ORIGINAL ARTICLE
Dynamic Contrast-Enhanced Magnetic Resonance Imaging for Pancreatic Ductal Adenocarcinoma at 3.0-T Magnetic Resonance: Correlation With Histopathology Kefu Liu, MD, PhD,*†‡ Ping Xie, MS,‡ Weijun Peng, MD, PhD,*† and Zhengrong Zhou, MD, PhD*† Purpose: The aim of this study was to discuss the correlation of quantitative dynamic contrast-enhanced magnetic resonance imaging (QDCEMRI) at 3.0-T magnetic resonance and histopathology for pancreatic ductal adenocarcinoma (PDA). Methods: Twenty-three patients with histopathologically proven PDA were included in this study after 75 cases of suspected pancreatic tumors had been performed by QDCE-MRI. The quantitative kinetic parameters analyzed by 2-compartment and 3-compartment models were calculated automatically, which included the volume transfer constant of the contrast agent, the rate constant (Kep), the volume as a percentage of the extravascular extracellular leakage space, the time of arrival of contrast agent, the time of peaking of contrast agent, the maximum slope of signal intensity ascent, and the contrast enhancement ratio. According to magnetic resonance images, tissue section were selected and stained for evaluating tumor differentiation, tumor fibrosis, tumor microvessel density, the expression of tumor vascular endothelial growth factor (VEGF) and Ki67. Subsequently, the relationship between the parameters of QDCE-MRI and histopathology of PDA was analyzed. Results: The tumor Kep and extravascular extracellular leakage space showed a statistically significant correlation with tumor fibrosis; the tumor volume transfer constant of the contrast agent 2-compartment showed a statistically significant correlation with the expressions of tumor VEGF; and the tumor Kep, maximum slope of signal intensity ascent, and contrast enhancement ratio showed a statistically significant correlation with the expression of tumor Ki67. Conclusions: The parameters of QDCE-MRI of PDA can be used to evaluate the degrees of tumor fibrosis and the expressions of VEGF and Ki67. Key Words: dynamic contrast-enhanced magnetic resonance imaging, Pancreatic ductal adenocarcinoma, Histopathology (J Comput Assist Tomogr 2015;39: 13–18)
Q
uantitative dynamic contrast-enhanced magnetic resonance imaging (QDCE-MRI) can be used as a new technique for assessing the function of a specific target tissue.1 Variables of QDCE-MRI are derived from a pharmacokinetic model analysis, and on the basis of the mathematical model used, they reflect the underlying perfusion and/or permeability of the target tissue. The 3-compartment (3C) model treats the vascular and the interstitial space as 2 separated compartments, whereas the 2-compartment
From the *Department of Radiology, Fudan University Shanghai Cancer Center, and †Department of Oncology, Shanghai Medical College, Fudan University, Shanghai; and ‡Department of Radiology, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, Jiangsu, China. Received for publication July 8, 2014; accepted September 18, 2014. Reprints: Zhengrong Zhou, MD, PhD, Department of Radiology, Cancer Hospital, Fudan University, No. 270, Dong An Rd, Shanghai 200032, China (e‐mail: [email protected]). This project was funded by Shanghai Important Technological Subject Fund of China (no. 08411954400) and Program for Shanghai Outstanding Academic Leader of China (no. 11XD1401600). The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
(2C) approach deals with the vascular and the interstitial space as a single entity.1 The past studies showed the parameters of QDCE-MRI are useful for the diagnosis of pancreatic ductal adenocarcinoma (PDA) and correlated well with glioma and breast cancer histopathology.2–5 However, the relationship between the parameters of QDCE-MRI and histopathology of PDA needs further study.
MATERIALS AND METHODS Ethics Statement The study was approved by the local institutional review board of Fudan University Shanghai Cancer Center, and all subjects provided written informed consent.
Patients Seventy-five patients with suspected pancreatic tumors underwent QDCE-MRI (5 patients had an unsuccessful QDCEMRI, of which 3 patients had contrast agent extravasation, and 2 patients had intolerance). In the 70 patients who underwent a successful QDCE-MRI, 23 patients with PDA proved by pathology were included in this study.
Magnetic Resonance Imaging Magnetic resonance studies were performed using a 3.0-T magnetic resonance scanner (Sigma HDx; GE Healthcare, Milwaukee, Wis) using a phased-array body coil. After a plain scan (T1 mapping), multiphase T1-weighted QDCE-MRI images were obtained using a spoiled gradient echo sequence (liver acquisition with volume acceleration) in the axial plane. Scan parameters were repetition time, 3.06 seconds; echo time, minimum; flip angle, 12; bandwidth, 83.33 kHz; field of view, 380 mm; slice thickness, 5 mm; and reconstruction matrix, 180 256. Multiple data sets were acquired every 4 seconds 45 times with a 4-second interval. An injection of 0.1-mmol/kg body weight of gadopentetate dimeglumine (Magnevist; Bayer Health-Care Pharmaceuticals Products, Berlin, Germany) was given at 2 mL/s using a power injector. The onset for contrast injection and data acquisition were trigged simultaneously. Images were transferred to a workstation (Advantage Workstation 4.3; GE Healthcare), and the quantitative kinetic parameters were analyzed with dedicated software (Cine Tool, GE Healthcare). The quantitative kinetic parameters analyzed by 2C and 3C models were calculated automatically, which included the volume transfer constant of the contrast agent (Ktrans), the rate constant (Kep), the volume as a percentage of the extravascular extracellular leakage space (Ve), the time of arrival of contrast agent (ArvIT ), the time of peaking of contrast agent (PeakT ), the maximum slope of signal intensity ascent (MxSIp), and the contrast enhancement ratio (CER). Regions of interest covering the PDA tissue with the exclusion of peripheral fat, artifact, necrosis, cystic necrosis/cystic component, and blood vessels were drawn over the tumor by a radiologist with 6 years of experience in gastrointestinal radiology; the quantitative
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
13
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Liu et al
FIGURE 1. Poorly differentiated PDA in tail of the pancreas in a 54-year-old woman. The parameters of QDCE-MRI of PDA were recorded (A). The tumor showed a mild fibrosis (B), VEGF expression score of 3 (C), MVD of 25 (D), and Ki67 labeling score of 10% (E). Figure 1 can be viewed online in color at www.jcat.org.
kinetic parameters were assessed 3 times within the same area, and an average was calculated (Fig. 1).
