- Email: [email protected]
Prostate cancer: correlation of intravoxel incoherent motion MR parameters with Gleason score
Prostate cancer: correlation of intravoxel incoherent motion MR parameters with Gleason score Dal Mo Yang, Hyun Cheol Kim, Sang Won Kim, Geon-Ho Jahng, Kyu Yeoun Won, Sung Jig Lim, Jang-Hoon Oh PII: DOI: Reference:
S0899-7071(16)00002-4 doi: 10.1016/j.clinimag.2016.01.001 JCT 7976
To appear in:
Journal of Clinical Imaging
Received date: Revised date: Accepted date:
13 October 2015 4 December 2015 6 January 2016
Please cite this article as: Yang Dal Mo, Kim Hyun Cheol, Kim Sang Won, Jahng Geon-Ho, Won Kyu Yeoun, Lim Sung Jig, Oh Jang-Hoon, Prostate cancer: correlation of intravoxel incoherent motion MR parameters with Gleason score, Journal of Clinical Imaging (2016), doi: 10.1016/j.clinimag.2016.01.001
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Prostate Cancer: Correlation of Intravoxel Incoherent Motion MR Parameters with Gleason Score Dal Mo Yang1*, Hyun Cheol Kim1, Sang Won Kim1, Geon-Ho Jahng1, Kyu Yeoun Won2, Sung Jig
RI P
T
Lim2, Jang-Hoon Oh3 1
Department of Radiology, Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea
2
SC
Department of Pathology, Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea
3
University, Yongin, Republic of Korea *
MA NU
Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee
Corresponding author. Department of Radiology, Kyung Hee University Hospital at Gangdong, 149,
Sangil-Dong, GAngdong-GU, Seoul, 134-727 Republic of Korea
AC
CE
PT
ED
E-mail: [email protected]
ACCEPTED MANUSCRIPT Prostate Cancer: Correlation of Intravoxel Incoherent Motion MR Parameters with Gleason Score
PT
ED
MA NU
SC
RI P
T
Purpose: To evaluate the potential of IVIM imaging to predict histological prognostic parameters by investigating whether IVIM parameters correlate with Gleason score. Materials and methods: The Institutional Review Board approved this retrospective study and informed consent was waived. A total of 41 patients with histologically proven prostate cancer who underwent prostate MR imaging using a 3T MRI machine were included. For 8 DWI b-values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2), a spin-echo echo-planar imaging (EPI) sequence was performed. D, f, D, and ADCfit values were compared among three groups of patients with prostate cancer: Gleason score 6 (n = 9), 7 (n = 16), or 8 or higher (n = 16). Receiver operating characteristic (ROC) curves were generated for D, f, D, and ADCfit to assess the ability of each parameter to distinguish cancers with low Gleason scores from cancers with intermediate or high Gleason scores. Results: Pearson’s coefficient analysis revealed significant negative correlations between Gleason score and ADCfit (r = -0.490, P = 0.001) and Gleason score and D values (r = -0.514, P = 0.001). Gleason score was poorly correlated with f (r = 0.168, P = 0.292) and D values (r = -0.108, P = 0.500). The ADCfit and D values of prostate cancers with Gleason scores 7 or 8 were significantly lower than values for prostate cancers with Gleason score 6 (P < 0.05). ROC curves were constructed to assess the ability of IVIM parameters to discriminate prostate cancers with Gleason score 6 from cancers with Gleason scores 7 or 8. Areas under the curve were 0.671 to 0.974. ADCfit and D yielded the highest Az value (0.960– 0.956), whereas f yielded the lowest Az value (0.633). Conclusions: The pure molecular diffusion parameter, D, was the IVIM parameter that best discriminated prostate cancers with low Gleason scores from prostate cancers with intermediate or high Gleason scores.
AC
CE
Keywords: Magnetic resonance imaging Diffusion-weighted imaging Incoherent intravoxel motion MR imaging Prostate Prostate cancer
ACCEPTED MANUSCRIPT 1.Introduction
PT
ED
MA NU
SC
RI P
T
Prostate cancer is the second most common cancer among men worldwide; moreover, the incidence of prostate cancer is rising in Asia and Eastern Europe [1]. The accurate assessment of prostate cancer aggressiveness is important for deciding the most appropriate initial treatment strategy. The Gleason score is an important preoperative predictor of response to treatment and patient outcomes [2]. However, biopsy determination of the Gleason grade is invasive and often does not accurately reflect the final Gleason score [3, 4]. Recently, functional MRI sequences such as DWI and dynamic contrast-enhanced (DCE) imaging have been shown to provide information about tumor aggressiveness [5-7]. The apparent diffusion coefficient (ADC) is negatively correlated with the Gleason score for prostate cancer [8-14]. Specifically, patients with high Gleason scores have tumors with lower ADCs. This relationship is likely because tumors with high Gleason scores exhibit increased cellularity [15]. Intravoxel incoherent motion (IVIM) MR imaging quantitatively assesses the microscopic translational motion that occurs in each image voxel on MR imaging scans [16-19]. Both pure molecular diffusion (D) and microcirculation-related diffusion, otherwise known as pseudodiffusion (D*), can be distinguished using IVIM MR imaging, as can the blood flow fraction (f) [17]. Recent studies with IVIM imaging showed that reduced ADCs for prostate cancer are due to decreased molecular diffusion and to a decreased perfusion fraction [2023]. However, studies on the relationship between IVIM parameters and Gleason scores are rare [24]. Therefore, this study evaluated the potential of IVIM imaging to predict histological and prognostic parameters by investigating whether IVIM parameters correlated with Gleason score. 2.Materials and Methods
CE
2.1.Patients
AC
Our Institutional Review Board approved this retrospective study; the requirement for informed patient consent was waived. We searched the archive of MR images at our institution and identified 47 consecutive patients who were seen between October 2012 and March 2014, had biopsy-proven prostate cancer, and underwent prostate MR imaging on a 3T MRI machine. Six patients were excluded because their lesions were not seen on T2weighted images, DCE and DWI. Gleason scores were 3 + 3 for five patients and 4 + 3 for one patient. Thus, the final study population comprised 41 patients (mean age, 71 years; range, 50–86 years). Of the 41 patients, 27 patients had the prostate cancers at only peripheral zone. In the remaining 14 patients, two had the prostate cancers at the transitional zone and 12 had the prostate cancers at both the peripheral zone and the transitional zone. The mean PSA level of the 41 patients was 21 ng/mL (range, 3.9–84.8 ng/mL). Prostate cancer was proven in 10 patients via radical prostatectomy and in 31 via systemic transrectal sonography-guided biopsy. The 31 patients diagnosed by systemic biopsy alone had 10–16 biopsies per session. Gleason scores were 6 for 9 patients, 7 for 16, 8 for 7, 9 for 7, and 10 for 2. A summary of patient characteristics is in Table 1. 2.2.MR imaging techniques MR imaging was performed with a 3-Tesla system (Achiva; Philips Health Care, Eindhoven, The Netherlands) with a six-channel sensitivity-encoding (SENSE) torso coil. All patients were initially examined with a routine MRI protocol for imaging the prostate, which included T2-weighted, T1-weighted, DCE and multi-b-value DWI images. For axial and coronal T2-weighted images, fast spin-echo (FSE) MR imaging was
ACCEPTED MANUSCRIPT
MA NU
SC
RI P
T
performed with the following parameters: repetition time (TR)/echo time (TE): 3500/90 ms; matrix: 340 × 240; field of view (FOV): 18 × 18 cm; acquisition of two signals; and section thickness: 3 mm with a 0.3 mm section gap. For sagittal T2-weighted images, FSE MR imaging was performed with the following parameters: TR/TE: 3500/90 ms; matrix: 460 × 320; FOV: 24 × 24 cm; acquisition of two signals; and section thickness: 5 mm with no section gap. For T1-weighted images, FSE MR imaging was performed with the following parameters: 550/10 ms; flip angle: 90°; matrix: 340 × 210; FOV: 18 × 18 cm; acquisition of two signals; and section thickness: 5 mm with a 2 mm section gap. DCE imaging after intravenous injection of 0.2 ml/kg of gadoterate meglumine (Doterem; Guerbet, Roissy, France) at 3 ml/s was performed with the following parameters: 3.1/1.2 ms; matrix: 240 x 220; FOV: 24 x 24 cm; acquisition of one signal; section thickness 4.8 mm with 2.4 mm section gap; number of temporal acquisitions: 48; and temporal resolutions: 5.5 s. For the eight DWI b-values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2), a spin-echo echo-planar imaging sequence was performed with the following parameters: TR/TE: 5000/90 ms; matrix: 112 × 100; FOV: 22 × 22 cm; acquisition of two signals; section thickness: 5 mm with 2 mm section gap; a transverse plane; SENSE parallel imaging factor: 2; and acquisition time: 3 minutes 45 seconds.
