JOURNAL OF MAGNETIC RESONANCE IMAGING 41:1608–1614 (2015)
Original Research
Dynamic Contrast-Enhanced MRI of Gastric Cancer: Correlation of the Perfusion Parameters With Pathological Prognostic Factors Ijin Joo, MD,1 Jeong Min Lee, MD,1,2* Joon Koo Han, MD,1,2 Han-Kwang Yang, MD,3 Hyuk-Joon Lee, MD,3 and Byung Ihn Choi, MD1,2 Key Words: gastric cancer; DCE-MRI; perfusion J. Magn. Reson. Imaging 2015;41:1608–1614. C 2014 Wiley Periodicals, Inc. V
Purpose: To investigate the feasibility of dynamic, contrast-enhanced, magnetic resonance imaging (DCEMRI) for perfusion quantification of gastric cancers, and to correlate the DCE-MRI parameters with the pathological prognostic factors. Materials and Methods: This prospective study was approved by our Institutional Review Board. Twentyseven patients with gastric cancers underwent DCE-MRI using a free-breathing, radial, gradient-echo (GRE) sequence with k-space weighted image contrast (KWIC) reconstruction on a 3T scanner. The DCE-MRI parameters (volume transfer coefficient [Ktrans], reverse reflux rate constant [Kep], extracellular extravascular volume fraction [Ve], and initial area under the gadolinium concentration curve during the first 60 seconds [iAUC]) of gastric cancer and normal wall were measured and compared with each other using the Wilcoxon signed rank test. The relationship between the DCE-MRI parameters of gastric cancer and the pathological prognostic factors were evaluated using the Mann–Whitney test or the Spearman rank correlation test. Results: DCE-MRIs were of diagnostic quality in 22 patients (81.5%). Ve and iAUC were significantly higher in gastric cancer than in normal gastric wall (P < 0.05). Ve showed significant positive correlation with T-staging of gastric cancers (P < 0.05). Ktrans was significantly correlated with the grades of epidermal growth-factor receptor expression (P < 0.05). Conclusion: DCE-MRI using a radial GRE with KWIC reconstruction is feasible for quantification of the perfusion dynamics of gastric cancers, and the DCE-MRI parameters of gastric cancers may provide prognostic information.
1 Department of Radiology, Seoul National University Hospital, Seoul, Korea. 2 Institute of Radiation Medicine, Seoul National University Hospital, Seoul, Korea. 3 Department of Surgery, Seoul National University College of Medicine, Korea. Contract grant sponsor: GE Healthcare. *Address reprint requests to: J.M.L., Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 110-744, Korea. E-mail: [email protected] Received March 31, 2014; Accepted July 3, 2014. DOI 10.1002/jmri.24711 View this article online at wileyonlinelibrary.com. C 2014 Wiley Periodicals, Inc. V
GASTRIC CANCER is one of the main causes of cancer death worldwide, and while surgical resection is the only potentially curative treatment, the long-term survival of patients with advanced gastric cancers (AGCs) remains poor even following surgical resection (1,2). In order to further improve the outcomes of patients with AGC, many clinical trials have been performed to evaluate the role of chemotherapy, radiation therapy, and combined-modality treatment (2). Given that there are various therapeutic options for AGC, identification of pretherapeutic, predictive markers for both the treatment response and prognosis is essential for a personalized cancer treatment (3). Pathologic TNM staging, the histological grades of tumor differentiation, and expression of molecular markers such as the epidermal growth-factor receptor (EGFR) are regarded as prognostic factors of gastric cancers (4–6). Several recent studies reported that the perfusion parameters of gastric cancers obtained by perfusion computed tomography (CT) or contrast-enhanced ultrasound may be useful for both tissue characterization and treatment monitoring (7–9). Dynamic, contrast-enhanced magnetic resonance imaging (DCE-MRI) is a currently evolving imaging method used to assess the functional features of a target tissue, including tissue blood flow, capillary permeability, and interstitial volume (10). DCE-MRI has actually become increasingly used recently for the characterization, differential diagnosis of various cancers, and for monitoring changes in tumor perfusion during anticancer treatment (11–13). Many oncologic studies have also assessed the potential role of DCEMRI as a prognostic indicator (14,15) as well as the correlation of DCE-MRI and other established or possible prognostic factors (16–18). However, in the acquisition of DCE-MRI data regarding abdominal lesions, the image quality may be degraded due to motion artifacts related to respiration and bowel peristalsis during the acquisition, and thus resulting in
1608
DCE-MRI of Gastric Cancer
interframe misregistration (19). To reduce motionrelated artifacts and improve the image registration, fast imaging sequences and new methods of sampling k-space have been developed and used in abdominal imaging (20–22). K-space weighted image contrast (KWIC) reconstruction is a kind of data-weighting system used for radially encoded MRI data in which the central k-space is oversampled and occupied by the selected radial views, while the peripheral k-space includes time-neighboring subsets of radial views and would, therefore, simultaneously provide high spatial and high temporal resolution (23,24). Using this concept, several previous studies have reported the feasibility of free-breathing radial DCE-MRI using a radial, 3D, gradient-echo (GRE) technique (radial volumetric interpolated, breath-hold examination [VIBE]; Siemens Healthcare, Erlangen, Germany) with KWIC reconstruction in the evaluation of abdominal lesions (25,26). However, there have been no previous studies that evaluated the technical feasibility and diagnostic value of DCE-MRI for patients with gastric cancers. Therefore, the purposes of our study were to investigate the feasibility of DCE-MRI for perfusion quantification of gastric cancers using a radial GRE sequence with KWIC reconstruction and to correlate the DCE-MRI parameters with the pathological prognostic factors.
MATERIALS AND METHODS Patients This prospective study was approved by the Institutional Review Board of Seoul National University Hospital, and informed consent was obtained from each patient. From January 2011 to September 2011, 28 consecutive patients who met the following criteria were enrolled in this study: 1) patients with endoscopic, biopsy-confirmed gastric cancers and 2) patients whose gastric cancers were suspected to be T2 or greater, any N, and M0 stage, based on the endoscopic examination and as well as on the preoperative MDCT scan. All patients underwent preoperative DCEMRI on a 3T scanner. Among the 28 patients, one patient who did not undergo prompt surgery due to active pulmonary tuberculosis was excluded. Therefore, 27 patients (21 males and six females) with an age range of 38–81 years were finally included. The time interval from DCE-MRI to surgery was within 2 weeks. MRI Data Acquisition MRI studies were performed on a 3T MR scanner (TIM Trio; Siemens Healthcare) with a standard, 24-channel, phased-array body coil and in the supine position. Each patient received an intravenous injection of 10 mg of hyoscine butylbromide (Buscopan; Boehringer Ingelheim Korea, Seoul, Korea) 5 minutes prior to the MRI study in order to reduce bowel motion and was asked to drink 500–1000 mL of water to distend the stomach immediately before beginning the MRI acquisition. Using the three-plane, true-fast imaging with steady-state precession (true-FISP) localizers, an axial MRI slab was placed from the dome of the liver to the
1609
right renal hilum. Unenhanced studies, including half-Fourier acquisition, single-shot, turbo spine-echo (HASTE), T2-weighted imaging with and without fat saturation, true-FISP, T1-weighted, 3D, GRE, in- and out-of-phase imaging, and T1-weighted fat-suppressed 3D GRE imaging were performed. For T1 mapping, unenhanced T1-weighted VIBE images were obtained using three different flip angles with the following parameters: relaxation time (TR) ¼ 5.3 msec; echo time (TE) ¼ 1.1 msec; flip angles (a) ¼ 2 , 8 , and 15 ; slice thickness ¼ 5 mm; number of excitations (NEX) ¼ 3; field of view (FOV) ¼ 38 38 cm2; matrix ¼ 128 96; and number of slices ¼ 40. T1 maps were subsequently generated using a threeparameter, nonlinear fitting algorithm to the MR data for each voxel (27). Thereafter, DCE-MRI sequences using free-breathing, radial 3D VIBE with KWIC reconstruction (25) performed over the same volume as the T1 maps, were performed after an intravenous bolus injection of 0.1 mmol/kg of gadolinium contrast agent (Omniscan, gadodiamide) at a rate of 3 mL/sec followed by a chasing bolus of 20 mL of normal saline administered using an MR-compatible power injector (SpectrisSolaris; Medrad, Indianola, PA). All patients were instructed to breathe as quietly as possible during the examination. The DCE-MRI parameters were as follows: TR ¼ 3.8 msec; TE ¼ 1.6 msec; flip angle ¼ 11 ; slice thickness ¼ 5 mm with a 50% slice resolution (10/5 mm, actual/interpolated slice thickness); NEX ¼ 1; FOV ¼ 38 38 cm2; matrix ¼ 256 256; and the number of slices ¼ 20. To reconstruct time-resolved, subframe images, the KWIC view-sharing technique which permits the manipulation of image contrast by careful filtering of the acquired views was used (26). Radial VIBE sequences were continuously scanned for 75 seconds, repeated volumetric sets of axial images at a 4.1-second intervals for 308 seconds. The spectral selection, attenuated inversion technique was used for fat suppression. DCE-MRI Analysis Postprocessing of DCE-MRIs was performed by one abdominal radiologist (I.J. with 7 years of clinical experience in abdominal MRI) who was aware that the patients had gastric cancers, but who was blinded to the pathological features. From the DCE-MRI data, voxelwise, parametric maps of the volume transfer coefficient (Ktrans), reverse reflux rate constant (Kep), extracellular extravascular volume fraction (Ve), and the initial area under the gadolinium concentration curve during the first 60 seconds (iAUC) were generated using a postprocessing platform with dedicated DCE-MRI software (Tissue4D; Siemens Healthcare) (26). Before analysis of the DCE-MRI data, motion correction of the subframe radial VIBE using KWIC reconstruction image sets was performed using the nonrigid registration technique of the software. A twocompartment Tofts model was applied (10,28), and the arterial input function was chosen according the fast sampling method (29). Using the software, by drawing a freehand region of interest (ROI), Ktrans, Kep, Ve, and iAUC values for the selected ROI were automatically calculated. Two ROIs were drawn on
1610
Joo et al.
