JOURNAL OF MAGNETIC RESONANCE IMAGING 40:99–105 (2014)
Original Research
Correlation Between Tissue Metabolism and Cellularity Assessed by Standardized Uptake Value and Apparent Diffusion Coefficient in Peritoneal Metastasis Xue Yu, BS,1 Elaine Yuen Phin Lee, FRCR,1* Vincent Lai, FRCR,1 and Queenie Chan, PhD2 Key Words: FDG-PET/CT; diffusion-weighted imaging; SUV; ADC; peritoneal metastases J. Magn. Reson. Imaging 2014;40:99–105. C 2013 Wiley Periodicals, Inc. V
Purpose: To evaluate the correlation between standardized uptake value (SUV) (tissue metabolism) and apparent diffusion coefficient (ADC) (water diffusivity) in peritoneal metastases. Materials and Methods: Patients with peritoneal dissemination detected on 18F-fluorodeoxyglucose positron emission tomography combined with computed tomography (FDG-PET/CT) were prospectively recruited for MRI examinations with informed consent and the study was approved by the local Institutional Review Board. FDGPET/CT, diffusion-weighted imaging (DWI), MRI, and DWI/MRI images were independently reviewed by two radiologists based on visual analysis. SUVmax/SUVmean and ADCmin/ADCmean were obtained manually by drawing ROIs over the peritoneal metastases on FDG-PET/CT and DWI, respectively. Diagnostic characteristics of each technique were evaluated. Pearson’s coefficient and McNemar and Kappa tests were used for statistical analysis.
METASTATIC PERITONEAL MALIGNANCIES, commonly arise from stomach, colon, pancreas, lungs, breasts, and gynecological cancers (1), are challenging to image, because of the extensive surface area and sometimes small volume of tumor (2). As new curative treatments emerge, it is crucial to detect and characterize peritoneal metastases at an early stage for precise patient selection (3), diagnosis and therapy outcome monitoring (4–6). 18 F-fluorodeoxyglucose positron emission tomography combined with computed tomography (FDG-PET/CT) is a functional technique based on glucose metabolism that has been used in peritoneal imaging (3,5–7) with reported sensitivity and specificity of 43–100% (3,7–9) and 90–100% (3,9), respectively. The standardized uptake value (SUV) derived from FDG-PET/CT is a diagnostic and prognostic parameter in assessing malignancy. MRI also plays an important role (2–5,8,10), especially with the use of gadolinium-enhanced MRI in depicting peritoneal metastases of small volume (4). The addition of diffusion-weighted imaging (DWI) has improved sensitivity, 74–90% as compared to 52–73% in conventional MRI (3,4,8). Apparent diffusion coefficient (ADC) measured by DWI has been reported to be inversely correlated with tissue cellularity (11). As ADC varies according to the changes of tissue microstructure and pathophysiological state, it can provide information on water diffusivity and tumor aggressiveness. The diagnostic performances of FDG-PET/CT and combined DWI/conventional MRI (DWI/MRI) have been compared in several types of carcinoma (10,12– 16) with mixed results. FDG-PET/CT performed better in breast and thyroid carcinoma (12,14), while DWI/ MRI was more superior in rectal cancer (13). Only two studies (3,5) have compared the utility of DWI/MRI and FDG-PET/CT in assessing peritoneal metastases.
Results: Eight patients were recruited for this prospective study and 34 peritoneal metastases were evaluated. ADCmean was significantly and negatively correlated with SUVmax (r ¼ 0.528, P ¼ 0.001) and SUVmean (r ¼ 0.548, P ¼ 0.001). ADCmin had similar correlation with SUVmax (r ¼ 0.508, P ¼ 0.002) and SUVmean (r ¼ 0.513, P ¼ 0.002). DWI/MRI had high diagnostic performance (accuracy ¼ 98%) comparable to FDG-PET/ CT, in peritoneal metastasis detection. Kappa values were excellent for all techniques. Conclusion: There was a significant inverse correlation between SUV and ADC.
1 Department of Diagnostic Radiology, University of Hong Kong, Hong Kong, China. 2 Philips Healthcare, Hong Kong, China. Contract grant sponsor: University of Hong Kong. *Address reprint requests to: E.Y.P.L., Room 406, Block K, Queen Mary Hospital, Pokfulam Road, Hong Kong. E-mail: [email protected] Received March 11, 2013; Accepted July 12, 2013. DOI 10.1002/jmri.24361 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
99
100
The aims of this study were to assess the correlation between SUV measured by FDG-PET/CT and ADC from DWI in peritoneal metastases and to evaluate the diagnostic performances of FDG-PET/CT and DWI/MRI in peritoneal metastases detection.