2.3.Image analysis
AC
CE
PT
ED
Three experienced radiologists (D.M.Y., H.C.K., and S.W.K.) retrospectively reviewed the MR images using a picture archiving and communications system. For each patient with prostate cancer, the three radiologists came to a consensus on a representative region to define as the cancer region-of-interest (ROI) on pathological reports and datasets of parametric T2WI, DCE, and multi-b-value DWI MR images. In details, the three radiologists reviewed the pathological reports to define the prostate cancer area. After that, MR images of T2WI, DCE, and DWI, and the corresponding ADC maps were reviewed to identify the prostate cancer area which was matched with the pathological report. Each radiologist had more than 10 years of experience in performing prostate MRI and interpreting prostate MR images. One radiologist (D.M.Y.) defined a three-dimensional ROI for each prostatic lesion using MRIcro software (http://www.cabiatl.com/micro/) to obtain a DWI signal intensity (SI) for the lesion (Fig. 1). ROIs were manually drawn as large as possible on ADC maps and positioned to avoid areas of necrosis and hemorrhage. Two different methods were used to obtain diffusion or flow-related parameters. In the first method, the ADC of each ROI was calculated by fitting the eight DWI SI values from eight different b values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2) to the following equation: SI/SI0 = Exp (-bADCfit), where ADCfit was the ADC obtained from the exponential fitting of DWI signal intensities. ADCfit was calculated using linear regression analysis with the natural logarithm of the intensity and the variable b from the eight images and was expressed in mm2/sec. Second, the flow rate, true diffusion coefficient, and pseudodiffusion coefficient were calculated according to Le Bihan et al [16] as follows: SI/SI0 = (1 – f) x exp (-bD) + f x exp (-b D), where D and D were the true diffusion coefficient and the pseudodiffusion coefficient, respectively, and f was the fractional volume occupied in the voxel by flowing spins. Using the Levenberg-Marquardt nonlinear least-squares algorithm, D, f, and D were calculated for each ROI of the prostatic lesions (Fig. 1). Eight b-values (0, 10, 20, 50, 100, 200, 500 and 800 sec/mm2) were used. D and D* were in mm2/sec and f was in percent (%). D was the true diffusion coefficient and represented the slow component of diffusion (i.e., pure molecular diffusion). D was the flow-related diffusion coefficient and represented the fast component of diffusion (i.e., via incoherent microcirculation).
ACCEPTED MANUSCRIPT
2.4.Statistical analysis
SC
RI P
T
The Kruskal-Wallis test was used to compare the three IVIM parameters (D, f, and D) and ADCfit among prostate cancer lesions with Gleason scores of 6, 7, or 8. Post hoc analysis with the Bonferroni correction was performed for all parameters identified as significant by the Kruskal-Wallis test. Spearman’s correlation analysis was used to characterize the association of ADCfit, D, f, and D for tumors with their corresponding Gleason scores. Receiver operating characteristics (ROC) curves were generated to determine the optimal cutoff values for significant IVIM parameters to discriminate low-grade from intermediate/high-grade lesions. All analyses were performed with SPSS software (version 18, Chicago, IL, USA). Probability less than 0.05 was considered statistically significant.
MA NU
3.Results
AC
CE
PT
ED
Mean ADCfit, D, f, and D for prostate cancer lesions with Gleason scores of 6, 7, and 8 are in Table 2. Pearson’s coefficients revealed a significant negative correlation between Gleason score and ADCfit (r = -0.490, p = 0.001) and Gleason score and D values (r = -0.514, p = 0.001) (Figs. 2a, 2b). In contrast, Gleason score correlated poorly with f (r = 0.168, P = 0.292) and D (r = -0.108, p = 0.500) (Figs. 2c, 2d). ADCfit and D for the prostate cancers with Gleason scores of 7 or 8 were significantly lower than ADCfit and D for cancers with Gleason score 6 (p < 0.05) (Figs. 2a, 2b). No significant differences in D or f were observed among prostate cancers with Gleason scores 6, 7, or 8 (p > 0.05) (Figs. 2c, 2d). ROC curves of the IVIM parameters yielded AUCs from 0.671 to 0.974 for discriminating prostate cancers with Gleason score 6 from prostate cancers with Gleason scores of 7 or 8 (Table 3). ADCfit and D yielded high Az values (0.960–0.956), whereas f yielded the lowest Az value (0.633) (Fig. 3). To optimize the diagnostic accuracy of each parameter, cutoff values of ADCfit < 1.29 x 10-3 mm2/s, D < 1.18 x 10-3 mm2/s, D < 27.97 x 10-3 mm2/s, and f < 18.42% were used. Diagnostic characteristics based on these cutoffs are in Table 3.
4.Discussion
In this analysis, we found that Gleason scores exhibited a significant negative correlation with ADCfit. In addition, ADCfit values for prostate cancers with Gleason scores of 7 or 8 were significantly lower than for prostate cancers with Gleason score 6. Thus, tumors with higher Gleason scores had lower ADCs. These results were consistent with previous studies [8-14]. Based on the IVIM model, the reduction of ADCs in prostate cancer was due to a decrease in molecular diffusion as well as partly decreased perfusion fraction [20-23]. In our study, Gleason scores were significantly negatively correlated with D value. In addition, D values of prostate cancers with Gleason scores of 7 or 8 were significantly lower than prostate cancers with Gleason score 6. However, Gleason scores were poorly correlated with both f and D values, which are related to perfusion-related diffusion. In addition, f values were not significantly different among prostate cancers with Gleason scores of 6, 7, and 8. These findings were in accordance with previous results [24]. Zhang et al. found that D values can be useful for discriminating low-grade tumors from intermediate/high-grade tumors [24]. However, no significant difference was observed between low-grade tumors and intermediate/high-grade tumor in analyses with both f and D [24].These results suggest that low ADCs observed in prostate cancers with high Gleason scores result from pure molecular diffusion of the prostate cancer rather than perfusion-related diffusion.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
Our results supplement those of recent studies showing that tumor cellularity is related to tumor aggressiveness [15]. Gibbs et al. found that mean cell density increased from 14.5% for prostate cancer with grade 6 to 21.9% for prostate cancer with GS 8 or over [25]. Similarly, Wang et al. found that ADC correlated inversely with tumor cellularity [26]. Moreover, Langer et al. found that ADC and T2 were inversely related to percentage area for both nuclei and cytoplasm, and positively related to percentage area of luminal space [27]. Conversely, in our results, flow-related diffusion parameters such as D and f were not significantly correlated with Gleason score. In a similar finding, Oto et al. reported that Gleason score was not significantly correlated with any examined DCE-MRI parameters [15]. Similarly, Huellner et al. found only a weak-to-moderate correlation between Gleason score and CT-perfusion parameters such as blood flow and blood volume [28]. However, several studies suggest that decreased perfusion fraction may be involved in reducing ADC in prostate cancer [20-23]. Interestingly, some studies revealed positive correlations between Gleason score and DCE-MRI parameters [29, 30]. Thus, the relationship between flowrelated diffusion parameters and Gleason score remains controversial. In our study, qualitative ROC curve analysis revealed that ADC and D distinguished lowgrade tumors (Gleason score 6) from intermediate/high-grade tumors (Gleason scores 7 and 8) (AUC range, 0.956–0.960). The cutoff values chosen to optimize the diagnostic accuracy of the parameters were ADCfit < 1.29 x 10-3mm2/s (sensitivity, 92%; specificity, 85.4%); and D < 1.18 x 10-3mm2/s (sensitivity, 96%; specificity, 82.9%). Using ROC curves, D might be useful for discriminating low-grade tumors from intermediate/high-grade tumor based on its high sensitivity of 96% and specificity of 82.9%. These results are similar to a previous study [24]. This study had some limitations. First, only 10 patients underwent radical prostatectomy. The remaining 31 patients were diagnosed by systemic biopsy alone. Thus, we cannot fully exclude the formal possibility that Gleason scores from prostate biopsy differed from grades from radical prostatectomy. In addition, the blind nature of transrectal sonography-guided biopsy may have resulted in the lack of correlation between lesion localization and pathological results. However, we matched IVIM parameters and histological findings using multiparametric MR imaging including T2WI, DWI and DCE in 31 patients. Multiparametric MR imaging is the most accurate imaging technique for prostate cancer detection [31-33]. Second, in our study, tumor sizes were larger and mean PSA levels were higher than in previous studies [13. 24] because of different inclusion criteria. Previous studies included only patients who underwent radical prostatectomy. We enrolled patients who did not undergo radical prostatectomy because we included patients with metastases to the lymph nodes or bone. Therefore, our patients had higher-risk cancers (Gleason scores 8–10) compared with previous studies [13. 24]. Third, we excluded six patients whose lesions were not visible on T2WI, DCE and DWI images because we could not correlate their biopsy results with the visible lesions. This might have resulted in selection bias because five of these six patients had low Gleason scores (3 + 3). Fourth, we used eight b values, although the number and choice of b values best suited for prostate cancer analyses are unknown. Studies using 10 or more b values provide informative analysis; however, this number was not clinically feasible for our study. In practice, six to eight b values are clinically acceptable [34]. Fifth, D showed poor quality due to a low signal-to-noise ratio. Further improvement of the IVIM analysis algorithm and DWI technique is needed for precise estimates of D. In conclusion, ADCfit and D strongly correlated with Gleason scores for prostate cancer. The best IVIM parameter for discriminating prostate cancers with low Gleason scores from cancers with intermediate or high Gleason scores was D, which expresses pure molecular diffusion.