Histopathology Mean (range) interval between magnetic resonance examination and operation was 11 (range, 1–23) days. All histopathologic analysis was performed by 1 senior pathologist with 10 years of experience, who was unaware of the radiologic results. The final diagnosis, the localization and the size of the lesion and the surgical procedures were recorded for each patient. According to magnetic resonance images, tissue sections of 5 mm were cut from each block and were to assess tumor differentiation, tumor fibrosis, microvessel density (MVD), the expression of tumor vascular endothelial growth factor (VEGF) and Ki67 by stained for hematoxylin-eosin (HE), anti-CD34 antibodies, anti-VEGF antibodies, and anti-Ki67 (Mib-1) antibodies (DAKO, Glostrup, Denmark). Tumor differentiation was assessed according to the World Health Organization classification.6 The extent of fibrosis was
scored according to the ratio of fibrosis in the tumor with HE staining, with a score of 1, less than 33%; 2, 33% to 66%; and 3, greater than 66%7 (Fig. 1). Microvessel density was determined according to the mean number of microvessels counted in the 5 hot spots at high magnification (200) after each CD34-stained slide was scanned at a low magnification (40) to determine 5 “hot spot” areas of the largest number of microvessels8 (Fig. 1). Vascular endothelial growth factor expression was calculated by combining an estimate of the percentage of immunoreactive cells (quantity score) with an estimate of the staining intensity (staining intensity score). For the quantity score, 0%, 1% to 25%, 26% to 50%, and greater than 50% of cells positively stained were scored as 0, 1, 2, and 3, respectively. Staining intensity was rated on a scale of 0 to 3, where 0 was no color, 1 was weak brown, 2 was moderate brown, and 3 was strong brown7,9 (Fig. 1). Ki67 labeling index was calculated as the percentage of positively stained nuclei among the 2000 cells at the areas of highest nuclear labeling7,9 (Fig. 1).
TABLE 1. Comparison of the Parameters of QDCE-MRI Among 3 Subgroups of Tumor Differentiation Tumor Differentiation Parameter ArvIT PeakT MxSIp CER Ktrans Kep Ve
14
Model
Poorly Differentiated
Moderately Differentiated
Well Differentiated
χ2
P
2C 3C 2C 3C 2C 3C
31.277 ± 9.657 87.933 ± 21.984 7.008 ± 4.006 0.878 ± 0.427 0.378 ± 0.255 0.376 ± 0.244 0.841 ± 0.500 0.840 ± 0.244 0.533 ± 0.234 0.523 ± 0..219
25.020 ± 5.160 68.260 ± 15.575 7.181 ± 2.775 0.729 ± 0.257 0.471 ± 0.163 0.464 ± 0.151 0.811 ± 0.338 0.818 ± 0.362 0.645 ± 0.282 0.645 ± 0.279
24.650 ± 3.833 83.550 ± 19.778 8.055 ± 3.148 1.123 ± 0.409 0.380 ± 0.153 0.365 ± 0.147 0.743 ± 0.228 0.726 ± 0.219 0.505 ± 0.092 0.501 ± 0.091
2.574 4.548 0.730 2.450 4.055 3.543 0.113 1.059 0.955 1.223
0.276 0.103 0.694 0.294 0.132 0.170 0.945 0.589 0.620 0.543
www.jcat.org
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
DCE-MRI for Pancreatic Ductal Adenocarcinoma
TABLE 2. Relationship of the Parameters of QDCE-MRI and Tumor Differentiation, Fibrosis, MVD, the Expression of VEGF and Ki67 Ktrans Tumor differentiation
r P r P r P r P r P
Fibrosis VEGF MVD Ki67
Kep
Ve
ArvIT
PeakT
MxSIp
CER
2C
3C
2C
3C
2C
3C
−0.305 0.157 −0.073 0.739 −0.172 0.432 −0.117 0.594 0.298 0.168
−0.169 0.440 −0.058 0.793 −0.199 0.364 0.111 0.613 0.171 0.434
0.179 0.414 −0.350 0.102 0.150 0.496 0.411 0.051 −0.538 0.008
0.201 0.358 −0.366 0.086 0.174 0.428 0.398 0.060 −0.482 0.020
0.198 0.364 −0.144 0.513 0.460 0.027 0.159 0.467 −0.222 0.310
0.176 0.422 −0.046 0.836 0.402 0.058 0.084 0.703 −0.209 0.338
−0.023 0.900 −0.545 0.007 0.155 0.481 0.276 0.202 −0.473 0.023
−0.043 0.846 −0.486 0.019 0.118 0.593 0.234 0.283 −0.444 0.034
0.032 0.886 0.446 0.033 0.057 0.796 −0.266 0.219 0.338 0.114
0.034 0.877 0.461 0.027 0.079 0.720 −0.263 0.225 0.350 0.101
Statistical Analysis Statistical analysis was calculated in Statistical Package for the Social Sciences16.0 (SPSS Inc, Chicago, Ill). The parameters of QDCE-MRI are shown as mean ± SD. Normal distribution was tested using Shapiro-Wilk test. Spearman correlation analysis was used to correlate the parameters of QDCE-MRI with tumor differentiation, tumor fibrosis, MVD, and the expression of tumor VEGF and Ki67. Parameters of QDCE-MRI were compared among all groups of tumor differentiation and tumor fibrosis by Kruskal-Wallis H test. In 2-tailed tests, P < 0.05 was considered statistically significant.