the dynamic, T1-weighted VIBE images: one ROI covered the gastric cancer by outlining the edges of the tumor on the section of central level of the tumor, which was referenced on the corresponding T2-weighted axial images, while the other ROI covered the normal gastric wall adjacent to the tumor (Fig. 1). Two abdominal radiologists (J.M.L. and I.J. with 22 years and 7 years, respectively, of clinical experience performing abdominal MRI) determined in consensus the image quality of the perfusion maps as nondiagnostic or diagnostic according to the following criteria: the perfusion map containing apparent errors in image registration probably due to motion was considered nondiagnostic; otherwise, the image quality was diagnostic (25). The DCE-MRI perfusion parameters in any patient showing nondiagnostic image quality were excluded in the analysis. To evaluate the test–retest repeatability of the DCE-MRI parameters of gastric cancers, the same radiologist (I.J.) who had performed the postprocessing of DCE-MRIs repeated the analysis of DCE-MRIs of gastric cancers at least 4 weeks after the initial analysis. Pathological Analysis The 27 patients who were finally included in this study underwent curative or palliative gastrectomy and lymph node (LN) dissection. Pathologic TNM staging of gastric cancers was reported based on the AJCC 7th guidelines (30). In patients with tubular adenocarcinoma, the histological grade was reported as well-, moderately, or poorly differentiated. The EGFR expression determined by immunohistochemistry was scored as follows: 0, no membrane staining; 1þ, faint or partial staining; 2þ, moderate staining; and 3þ, strong staining (31). Statistical Analysis To compare the DCE-MRI parameters between gastric cancer and normal wall, the Wilcoxon signed rank test was performed. The DCE-MRI parameters were compared between the N-negative and N-positive stage of gastric cancers and between well- or moderately differentiated and poorly differentiated tubular adenocarcinomas using the Mann–Whitney test. The Spearman rank correlation test was performed to evaluate the correlation of the DCE-MRI parameters of gastric cancers with T-staging (T2, T3, T4) and the grades of EGFR expression (0, 1þ, 2þ). P < 0.05 was considered to indicate statistical significance. To obtain the test–retest repeatability of DCE-MRI parameters of gastric cancers, coefficients of variation (CVs ¼ standard deviation divided by mean) within subjects were calculated (32). All statistical analyses were performed using MedCalc software v. 12.4.0.0 (Mariakerke, Belgium). RESULTS DCE-MRI in Patients With Gastric Cancer As the perfusion maps from the DCE-MRI data were of diagnostic quality in 22 (81.5%) of 27 patients, those 22 patients were finally included in the statistical anal-
Figure 1. MR images of a 69-year-old male with a histopathologically confirmed stage T4aN3b gastric cancer. a: On the contrast-enhanced axial T1-weighted image, an ulceroinfiltrative mass with transmural enhancement (arrows) is seen in the gastric antrum. On the corresponding parametric map of Ktrans (b) displayed in colors ranging from blue to red, gastric cancer (green ROI) shows a wide range of Ktrans, while normal gastric wall (yellow ROI) shows homogeneously low Ktrans. Parametric map of Ve (c) shows relatively higher Ve in the gastric cancer (green ROI) than in the normal gastric wall (yellow ROI).
ysis. Five patients were excluded because their perfusion maps were nondiagnostic and showed apparent in-plane misregistration from the serial DCE-MRI data caused by severe peristaltic movement of the stomach.
DCE-MRI of Gastric Cancer
1611
Table 1 DCE-MRI Parameters of Gastric Cancer and Normal Gastric Wall DCE-MRI Parameters Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
Gastric Cancer 0.046 0.623 0.133 5.533
(0.042-0.065) (0.522-0.753) (0.088-0.173) (4.907-7.734)
Normal Wall 0.037 0.785 0.063 3.894
(0.028-0.053) (0.522-1.048) (0.044-0.083) (3.214-5.613)
P value 0.085 0.200 0.004* 0.039*
Data are medians (95% confidence intervals). All data were tested using the Wilcoxon signed rank test. *Significant value, P < 0.05.
T-staging, N-staging, histological grades, and grades of EGFR expression, are shown in Tables (2–4). Among the DCE-MRI parameters, Ve showed a significant correlation with T-staging (rho ¼ 0.483, P ¼ 0.023) (Table 2). None of the DCE-MRI parameters of gastric cancer showed significant differences in Nnegative and N-positive cases (P > 0.05). In cases of tubular adenocarcinoma, poorly differentiated carcinomas showed a higher Ve than moderately differentiated carcinomas, however, the difference was not statistically significant (P > 0.05) (Table 3). Ktrans was significantly correlated with the grades of EGFR expression (rho ¼ 0.460, P ¼ 0.031) (Table 4).
Test–Retest Repeatability of DCE-MRI Parameters of Gastric Cancers In 22 patients whose perfusion maps were of diagnostic quality, DCE-MRI parameters of gastric cancers showed good or moderate test–retest repeatability with CVs of Ktrans, Kep, Ve, and iAUC of 12.3%, 25.7%, 20.8%, and 17.9%, respectively. Pathological Results T-stages in five excluded patients with nondiagnostic image quality of the perfusion maps were confirmed as follows: T1b, n ¼ 2; T2, n ¼ 1; T3, n ¼ 1; and T4a, n ¼ 1. Among the 22 study patients, two were confirmed as having stage T1b cancer, eight with T2, six with T3, five with T4a, and one with T4b (T2, n ¼ 10; T3, n ¼ 6; and T4, n ¼ 6), and eight were N-negative and 14 N-positive. One patient was pathologically confirmed to have M1 stage peritoneal seeding, while the other 21 patients had M0 stage. The histologic patterns of the gastric cancers according to the 2010 WHO classification were as follows: 17 tubular adenocarcinomas; one papillary adenocarcinoma; three mucinous and poorly cohesive carcinomas (including signet-ring-cell carcinoma); and one with uncommon histologic variants (carcinoma with lymphoid stroma) (33). The histological grades of the 17 tubular adenocarcinomas were confirmed as moderately differentiated carcinomas in nine patients and poorly differentiated carcinomas in eight patients. The grades of EGFR expression were 0 in six patients, 1þ in eight, and 2þ or 3þ in eight patients. Comparison of the DCE-MRI Parameters Between Gastric Cancer and Normal Wall The DCE-MRI parameters of gastric cancer and normal wall are summarized in Table 1. Among the DCEMRI parameters, the medians of Ve and iAUC were significantly higher in gastric cancer than those in normal wall (Ve, 0.133 vs. 0.063; iAUC, 5.533 vs. 3.894 mM/sec, respectively) (P < 0.05) (Fig. 1). Correlation of the DCE-MRI Parameters of Gastric Cancer With the Pathological Prognostic Factors Correlation of the DCE-MRI parameters of gastric cancer with the pathological prognostic factors, including
DISCUSSION Our study demonstrates that DCE-MRI is feasible for the evaluation of gastric cancers, using the freebreathing radial 3D-GRE technique with KWIC reconstruction. Some of the DCE-MRI parameters of gastric cancer also showed a significant correlation with the pathological prognostic factors, including T-staging, and grades of EGFR-expression, ie, Ve showed significant positive correlation with T-staging and Ktrans was significantly correlated with the grades of EGFR expression. These results suggest that DCE-MRI can be used to characterize the tumor biology of gastric cancer, and may have the potential to provide additional prognostic information regarding patients with gastric cancer. Our study results are in good agreement with those of several recent studies using perfusion CT and in which the tumor perfusion measurement of gastric cancer and its possible role in clinical practice have been reported (7,8,34). However, considering that the clinical use of perfusion CT may be limited by its relatively high radiation dose (15– 25 mSv), if our study results are validated by larger studies, DCE-MRI may provide additional benefits over those of perfusion CT. Furthermore, the multiparametric capability of MR for oncologic evaluation such as diffusion-weighted imaging, T1 or T2 relaxometry, or MR elastography, could be an additional advantage over the use of CT (35). In our study, perfusion parametric maps of diagnostic quality were successfully obtained in 81.5% of the patients from the DCE-MR image sets using the free-breathing, 3D-radial VIBE with KWIC technique. In fact, there have been no previous reports of DCEMRI used for the evaluation of gastric cancers. This may be due to the poor temporal resolution and susceptibility to motion of T1-weighted sequences used for DCE-MRI, compared with CT scans, and which may result in inevitable respiratory and bowel motionrelated artifacts (22). Our study result may be primarily due to the advantages of free-breathing radial VIBE with KWIC reconstruction, which provides high spatial and temporal resolution by oversampling of the central k-space region with low spatial frequency data and fewer respiration-related artifacts due to its inherent motion insensitivity (25,26). Furthermore, in our study hyoscine butylbromide was injected before MRI scanning in order to reduce bowel peristalsis.