ACCEPTED MANUSCRIPT References
AC
CE
PT
ED
MA NU
SC
RI P
T
1. Jemal A, Center MM, DeSantis C, Ward EM. Global patterns of cancer incidence and mortality rates and trends. Cancer Epidemiol Biomarkers Prev 2010;19:1893-1907 2. Herman CM, Kattan MW, Ohori M, Scardino PT, Wheeler TM. Primary Gleason pattern as a predictor of disease progression in Gleason score 7 prostate cancer: a multivariate analysis of 823 men treated with radical prostatectomy. Am J Surg Pathol 2001;25:657-60. 3. Cookson MS, Fleshner NE, Soloway SM, Fair WR. Correlation between Gleason score of needle biopsy and radical prostatectomy specimen: accuracy and clinical implications. J Urol 1997;157:559-62 4. Divrik RT, Eroglu A, Sahin A, Zorlu F, Ozen H. Increasing the number of biopsies increases the concordance of Gleason scores of needle biopsies and prostatectomy specimens. Urol Oncol 2007;25:376-82 5. Tamada T, Sone T, Jo Y, Toshimitsu S, Yamashita T, Yamamoto A, et al. Apparent diffusion coefficient values in peripheral and transition zones of the prostate: comparison between normal and malignant prostatic tissues and correlation with histologic grade. J Magn Reson Imaging 2008;28:720-6. 6. Hambrock T, Somford DM, Huisman HJ, van Oort I, Witjes JA, Hulsbergen-van de Kaa CA, et al. Relationship between apparent diffusion coefficients at 3.0-T MR imaging and Gleason grade in peripheral zone prostate cancer. Radiology 2011;259:453-61. 7. Peng Y, Jiang Y, Yang C, Brown JB, Antic T, Sethi I, et al. Quantitative analysis of multiparametric prostate MR images: differentiation between prostate cancer and normal tissue and correlation with Gleason score--a computer-aided diagnosis development study. Radiology 2013;267:787-96. 8. Woodfield CA, Tung GA, Grand DJ, Pezzullo JA, Machan JT, Renzulli JF 2nd Diffusion-weighted MRI of peripheral zone prostate cancer: comparison of tumor apparent diffusion coefficient with Gleason score and percentage of tumor on core biopsy. AJR Am J Roentgenol 2010;194:W316-22. 9. Nagarajan R, Margolis D, Raman S, Sarma MK, Sheng K, King CR, et al. MR spectroscopic imaging and diffusion-weighted imaging of prostate cancer with Gleason scores. J Magn Reson Imaging 2012;36:697-703. 10. Verma S, Rajesh A, Morales H, Lemen L, Bills G, Delworth M, et al. Assessment of aggressiveness of prostate cancer: correlation of apparent diffusion coefficient with histologic grade after radical prostatectomy. AJR Am J Roentgenol 2011;196:374-81. 11. Itou Y, Nakanishi K, Narumi Y, Nishizawa Y, Tsukuma H. Clinical utility of apparent diffusion coefficient (ADC) values in patients with prostate cancer: can ADC values contribute to assess the aggressiveness of prostate cancer? J Magn Reson Imaging 2011;33:167-72. 12. Peng Y, Jiang Y, Antic T, Giger ML, Eggener SE, Oto A. Validation of quantitative analysis of multiparametric prostate MR images for prostate cancer detection and aggressiveness assessment: a cross-imager study. Radiology 2014;271:461-71. 13. Vargas HA, Akin O, Franiel T, Mazaheri Y, Zheng J, Moskowitz C, et al. Diffusionweighted endorectal MR imaging at 3 T for prostate cancer: tumor detection and assessment of aggressiveness. Radiology 2011;259:775-84. 14. Jung SI, Donati OF, Vargas HA, Goldman D, Hricak H, Akin O. Transition zone prostate cancer: incremental value of diffusion-weighted endorectal MR imaging in tumor detection and assessment of aggressiveness. Radiology 2013;269:493-503. 15. Oto A, Yang C, Kayhan A, Tretiakova M, Antic T, Schmid-Tannwald C, et al. Diffusion-weighted and dynamic contrast-enhanced MRI of prostate cancer: correlation of quantitative MR parameters with Gleason score and tumor angiogenesis. AJR Am J Roentgenol 2011;197:1382-90. 16. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Jeantet ML. Separation of
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168:497-505. 17. Le Bihan D, Turner R, MacFall JR. Effects of intravoxel incoherent motions (IVIM) in steady-state free precession (SSFP) imaging: application to molecular diffusion imaging. Magn Reson Med 1989;10:324-37. 18. Turner R, Le Bian D, Maier J, Vavrek R, Hedges LK, Pekar J. Echo-planar imaging of intravoxel incoherent motions. Radiology 1990;177:407-14. 19. Dixon WT. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging: a modest proposal with tremendous potential. Radiology 1988;168:566-7. 20. Shinmoto H, Tamura C, Soga S, Shiomi E, Yoshihara N, Kaji T, et al. An intravoxel incoherent motion diffusion-weighted imaging study of prostate cancer. AJR Am J Roentgenol 2012;199:W496-500. 21. Pang Y, Turkbey B, Bernardo M, Kruecker J, Kadoury S, Merino ML, et al. Intravoxel incoherent motion MR imaging for prostate cancer: an evaluation of perfusion fraction and diffusion coefficient derived from different b-value combinations. Magn Reson Med 2013;69:553-62. 22. Döpfert J, Lemke A, Weidner A, Schad LR. Investigation of prostate cancer using diffusion-weighted intravoxel incoherent motion imaging. Magn Reson Imaging 2011;29:1053-8. 23. Kuru TH, Roethke MC, Stieltjes B, Maier-Hein K, Schlemmer HP, Hadaschik BA, et al. Intravoxel Incoherent Motion (IVIM) Diffusion Imaging in Prostate Cancer - What Does It Add? J Comput Assist Tomogr 2014;38:558-64. 24. Zhang YD, Wang Q, Wu CJ, Wang XN, Zhang J, Liu H, et al. The histogram analysis of diffusion-weighted intravoxel incoherent motion (IVIM) imaging for differentiating the Gleason grade of prostate cancer. Eur Radiol 2015;25:994-1004. 25. Gibbs P, Liney GP, Pickles MD, Zelhof B, Rodrigues G, Turnbull LW. Correlation of ADC and T2 measurements with cell density in prostate cancer at 3.0 Tesla. Invest Radiol 2009;44:572-76. 