral microvessel density (MVD), which is considered a marker for neo-angiogenesis in malignant tissue. As further studies have described, MVD correlates with clinical/pathological stage, metastasis and histological grade, as well as disease-specific survival and progression after treatment (30, 31). In this context, we conclude that an increased MVD, as reflected by higher PF values and corresponding MTT shortening, also correlates with a higher probability of CI and ECE.
Our study has several limitations. For one, the relatively small number of patients limits the statistical relevance of the results. In addition, all examinations were performed using an air-filled ERC. Although this affords a greater signal-to-noise ratio, use of such a coil is associated with a number of disadvantages such as gland deformation and, especially for DWI, susceptibility-related artifacts. The latter may have contributed to the limited diagnostic accuracy observed with ADC calculations.
In conclusion, CI and ECE can be assessed by quantitative DCE-MRI analysis with a high-level of diagnostic confidence. Thus, quantitative DCE-MRI parameters have the potential to serve as a reader-independent risk estimator in the diagnosis of ECE, evaluating ECE with a greater SS and SP than ADC or T2-weighted image analysis. The statistical relationship between high GS and high PF also suggests that ECE correlates with increased MVD.
Footnotes
↵* These Authors contributed equally to this study.
Conflicts of Interest
The Authors declare that there is no conflict of interest.
- Received October 16, 2017.
- Revision received November 6, 2017.
- Accepted November 8, 2017.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.