Joo et al. 0.608 0.973 0.759 0.402
Table 3 DCE-MRI Parameters According to the Histological Grades of Tumor Differentiation in Tubular Adenocarcinoma of the Stomach
(0.031-0.072) (0.356-0.953) (0.078-0.178) (4.055-7.927)
DCE-MRI Parameters
0.043 0.596 0.139 5.242
Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
Moderately Differentiated (n ¼ 9) 0.044 0.568 0.105 5.537
(0.031-0.059) (0.383-0.747) (0.060-0.141) (4.184-7.778)
Poorly Differentiated (n ¼ 7) 0.065 0.653 0.138 7.295
(0.042-0.139) (0.443-1.047) (0.098-0.291) (4.939-13.256)
P value* 0.186 0.408 0.142 0.470
0.199 0.400 0.483 0.158
0.374 0.065 0.023z 0.484
0.057 0.637 0.107 6.639
(0.039-0.080) (0.537-0.724) (0.082-0.194) (4.511-9.533)
Data are medians (95% confidence intervals). All data were tested using the Mann-Whitney test. *Significant value, P < 0.05.
Data are medians (95% confidence intervals). *Data were tested using the Spearman rank correlation test. y Data were tested using the Mann-Whitney test. z Significant value, P < 0.05.
0.050 (0.034-0.135) 0.419 (0.235-1.028) 0.187 (0.136-0.297) 5.755 (4.128-12.189) (0.030-0.094) (0.508-1.549) (0.061-0.168) (4.266-9.012) 0.054 0.596 0.107 6.370 (0.027-0.068) (0.595-0.844) (0.047-0.157) (2.418-7.566) 0.043 0.687 0.097 5.461 Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
N-positive (n 5 14) DCE-MRI Parameters
T2 (n 5 10)
T3 (n 5 6)
T-staging
Table 2 DCE-MRI Parameters of Gastric Cancer According to T- and N-staging
T4 (n 5 6)
Correlation coefficient*
P value*
N-negative (n 5 8)
N-staging
P valuey
1612
However, although hyoscine butylbromide was used, as five patients showed severe bowel peristalsis during DCE-MRI, generation of perfusion maps of diagnostic quality failed in those patients. Our results, which show correlations between the DCE-MRI parameters and the pathologic features of gastric cancer, would suggest that DCE-MRI parameters can be used as imaging biomarkers to provide the tumor microenvironment of gastric cancer, and which may be helpful in predicting a patient’s prognosis, although further studies using a large patient cohort will be required for their validation. Furthermore, as the DCE-MRI parameters reflect the function and dynamic microenvironment of tissue, DCE-MRI is expected to provide additional information in addition to the histological features (36). Ve, which represents the volume of extracellular extravascular space (EES) per unit volume of tissue, may be enlarged in many tumors; however, it can differ substantially according to tumor types and tumor aggressiveness (36,37). In our study, Ve was significantly higher in gastric cancer compared to the normal gastric wall, had a significant positive correlation with the pathologic T-staging of gastric cancer, and was higher in poorly differentiated tubular adenocarcinoma than in moderately differentiated adenocarcinoma. In addition, iAUC, a model-free parameter of DCE-MRI reflecting tissue perfusion characteristics (38), was significantly higher in gastric cancer than in normal gastric wall, and which agrees with the previous perfusion CT study by Yao et al (39), which reported that the blood volume was significantly higher in gastric cancer than in normal gastric wall. Ktrans was also higher in gastric cancer than in normal gastric wall in our study; however, the difference was not statistically significant. Our study is the first attempt to assess the DCE-MRI parameters in gastric cancers, although there have been several published studies evaluating the DCEMRI parameters of esophageal and colorectal cancers. In esophageal cancer, squamous-cell carcinoma has been reported to show lower a Ve and higher Ktrans compared to those of adenocarcinoma (40). In rectal cancer, Kim et al (18) reported that the Ktrans was significantly higher in rectal cancer than in the normal rectal wall and that none of the DCE-MRI parameters, including Ktrans and Ve, were correlated with the Tand N-staging. Yao et al (41) reported that Ktrans
DCE-MRI of Gastric Cancer
1613
Table 4 Correlation of the DCE-MRI Parameters of Gastric Cancer With the Grades of EGFR Expression EGFR Expression DCE-MRI Parameters trans
1
(min ) K Kep (min1) Ve iAUC (mM/sec)
0 (n 5 6) 0.037 0.572 0.107 4.727
(0.025-0.061) (0.235-0.938) (0.044-0.196) (2.531-6.952)
1þ (n 5 8) 0.052 0.705 0.110 6.634
(0.028-0.074) (0.467-1.090) (0.050-0.175) (2.419-8.625)
2þ or 3þ (n 5 8) 0.064 0.598 0.140 6.682
(0.042-0.134) (0.455-0.949) (0.087-0.280) (4.945-12.824)
Correlation coefficient
P value
0.460 0.091 0.306 0.422
0.031* 0.686 0.166 0.050
Data are medians (95% confidence intervals). All data were tested using the Spearman rank correlation test. *Significant value, P < 0.05.
showed a significant correlation with the pathologic TNM staging in rectal cancer. In our study, Ktrans was positively correlated with the grades of EGFR expression in gastric cancers. EGFR is a kind of membrane receptor that is overexpressed in many different cancer types and it initiates an intracellular signal pathway that promotes cancercell proliferation, cell migration, and angiogenesis (6). In many human malignancies, including gastric cancer, an elevated level of EGFR has been reported to be associated with a poor patient prognosis (42). As Ktrans estimates the combination of blood flow and permeability properties (43), our study results would suggest that gastric cancers with a higher grade of EGFR expression may have higher tissue perfusion than those with a lower grade of EGFR expression. These results show good agreement with previous rectal cancer studies which have shown that EGFRpositive cancer displayed a higher Ktrans value than EGFR-negative cancer (44). Our study has a number of limitations. First, as it only included patients whose gastric cancers were suspected to be T2 or greater, as seen on the preoperative MDCT scan, there were no pathologically confirmed T1a cancers or well-differentiated adenocarcinomas in our patient study population. As these inclusion criteria were applied because we intended to evaluate the tumor perfusion characteristics of gastric cancers using DCE-MRI as the initial clinical experience, early gastric cancers which usually appear as subtle wall thickening or even as undetectable lesions on conventional imaging studies would not have been suitable for inclusion in this study. Second, the DCE-MRI parameters were measured on one selected axial section of the perfusion maps, which may be less representative of the entire tumor considering the 3D tumor structure. Third, among our study population five patients (18.5%), whose perfusion maps were of nondiagnostic quality, were excluded in the statistical analysis. This nonrandom loss of data potentially resulted in a bias in the interpretation of DCE-MRI data. In addition, there was a gender disparity (21 males and six females) with a wide age range (38–81 years old) in our study. The potential effects of those demographics of patients as confounding factors were not evaluated considering the small sample size of our study. Fourth, although we evaluated the test–retest intratester repeatability, we did not assess the reproducibility of the timing bolus of DCE-MRI as well as the interobserver vari-
ability when analyzing the perfusion parameters of DCE-MRI. Finally, we did not assess whether there is a direct correlation between the DCE-MRI parameters and the patient prognosis, although there were correlations between the DCE-MRI parameters and the pathological prognostic markers. Therefore, further prospective studies with a long-term follow-up are warranted in order to evaluate the role of the DCEMRI parameters as prognostic markers in patients with gastric cancer. In conclusion, DCE-MRI with radial VIBE with KWIC reconstruction was technically feasible for quantification of the perfusion dynamics of gastric cancers, and the DCE-MRI parameters of gastric cancers may provide prognostic information correlated with pathological prognostic markers.