26. Wang XZ, Wang B, Gao ZQ, Liu JG, Liu ZQ, Niu QL, et al. Diffusion-weighted imaging of prostate cancer: correlation between apparent diffusion coefficient values and tumor proliferation. J Magn Reson Imaging 2009;29:1360-6. 27. Langer DL, van der Kwast TH, Evans AJ, Plotkin A, Trachtenberg J, Wilson BC, et al. Prostate tissue composition and MR meaurements: investigating the relationships between ADC, T2, K-trans, v(e), and corresponding histologic features. Radiology 2010;255:485-94. 28. Huellener MW, Mattei A, Ross S, Butea-Bocu M, Vosbeck J, Pauli C, et al. Integrated CT-perfusion shows no meaningful correlation with PSA and presurgical Gleason score in patients with early prostate cancer. Clinical Imaging 2014;38:850-7. 29. Vos EK, Litjens GJ, Kobus T, Hambrock T, Hulsbergen-van de Kaa CA, Barentsz JO, et al. Assessment of prostate cancer aggressiveness using dynamic contrastenhanced magnetic resonance imaging at 3T. Eur Urol 2013;64:448-55. 30. Moradi M, Salcudean SE, Chang SD, Jones EC, Buchan N, Casey RG, et al. Multiparametric MRI maps for detection and grading of dominant prostate tumors. J Magn Reson Imaging 2012;35;1403-13. 31. Padhani AR, Miles KA. Multiparametric imaging of tumor response to therapy. Radiology 2010;256:348-64. 32. Cornud F, Delongchamps NB, Mozer P, Beuvon F, Schull A, Muradyan N, et al. Value of multiparametric MRI in the work-up of prostate cancer. Curr Urol Rep 2012;13:82-92. 33. Langer DL, van der Kwast TH, Evans AJ, Trachtenberg J, Wilson BC, Haider MA. Prostate cancer detection with multiparametric MRI; logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast-enhanced MRI. J Magn Reson Imaging 2009;30:327-34.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
34. Koh DM, Collins DJ, Orton MR. Intravoxel incoherent motion on body diffusionweighted MRI: reality and challenges . AJR Am J Roentgenol 2011;196:1351-81.
ACCEPTED MANUSCRIPT Figure Legends
RI P
T
Fig. 1. Parametric maps generated from IVIM MR images of a 75-year-old man with Gleason score 7 (4 + 3) prostate cancer and PSA 12 ng/ml. (a) T2-weighted axial image showing a focal, hypointense lesion in the peripheral zone of the middle of the prostate. (b) ADC map showing diffusion restriction. White outline, ROI (ADC = 0.83 x 10-3 mm2/sec). (c-e) D, 0.66 x 10-3 mm2/sec; D, 36.63 x 10-3 mm2/sec, and f, 19.7%. (f) Photomicrograph of specimen showing prostate cancer in left peripheral zone. (Hematoxylin-eosin stain; original magnification, X 1).
ED
MA NU
SC
Fig. 2. Box-whisker plots of ADCfit, D, f, and D for cancer lesions according to Gleason score (6, 7, and 8 or higher). Top of each box, 25th percentile; bottom 75th percentile. Box lengths represent interquartile range within the 50th percentile. Horizontal line inside boxes indicate medians. Points outside boxes are outliers, defined as values smaller than the lower quartile minus 1.5 times the interquartile range or larger than the upper quartile plus 1.5 times the interquartile range. (a, b) ADCfit and D for prostate cancers with Gleason scores 7 or 8 were significantly lower than for prostate cancers with Gleason score 6 (P < 0.05). Pearson’s coefficient analysis revealed significant negative correlation between Gleason score and ADCfit (r = -0.490, P = 0.001) and Gleason score and D value (r = -0.514, P = 0.001). (c, d) D and f were not significantly different among prostate cancers with Gleason scores 6, 7, and 8 (P > 0.05). Gleason score was poorly correlated with f (r = 0.168, P = 0.292) and D (r = -0.108, P = 0.500).
AC
CE
PT
Fig. 3. ROC curves. Curves assessed the ability of IVIM parameters to discriminate prostate cancers with Gleason score 6 from prostate cancers with Gleason scores 7 or 8. ADCfit and D yielded the highest Az value (0.960–0.956); f yielded the lowest Az value (0.633).
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1a
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1b
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1c
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1d
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1e
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1f
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2a
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 2b
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2c
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2d
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 3
ACCEPTED MANUSCRIPT Table 1. Patient Characteristics Value
Age (y)
71 (50-86)*
Prostate specific antigen (mg/mL)
21 (3.9-84.8)*
Tumor size (mL)
5.3 (0.5-28.2)*
Time between MR imaging and biopsy (d) (n=41)
7 (1-43)*
Gleason Score
MA NU
Biopsy only (n=31) 3+3 3+4 4+3
4+5
PT
5+4 5+5 Prostatectomy (n=10) 3+3
CE
AC
4+4
ED
4+4
4+3
28 (6-52)*
SC
Time between MR imaging and prostatectomy (d) (n=10)
3+4
RI P
T
Characteristics
7 (22%) 4 (13%) 5 (16%) 6 (19%) 5 (16%) 2 (7%) 2 (7%) 2 (20%) 5 (50%) 2 (20%) 1 (10%)
Pathologic characteristics (n=10) 1T2b
1 (10%)
1T2c
5 (50%)
1T3a
3 (30%)
1T3b
1 (10%)
*Data are mean with range
ACCEPTED MANUSCRIPT Table 2. Mean ADCfit and D, D*, and f parameters accessed in 41 patients
ADCfit (×10-3mm²/sec) -3
†
1.13 0.20
0.92 0.25
†
1.03 0.19
0.82 0.20
D
(×10 mm²/sec)
D*
(×10 mm²/sec)
37.24 11.78
f
(%)
12.71 3.71
-3
GS 8 ≦ (n=16)
GS 7 (n=16)
0.82 0.20
T
GS 6 (n=9)
p value 0.006
0.73 0.18
0.004
41.07 16.80
41.39 10.69
0.550
13.99 4.78
11.87 3.48
0.423
RI P
Parameters
SC
Data are expressed in mean SD ADC apparent diffusion coefficient, D pure molecular diffusion, D* pseudo diffusion, f perfusion fraction, n number of cases †
AC
CE
PT
ED
MA NU
The post-hoc test showed statistically significant differences between GS 6 and GS 7 or GS 8≦ after the Bonferroni correction (p< 0.05).