RESULTS In 23 cases of our study, there were 8 poorly differentiated, 11 moderately differentiated, and 4 well-differentiated PDAs. The parameters of QDCE-MRI did not show significant difference among 3 groups of tumor differentiation. The parameters of QDCE-MRI did not show a statistically significant correlation with tumor differentiation (Tables 1 and 2). In 23 cases of our study, there were 6 PDAs with mild fibrosis, 8 PDAs with moderate fibrosis, and 8 PDAs with severe fibrosis. The Kep-2C of the PDA tissue was significantly different among subgroups of tumor fibrosis. The Kep showed the significant negative correlation with tumor fibrosis, whereas Ve showed the significant positive correlation (Tables 2 and 3, Fig. 2).
The Ktrans-2C showed a positive statistical correlation with the expression of tumor VEGF (Table 2, Fig. 3). The parameters of QDCE-MRI did not show a statistically significant correlation with tumor MVD (Table 2). The Kep, MxSIp, and CER of PDA showed a negative statistical correlation with Ki67 of PDA (Table 2, Fig. 4).
DISCUSSION The past studies showed the Ktrans of glioma and breast cancer correlated well with histopathologic grade; moreover, Ktrans and Kep of breast cancer correlated well with the expression of estrogen receptor and progesterone receptor.2,3 However, in our study, the parameters of QDCE-MRI did not show significant difference among 3 groups of tumor differentiation and did not show a statistically significant correlation with tumor differentiation. We supposed the heterogeneity of PDA caused by fibrosis, glandular formation, and mucin production may be the reason that indicates the complexity of efforts to correlate the parameters of QDCE-MRI.10–12 In our study, The Kep-2C of the PDA tissue was significantly different among subgroups of tumor fibrosis, and Kep showed a significant negative correlation with tumor fibrosis, whereas Ve showed a significant positive correlation. Bali et al1 reported Ve and Ktrans had correlation with fibrosis, but did not discuss the relation between Kep and fibrosis. We supposed that the more fibrosis, the more extravascular extracellular
TABLE 3. Comparison of the Parameters of QDCE-MRI Among 3 Subgroups of PDA Fibrosis Tumor Fibrosis Parameter ArvIT PeakT MxSIp CER Ktrans Kep Ve
Model
Mild
Moderate
Severe
χ2
P
2C 3C 2C 3C 2C 3C
27.817 ± 9.188 80.967 ± 11.549 7.572 ± 1.549 1.000 ± 0.365 0.440 ± 0.199 0.402 ± 0.155 1.140 ± 0.330 1.086 ± 0.311 0.417 ± 0.222 0.406 ± 0.213
26.211 ± 5.624 76.156 ± 19.261 7.826 ± 3.501 0.978 ± 0.511 0.404 ± 0.144 0.411 ± 0.140 0.723 ± 0.166 0.751 ± 0.217 0.583 ± 0.240 0.586 ± 0.235
28.436 ± 8.900 84.375 ± 24.242 5.964 ± 3.163 0.670 ± 0.179 0.419 ± 0.269 0.422 ± 0.271 0.663 ± 0.470 0.672 ± 0.501 0.690 ± 0.206 0.681 ± 0.197
0.139 0.573 2.822 2.992 0.454 0.220 6.944 5.502 4.481 4.906
0.933 0.751 0.244 0.224 0.797 0.896 0.031 0.064 0.106 0.086
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
15
Liu et al
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
FIGURE 2. Scatterplot of the tumor Kep and Ve values versus tumor fibrosis. The tumor Kep-2C, Kep-3C, Ve-2C, and Ve-3C showed a statistically significant correlation with tumor fibrosis (r = −0.545, P = 0.007; r = −0.486, P = 0.019; r = 0.446, P = 0.033; r = 0.461, P = 0.027). Figure 2 can be viewed online in color at www.jcat.org.
space (EES) in tumor, which resulted in the increase in Ve. Otherwise, we too supposed the backflow of contrast agent was obstructed by the fibrosis of stroma, which resulted in the decrease in Kep. In our study, the Ktrans-2C showed a positive statistical correlation with the expression of tumor VEGF. Vascular endothelial growth factor is a key regulator of tumor angiogenesis, which can result in the increase in Ktrans by increased angiogenesis and permeability of tumor.13 Tumor angiogenesis is closely related with development and metastasis.14,15 Ktrans and Kep tend to have a reverse correlation with MVD in this colon cancer mouse model16; there is also a moderate correlation between Kep and MVD of prostate cancer,17 and quantitative assessment of tumor angiogenesis can be used to evaluate the therapeutic effect in malignant solid tumors.18 But our study showed the parameters of QDCE-MRI did not show a statistically significant correlation with tumor MVD; the reasons may include that PDA is a type of hypovascular and flow-limited tumor, and MVD only reflects the morphologic feature rather than the functional features.1 There were a few studies discussing the relationship between Ki67 and parameters of QDCE-MRI. The Ktrans changes in the peripheral region of pancreatic cancer after 3 days of therapy were linearly correlated with proliferating cell densities.19 In our study, the Ki67 labeling index of PDA showed a negative statistical correlation with the Kep, MxSIp, and CER of PDA. Ki67 is associated with tumor cell proliferation,20 which presumably resulted in the restriction of material exchange in EES, then agglomeration and reflux of contrast agent in EES slowed simultaneously. These may be the reasons of decrease in Kep, MxSIp, and CER. Our study had some limitations. First, the size of the sample is still small, and further studies with an expanded sample size are
16
www.jcat.org
needed. Second, we focused on only 1 type of pancreatic lesions, the PDA. Third, although tissue sections were chosen according to magnetic resonance images, exact colocalization between histological sections and QDCE-MRI tumor planes for region-ofinterest analyses was not possible. Fourth, the time resolution of scanning of QDCE-MRI needs further improvement for a better assessment, although the time resolution in recent other studies were about the same.5,21 In conclusion, the parameters of QDCE-MRI of PDA can be used to evaluate the degrees of PDA fibrosis and the expressions of VEGF and Ki67.