ACKNOWLEDGMENT We thank Bonnie Hami for editing and Berthold Kiefer, PhD, and Yo Han Son, BA, for technical advice on postprocessing of DCE-MRI data using tissue 4D. REFERENCES 1. Soerjomataram I, Lortet-Tieulent J, Parkin DM, et al. Global burden of cancer in 2008: a systematic analysis of disabilityadjusted life-years in 12 world regions. Lancet 2012;380:1840– 1850. 2. De Vita F, Giuliani F, Galizia G, et al. Neo-adjuvant and adjuvant chemotherapy of gastric cancer. Ann Oncol 2007;18(Suppl 6): vi120–123. 3. Lorenzen S, Blank S, Lordick F, Siewert JR, Ott K. Prediction of response and prognosis by a score including only pretherapeutic parameters in 410 neoadjuvant treated gastric cancer patients. Ann Surg Oncol 2012;19:2119–2127. 4. Chae S, Lee A, Lee JH. The effectiveness of the new (7th) UICC N classification in the prognosis evaluation of gastric cancer patients: a comparative study between the 5th/6th and 7th UICC N classification. Gastric Cancer 2011;14:166–171. 5. Yasui W, Oue N, Aung PP, Matsumura S, Shutoh M, Nakayama H. Molecular-pathological prognostic factors of gastric cancer: a review. Gastric Cancer 2005;8:86–94. 6. Galizia G, Lieto E, Orditura M, et al. Epidermal growth factor receptor (EGFR) expression is associated with a worse prognosis in gastric cancer patients undergoing curative surgery. World J Surg 2007;31:1458–1468. 7. Satoh A, Shuto K, Okazumi S, et al. Role of perfusion CT in assessing tumor blood flow and malignancy level of gastric cancer. Dig Surg 2010;27:253–260. 8. Zhang H, Pan Z, Du L, et al. Advanced gastric cancer and perfusion imaging using a multidetector row computed tomography: correlation with prognostic determinants. Korean J Radiol 2008; 9:119–127.
1614 9. Ang J, Hu L, Huang PT, et al. Contrast-enhanced ultrasonography assessment of gastric cancer response to neoadjuvant chemotherapy. World J Gastroenterol 2012;18:7026–7032. 10. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging 1999;10:223–232. 11. Choyke PL, Dwyer AJ, Knopp MV. Functional tumor imaging with dynamic contrast-enhanced magnetic resonance imaging. J Magn Reson Imaging 2003;17:509–520. 12. O’Connor JP, Jackson A, Parker GJ, Jayson GC. DCE-MRI biomarkers in the clinical evaluation of antiangiogenic and vascular disrupting agents. Br J Cancer 2007;96:189–195. 13. Liu G, Rugo HS, Wilding G, et al. Dynamic contrast-enhanced magnetic resonance imaging as a pharmacodynamic measure of response after acute dosing of AG-013736, an oral angiogenesis inhibitor, in patients with advanced solid tumors: results from a phase I study. J Clin Oncol 2005;23:5464–5473. 14. Hylton N. Dynamic contrast-enhanced magnetic resonance imaging as an imaging biomarker. J Clin Oncol 2006;24:3293–3298. 15. Yuh WT, Mayr NA, Jarjoura D, et al. Predicting control of primary tumor and survival by DCE MRI during early therapy in cervical cancer. Invest Radiol 2009;44:343–350. 16. 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. 17. Tuncbilek N, Karakas HM, Altaner S. Dynamic MRI in indirect estimation of microvessel density, histologic grade, and prognosis in colorectal adenocarcinomas. Abdom Imaging 2004;29:166–172. 18. Kim YE, Lim JS, Choi J, et al. Perfusion parameters of dynamic contrast-enhanced magnetic resonance imaging in patients with rectal cancer: correlation with microvascular density and vascular endothelial growth factor expression. Korean J Radiol 2013; 14:878–885. 19. Orton MR, Miyazaki K, Koh DM, et al. Optimizing functional parameter accuracy for breath-hold DCE-MRI of liver tumours. Phys Med Biol 2009;54:2197–2215. 20. Yang RK, Roth CG, Ward RJ, deJesus JO, Mitchell DG. Optimizing abdominal MR imaging: approaches to common problems. Radiographics 2010;30:185–199. 21. Lane BF, Vandermeer FQ, Oz RC, Irwin EW, McMillan AB, WongYou-Cheong JJ. Comparison of sagittal T2-weighted BLADE and fast spin-echo MRI of the female pelvis for motion artifact and lesion detection. AJR Am J Roentgenol 2011;197:W307–313. 22. Hirokawa Y, Isoda H, Maetani YS, Arizono S, Shimada K, Togashi K. MRI artifact reduction and quality improvement in the upper abdomen with PROPELLER and prospective acquisition correction (PACE) technique. AJR Am J Roentgenol 2008;191:1154– 1158. 23. Song HK, Dougherty L. k-space weighted image contrast (KWIC) for contrast manipulation in projection reconstruction MRI. Magn Reson Med 2000;44:825–832. 24. Song HK, Yan L, Smith RX, et al. Noncontrast enhanced fourdimensional dynamic MRA with golden angle radial acquisition and k-space weighted image contrast (KWIC) reconstruction. Magn Reson Med 2013 [Epub ahead of print]. 25. Kim KW, Lee JM, Jeon YS, et al. Free-breathing dynamic contrast-enhanced MRI of the abdomen and chest using a radial gradient echo sequence with k-space weighted image contrast (KWIC). Eur Radiol 2013;23:1352–1360.
Joo et al. 26. 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. 27. Priest AN, Gill AB, Kataoka M, et al. Dynamic contrast-enhanced MRI in ovarian cancer: Initial experience at 3 Tesla in primary and metastatic disease. Magn Reson Med 2010;63:1044–1049. 28. Tofts PS, Kermode AG. Measurement of the blood-brain barrier permeability and leakage space using dynamic MR imaging. 1. Fundamental concepts. Magn Reson Med 1991;17:357–367. 29. Orton MR, d’Arcy JA, Walker-Samuel S, et al. Computationally efficient vascular input function models for quantitative kinetic modelling using DCE-MRI. Phys Med Biol 2008;53:1225–1239. 30. Washington K. 7th edition of the AJCC cancer staging manual: stomach. Ann Surg Oncol 2010;17:3077–3079. 31. Brevet M, Arcila M, Ladanyi M. Assessment of EGFR mutation status in lung adenocarcinoma by immunohistochemistry using antibodies specific to the two major forms of mutant EGFR. J Mol Diagn 2010;12:169–176. 32. Chinn S. Statistics in respiratory medicine. 2. Repeatability and method comparison. Thorax 1991;46:454–456. 33. Hu B, El Hajj N, Sittler S, Lammert N, Barnes R, Meloni-Ehrig A. Gastric cancer: classification, histology and application of molecular pathology. J Gastrointest Oncol 2012;3:251–261. 34. Yao J, Yang ZG, Chen HJ, Chen TW, Huang J. Gastric adenocarcinoma: can perfusion CT help to noninvasively evaluate tumor angiogenesis? Abdom Imaging 2011;36:15–21. 35. Banerjee R, Pavlides M, Tunnicliffe EM, et al. Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease. J Hepatol 2014;60:69–77. 36. Aref M, Chaudhari AR, Bailey KL, Aref S, Wiener EC. Comparison of tumor histology to dynamic contrast enhanced magnetic resonance imaging-based physiological estimates. Magn Reson Imaging 2008;26:1279–1293. 37. Benjaminsen IC, Brurberg KG, Ruud EB, Rofstad EK. Assessment of extravascular extracellular space fraction in human melanoma xenografts by DCE-MRI and kinetic modeling. Magn Reson Imaging 2008;26:160–170. 38. Roberts C, Issa B, Stone A, Jackson A, Waterton JC, Parker GJ. Comparative study into the robustness of compartmental modeling and model-free analysis in DCE-MRI studies. J Magn Reson Imaging 2006;23:554–563. 39. Yao J, Yang ZG, Chen TW, Li Y, Yang L. Perfusion changes in gastric adenocarcinoma: evaluation with 64-section MDCT. Abdom Imaging 2010;35:195–202. 40. Oberholzer K, Pohlmann A, Schreiber W, et al. Assessment of tumor microcirculation with dynamic contrast-enhanced MRI in patients with esophageal cancer: initial experience. J Magn Reson Imaging 2008;27:1296–1301. 41. Yao WW, Zhang H, Ding B, et al. Rectal cancer: 3D dynamic contrast-enhanced MRI; correlation with microvascular density and clinicopathological features. Radiol Med 2011;116:366–374. 42. Nicholson RI, Gee JM, Harper ME. EGFR and cancer prognosis. Eur J Cancer 2001;37 Suppl 4:S9–15. 43. Turkbey B, Thomasson D, Pang Y, Bernardo M, Choyke PL. The role of dynamic contrast-enhanced MRI in cancer diagnosis and treatment. Diagn Interv Radiol 2010;16:186–192. 44. Yeo DM, Oh SN, Jung CK, et al. Correlation of dynamic contrastenhanced MRI perfusion parameters with angiogenesis and biologic aggressiveness of rectal cancer: preliminary results. J Magn Reson Imaging 2013 [Epub ahead of print].