ACCEPTED MANUSCRIPT Table 3. Diagnostic characteristics of ADCfit and D, D*, and f parameters in the differentiating prostate cancer with GS 6 from prostate cancer with ≧ GS 7
92.0
96.0
Specificity (%)
85.4
82.9
PPV (%)
79.3
77.4
NPV (%)
94.6
97.1
88.7
89.5
Accuracy (%)
0.960
0.956
MA NU
AUC (%)
SC
Sensitivity (%)
D* -3 (>27.97×10 mm²/sec)
T
D -3 (≦1.18×10 mm²/sec)
RI P
ADCfit -3 (≦1.29×10 mm²/sec)
f (≦18.42%)
88.0
84.0
48.8
39.0
51.2
45.7
87.0
80.0
68.4
61.5
0.725
AC
CE
PT
ED
ADC apparent diffusion coefficient, D pure molecular diffusion, D* pseudo diffusion, f perfusion fraction PPV positive predictive value NPV negative predictive value AUC area under the curve
0.633
S0899-7071(16)00002-4 doi: 10.1016/j.clinimag.2016.01.001 JCT 7976
To appear in:
Journal of Clinical Imaging
Received date: Revised date: Accepted date:
13 October 2015 4 December 2015 6 January 2016
Please cite this article as: Yang Dal Mo, Kim Hyun Cheol, Kim Sang Won, Jahng Geon-Ho, Won Kyu Yeoun, Lim Sung Jig, Oh Jang-Hoon, Prostate cancer: correlation of intravoxel incoherent motion MR parameters with Gleason score, Journal of Clinical Imaging (2016), doi: 10.1016/j.clinimag.2016.01.001
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Prostate Cancer: Correlation of Intravoxel Incoherent Motion MR Parameters with Gleason Score Dal Mo Yang1*, Hyun Cheol Kim1, Sang Won Kim1, Geon-Ho Jahng1, Kyu Yeoun Won2, Sung Jig
RI P
T
Lim2, Jang-Hoon Oh3 1
Department of Radiology, Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea
2
SC
Department of Pathology, Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea
3
University, Yongin, Republic of Korea *
MA NU
Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee
Corresponding author. Department of Radiology, Kyung Hee University Hospital at Gangdong, 149,
Sangil-Dong, GAngdong-GU, Seoul, 134-727 Republic of Korea
AC
CE
PT
ED
E-mail: [email protected]
ACCEPTED MANUSCRIPT Prostate Cancer: Correlation of Intravoxel Incoherent Motion MR Parameters with Gleason Score
PT
ED
MA NU
SC
RI P
T
Purpose: To evaluate the potential of IVIM imaging to predict histological prognostic parameters by investigating whether IVIM parameters correlate with Gleason score. Materials and methods: The Institutional Review Board approved this retrospective study and informed consent was waived. A total of 41 patients with histologically proven prostate cancer who underwent prostate MR imaging using a 3T MRI machine were included. For 8 DWI b-values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2), a spin-echo echo-planar imaging (EPI) sequence was performed. D, f, D, and ADCfit values were compared among three groups of patients with prostate cancer: Gleason score 6 (n = 9), 7 (n = 16), or 8 or higher (n = 16). Receiver operating characteristic (ROC) curves were generated for D, f, D, and ADCfit to assess the ability of each parameter to distinguish cancers with low Gleason scores from cancers with intermediate or high Gleason scores. Results: Pearson’s coefficient analysis revealed significant negative correlations between Gleason score and ADCfit (r = -0.490, P = 0.001) and Gleason score and D values (r = -0.514, P = 0.001). Gleason score was poorly correlated with f (r = 0.168, P = 0.292) and D values (r = -0.108, P = 0.500). The ADCfit and D values of prostate cancers with Gleason scores 7 or 8 were significantly lower than values for prostate cancers with Gleason score 6 (P < 0.05). ROC curves were constructed to assess the ability of IVIM parameters to discriminate prostate cancers with Gleason score 6 from cancers with Gleason scores 7 or 8. Areas under the curve were 0.671 to 0.974. ADCfit and D yielded the highest Az value (0.960– 0.956), whereas f yielded the lowest Az value (0.633). Conclusions: The pure molecular diffusion parameter, D, was the IVIM parameter that best discriminated prostate cancers with low Gleason scores from prostate cancers with intermediate or high Gleason scores.
AC
CE
Keywords: Magnetic resonance imaging Diffusion-weighted imaging Incoherent intravoxel motion MR imaging Prostate Prostate cancer
ACCEPTED MANUSCRIPT 1.Introduction
PT
ED
MA NU
SC
RI P
T
Prostate cancer is the second most common cancer among men worldwide; moreover, the incidence of prostate cancer is rising in Asia and Eastern Europe [1]. The accurate assessment of prostate cancer aggressiveness is important for deciding the most appropriate initial treatment strategy. The Gleason score is an important preoperative predictor of response to treatment and patient outcomes [2]. However, biopsy determination of the Gleason grade is invasive and often does not accurately reflect the final Gleason score [3, 4]. Recently, functional MRI sequences such as DWI and dynamic contrast-enhanced (DCE) imaging have been shown to provide information about tumor aggressiveness [5-7]. The apparent diffusion coefficient (ADC) is negatively correlated with the Gleason score for prostate cancer [8-14]. Specifically, patients with high Gleason scores have tumors with lower ADCs. This relationship is likely because tumors with high Gleason scores exhibit increased cellularity [15]. Intravoxel incoherent motion (IVIM) MR imaging quantitatively assesses the microscopic translational motion that occurs in each image voxel on MR imaging scans [16-19]. Both pure molecular diffusion (D) and microcirculation-related diffusion, otherwise known as pseudodiffusion (D*), can be distinguished using IVIM MR imaging, as can the blood flow fraction (f) [17]. Recent studies with IVIM imaging showed that reduced ADCs for prostate cancer are due to decreased molecular diffusion and to a decreased perfusion fraction [2023]. However, studies on the relationship between IVIM parameters and Gleason scores are rare [24]. Therefore, this study evaluated the potential of IVIM imaging to predict histological and prognostic parameters by investigating whether IVIM parameters correlated with Gleason score. 2.Materials and Methods
CE
2.1.Patients
AC
Our Institutional Review Board approved this retrospective study; the requirement for informed patient consent was waived. We searched the archive of MR images at our institution and identified 47 consecutive patients who were seen between October 2012 and March 2014, had biopsy-proven prostate cancer, and underwent prostate MR imaging on a 3T MRI machine. Six patients were excluded because their lesions were not seen on T2weighted images, DCE and DWI. Gleason scores were 3 + 3 for five patients and 4 + 3 for one patient. Thus, the final study population comprised 41 patients (mean age, 71 years; range, 50–86 years). Of the 41 patients, 27 patients had the prostate cancers at only peripheral zone. In the remaining 14 patients, two had the prostate cancers at the transitional zone and 12 had the prostate cancers at both the peripheral zone and the transitional zone. The mean PSA level of the 41 patients was 21 ng/mL (range, 3.9–84.8 ng/mL). Prostate cancer was proven in 10 patients via radical prostatectomy and in 31 via systemic transrectal sonography-guided biopsy. The 31 patients diagnosed by systemic biopsy alone had 10–16 biopsies per session. Gleason scores were 6 for 9 patients, 7 for 16, 8 for 7, 9 for 7, and 10 for 2. A summary of patient characteristics is in Table 1. 2.2.MR imaging techniques MR imaging was performed with a 3-Tesla system (Achiva; Philips Health Care, Eindhoven, The Netherlands) with a six-channel sensitivity-encoding (SENSE) torso coil. All patients were initially examined with a routine MRI protocol for imaging the prostate, which included T2-weighted, T1-weighted, DCE and multi-b-value DWI images. For axial and coronal T2-weighted images, fast spin-echo (FSE) MR imaging was
ACCEPTED MANUSCRIPT
MA NU
SC
RI P
T
performed with the following parameters: repetition time (TR)/echo time (TE): 3500/90 ms; matrix: 340 × 240; field of view (FOV): 18 × 18 cm; acquisition of two signals; and section thickness: 3 mm with a 0.3 mm section gap. For sagittal T2-weighted images, FSE MR imaging was performed with the following parameters: TR/TE: 3500/90 ms; matrix: 460 × 320; FOV: 24 × 24 cm; acquisition of two signals; and section thickness: 5 mm with no section gap. For T1-weighted images, FSE MR imaging was performed with the following parameters: 550/10 ms; flip angle: 90°; matrix: 340 × 210; FOV: 18 × 18 cm; acquisition of two signals; and section thickness: 5 mm with a 2 mm section gap. DCE imaging after intravenous injection of 0.2 ml/kg of gadoterate meglumine (Doterem; Guerbet, Roissy, France) at 3 ml/s was performed with the following parameters: 3.1/1.2 ms; matrix: 240 x 220; FOV: 24 x 24 cm; acquisition of one signal; section thickness 4.8 mm with 2.4 mm section gap; number of temporal acquisitions: 48; and temporal resolutions: 5.5 s. For the eight DWI b-values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2), a spin-echo echo-planar imaging sequence was performed with the following parameters: TR/TE: 5000/90 ms; matrix: 112 × 100; FOV: 22 × 22 cm; acquisition of two signals; section thickness: 5 mm with 2 mm section gap; a transverse plane; SENSE parallel imaging factor: 2; and acquisition time: 3 minutes 45 seconds.