FIGURE 3. Scatterplot of the tumor Ktrans-2C value versus the expressions of tumor VEGF. The tumor Ktrans-2C showed a statistically significant correlation with the expressions of tumor VEGF (r = 0.460, P = 0.027). Figure 3 can be viewed online in color at www.jcat.org. © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
DCE-MRI for Pancreatic Ductal Adenocarcinoma
FIGURE 4. Scatterplot of the tumor Kep, MxSIp, and CER value versus the expressions of tumor Ki67. The tumor Kep-2C, Kep-3C, MxSIp, and CER showed a statistically significant correlation with the expression of tumor Ki67 (r = −0.473, P = 0.023; r = −0.444, P = 0.034; r = −0.538, P = 0.008; r = −0.482, P = 0.020). Figure 4 can be viewed online in color at www.jcat.org.
REFERENCES 1. Ma Bali, Metens T, Denolin V, et al. Tumoral and nontumoral pancreas: correlation between quantitative dynamic contrast-enhanced MR imaging and histopathologic parameters. Radiology. 2011;261: 456–466. 2. Law M, Young R, Babb J, et al. Comparing perfusion metrics obtained from a single compartment versus pharmacokinetic modeling methods using dynamic susceptibility contrast-enhanced perfusion MR imaging with glioma grade. AJNR Am J Neuroradiol. 2006;27: 1975–1982. 3. Koo HR, Cho N, Song IC, et al. Correlation of perfusion parameters on dynamic contrast-enhanced MRI with prognostic factors and subtypes of breast cancers. J Magn Reson Imaging. 2012;36:145–151. 4. Liu K, Xie P, Peng W, et al. Assessment of dynamic contrast-enhanced magnetic resonance imaging in the differentiation of pancreatic ductal adenocarcinoma from other pancreatic solid lesions. J Comput Assist Tomogr. 2014;38:681–686. 5. Kim JH, Lee JM, Park JH, et al. Solid pancreatic lesions: characterization by using timing bolus dynamic contrast-enhanced MR imaging assessment—a preliminary study. Radiology. 2013;266:185–196. 6. Bosman FT. World Health Organization, International Agency for Research on Cancer. WHO classification of tumours of the digestive system. Lyon, France: International Agency for Research on Cancer; 2010. 7. Wang Y, Chen ZE, Yaghmai V, et al. Diffusion-weighted MR imaging in pancreatic endocrine tumors correlated with histopathologic characteristics. J Magn Reson Imaging. 2011;33:1071–1079. 8. Bosari S, Lee AK, DeLellis RA, et al. Microvessel quantitation and prognosis in invasive breast carcinoma. Hum Pathol. 1992;23: 755–761.
9. Calvar JA, Meli FJ, Romero C, et al. Characterization of brain tumors by MRS, DWI and Ki-67 labeling index. J Neurooncol. 2005;72: 273–280. 10. Rosenkrantz AB, Matza BW, Sabach A, et al. Pancreatic cancer: lack of association between apparent diffusion coefficient values and adverse pathological features. Clin Radiol. 2013;68: e191–e197. 11. Muraoka N, Uematsu H, Kimura H, et al. Apparent diffusion coefficient in pancreatic cancer: characterization and histopathological correlations. J Magn Reson Imaging. 2008;27: 1302–1308. 12. Wang Y, Chen ZE, Nikolaidis P, et al. Diffusion-weighted magnetic resonance imaging of pancreatic adenocarcinomas: association with histopathology and tumor grade. J Magn Reson Imaging. 2011;33: 136–142. 13. Dobrila-Dintinjana R, Vanis N, Dintinjana M, et al. Etiology and oncogenesis of pancreatic carcinoma. Coll Antropol. 2012;36: 1063–1067. 14. Sunamura M, Duda DG, Ghattas MH, et al. Heme oxygenase-1 accelerates tumor angiogenesis of human pancreatic cancer. Angiogenesis. 2003;6:15–24. 15. Takagi K, Takada T, Amano H. A high peripheral microvessel density count correlates with a poor prognosis in pancreatic cancer. J Gastroenterol. 2005;40:402–408. 16. Ahn SJ, An CS, Koom WS, et al. Correlations of 3T DCE-MRI quantitative parameters with microvessel density in a human-colorectal-cancer xenograft mouse model. Korean J Radiol. 2011; 12:722–730. 17. Oto A, Yang C, Kayhan A, et al. Diffusion-weighted and dynamic contrast-enhanced MRI of prostate cancer: correlation of quantitative MR
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
17
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Liu et al
parameters with Gleason score and tumor angiogenesis. AJR Am J Roentgenol. 2011;197:1382–1390. 18. Zhang J, Wang R, Lou H, et al. Functional computed tomographic quantification of angiogenesis in rabbit VX2 soft-tissue tumor before and after interventional therapy. J Comput Assist Tomogr. 2008;32: 697–705. 19. Kim H, Folks KD, Guo L, et al. Early therapy evaluation of combined cetuximab and irinotecan in orthotopic pancreatic tumor xenografts by
18
www.jcat.org
dynamic contrast-enhanced magnetic resonance imaging. Mol Imaging. 2011;10:153–167. 20. Handra-Luca A, Lesty C, Hammel P, et al. Biological and prognostic relevance of mitogen-activated protein kinases in pancreatic adenocarcinoma. Pancreas. 2012;41:416–421. 21. Akisik MF, Sandrasegaran K, Bu G, et al. Pancreatic cancer: utility of dynamic contrast-enhanced MR imaging in assessment of antiangiogenic therapy. Radiology. 2010;256:441–449.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Dynamic Contrast-Enhanced Magnetic Resonance Imaging for Pancreatic Ductal Adenocarcinoma at 3.0-T Magnetic Resonance: Correlation With Histopathology Kefu Liu, MD, PhD,*†‡ Ping Xie, MS,‡ Weijun Peng, MD, PhD,*† and Zhengrong Zhou, MD, PhD*† Purpose: The aim of this study was to discuss the correlation of quantitative dynamic contrast-enhanced magnetic resonance imaging (QDCEMRI) at 3.0-T magnetic resonance and histopathology for pancreatic ductal adenocarcinoma (PDA). Methods: Twenty-three patients with histopathologically proven PDA were included in this study after 75 cases of suspected pancreatic tumors had been performed by QDCE-MRI. The quantitative kinetic parameters analyzed by 2-compartment and 3-compartment models were calculated automatically, which included the volume transfer constant of the contrast agent, the rate constant (Kep), the volume