Original Research
Dynamic Contrast-Enhanced MRI of Gastric Cancer: Correlation of the Perfusion Parameters With Pathological Prognostic Factors Ijin Joo, MD,1 Jeong Min Lee, MD,1,2* Joon Koo Han, MD,1,2 Han-Kwang Yang, MD,3 Hyuk-Joon Lee, MD,3 and Byung Ihn Choi, MD1,2 Key Words: gastric cancer; DCE-MRI; perfusion J. Magn. Reson. Imaging 2015;41:1608–1614. C 2014 Wiley Periodicals, Inc. V
Purpose: To investigate the feasibility of dynamic, contrast-enhanced, magnetic resonance imaging (DCEMRI) for perfusion quantification of gastric cancers, and to correlate the DCE-MRI parameters with the pathological prognostic factors. Materials and Methods: This prospective study was approved by our Institutional Review Board. Twentyseven patients with gastric cancers underwent DCE-MRI using a free-breathing, radial, gradient-echo (GRE) sequence with k-space weighted image contrast (KWIC) reconstruction on a 3T scanner. The DCE-MRI parameters (volume transfer coefficient [Ktrans], reverse reflux rate constant [Kep], extracellular extravascular volume fraction [Ve], and initial area under the gadolinium concentration curve during the first 60 seconds [iAUC]) of gastric cancer and normal wall were measured and compared with each other using the Wilcoxon signed rank test. The relationship between the DCE-MRI parameters of gastric cancer and the pathological prognostic factors were evaluated using the Mann–Whitney test or the Spearman rank correlation test. Results: DCE-MRIs were of diagnostic quality in 22 patients (81.5%). Ve and iAUC were significantly higher in gastric cancer than in normal gastric wall (P < 0.05). Ve showed significant positive correlation with T-staging of gastric cancers (P < 0.05). Ktrans was significantly correlated with the grades of epidermal growth-factor receptor expression (P < 0.05). Conclusion: DCE-MRI using a radial GRE with KWIC reconstruction is feasible for quantification of the perfusion dynamics of gastric cancers, and the DCE-MRI parameters of gastric cancers may provide prognostic information.
1 Department of Radiology, Seoul National University Hospital, Seoul, Korea. 2 Institute of Radiation Medicine, Seoul National University Hospital, Seoul, Korea. 3 Department of Surgery, Seoul National University College of Medicine, Korea. Contract grant sponsor: GE Healthcare. *Address reprint requests to: J.M.L., Department of Radiology, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 110-744, Korea. E-mail: [email protected] Received March 31, 2014; Accepted July 3, 2014. DOI 10.1002/jmri.24711 View this article online at wileyonlinelibrary.com. C 2014 Wiley Periodicals, Inc. V
GASTRIC CANCER is one of the main causes of cancer death worldwide, and while surgical resection is the only potentially curative treatment, the long-term survival of patients with advanced gastric cancers (AGCs) remains poor even following surgical resection (1,2). In order to further improve the outcomes of patients with AGC, many clinical trials have been performed to evaluate the role of chemotherapy, radiation therapy, and combined-modality treatment (2). Given that there are various therapeutic options for AGC, identification of pretherapeutic, predictive markers for both the treatment response and prognosis is essential for a personalized cancer treatment (3). Pathologic TNM staging, the histological grades of tumor differentiation, and expression of molecular markers such as the epidermal growth-factor receptor (EGFR) are regarded as prognostic factors of gastric cancers (4–6). Several recent studies reported that the perfusion parameters of gastric cancers obtained by perfusion computed tomography (CT) or contrast-enhanced ultrasound may be useful for both tissue characterization and treatment monitoring (7–9). Dynamic, contrast-enhanced magnetic resonance imaging (DCE-MRI) is a currently evolving imaging method used to assess the functional features of a target tissue, including tissue blood flow, capillary permeability, and interstitial volume (10). DCE-MRI has actually become increasingly used recently for the characterization, differential diagnosis of various cancers, and for monitoring changes in tumor perfusion during anticancer treatment (11–13). Many oncologic studies have also assessed the potential role of DCEMRI as a prognostic indicator (14,15) as well as the correlation of DCE-MRI and other established or possible prognostic factors (16–18). However, in the acquisition of DCE-MRI data regarding abdominal lesions, the image quality may be degraded due to motion artifacts related to respiration and bowel peristalsis during the acquisition, and thus resulting in
1608
DCE-MRI of Gastric Cancer
interframe misregistration (19). To reduce motionrelated artifacts and improve the image registration, fast imaging sequences and new methods of sampling k-space have been developed and used in abdominal imaging (20–22). K-space weighted image contrast (KWIC) reconstruction is a kind of data-weighting system used for radially encoded MRI data in which the central k-space is oversampled and occupied by the selected radial views, while the peripheral k-space includes time-neighboring subsets of radial views and would, therefore, simultaneously provide high spatial and high temporal resolution (23,24). Using this concept, several previous studies have reported the feasibility of free-breathing radial DCE-MRI using a radial, 3D, gradient-echo (GRE) technique (radial volumetric interpolated, breath-hold examination [VIBE]; Siemens Healthcare, Erlangen, Germany) with KWIC reconstruction in the evaluation of abdominal lesions (25,26). However, there have been no previous studies that evaluated the technical feasibility and diagnostic value of DCE-MRI for patients with gastric cancers. Therefore, the purposes of our study were to investigate the feasibility of DCE-MRI for perfusion quantification of gastric cancers using a radial GRE sequence with KWIC reconstruction and to correlate the DCE-MRI parameters with the pathological prognostic factors.
MATERIALS AND METHODS Patients This prospective study was approved by the Institutional Review Board of Seoul National University Hospital, and informed consent was obtained from each patient. From January 2011 to September 2011, 28 consecutive patients who met the following criteria were enrolled in this study: 1) patients with endoscopic, biopsy-confirmed gastric cancers and 2) patients whose gastric cancers were suspected to be T2 or greater, any N, and M0 stage, based on the endoscopic examination and as well as on the preoperative MDCT scan. All patients underwent preoperative DCEMRI on a 3T scanner. Among the 28 patients, one patient who did not undergo prompt surgery due to active pulmonary tuberculosis was excluded. Therefore, 27 patients (21 males and six females) with an age range of 38–81 years were finally included. The time interval from DCE-MRI to surgery was within 2 weeks. MRI Data Acquisition MRI studies were performed on a 3T MR scanner (TIM Trio; Siemens Healthcare) with a standard, 24-channel, phased-array body coil and in the supine position. Each patient received an intravenous injection of 10 mg of hyoscine butylbromide (Buscopan; Boehringer Ingelheim Korea, Seoul, Korea) 5 minutes prior to the MRI study in order to reduce bowel motion and was asked to drink 500–1000 mL of water to distend the stomach immediately before beginning the MRI acquisition. Using the three-plane, true-fast imaging with steady-state precession (true-FISP) localizers, an axial MRI slab was placed from the dome of the liver to the
1609
right renal hilum. Unenhanced studies, including half-Fourier acquisition, single-shot, turbo spine-echo (HASTE), T2-weighted imaging with and without fat saturation, true-FISP, T1-weighted, 3D, GRE, in- and out-of-phase imaging, and T1-weighted fat-suppressed 3D GRE imaging were performed. For T1 mapping, unenhanced T1-weighted VIBE images were obtained using three different flip angles with the following parameters: relaxation time (TR) ¼ 5.3 msec; echo time (TE) ¼ 1.1 msec; flip angles (a) ¼ 2 , 8 , and 15 ; slice thickness ¼ 5 mm; number of excitations (NEX) ¼ 3; field of view (FOV) ¼ 38 38 cm2; matrix ¼ 128 96; and number of slices ¼ 40. T1 maps were subsequently generated using a threeparameter, nonlinear fitting algorithm to the MR data for each voxel (27). Thereafter, DCE-MRI sequences using free-breathing, radial 3D VIBE with KWIC reconstruction (25) performed over the same volume as the T1 maps, were performed after an intravenous bolus injection of 0.1 mmol/kg of gadolinium contrast agent (Omniscan, gadodiamide) at a rate of 3 mL/sec followed by a chasing bolus of 20 mL of normal saline administered using an MR-compatible power injector (SpectrisSolaris; Medrad, Indianola, PA). All patients were instructed to breathe as quietly as possible during the examination. The DCE-MRI parameters were as follows: TR ¼ 3.8 msec; TE ¼ 1.6 msec; flip angle ¼ 11 ; slice thickness ¼ 5 mm with a 50% slice resolution (10/5 mm, actual/interpolated slice thickness); NEX ¼ 1; FOV ¼ 38 38 cm2; matrix ¼ 256 256; and the number of slices ¼ 20. To reconstruct time-resolved, subframe images, the KWIC view-sharing technique which permits the manipulation of image contrast by careful filtering of the acquired views was used (26). Radial VIBE sequences were continuously scanned for 75 seconds, repeated volumetric sets of axial images at a 4.1-second intervals for 308 seconds. The spectral selection, attenuated inversion technique was used for fat suppression. DCE-MRI Analysis Postprocessing of DCE-MRIs was performed by one abdominal radiologist (I.J. with 7 years of clinical experience in abdominal MRI) who was aware that the patients had gastric cancers, but who was blinded to the pathological features. From the DCE-MRI data, voxelwise, parametric maps of the volume transfer coefficient (Ktrans), reverse reflux rate constant (Kep), extracellular extravascular volume fraction (Ve), and the initial area under the gadolinium concentration curve during the first 60 seconds (iAUC) were generated using a postprocessing platform with dedicated DCE-MRI software (Tissue4D; Siemens Healthcare) (26). Before analysis of the DCE-MRI data, motion correction of the subframe radial VIBE using KWIC reconstruction image sets was performed using the nonrigid registration technique of the software. A twocompartment Tofts model was applied (10,28), and the arterial input function was chosen according the fast sampling method (29). Using the software, by drawing a freehand region of interest (ROI), Ktrans, Kep, Ve, and iAUC values for the selected ROI were automatically calculated. Two ROIs were drawn on
1610
Joo et al.