2.3.Image analysis
AC
CE
PT
ED
Three experienced radiologists (D.M.Y., H.C.K., and S.W.K.) retrospectively reviewed the MR images using a picture archiving and communications system. For each patient with prostate cancer, the three radiologists came to a consensus on a representative region to define as the cancer region-of-interest (ROI) on pathological reports and datasets of parametric T2WI, DCE, and multi-b-value DWI MR images. In details, the three radiologists reviewed the pathological reports to define the prostate cancer area. After that, MR images of T2WI, DCE, and DWI, and the corresponding ADC maps were reviewed to identify the prostate cancer area which was matched with the pathological report. Each radiologist had more than 10 years of experience in performing prostate MRI and interpreting prostate MR images. One radiologist (D.M.Y.) defined a three-dimensional ROI for each prostatic lesion using MRIcro software (http://www.cabiatl.com/micro/) to obtain a DWI signal intensity (SI) for the lesion (Fig. 1). ROIs were manually drawn as large as possible on ADC maps and positioned to avoid areas of necrosis and hemorrhage. Two different methods were used to obtain diffusion or flow-related parameters. In the first method, the ADC of each ROI was calculated by fitting the eight DWI SI values from eight different b values (0, 10, 20, 50, 100, 200, 500, and 800 sec/mm2) to the following equation: SI/SI0 = Exp (-bADCfit), where ADCfit was the ADC obtained from the exponential fitting of DWI signal intensities. ADCfit was calculated using linear regression analysis with the natural logarithm of the intensity and the variable b from the eight images and was expressed in mm2/sec. Second, the flow rate, true diffusion coefficient, and pseudodiffusion coefficient were calculated according to Le Bihan et al [16] as follows: SI/SI0 = (1 – f) x exp (-bD) + f x exp (-b D), where D and D were the true diffusion coefficient and the pseudodiffusion coefficient, respectively, and f was the fractional volume occupied in the voxel by flowing spins. Using the Levenberg-Marquardt nonlinear least-squares algorithm, D, f, and D were calculated for each ROI of the prostatic lesions (Fig. 1). Eight b-values (0, 10, 20, 50, 100, 200, 500 and 800 sec/mm2) were used. D and D* were in mm2/sec and f was in percent (%). D was the true diffusion coefficient and represented the slow component of diffusion (i.e., pure molecular diffusion). D was the flow-related diffusion coefficient and represented the fast component of diffusion (i.e., via incoherent microcirculation).
ACCEPTED MANUSCRIPT
2.4.Statistical analysis
SC
RI P
T
The Kruskal-Wallis test was used to compare the three IVIM parameters (D, f, and D) and ADCfit among prostate cancer lesions with Gleason scores of 6, 7, or 8. Post hoc analysis with the Bonferroni correction was performed for all parameters identified as significant by the Kruskal-Wallis test. Spearman’s correlation analysis was used to characterize the association of ADCfit, D, f, and D for tumors with their corresponding Gleason scores. Receiver operating characteristics (ROC) curves were generated to determine the optimal cutoff values for significant IVIM parameters to discriminate low-grade from intermediate/high-grade lesions. All analyses were performed with SPSS software (version 18, Chicago, IL, USA). Probability less than 0.05 was considered statistically significant.
MA NU
3.Results
AC
CE
PT
ED
Mean ADCfit, D, f, and D for prostate cancer lesions with Gleason scores of 6, 7, and 8 are in Table 2. Pearson’s coefficients revealed a significant negative correlation between Gleason score and ADCfit (r = -0.490, p = 0.001) and Gleason score and D values (r = -0.514, p = 0.001) (Figs. 2a, 2b). In contrast, Gleason score correlated poorly with f (r = 0.168, P = 0.292) and D (r = -0.108, p = 0.500) (Figs. 2c, 2d). ADCfit and D for the prostate cancers with Gleason scores of 7 or 8 were significantly lower than ADCfit and D for cancers with Gleason score 6 (p < 0.05) (Figs. 2a, 2b). No significant differences in D or f were observed among prostate cancers with Gleason scores 6, 7, or 8 (p > 0.05) (Figs. 2c, 2d). ROC curves of the IVIM parameters yielded AUCs from 0.671 to 0.974 for discriminating prostate cancers with Gleason score 6 from prostate cancers with Gleason scores of 7 or 8 (Table 3). ADCfit and D yielded high Az values (0.960–0.956), whereas f yielded the lowest Az value (0.633) (Fig. 3). To optimize the diagnostic accuracy of each parameter, cutoff values of ADCfit < 1.29 x 10-3 mm2/s, D < 1.18 x 10-3 mm2/s, D < 27.97 x 10-3 mm2/s, and f < 18.42% were used. Diagnostic characteristics based on these cutoffs are in Table 3.
4.Discussion
In this analysis, we found that Gleason scores exhibited a significant negative correlation with ADCfit. In addition, ADCfit values for prostate cancers with Gleason scores of 7 or 8 were significantly lower than for prostate cancers with Gleason score 6. Thus, tumors with higher Gleason scores had lower ADCs. These results were consistent with previous studies [8-14]. Based on the IVIM model, the reduction of ADCs in prostate cancer was due to a decrease in molecular diffusion as well as partly decreased perfusion fraction [20-23]. In our study, Gleason scores were significantly negatively correlated with D value. In addition, D values of prostate cancers with Gleason scores of 7 or 8 were significantly lower than prostate cancers with Gleason score 6. However, Gleason scores were poorly correlated with both f and D values, which are related to perfusion-related diffusion. In addition, f values were not significantly different among prostate cancers with Gleason scores of 6, 7, and 8. These findings were in accordance with previous results [24]. Zhang et al. found that D values can be useful for discriminating low-grade tumors from intermediate/high-grade tumors [24]. However, no significant difference was observed between low-grade tumors and intermediate/high-grade tumor in analyses with both f and D [24].These results suggest that low ADCs observed in prostate cancers with high Gleason scores result from pure molecular diffusion of the prostate cancer rather than perfusion-related diffusion.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
Our results supplement those of recent studies showing that tumor cellularity is related to tumor aggressiveness [15]. Gibbs et al. found that mean cell density increased from 14.5% for prostate cancer with grade 6 to 21.9% for prostate cancer with GS 8 or over [25]. Similarly, Wang et al. found that ADC correlated inversely with tumor cellularity [26]. Moreover, Langer et al. found that ADC and T2 were inversely related to percentage area for both nuclei and cytoplasm, and positively related to percentage area of luminal space [27]. Conversely, in our results, flow-related diffusion parameters such as D and f were not significantly correlated with Gleason score. In a similar finding, Oto et al. reported that Gleason score was not significantly correlated with any examined DCE-MRI parameters [15]. Similarly, Huellner et al. found only a weak-to-moderate correlation between Gleason score and CT-perfusion parameters such as blood flow and blood volume [28]. However, several studies suggest that decreased perfusion fraction may be involved in reducing ADC in prostate cancer [20-23]. Interestingly, some studies revealed positive correlations between Gleason score and DCE-MRI parameters [29, 30]. Thus, the relationship between flowrelated diffusion parameters and Gleason score remains controversial. In our study, qualitative ROC curve analysis revealed that ADC and D distinguished lowgrade tumors (Gleason score 6) from intermediate/high-grade tumors (Gleason scores 7 and 8) (AUC range, 0.956–0.960). The cutoff values chosen to optimize the diagnostic accuracy of the parameters were ADCfit < 1.29 x 10-3mm2/s (sensitivity, 92%; specificity, 85.4%); and D < 1.18 x 10-3mm2/s (sensitivity, 96%; specificity, 82.9%). Using ROC curves, D might be useful for discriminating low-grade tumors from intermediate/high-grade tumor based on its high sensitivity of 96% and specificity of 82.9%. These results are similar to a previous study [24]. This study had some limitations. First, only 10 patients underwent radical prostatectomy. The remaining 31 patients were diagnosed by systemic biopsy alone. Thus, we cannot fully exclude the formal possibility that Gleason scores from prostate biopsy differed from grades from radical prostatectomy. In addition, the blind nature of transrectal sonography-guided biopsy may have resulted in the lack of correlation between lesion localization and pathological results. However, we matched IVIM parameters and histological findings using multiparametric MR imaging including T2WI, DWI and DCE in 31 patients. Multiparametric MR imaging is the most accurate imaging technique for prostate cancer detection [31-33]. Second, in our study, tumor sizes were larger and mean PSA levels were higher than in previous studies [13. 24] because of different inclusion criteria. Previous studies included only patients who underwent radical prostatectomy. We enrolled patients who did not undergo radical prostatectomy because we included patients with metastases to the lymph nodes or bone. Therefore, our patients had higher-risk cancers (Gleason scores 8–10) compared with previous studies [13. 24]. Third, we excluded six patients whose lesions were not visible on T2WI, DCE and DWI images because we could not correlate their biopsy results with the visible lesions. This might have resulted in selection bias because five of these six patients had low Gleason scores (3 + 3). Fourth, we used eight b values, although the number and choice of b values best suited for prostate cancer analyses are unknown. Studies using 10 or more b values provide informative analysis; however, this number was not clinically feasible for our study. In practice, six to eight b values are clinically acceptable [34]. Fifth, D showed poor quality due to a low signal-to-noise ratio. Further improvement of the IVIM analysis algorithm and DWI technique is needed for precise estimates of D. In conclusion, ADCfit and D strongly correlated with Gleason scores for prostate cancer. The best IVIM parameter for discriminating prostate cancers with low Gleason scores from cancers with intermediate or high Gleason scores was D, which expresses pure molecular diffusion.