as a percentage of the extravascular extracellular leakage space, the time of arrival of contrast agent, the time of peaking of contrast agent, the maximum slope of signal intensity ascent, and the contrast enhancement ratio. According to magnetic resonance images, tissue section were selected and stained for evaluating tumor differentiation, tumor fibrosis, tumor microvessel density, the expression of tumor vascular endothelial growth factor (VEGF) and Ki67. Subsequently, the relationship between the parameters of QDCE-MRI and histopathology of PDA was analyzed. Results: The tumor Kep and extravascular extracellular leakage space showed a statistically significant correlation with tumor fibrosis; the tumor volume transfer constant of the contrast agent 2-compartment showed a statistically significant correlation with the expressions of tumor VEGF; and the tumor Kep, maximum slope of signal intensity ascent, and contrast enhancement ratio showed a statistically significant correlation with the expression of tumor Ki67. Conclusions: The parameters of QDCE-MRI of PDA can be used to evaluate the degrees of tumor fibrosis and the expressions of VEGF and Ki67. Key Words: dynamic contrast-enhanced magnetic resonance imaging, Pancreatic ductal adenocarcinoma, Histopathology (J Comput Assist Tomogr 2015;39: 13–18)
Q
uantitative dynamic contrast-enhanced magnetic resonance imaging (QDCE-MRI) can be used as a new technique for assessing the function of a specific target tissue.1 Variables of QDCE-MRI are derived from a pharmacokinetic model analysis, and on the basis of the mathematical model used, they reflect the underlying perfusion and/or permeability of the target tissue. The 3-compartment (3C) model treats the vascular and the interstitial space as 2 separated compartments, whereas the 2-compartment
From the *Department of Radiology, Fudan University Shanghai Cancer Center, and †Department of Oncology, Shanghai Medical College, Fudan University, Shanghai; and ‡Department of Radiology, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, Jiangsu, China. Received for publication July 8, 2014; accepted September 18, 2014. Reprints: Zhengrong Zhou, MD, PhD, Department of Radiology, Cancer Hospital, Fudan University, No. 270, Dong An Rd, Shanghai 200032, China (e‐mail: [email protected]). This project was funded by Shanghai Important Technological Subject Fund of China (no. 08411954400) and Program for Shanghai Outstanding Academic Leader of China (no. 11XD1401600). The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
(2C) approach deals with the vascular and the interstitial space as a single entity.1 The past studies showed the parameters of QDCE-MRI are useful for the diagnosis of pancreatic ductal adenocarcinoma (PDA) and correlated well with glioma and breast cancer histopathology.2–5 However, the relationship between the parameters of QDCE-MRI and histopathology of PDA needs further study.
MATERIALS AND METHODS Ethics Statement The study was approved by the local institutional review board of Fudan University Shanghai Cancer Center, and all subjects provided written informed consent.
Patients Seventy-five patients with suspected pancreatic tumors underwent QDCE-MRI (5 patients had an unsuccessful QDCEMRI, of which 3 patients had contrast agent extravasation, and 2 patients had intolerance). In the 70 patients who underwent a successful QDCE-MRI, 23 patients with PDA proved by pathology were included in this study.
Magnetic Resonance Imaging Magnetic resonance studies were performed using a 3.0-T magnetic resonance scanner (Sigma HDx; GE Healthcare, Milwaukee, Wis) using a phased-array body coil. After a plain scan (T1 mapping), multiphase T1-weighted QDCE-MRI images were obtained using a spoiled gradient echo sequence (liver acquisition with volume acceleration) in the axial plane. Scan parameters were repetition time, 3.06 seconds; echo time, minimum; flip angle, 12; bandwidth, 83.33 kHz; field of view, 380 mm; slice thickness, 5 mm; and reconstruction matrix, 180 256. Multiple data sets were acquired every 4 seconds 45 times with a 4-second interval. An injection of 0.1-mmol/kg body weight of gadopentetate dimeglumine (Magnevist; Bayer Health-Care Pharmaceuticals Products, Berlin, Germany) was given at 2 mL/s using a power injector. The onset for contrast injection and data acquisition were trigged simultaneously. Images were transferred to a workstation (Advantage Workstation 4.3; GE Healthcare), and the quantitative kinetic parameters were analyzed with dedicated software (Cine Tool, GE Healthcare). The quantitative kinetic parameters analyzed by 2C and 3C models were calculated automatically, which included the volume transfer constant of the contrast agent (Ktrans), the rate constant (Kep), the volume as a percentage of the extravascular extracellular leakage space (Ve), the time of arrival of contrast agent (ArvIT ), the time of peaking of contrast agent (PeakT ), the maximum slope of signal intensity ascent (MxSIp), and the contrast enhancement ratio (CER). Regions of interest covering the PDA tissue with the exclusion of peripheral fat, artifact, necrosis, cystic necrosis/cystic component, and blood vessels were drawn over the tumor by a radiologist with 6 years of experience in gastrointestinal radiology; the quantitative
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
13
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Liu et al
FIGURE 1. Poorly differentiated PDA in tail of the pancreas in a 54-year-old woman. The parameters of QDCE-MRI of PDA were recorded (A). The tumor showed a mild fibrosis (B), VEGF expression score of 3 (C), MVD of 25 (D), and Ki67 labeling score of 10% (E). Figure 1 can be viewed online in color at www.jcat.org.
kinetic parameters were assessed 3 times within the same area, and an average was calculated (Fig. 1).