the dynamic, T1-weighted VIBE images: one ROI covered the gastric cancer by outlining the edges of the tumor on the section of central level of the tumor, which was referenced on the corresponding T2-weighted axial images, while the other ROI covered the normal gastric wall adjacent to the tumor (Fig. 1). Two abdominal radiologists (J.M.L. and I.J. with 22 years and 7 years, respectively, of clinical experience performing abdominal MRI) determined in consensus the image quality of the perfusion maps as nondiagnostic or diagnostic according to the following criteria: the perfusion map containing apparent errors in image registration probably due to motion was considered nondiagnostic; otherwise, the image quality was diagnostic (25). The DCE-MRI perfusion parameters in any patient showing nondiagnostic image quality were excluded in the analysis. To evaluate the test–retest repeatability of the DCE-MRI parameters of gastric cancers, the same radiologist (I.J.) who had performed the postprocessing of DCE-MRIs repeated the analysis of DCE-MRIs of gastric cancers at least 4 weeks after the initial analysis. Pathological Analysis The 27 patients who were finally included in this study underwent curative or palliative gastrectomy and lymph node (LN) dissection. Pathologic TNM staging of gastric cancers was reported based on the AJCC 7th guidelines (30). In patients with tubular adenocarcinoma, the histological grade was reported as well-, moderately, or poorly differentiated. The EGFR expression determined by immunohistochemistry was scored as follows: 0, no membrane staining; 1þ, faint or partial staining; 2þ, moderate staining; and 3þ, strong staining (31). Statistical Analysis To compare the DCE-MRI parameters between gastric cancer and normal wall, the Wilcoxon signed rank test was performed. The DCE-MRI parameters were compared between the N-negative and N-positive stage of gastric cancers and between well- or moderately differentiated and poorly differentiated tubular adenocarcinomas using the Mann–Whitney test. The Spearman rank correlation test was performed to evaluate the correlation of the DCE-MRI parameters of gastric cancers with T-staging (T2, T3, T4) and the grades of EGFR expression (0, 1þ, 2þ). P < 0.05 was considered to indicate statistical significance. To obtain the test–retest repeatability of DCE-MRI parameters of gastric cancers, coefficients of variation (CVs ¼ standard deviation divided by mean) within subjects were calculated (32). All statistical analyses were performed using MedCalc software v. 12.4.0.0 (Mariakerke, Belgium). RESULTS DCE-MRI in Patients With Gastric Cancer As the perfusion maps from the DCE-MRI data were of diagnostic quality in 22 (81.5%) of 27 patients, those 22 patients were finally included in the statistical anal-
Figure 1. MR images of a 69-year-old male with a histopathologically confirmed stage T4aN3b gastric cancer. a: On the contrast-enhanced axial T1-weighted image, an ulceroinfiltrative mass with transmural enhancement (arrows) is seen in the gastric antrum. On the corresponding parametric map of Ktrans (b) displayed in colors ranging from blue to red, gastric cancer (green ROI) shows a wide range of Ktrans, while normal gastric wall (yellow ROI) shows homogeneously low Ktrans. Parametric map of Ve (c) shows relatively higher Ve in the gastric cancer (green ROI) than in the normal gastric wall (yellow ROI).
ysis. Five patients were excluded because their perfusion maps were nondiagnostic and showed apparent in-plane misregistration from the serial DCE-MRI data caused by severe peristaltic movement of the stomach.
DCE-MRI of Gastric Cancer
1611
Table 1 DCE-MRI Parameters of Gastric Cancer and Normal Gastric Wall DCE-MRI Parameters Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
Gastric Cancer 0.046 0.623 0.133 5.533
(0.042-0.065) (0.522-0.753) (0.088-0.173) (4.907-7.734)
Normal Wall 0.037 0.785 0.063 3.894
(0.028-0.053) (0.522-1.048) (0.044-0.083) (3.214-5.613)
P value 0.085 0.200 0.004* 0.039*
Data are medians (95% confidence intervals). All data were tested using the Wilcoxon signed rank test. *Significant value, P < 0.05.
T-staging, N-staging, histological grades, and grades of EGFR expression, are shown in Tables (2–4). Among the DCE-MRI parameters, Ve showed a significant correlation with T-staging (rho ¼ 0.483, P ¼ 0.023) (Table 2). None of the DCE-MRI parameters of gastric cancer showed significant differences in Nnegative and N-positive cases (P > 0.05). In cases of tubular adenocarcinoma, poorly differentiated carcinomas showed a higher Ve than moderately differentiated carcinomas, however, the difference was not statistically significant (P > 0.05) (Table 3). Ktrans was significantly correlated with the grades of EGFR expression (rho ¼ 0.460, P ¼ 0.031) (Table 4).
Test–Retest Repeatability of DCE-MRI Parameters of Gastric Cancers In 22 patients whose perfusion maps were of diagnostic quality, DCE-MRI parameters of gastric cancers showed good or moderate test–retest repeatability with CVs of Ktrans, Kep, Ve, and iAUC of 12.3%, 25.7%, 20.8%, and 17.9%, respectively. Pathological Results T-stages in five excluded patients with nondiagnostic image quality of the perfusion maps were confirmed as follows: T1b, n ¼ 2; T2, n ¼ 1; T3, n ¼ 1; and T4a, n ¼ 1. Among the 22 study patients, two were confirmed as having stage T1b cancer, eight with T2, six with T3, five with T4a, and one with T4b (T2, n ¼ 10; T3, n ¼ 6; and T4, n ¼ 6), and eight were N-negative and 14 N-positive. One patient was pathologically confirmed to have M1 stage peritoneal seeding, while the other 21 patients had M0 stage. The histologic patterns of the gastric cancers according to the 2010 WHO classification were as follows: 17 tubular adenocarcinomas; one papillary adenocarcinoma; three mucinous and poorly cohesive carcinomas (including signet-ring-cell carcinoma); and one with uncommon histologic variants (carcinoma with lymphoid stroma) (33). The histological grades of the 17 tubular adenocarcinomas were confirmed as moderately differentiated carcinomas in nine patients and poorly differentiated carcinomas in eight patients. The grades of EGFR expression were 0 in six patients, 1þ in eight, and 2þ or 3þ in eight patients. Comparison of the DCE-MRI Parameters Between Gastric Cancer and Normal Wall The DCE-MRI parameters of gastric cancer and normal wall are summarized in Table 1. Among the DCEMRI parameters, the medians of Ve and iAUC were significantly higher in gastric cancer than those in normal wall (Ve, 0.133 vs. 0.063; iAUC, 5.533 vs. 3.894 mM/sec, respectively) (P < 0.05) (Fig. 1). Correlation of the DCE-MRI Parameters of Gastric Cancer With the Pathological Prognostic Factors Correlation of the DCE-MRI parameters of gastric cancer with the pathological prognostic factors, including
DISCUSSION Our study demonstrates that DCE-MRI is feasible for the evaluation of gastric cancers, using the freebreathing radial 3D-GRE technique with KWIC reconstruction. Some of the DCE-MRI parameters of gastric cancer also showed a significant correlation with the pathological prognostic factors, including T-staging, and grades of EGFR-expression, ie, Ve showed significant positive correlation with T-staging and Ktrans was significantly correlated with the grades of EGFR expression. These results suggest that DCE-MRI can be used to characterize the tumor biology of gastric cancer, and may have the potential to provide additional prognostic information regarding patients with gastric cancer. Our study results are in good agreement with those of several recent studies using perfusion CT and in which the tumor perfusion measurement of gastric cancer and its possible role in clinical practice have been reported (7,8,34). However, considering that the clinical use of perfusion CT may be limited by its relatively high radiation dose (15– 25 mSv), if our study results are validated by larger studies, DCE-MRI may provide additional benefits over those of perfusion CT. Furthermore, the multiparametric capability of MR for oncologic evaluation such as diffusion-weighted imaging, T1 or T2 relaxometry, or MR elastography, could be an additional advantage over the use of CT (35). In our study, perfusion parametric maps of diagnostic quality were successfully obtained in 81.5% of the patients from the DCE-MR image sets using the free-breathing, 3D-radial VIBE with KWIC technique. In fact, there have been no previous reports of DCEMRI used for the evaluation of gastric cancers. This may be due to the poor temporal resolution and susceptibility to motion of T1-weighted sequences used for DCE-MRI, compared with CT scans, and which may result in inevitable respiratory and bowel motionrelated artifacts (22). Our study result may be primarily due to the advantages of free-breathing radial VIBE with KWIC reconstruction, which provides high spatial and temporal resolution by oversampling of the central k-space region with low spatial frequency data and fewer respiration-related artifacts due to its inherent motion insensitivity (25,26). Furthermore, in our study hyoscine butylbromide was injected before MRI scanning in order to reduce bowel peristalsis.