ACCEPTED MANUSCRIPT References
AC
CE
PT
ED
MA NU
SC
RI P
T
1. Jemal A, Center MM, DeSantis C, Ward EM. Global patterns of cancer incidence and mortality rates and trends. Cancer Epidemiol Biomarkers Prev 2010;19:1893-1907 2. Herman CM, Kattan MW, Ohori M, Scardino PT, Wheeler TM. Primary Gleason pattern as a predictor of disease progression in Gleason score 7 prostate cancer: a multivariate analysis of 823 men treated with radical prostatectomy. Am J Surg Pathol 2001;25:657-60. 3. Cookson MS, Fleshner NE, Soloway SM, Fair WR. Correlation between Gleason score of needle biopsy and radical prostatectomy specimen: accuracy and clinical implications. J Urol 1997;157:559-62 4. Divrik RT, Eroglu A, Sahin A, Zorlu F, Ozen H. Increasing the number of biopsies increases the concordance of Gleason scores of needle biopsies and prostatectomy specimens. Urol Oncol 2007;25:376-82 5. Tamada T, Sone T, Jo Y, Toshimitsu S, Yamashita T, Yamamoto A, et al. Apparent diffusion coefficient values in peripheral and transition zones of the prostate: comparison between normal and malignant prostatic tissues and correlation with histologic grade. J Magn Reson Imaging 2008;28:720-6. 6. Hambrock T, Somford DM, Huisman HJ, van Oort I, Witjes JA, Hulsbergen-van de Kaa CA, et al. Relationship between apparent diffusion coefficients at 3.0-T MR imaging and Gleason grade in peripheral zone prostate cancer. Radiology 2011;259:453-61. 7. Peng Y, Jiang Y, Yang C, Brown JB, Antic T, Sethi I, et al. Quantitative analysis of multiparametric prostate MR images: differentiation between prostate cancer and normal tissue and correlation with Gleason score--a computer-aided diagnosis development study. Radiology 2013;267:787-96. 8. Woodfield CA, Tung GA, Grand DJ, Pezzullo JA, Machan JT, Renzulli JF 2nd Diffusion-weighted MRI of peripheral zone prostate cancer: comparison of tumor apparent diffusion coefficient with Gleason score and percentage of tumor on core biopsy. AJR Am J Roentgenol 2010;194:W316-22. 9. Nagarajan R, Margolis D, Raman S, Sarma MK, Sheng K, King CR, et al. MR spectroscopic imaging and diffusion-weighted imaging of prostate cancer with Gleason scores. J Magn Reson Imaging 2012;36:697-703. 10. Verma S, Rajesh A, Morales H, Lemen L, Bills G, Delworth M, et al. Assessment of aggressiveness of prostate cancer: correlation of apparent diffusion coefficient with histologic grade after radical prostatectomy. AJR Am J Roentgenol 2011;196:374-81. 11. Itou Y, Nakanishi K, Narumi Y, Nishizawa Y, Tsukuma H. Clinical utility of apparent diffusion coefficient (ADC) values in patients with prostate cancer: can ADC values contribute to assess the aggressiveness of prostate cancer? J Magn Reson Imaging 2011;33:167-72. 12. Peng Y, Jiang Y, Antic T, Giger ML, Eggener SE, Oto A. Validation of quantitative analysis of multiparametric prostate MR images for prostate cancer detection and aggressiveness assessment: a cross-imager study. Radiology 2014;271:461-71. 13. Vargas HA, Akin O, Franiel T, Mazaheri Y, Zheng J, Moskowitz C, et al. Diffusionweighted endorectal MR imaging at 3 T for prostate cancer: tumor detection and assessment of aggressiveness. Radiology 2011;259:775-84. 14. Jung SI, Donati OF, Vargas HA, Goldman D, Hricak H, Akin O. Transition zone prostate cancer: incremental value of diffusion-weighted endorectal MR imaging in tumor detection and assessment of aggressiveness. Radiology 2013;269:493-503. 15. Oto A, Yang C, Kayhan A, Tretiakova M, Antic T, Schmid-Tannwald C, et al. Diffusion-weighted and dynamic contrast-enhanced MRI of prostate cancer: correlation of quantitative MR parameters with Gleason score and tumor angiogenesis. AJR Am J Roentgenol 2011;197:1382-90. 16. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Jeantet ML. Separation of
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168:497-505. 17. Le Bihan D, Turner R, MacFall JR. Effects of intravoxel incoherent motions (IVIM) in steady-state free precession (SSFP) imaging: application to molecular diffusion imaging. Magn Reson Med 1989;10:324-37. 18. Turner R, Le Bian D, Maier J, Vavrek R, Hedges LK, Pekar J. Echo-planar imaging of intravoxel incoherent motions. Radiology 1990;177:407-14. 19. Dixon WT. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging: a modest proposal with tremendous potential. Radiology 1988;168:566-7. 20. Shinmoto H, Tamura C, Soga S, Shiomi E, Yoshihara N, Kaji T, et al. An intravoxel incoherent motion diffusion-weighted imaging study of prostate cancer. AJR Am J Roentgenol 2012;199:W496-500. 21. Pang Y, Turkbey B, Bernardo M, Kruecker J, Kadoury S, Merino ML, et al. Intravoxel incoherent motion MR imaging for prostate cancer: an evaluation of perfusion fraction and diffusion coefficient derived from different b-value combinations. Magn Reson Med 2013;69:553-62. 22. Döpfert J, Lemke A, Weidner A, Schad LR. Investigation of prostate cancer using diffusion-weighted intravoxel incoherent motion imaging. Magn Reson Imaging 2011;29:1053-8. 23. Kuru TH, Roethke MC, Stieltjes B, Maier-Hein K, Schlemmer HP, Hadaschik BA, et al. Intravoxel Incoherent Motion (IVIM) Diffusion Imaging in Prostate Cancer - What Does It Add? J Comput Assist Tomogr 2014;38:558-64. 24. Zhang YD, Wang Q, Wu CJ, Wang XN, Zhang J, Liu H, et al. The histogram analysis of diffusion-weighted intravoxel incoherent motion (IVIM) imaging for differentiating the Gleason grade of prostate cancer. Eur Radiol 2015;25:994-1004. 25. Gibbs P, Liney GP, Pickles MD, Zelhof B, Rodrigues G, Turnbull LW. Correlation of ADC and T2 measurements with cell density in prostate cancer at 3.0 Tesla. Invest Radiol 2009;44:572-76. 26. Wang XZ, Wang B, Gao ZQ, Liu JG, Liu ZQ, Niu QL, et al. Diffusion-weighted imaging of prostate cancer: correlation between apparent diffusion coefficient values and tumor proliferation. J Magn Reson Imaging 2009;29:1360-6. 27. Langer DL, van der Kwast TH, Evans AJ, Plotkin A, Trachtenberg J, Wilson BC, et al. Prostate tissue composition and MR meaurements: investigating the relationships between ADC, T2, K-trans, v(e), and corresponding histologic features. Radiology 2010;255:485-94. 