Histopathology Mean (range) interval between magnetic resonance examination and operation was 11 (range, 1–23) days. All histopathologic analysis was performed by 1 senior pathologist with 10 years of experience, who was unaware of the radiologic results. The final diagnosis, the localization and the size of the lesion and the surgical procedures were recorded for each patient. According to magnetic resonance images, tissue sections of 5 mm were cut from each block and were to assess tumor differentiation, tumor fibrosis, microvessel density (MVD), the expression of tumor vascular endothelial growth factor (VEGF) and Ki67 by stained for hematoxylin-eosin (HE), anti-CD34 antibodies, anti-VEGF antibodies, and anti-Ki67 (Mib-1) antibodies (DAKO, Glostrup, Denmark). Tumor differentiation was assessed according to the World Health Organization classification.6 The extent of fibrosis was
scored according to the ratio of fibrosis in the tumor with HE staining, with a score of 1, less than 33%; 2, 33% to 66%; and 3, greater than 66%7 (Fig. 1). Microvessel density was determined according to the mean number of microvessels counted in the 5 hot spots at high magnification (200) after each CD34-stained slide was scanned at a low magnification (40) to determine 5 “hot spot” areas of the largest number of microvessels8 (Fig. 1). Vascular endothelial growth factor expression was calculated by combining an estimate of the percentage of immunoreactive cells (quantity score) with an estimate of the staining intensity (staining intensity score). For the quantity score, 0%, 1% to 25%, 26% to 50%, and greater than 50% of cells positively stained were scored as 0, 1, 2, and 3, respectively. Staining intensity was rated on a scale of 0 to 3, where 0 was no color, 1 was weak brown, 2 was moderate brown, and 3 was strong brown7,9 (Fig. 1). Ki67 labeling index was calculated as the percentage of positively stained nuclei among the 2000 cells at the areas of highest nuclear labeling7,9 (Fig. 1).
TABLE 1. Comparison of the Parameters of QDCE-MRI Among 3 Subgroups of Tumor Differentiation Tumor Differentiation Parameter ArvIT PeakT MxSIp CER Ktrans Kep Ve
14
Model
Poorly Differentiated
Moderately Differentiated
Well Differentiated
χ2
P
2C 3C 2C 3C 2C 3C
31.277 ± 9.657 87.933 ± 21.984 7.008 ± 4.006 0.878 ± 0.427 0.378 ± 0.255 0.376 ± 0.244 0.841 ± 0.500 0.840 ± 0.244 0.533 ± 0.234 0.523 ± 0..219
25.020 ± 5.160 68.260 ± 15.575 7.181 ± 2.775 0.729 ± 0.257 0.471 ± 0.163 0.464 ± 0.151 0.811 ± 0.338 0.818 ± 0.362 0.645 ± 0.282 0.645 ± 0.279
24.650 ± 3.833 83.550 ± 19.778 8.055 ± 3.148 1.123 ± 0.409 0.380 ± 0.153 0.365 ± 0.147 0.743 ± 0.228 0.726 ± 0.219 0.505 ± 0.092 0.501 ± 0.091
2.574 4.548 0.730 2.450 4.055 3.543 0.113 1.059 0.955 1.223
0.276 0.103 0.694 0.294 0.132 0.170 0.945 0.589 0.620 0.543
www.jcat.org
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
DCE-MRI for Pancreatic Ductal Adenocarcinoma
TABLE 2. Relationship of the Parameters of QDCE-MRI and Tumor Differentiation, Fibrosis, MVD, the Expression of VEGF and Ki67 Ktrans Tumor differentiation
r P r P r P r P r P
Fibrosis VEGF MVD Ki67
Kep
Ve
ArvIT
PeakT
MxSIp
CER
2C
3C
2C
3C
2C
3C
−0.305 0.157 −0.073 0.739 −0.172 0.432 −0.117 0.594 0.298 0.168
−0.169 0.440 −0.058 0.793 −0.199 0.364 0.111 0.613 0.171 0.434
0.179 0.414 −0.350 0.102 0.150 0.496 0.411 0.051 −0.538 0.008
0.201 0.358 −0.366 0.086 0.174 0.428 0.398 0.060 −0.482 0.020
0.198 0.364 −0.144 0.513 0.460 0.027 0.159 0.467 −0.222 0.310
0.176 0.422 −0.046 0.836 0.402 0.058 0.084 0.703 −0.209 0.338
−0.023 0.900 −0.545 0.007 0.155 0.481 0.276 0.202 −0.473 0.023
−0.043 0.846 −0.486 0.019 0.118 0.593 0.234 0.283 −0.444 0.034
0.032 0.886 0.446 0.033 0.057 0.796 −0.266 0.219 0.338 0.114
0.034 0.877 0.461 0.027 0.079 0.720 −0.263 0.225 0.350 0.101
Statistical Analysis Statistical analysis was calculated in Statistical Package for the Social Sciences16.0 (SPSS Inc, Chicago, Ill). The parameters of QDCE-MRI are shown as mean ± SD. Normal distribution was tested using Shapiro-Wilk test. Spearman correlation analysis was used to correlate the parameters of QDCE-MRI with tumor differentiation, tumor fibrosis, MVD, and the expression of tumor VEGF and Ki67. Parameters of QDCE-MRI were compared among all groups of tumor differentiation and tumor fibrosis by Kruskal-Wallis H test. In 2-tailed tests, P < 0.05 was considered statistically significant.