Joo et al. 0.608 0.973 0.759 0.402
Table 3 DCE-MRI Parameters According to the Histological Grades of Tumor Differentiation in Tubular Adenocarcinoma of the Stomach
(0.031-0.072) (0.356-0.953) (0.078-0.178) (4.055-7.927)
DCE-MRI Parameters
0.043 0.596 0.139 5.242
Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
Moderately Differentiated (n ¼ 9) 0.044 0.568 0.105 5.537
(0.031-0.059) (0.383-0.747) (0.060-0.141) (4.184-7.778)
Poorly Differentiated (n ¼ 7) 0.065 0.653 0.138 7.295
(0.042-0.139) (0.443-1.047) (0.098-0.291) (4.939-13.256)
P value* 0.186 0.408 0.142 0.470
0.199 0.400 0.483 0.158
0.374 0.065 0.023z 0.484
0.057 0.637 0.107 6.639
(0.039-0.080) (0.537-0.724) (0.082-0.194) (4.511-9.533)
Data are medians (95% confidence intervals). All data were tested using the Mann-Whitney test. *Significant value, P < 0.05.
Data are medians (95% confidence intervals). *Data were tested using the Spearman rank correlation test. y Data were tested using the Mann-Whitney test. z Significant value, P < 0.05.
0.050 (0.034-0.135) 0.419 (0.235-1.028) 0.187 (0.136-0.297) 5.755 (4.128-12.189) (0.030-0.094) (0.508-1.549) (0.061-0.168) (4.266-9.012) 0.054 0.596 0.107 6.370 (0.027-0.068) (0.595-0.844) (0.047-0.157) (2.418-7.566) 0.043 0.687 0.097 5.461 Ktrans (min1) Kep (min1) Ve iAUC (mM/sec)
N-positive (n 5 14) DCE-MRI Parameters
T2 (n 5 10)
T3 (n 5 6)
T-staging
Table 2 DCE-MRI Parameters of Gastric Cancer According to T- and N-staging
T4 (n 5 6)
Correlation coefficient*
P value*
N-negative (n 5 8)
N-staging
P valuey
1612
However, although hyoscine butylbromide was used, as five patients showed severe bowel peristalsis during DCE-MRI, generation of perfusion maps of diagnostic quality failed in those patients. Our results, which show correlations between the DCE-MRI parameters and the pathologic features of gastric cancer, would suggest that DCE-MRI parameters can be used as imaging biomarkers to provide the tumor microenvironment of gastric cancer, and which may be helpful in predicting a patient’s prognosis, although further studies using a large patient cohort will be required for their validation. Furthermore, as the DCE-MRI parameters reflect the function and dynamic microenvironment of tissue, DCE-MRI is expected to provide additional information in addition to the histological features (36). Ve, which represents the volume of extracellular extravascular space (EES) per unit volume of tissue, may be enlarged in many tumors; however, it can differ substantially according to tumor types and tumor aggressiveness (36,37). In our study, Ve was significantly higher in gastric cancer compared to the normal gastric wall, had a significant positive correlation with the pathologic T-staging of gastric cancer, and was higher in poorly differentiated tubular adenocarcinoma than in moderately differentiated adenocarcinoma. In addition, iAUC, a model-free parameter of DCE-MRI reflecting tissue perfusion characteristics (38), was significantly higher in gastric cancer than in normal gastric wall, and which agrees with the previous perfusion CT study by Yao et al (39), which reported that the blood volume was significantly higher in gastric cancer than in normal gastric wall. Ktrans was also higher in gastric cancer than in normal gastric wall in our study; however, the difference was not statistically significant. Our study is the first attempt to assess the DCE-MRI parameters in gastric cancers, although there have been several published studies evaluating the DCEMRI parameters of esophageal and colorectal cancers. In esophageal cancer, squamous-cell carcinoma has been reported to show lower a Ve and higher Ktrans compared to those of adenocarcinoma (40). In rectal cancer, Kim et al (18) reported that the Ktrans was significantly higher in rectal cancer than in the normal rectal wall and that none of the DCE-MRI parameters, including Ktrans and Ve, were correlated with the Tand N-staging. Yao et al (41) reported that Ktrans
DCE-MRI of Gastric Cancer
1613
Table 4 Correlation of the DCE-MRI Parameters of Gastric Cancer With the Grades of EGFR Expression EGFR Expression DCE-MRI Parameters trans
1
(min ) K Kep (min1) Ve iAUC (mM/sec)
0 (n 5 6) 0.037 0.572 0.107 4.727
(0.025-0.061) (0.235-0.938) (0.044-0.196) (2.531-6.952)
1þ (n 5 8) 0.052 0.705 0.110 6.634
(0.028-0.074) (0.467-1.090) (0.050-0.175) (2.419-8.625)
2þ or 3þ (n 5 8) 0.064 0.598 0.140 6.682
(0.042-0.134) (0.455-0.949) (0.087-0.280) (4.945-12.824)
Correlation coefficient
P value
0.460 0.091 0.306 0.422
0.031* 0.686 0.166 0.050
Data are medians (95% confidence intervals). All data were tested using the Spearman rank correlation test. *Significant value, P < 0.05.
showed a significant correlation with the pathologic TNM staging in rectal cancer. In our study, Ktrans was positively correlated with the grades of EGFR expression in gastric cancers. EGFR is a kind of membrane receptor that is overexpressed in many different cancer types and it initiates an intracellular signal pathway that promotes cancercell proliferation, cell migration, and angiogenesis (6). In many human malignancies, including gastric cancer, an elevated level of EGFR has been reported to be associated with a poor patient prognosis (42). As Ktrans estimates the combination of blood flow and permeability properties (43), our study results would suggest that gastric cancers with a higher grade of EGFR expression may have higher tissue perfusion than those with a lower grade of EGFR expression. These results show good agreement with previous rectal cancer studies which have shown that EGFRpositive cancer displayed a higher Ktrans value than EGFR-negative cancer (44). Our study has a number of limitations. First, as it only included patients whose gastric cancers were suspected to be T2 or greater, as seen on the preoperative MDCT scan, there were no pathologically confirmed T1a cancers or well-differentiated adenocarcinomas in our patient study population. As these inclusion criteria were applied because we intended to evaluate the tumor perfusion characteristics of gastric cancers using DCE-MRI as the initial clinical experience, early gastric cancers which usually appear as subtle wall thickening or even as undetectable lesions on conventional imaging studies would not have been suitable for inclusion in this study. Second, the DCE-MRI parameters were measured on one selected axial section of the perfusion maps, which may be less representative of the entire tumor considering the 3D tumor structure. Third, among our study population five patients (18.5%), whose perfusion maps were of nondiagnostic quality, were excluded in the statistical analysis. This nonrandom loss of data potentially resulted in a bias in the interpretation of DCE-MRI data. In addition, there was a gender disparity (21 males and six females) with a wide age range (38–81 years old) in our study. The potential effects of those demographics of patients as confounding factors were not evaluated considering the small sample size of our study. Fourth, although we evaluated the test–retest intratester repeatability, we did not assess the reproducibility of the timing bolus of DCE-MRI as well as the interobserver vari-
ability when analyzing the perfusion parameters of DCE-MRI. Finally, we did not assess whether there is a direct correlation between the DCE-MRI parameters and the patient prognosis, although there were correlations between the DCE-MRI parameters and the pathological prognostic markers. Therefore, further prospective studies with a long-term follow-up are warranted in order to evaluate the role of the DCEMRI parameters as prognostic markers in patients with gastric cancer. In conclusion, DCE-MRI with radial VIBE with KWIC reconstruction was technically feasible for quantification of the perfusion dynamics of gastric cancers, and the DCE-MRI parameters of gastric cancers may provide prognostic information correlated with pathological prognostic markers.