28. Huellener MW, Mattei A, Ross S, Butea-Bocu M, Vosbeck J, Pauli C, et al. Integrated CT-perfusion shows no meaningful correlation with PSA and presurgical Gleason score in patients with early prostate cancer. Clinical Imaging 2014;38:850-7. 29. Vos EK, Litjens GJ, Kobus T, Hambrock T, Hulsbergen-van de Kaa CA, Barentsz JO, et al. Assessment of prostate cancer aggressiveness using dynamic contrastenhanced magnetic resonance imaging at 3T. Eur Urol 2013;64:448-55. 30. Moradi M, Salcudean SE, Chang SD, Jones EC, Buchan N, Casey RG, et al. Multiparametric MRI maps for detection and grading of dominant prostate tumors. J Magn Reson Imaging 2012;35;1403-13. 31. Padhani AR, Miles KA. Multiparametric imaging of tumor response to therapy. Radiology 2010;256:348-64. 32. Cornud F, Delongchamps NB, Mozer P, Beuvon F, Schull A, Muradyan N, et al. Value of multiparametric MRI in the work-up of prostate cancer. Curr Urol Rep 2012;13:82-92. 33. Langer DL, van der Kwast TH, Evans AJ, Trachtenberg J, Wilson BC, Haider MA. Prostate cancer detection with multiparametric MRI; logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast-enhanced MRI. J Magn Reson Imaging 2009;30:327-34.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU
SC
RI P
T
34. Koh DM, Collins DJ, Orton MR. Intravoxel incoherent motion on body diffusionweighted MRI: reality and challenges . AJR Am J Roentgenol 2011;196:1351-81.
ACCEPTED MANUSCRIPT Figure Legends
RI P
T
Fig. 1. Parametric maps generated from IVIM MR images of a 75-year-old man with Gleason score 7 (4 + 3) prostate cancer and PSA 12 ng/ml. (a) T2-weighted axial image showing a focal, hypointense lesion in the peripheral zone of the middle of the prostate. (b) ADC map showing diffusion restriction. White outline, ROI (ADC = 0.83 x 10-3 mm2/sec). (c-e) D, 0.66 x 10-3 mm2/sec; D, 36.63 x 10-3 mm2/sec, and f, 19.7%. (f) Photomicrograph of specimen showing prostate cancer in left peripheral zone. (Hematoxylin-eosin stain; original magnification, X 1).
ED
MA NU
SC
Fig. 2. Box-whisker plots of ADCfit, D, f, and D for cancer lesions according to Gleason score (6, 7, and 8 or higher). Top of each box, 25th percentile; bottom 75th percentile. Box lengths represent interquartile range within the 50th percentile. Horizontal line inside boxes indicate medians. Points outside boxes are outliers, defined as values smaller than the lower quartile minus 1.5 times the interquartile range or larger than the upper quartile plus 1.5 times the interquartile range. (a, b) ADCfit and D for prostate cancers with Gleason scores 7 or 8 were significantly lower than for prostate cancers with Gleason score 6 (P < 0.05). Pearson’s coefficient analysis revealed significant negative correlation between Gleason score and ADCfit (r = -0.490, P = 0.001) and Gleason score and D value (r = -0.514, P = 0.001). (c, d) D and f were not significantly different among prostate cancers with Gleason scores 6, 7, and 8 (P > 0.05). Gleason score was poorly correlated with f (r = 0.168, P = 0.292) and D (r = -0.108, P = 0.500).
AC
CE
PT
Fig. 3. ROC curves. Curves assessed the ability of IVIM parameters to discriminate prostate cancers with Gleason score 6 from prostate cancers with Gleason scores 7 or 8. ADCfit and D yielded the highest Az value (0.960–0.956); f yielded the lowest Az value (0.633).
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1a
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1b
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1c
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1d
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1e
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1f
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2a
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 2b
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2c
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 2d
ED
MA NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 3
ACCEPTED MANUSCRIPT Table 1. Patient Characteristics Value
Age (y)
71 (50-86)*
Prostate specific antigen (mg/mL)
21 (3.9-84.8)*
Tumor size (mL)
5.3 (0.5-28.2)*
Time between MR imaging and biopsy (d) (n=41)
7 (1-43)*
Gleason Score
MA NU
Biopsy only (n=31) 3+3 3+4 4+3
4+5
PT
5+4 5+5 Prostatectomy (n=10) 3+3
CE
AC
4+4
ED
4+4
4+3
28 (6-52)*
SC
Time between MR imaging and prostatectomy (d) (n=10)
3+4
RI P
T
Characteristics
7 (22%) 4 (13%) 5 (16%) 6 (19%) 5 (16%) 2 (7%) 2 (7%) 2 (20%) 5 (50%) 2 (20%) 1 (10%)
Pathologic characteristics (n=10) 1T2b
1 (10%)
1T2c
5 (50%)
1T3a
3 (30%)
1T3b
1 (10%)
*Data are mean with range
ACCEPTED MANUSCRIPT Table 2. Mean ADCfit and D, D*, and f parameters accessed in 41 patients
ADCfit (×10-3mm²/sec) -3
†
1.13 0.20
0.92 0.25
†
1.03 0.19
0.82 0.20
D
(×10 mm²/sec)
D*
(×10 mm²/sec)
37.24 11.78
f
(%)
12.71 3.71
-3
GS 8 ≦ (n=16)
GS 7 (n=16)
0.82 0.20
T
GS 6 (n=9)
p value 0.006
0.73 0.18
0.004
41.07 16.80
41.39 10.69
0.550
13.99 4.78
11.87 3.48
0.423
RI P
Parameters
SC
Data are expressed in mean SD ADC apparent diffusion coefficient, D pure molecular diffusion, D* pseudo diffusion, f perfusion fraction, n number of cases †
AC
CE
PT
ED
MA NU
The post-hoc test showed statistically significant differences between GS 6 and GS 7 or GS 8≦ after the Bonferroni correction (p< 0.05).
ACCEPTED MANUSCRIPT Table 3. Diagnostic characteristics of ADCfit and D, D*, and f parameters in the differentiating prostate cancer with GS 6 from prostate cancer with ≧ GS 7
92.0
96.0
Specificity (%)
85.4
82.9
PPV (%)
79.3
77.4
NPV (%)
94.6
97.1
88.7
89.5
Accuracy (%)
0.960
0.956
MA NU
AUC (%)
SC
Sensitivity (%)
D* -3 (>27.97×10 mm²/sec)
T
D -3 (≦1.18×10 mm²/sec)
RI P
ADCfit -3 (≦1.29×10 mm²/sec)
f (≦18.42%)
88.0
84.0
48.8
39.0
51.2
45.7
87.0
80.0
68.4
61.5
0.725
AC
CE
PT
ED
ADC apparent diffusion coefficient, D pure molecular diffusion, D* pseudo diffusion, f perfusion fraction PPV positive predictive value NPV negative predictive value AUC area under the curve
0.633