MATERIALS AND METHODS This study was approved by the local Institutional Review Board. Patients From September 2011 to June 2012, 8 patients (mean age, 56 6 6 years old; range, 45–63 years), with peritoneal metastases found on FDG-PET/CT were prospectively recruited for MRI. Informed consent was obtained from each patient. Patients who had treatment interventions between the two examinations were excluded. The temporal difference between FDGPET/CT and MRI was 9 6 8 days. FDG-PET/CT FDG-PET/CT was performed using an integrated PET/CT scanner (Discovery VCT, GE Healthcare, BioScience Corp., NJ). All patients were required to fast for at least 6 hours. After verifying that the patient’s glucose level was below 8 mmol/L, FDG (4.8 MBq/kg body weight) was injected intravenously. One hour after administration of FDG, CT was obtained from skull base to proximal thigh (imaging parameters: 80– 100 mAs, 120 kV, 0.5 s/rotation, 2.5 mm slice thickness, pitch of 0.984:1, field of view 50 cm), followed by PET acquisition. Each bed position took 2 min 30 s with approximately 6 bed positions per patient. CT was used for attenuation correction. PET images were reconstructed using an ordered-subset expectation maximization algorithm with 14 subsets and 2 iterations. MRI All patients were required to fast for at least 6 hours before the MRI examination to minimize bowel peristalsis. Oral contrast (Metamucil) was administrated to distend the bowel loops to facilitate serosal metastases detection. Metamucil (The Proctor and Gamble Co., Cincinnati, OH) was mixed with water (50 g per liter), and patients were requested to drink 1.5–2 liters of the mixture 2 h before scanning. MRI was performed on a 3 Tesla (T) Achieva scanner (Philips Healthcare, Best, the Netherlands) with a torso coil for both abdominal and pelvic imaging. Conventional MRI included coronal T2 weighted turbo spin echo (repetition time/echo time [TR/TE], 3000/ 80 ms; FOV, 230 322 mm2; matrix, 152 188; turbo spin echo factor, 65), and 3D e-THRIVE dynamic contrast-enhanced imaging (TR/TE, 3/1.38 ms; FOV, 350249 mm2; matrix, 204 145; 4 sets of axial images obtained at 0, 25 s, 90 s, and 3 min after intravenous injection of 0.15 mmol/kg of gadolinium).
Yu et al.
Free-breathing DWI with background body signal suppression was acquired in the axial plane using single short spin echo sequence with 3 b-factors of 0, 400, and 800 s/mm2 before gadolinium injection. The scanning parameters for DWI were: TR/TE, 7600/60 ms; field of view (FOV), 350 318 mm2; matrix, 164 150; slice thickness, 5–7 mm; gap, 0; number of excitations (NEX), 2; acquisition time, 5–6 min. The ADC map was generated automatically by manufacturer’s software (Achieva, Philips Healthcare, Best, the Netherlands). Image Evaluation and Data Analysis Two radiologists (E.L. and V.L.) with 5-year and 6year radiology experience independently and qualitatively reviewed the FDG-PET/CT and MRI, on four separate sessions in the following order: DWI, MRI, DWI/MRI, and FDG-PET/CT with temporal difference of 2 weeks. Both radiologists were blinded to the patients’ information and results of previous examinations, and were unaware of each other’s readings. In each review session, the sequence of patients was randomized to avoid recall bias from prior review session. Previously described 16 anatomic sites were scrutinized for the presence and absence of peritoneal metastasis: left and right subphrenic spaces, right subhepatic space, left and right paracolic gutters, greater and lesser omentum, lesser sac, stomach, small bowel, small bowel mesenteries, colon, pelvis, ovaries, uterine serosa, and bladder (4). Given that only patients with positive FDG-PET/CT were selected as an inclusion criterion, diagnostic performance based on patient-based analysis will be biased by selection criterion and therefore only lesion-based analysis was performed. FDG-PET/CT was analyzed on dedicated workstation (Advantage Workstation, 4.3; GE healthcare, NJ). Peritoneal metastases were depicted as focal abnormal hypermetabolic