RESULTS In 23 cases of our study, there were 8 poorly differentiated, 11 moderately differentiated, and 4 well-differentiated PDAs. The parameters of QDCE-MRI did not show significant difference among 3 groups of tumor differentiation. The parameters of QDCE-MRI did not show a statistically significant correlation with tumor differentiation (Tables 1 and 2). In 23 cases of our study, there were 6 PDAs with mild fibrosis, 8 PDAs with moderate fibrosis, and 8 PDAs with severe fibrosis. The Kep-2C of the PDA tissue was significantly different among subgroups of tumor fibrosis. The Kep showed the significant negative correlation with tumor fibrosis, whereas Ve showed the significant positive correlation (Tables 2 and 3, Fig. 2).
The Ktrans-2C showed a positive statistical correlation with the expression of tumor VEGF (Table 2, Fig. 3). The parameters of QDCE-MRI did not show a statistically significant correlation with tumor MVD (Table 2). The Kep, MxSIp, and CER of PDA showed a negative statistical correlation with Ki67 of PDA (Table 2, Fig. 4).
DISCUSSION The past studies showed the Ktrans of glioma and breast cancer correlated well with histopathologic grade; moreover, Ktrans and Kep of breast cancer correlated well with the expression of estrogen receptor and progesterone receptor.2,3 However, in our study, the parameters of QDCE-MRI did not show significant difference among 3 groups of tumor differentiation and did not show a statistically significant correlation with tumor differentiation. We supposed the heterogeneity of PDA caused by fibrosis, glandular formation, and mucin production may be the reason that indicates the complexity of efforts to correlate the parameters of QDCE-MRI.10–12 In our study, The Kep-2C of the PDA tissue was significantly different among subgroups of tumor fibrosis, and Kep showed a significant negative correlation with tumor fibrosis, whereas Ve showed a significant positive correlation. Bali et al1 reported Ve and Ktrans had correlation with fibrosis, but did not discuss the relation between Kep and fibrosis. We supposed that the more fibrosis, the more extravascular extracellular
TABLE 3. Comparison of the Parameters of QDCE-MRI Among 3 Subgroups of PDA Fibrosis Tumor Fibrosis Parameter ArvIT PeakT MxSIp CER Ktrans Kep Ve
Model
Mild
Moderate
Severe
χ2
P
2C 3C 2C 3C 2C 3C
27.817 ± 9.188 80.967 ± 11.549 7.572 ± 1.549 1.000 ± 0.365 0.440 ± 0.199 0.402 ± 0.155 1.140 ± 0.330 1.086 ± 0.311 0.417 ± 0.222 0.406 ± 0.213
26.211 ± 5.624 76.156 ± 19.261 7.826 ± 3.501 0.978 ± 0.511 0.404 ± 0.144 0.411 ± 0.140 0.723 ± 0.166 0.751 ± 0.217 0.583 ± 0.240 0.586 ± 0.235
28.436 ± 8.900 84.375 ± 24.242 5.964 ± 3.163 0.670 ± 0.179 0.419 ± 0.269 0.422 ± 0.271 0.663 ± 0.470 0.672 ± 0.501 0.690 ± 0.206 0.681 ± 0.197
0.139 0.573 2.822 2.992 0.454 0.220 6.944 5.502 4.481 4.906
0.933 0.751 0.244 0.224 0.797 0.896 0.031 0.064 0.106 0.086
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
15
Liu et al
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
FIGURE 2. Scatterplot of the tumor Kep and Ve values versus tumor fibrosis. The tumor Kep-2C, Kep-3C, Ve-2C, and Ve-3C showed a statistically significant correlation with tumor fibrosis (r = −0.545, P = 0.007; r = −0.486, P = 0.019; r = 0.446, P = 0.033; r = 0.461, P = 0.027). Figure 2 can be viewed online in color at www.jcat.org.
space (EES) in tumor, which resulted in the increase in Ve. Otherwise, we too supposed the backflow of contrast agent was obstructed by the fibrosis of stroma, which resulted in the decrease in Kep. In our study, the Ktrans-2C showed a positive statistical correlation with the expression of tumor VEGF. Vascular endothelial growth factor is a key regulator of tumor angiogenesis, which can result in the increase in Ktrans by increased angiogenesis and permeability of tumor.13 Tumor angiogenesis is closely related with development and metastasis.14,15 Ktrans and Kep tend to have a reverse correlation with MVD in this colon cancer mouse model16; there is also a moderate correlation between Kep and MVD of prostate cancer,17 and quantitative assessment of tumor angiogenesis can be used to evaluate the therapeutic effect in malignant solid tumors.18 But our study showed the parameters of QDCE-MRI did not show a statistically significant correlation with tumor MVD; the reasons may include that PDA is a type of hypovascular and flow-limited tumor, and MVD only reflects the morphologic feature rather than the functional features.1 There were a few studies discussing the relationship between Ki67 and parameters of QDCE-MRI. The Ktrans changes in the peripheral region of pancreatic cancer after 3 days of therapy were linearly correlated with proliferating cell densities.19 In our study, the Ki67 labeling index of PDA showed a negative statistical correlation with the Kep, MxSIp, and CER of PDA. Ki67 is associated with tumor cell proliferation,20 which presumably resulted in the restriction of material exchange in EES, then agglomeration and reflux of contrast agent in EES slowed simultaneously. These may be the reasons of decrease in Kep, MxSIp, and CER. Our study had some limitations. First, the size of the sample is still small, and further studies with an expanded sample size are
16
www.jcat.org
needed. Second, we focused on only 1 type of pancreatic lesions, the PDA. Third, although tissue sections were chosen according to magnetic resonance images, exact colocalization between histological sections and QDCE-MRI tumor planes for region-ofinterest analyses was not possible. Fourth, the time resolution of scanning of QDCE-MRI needs further improvement for a better assessment, although the time resolution in recent other studies were about the same.5,21 In conclusion, the parameters of QDCE-MRI of PDA can be used to evaluate the degrees of PDA fibrosis and the expressions of VEGF and Ki67.