ACKNOWLEDGMENT We thank Bonnie Hami for editing and Berthold Kiefer, PhD, and Yo Han Son, BA, for technical advice on postprocessing of DCE-MRI data using tissue 4D. REFERENCES 1. Soerjomataram I, Lortet-Tieulent J, Parkin DM, et al. Global burden of cancer in 2008: a systematic analysis of disabilityadjusted life-years in 12 world regions. Lancet 2012;380:1840– 1850. 2. De Vita F, Giuliani F, Galizia G, et al. Neo-adjuvant and adjuvant chemotherapy of gastric cancer. Ann Oncol 2007;18(Suppl 6): vi120–123. 3. Lorenzen S, Blank S, Lordick F, Siewert JR, Ott K. Prediction of response and prognosis by a score including only pretherapeutic parameters in 410 neoadjuvant treated gastric cancer patients. Ann Surg Oncol 2012;19:2119–2127. 4. Chae S, Lee A, Lee JH. The effectiveness of the new (7th) UICC N classification in the prognosis evaluation of gastric cancer patients: a comparative study between the 5th/6th and 7th UICC N classification. Gastric Cancer 2011;14:166–171. 5. Yasui W, Oue N, Aung PP, Matsumura S, Shutoh M, Nakayama H. Molecular-pathological prognostic factors of gastric cancer: a review. Gastric Cancer 2005;8:86–94. 6. Galizia G, Lieto E, Orditura M, et al. Epidermal growth factor receptor (EGFR) expression is associated with a worse prognosis in gastric cancer patients undergoing curative surgery. World J Surg 2007;31:1458–1468. 7. Satoh A, Shuto K, Okazumi S, et al. Role of perfusion CT in assessing tumor blood flow and malignancy level of gastric cancer. Dig Surg 2010;27:253–260. 8. Zhang H, Pan Z, Du L, et al. Advanced gastric cancer and perfusion imaging using a multidetector row computed tomography: correlation with prognostic determinants. Korean J Radiol 2008; 9:119–127.
1614 9. Ang J, Hu L, Huang PT, et al. Contrast-enhanced ultrasonography assessment of gastric cancer response to neoadjuvant chemotherapy. World J Gastroenterol 2012;18:7026–7032. 10. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging 1999;10:223–232. 11. Choyke PL, Dwyer AJ, Knopp MV. Functional tumor imaging with dynamic contrast-enhanced magnetic resonance imaging. J Magn Reson Imaging 2003;17:509–520. 12. O’Connor JP, Jackson A, Parker GJ, Jayson GC. DCE-MRI biomarkers in the clinical evaluation of antiangiogenic and vascular disrupting agents. Br J Cancer 2007;96:189–195. 13. Liu G, Rugo HS, Wilding G, et al. Dynamic contrast-enhanced magnetic resonance imaging as a pharmacodynamic measure of response after acute dosing of AG-013736, an oral angiogenesis inhibitor, in patients with advanced solid tumors: results from a phase I study. J Clin Oncol 2005;23:5464–5473. 14. Hylton N. Dynamic contrast-enhanced magnetic resonance imaging as an imaging biomarker. J Clin Oncol 2006;24:3293–3298. 15. Yuh WT, Mayr NA, Jarjoura D, et al. Predicting control of primary tumor and survival by DCE MRI during early therapy in cervical cancer. Invest Radiol 2009;44:343–350. 16. 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. 17. Tuncbilek N, Karakas HM, Altaner S. Dynamic MRI in indirect estimation of microvessel density, histologic grade, and prognosis in colorectal adenocarcinomas. Abdom Imaging 2004;29:166–172. 18. Kim YE, Lim JS, Choi J, et al. Perfusion parameters of dynamic contrast-enhanced magnetic resonance imaging in patients with rectal cancer: correlation with microvascular density and vascular endothelial growth factor expression. Korean J Radiol 2013; 14:878–885. 19. Orton MR, Miyazaki K, Koh DM, et al. Optimizing functional parameter accuracy for breath-hold DCE-MRI of liver tumours. Phys Med Biol 2009;54:2197–2215. 20. Yang RK, Roth CG, Ward RJ, deJesus JO, Mitchell DG. Optimizing abdominal MR imaging: approaches to common problems. Radiographics 2010;30:185–199. 21. Lane BF, Vandermeer FQ, Oz RC, Irwin EW, McMillan AB, WongYou-Cheong JJ. Comparison of sagittal T2-weighted BLADE and fast spin-echo MRI of the female pelvis for motion artifact and lesion detection. AJR Am J Roentgenol 2011;197:W307–313. 22. Hirokawa Y, Isoda H, Maetani YS, Arizono S, Shimada K, Togashi K. MRI artifact reduction and quality improvement in the upper abdomen with PROPELLER and prospective acquisition correction (PACE) technique. AJR Am J Roentgenol 2008;191:1154– 1158. 23. Song HK, Dougherty L. k-space weighted image contrast (KWIC) for contrast manipulation in projection reconstruction MRI. Magn Reson Med 2000;44:825–832. 24. Song HK, Yan L, Smith RX, et al. Noncontrast enhanced fourdimensional dynamic MRA with golden angle radial acquisition and k-space weighted image contrast (KWIC) reconstruction. Magn Reson Med 2013 [Epub ahead of print]. 25. Kim KW, Lee JM, Jeon YS, et al. Free-breathing dynamic contrast-enhanced MRI of the abdomen and chest using a radial gradient echo sequence with k-space weighted image contrast (KWIC). Eur Radiol 2013;23:1352–1360.
Joo et al. 26. 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. 27. Priest AN, Gill AB, Kataoka M, et al. Dynamic contrast-enhanced MRI in ovarian cancer: Initial experience at 3 Tesla in primary and metastatic disease. Magn Reson Med 2010;63:1044–1049. 28. Tofts PS, Kermode AG. Measurement of the blood-brain barrier permeability and leakage space using dynamic MR imaging. 1. Fundamental concepts. Magn Reson Med 1991;17:357–367. 29. Orton MR, d’Arcy JA, Walker-Samuel S, et al. Computationally efficient vascular input function models for quantitative kinetic modelling using DCE-MRI. Phys Med Biol 2008;53:1225–1239. 30. Washington K. 7th edition of the AJCC cancer staging manual: stomach. Ann Surg Oncol 2010;17:3077–3079. 31. Brevet M, Arcila M, Ladanyi M. Assessment of EGFR mutation status in lung adenocarcinoma by immunohistochemistry using antibodies specific to the two major forms of mutant EGFR. J Mol Diagn 2010;12:169–176. 32. Chinn S. Statistics in respiratory medicine. 2. Repeatability and method comparison. Thorax 1991;46:454–456. 33. Hu B, El Hajj N, Sittler S, Lammert N, Barnes R, Meloni-Ehrig A. Gastric cancer: classification, histology and application of molecular pathology. J Gastrointest Oncol 2012;3:251–261. 34. Yao J, Yang ZG, Chen HJ, Chen TW, Huang J. Gastric adenocarcinoma: can perfusion CT help to noninvasively evaluate tumor angiogenesis? Abdom Imaging 2011;36:15–21. 35. Banerjee R, Pavlides M, Tunnicliffe EM, et al. Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease. J Hepatol 2014;60:69–77. 36. Aref M, Chaudhari AR, Bailey KL, Aref S, Wiener EC. Comparison of tumor histology to dynamic contrast enhanced magnetic resonance imaging-based physiological estimates. Magn Reson Imaging 2008;26:1279–1293. 37. Benjaminsen IC, Brurberg KG, Ruud EB, Rofstad EK. Assessment of extravascular extracellular space fraction in human melanoma xenografts by DCE-MRI and kinetic modeling. Magn Reson Imaging 2008;26:160–170. 38. Roberts C, Issa B, Stone A, Jackson A, Waterton JC, Parker GJ. Comparative study into the robustness of compartmental modeling and model-free analysis in DCE-MRI studies. J Magn Reson Imaging 2006;23:554–563. 39. Yao J, Yang ZG, Chen TW, Li Y, Yang L. Perfusion changes in gastric adenocarcinoma: evaluation with 64-section MDCT. Abdom Imaging 2010;35:195–202. 40. Oberholzer K, Pohlmann A, Schreiber W, et al. Assessment of tumor microcirculation with dynamic contrast-enhanced MRI in patients with esophageal cancer: initial experience. J Magn Reson Imaging 2008;27:1296–1301. 41. Yao WW, Zhang H, Ding B, et al. Rectal cancer: 3D dynamic contrast-enhanced MRI; correlation with microvascular density and clinicopathological features. Radiol Med 2011;116:366–374. 42. Nicholson RI, Gee JM, Harper ME. EGFR and cancer prognosis. Eur J Cancer 2001;37 Suppl 4:S9–15. 43. Turkbey B, Thomasson D, Pang Y, Bernardo M, Choyke PL. The role of dynamic contrast-enhanced MRI in cancer diagnosis and treatment. Diagn Interv Radiol 2010;16:186–192. 44. Yeo DM, Oh SN, Jung CK, et al. Correlation of dynamic contrastenhanced MRI perfusion parameters with angiogenesis and biologic aggressiveness of rectal cancer: preliminary results. J Magn Reson Imaging 2013 [Epub ahead of print].