FIGURE 3. Scatterplot of the tumor Ktrans-2C value versus the expressions of tumor VEGF. The tumor Ktrans-2C showed a statistically significant correlation with the expressions of tumor VEGF (r = 0.460, P = 0.027). Figure 3 can be viewed online in color at www.jcat.org. © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
DCE-MRI for Pancreatic Ductal Adenocarcinoma
FIGURE 4. Scatterplot of the tumor Kep, MxSIp, and CER value versus the expressions of tumor Ki67. The tumor Kep-2C, Kep-3C, MxSIp, and CER showed a statistically significant correlation with the expression of tumor Ki67 (r = −0.473, P = 0.023; r = −0.444, P = 0.034; r = −0.538, P = 0.008; r = −0.482, P = 0.020). Figure 4 can be viewed online in color at www.jcat.org.
REFERENCES 1. Ma Bali, Metens T, Denolin V, et al. Tumoral and nontumoral pancreas: correlation between quantitative dynamic contrast-enhanced MR imaging and histopathologic parameters. Radiology. 2011;261: 456–466. 2. Law M, Young R, Babb J, et al. Comparing perfusion metrics obtained from a single compartment versus pharmacokinetic modeling methods using dynamic susceptibility contrast-enhanced perfusion MR imaging with glioma grade. AJNR Am J Neuroradiol. 2006;27: 1975–1982. 3. Koo HR, Cho N, Song IC, et al. Correlation of perfusion parameters on dynamic contrast-enhanced MRI with prognostic factors and subtypes of breast cancers. J Magn Reson Imaging. 2012;36:145–151. 4. Liu K, Xie P, Peng W, et al. Assessment of dynamic contrast-enhanced magnetic resonance imaging in the differentiation of pancreatic ductal adenocarcinoma from other pancreatic solid lesions. J Comput Assist Tomogr. 2014;38:681–686. 5. Kim JH, Lee JM, Park JH, et al. Solid pancreatic lesions: characterization by using timing bolus dynamic contrast-enhanced MR imaging assessment—a preliminary study. Radiology. 2013;266:185–196. 6. Bosman FT. World Health Organization, International Agency for Research on Cancer. WHO classification of tumours of the digestive system. Lyon, France: International Agency for Research on Cancer; 2010. 7. Wang Y, Chen ZE, Yaghmai V, et al. Diffusion-weighted MR imaging in pancreatic endocrine tumors correlated with histopathologic characteristics. J Magn Reson Imaging. 2011;33:1071–1079. 8. Bosari S, Lee AK, DeLellis RA, et al. Microvessel quantitation and prognosis in invasive breast carcinoma. Hum Pathol. 1992;23: 755–761.
9. Calvar JA, Meli FJ, Romero C, et al. Characterization of brain tumors by MRS, DWI and Ki-67 labeling index. J Neurooncol. 2005;72: 273–280. 10. Rosenkrantz AB, Matza BW, Sabach A, et al. Pancreatic cancer: lack of association between apparent diffusion coefficient values and adverse pathological features. Clin Radiol. 2013;68: e191–e197. 11. Muraoka N, Uematsu H, Kimura H, et al. Apparent diffusion coefficient in pancreatic cancer: characterization and histopathological correlations. J Magn Reson Imaging. 2008;27: 1302–1308. 12. Wang Y, Chen ZE, Nikolaidis P, et al. Diffusion-weighted magnetic resonance imaging of pancreatic adenocarcinomas: association with histopathology and tumor grade. J Magn Reson Imaging. 2011;33: 136–142. 13. Dobrila-Dintinjana R, Vanis N, Dintinjana M, et al. Etiology and oncogenesis of pancreatic carcinoma. Coll Antropol. 2012;36: 1063–1067. 14. Sunamura M, Duda DG, Ghattas MH, et al. Heme oxygenase-1 accelerates tumor angiogenesis of human pancreatic cancer. Angiogenesis. 2003;6:15–24. 15. Takagi K, Takada T, Amano H. A high peripheral microvessel density count correlates with a poor prognosis in pancreatic cancer. J Gastroenterol. 2005;40:402–408. 16. Ahn SJ, An CS, Koom WS, et al. Correlations of 3T DCE-MRI quantitative parameters with microvessel density in a human-colorectal-cancer xenograft mouse model. Korean J Radiol. 2011; 12:722–730. 17. Oto A, Yang C, Kayhan A, et al. Diffusion-weighted and dynamic contrast-enhanced MRI of prostate cancer: correlation of quantitative MR
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
17
J Comput Assist Tomogr • Volume 39, Number 1, January/February 2015
Liu et al
parameters with Gleason score and tumor angiogenesis. AJR Am J Roentgenol. 2011;197:1382–1390. 18. Zhang J, Wang R, Lou H, et al. Functional computed tomographic quantification of angiogenesis in rabbit VX2 soft-tissue tumor before and after interventional therapy. J Comput Assist Tomogr. 2008;32: 697–705. 19. Kim H, Folks KD, Guo L, et al. Early therapy evaluation of combined cetuximab and irinotecan in orthotopic pancreatic tumor xenografts by
18
www.jcat.org
dynamic contrast-enhanced magnetic resonance imaging. Mol Imaging. 2011;10:153–167. 20. Handra-Luca A, Lesty C, Hammel P, et al. Biological and prognostic relevance of mitogen-activated protein kinases in pancreatic adenocarcinoma. Pancreas. 2012;41:416–421. 21. Akisik MF, Sandrasegaran K, Bu G, et al. Pancreatic cancer: utility of dynamic contrast-enhanced MR imaging in assessment of antiangiogenic therapy. Radiology. 2010;256:441–449.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