lesions, which were above background activity and physiological uptake. The locations of peritoneal metastases were coregistered on CT images. All peritoneal metastases were documented according to the prior 16 anatomical sites. For measurable lesions (lesions more than 1.0 cm in the shortest axis), maximum SUV (SUVmax) was derived using a rectangular three-dimensional (3D) region of interest (ROI) drawn on PET images to cover the entire tumor, but carefully avoiding surrounding tissue outside the lesion. SUVmean was calculated as the average value of SUV within threshold-defined ROI. The threshold of 40%, 45%, and 50% of the SUVmax were tested, with the threshold of 45% giving the smallest variance between the tumor volumes measured on CT and threshold-defined PET, and, therefore, was chosen. MRI images were analyzed on dedicated console (Achieva, Philips Healthcare, Best, the Netherlands). Quantitative analysis was performed by in-house program written in MatLab (Mathworks). The ADC map was reviewed side by side with DWI. Hyperintense lesion on DWI with suppressed signal intensity on the ADC map, in these 16 anatomical sites, was
SUV and ADC in Peritoneal Metastasis
101
Table 1 Patient Characteristics, SUV, and ADC Values Lesion
Patient
Primary cancer
1
PM_1
2 3 4 5 6 7 8 9 10 11 12 13 14
PM_2
Squamous cell carcinoma of cervix Endometrioid carcinoma of uterine corpus Adenocarcinoma of lung
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
PM_5
PM_3
PM_4
PM_6 PM_7
PM_8
G3 endometrioid serous adenocarcinoma of uterine corpus Adenocarcinoma of colon
Clear cell carcinoma of ovaries High-grade serous adenocarcinoma of ovaries Serous adenocarcinoma of peritoneum
Lesion Location
Confirmation
SUVmean (g/ml)
ADCmin (103 mm2/s)
ADCmean (103 mm2/s)
5.4
3.6
0.6
1.1
SUVmax (g/ml)
Pelvis
CT
Right subphrenic space Right subhepatic space Pelvis Right paracolic gutter Right paracolic gutter Right paracolic gutter Pelvis Pelvis Right paracolic gutter Pelvis Pelvis Pelvis Pelvis
PET/CT PET/CT PET/CT CT CT CT CT CT Histology Histology Histology Histology Histology
7.1 8.4 6.4 3.5 5.0 8.0 10.3 2.4 7.7 3.3 10 7.9 4.5
4.6 5.3 4.2 1.9 3.2 5.3 6.4 1.5 5 2.2 6.2 5.1 2.6
0.4 0.4 0.4 0.5 0.9 0.5 0.5 0.5 0.5 0.5 0.5 0.3 0.3
1.4 0.9 0.7 2.1 1.8 1.2 1.4 1.5 1.4 1.4 1.4 0.9 1.1
Pelvis Right subphrenic space Right subphrenic space Left paracolic gutter Right subphrenic space Right subphrenic space Pelvis Ovaries Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Left subhepatic space Pelvis Pelvis
CT CT CT CT CT CT Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology
3.6 3.4 2.9 3.0 3.3 2.8 8.0 6.5 8.6 5.9 5.9 7.1 3.4 4.4 3.9 8.4 7.8 6.1 6.3 7.4
2.4 2 1.8 1.8 2 1.9 5 4.3 5.9 3.7 3.7 4.5 2.1 3.2 2.3 5.2 4.8 3.6 3.9 4.5
0.7 0.7 0.7 0.9 0.8 0.7 0.5 0.5 0.5 0.5 0.4 0.5 0.5 0.5 0.6 0.5 0.5 0.5 0.5 0.5
1.8 1.9 1.9 2.0 2.1 1.8 1.3 1.4 0.8 1.0 0.7 0.9 1.2 0.9 1.2 0.8 1.0 0.9 0.8 1.8
documented as suspicious peritoneal metastasis. Measurable tumor was contoured manually using irregular shaped ROIs on every slice of the DWI (b ¼ 0 s/mm2) images containing the lesion, and then automatically transferred to the ADC map by the in-house program. The mean ADC (ADCmean) was measured by averaging the ADC for each voxel within the ROI. A patient-specific threshold was applied to remove poorly fitted voxels induced by noise for the calculation of minimum ADC (ADCmin) and ADCmean. Noise was taken by averaging the mean signal values of 5 ROIs (1010 pixels) placed outside the body on 5 different slices (17,18). Conventional MRI sequences were scrutinized for presence of peritoneal metastases based on the previously mentioned 16 anatomical sites. In the quantitative part of the study, only measurable lesions were included for analysis with their respective SUV and ADC. All ROI were placed by X.Y., and verified by E.L.. The same exercise was repeated 6 months later in 10 randomly selected measurable lesions to assess for reproducibility of SUV and ADC using Bland-Altman plot through visual analysis. Tumor volumes measured on PET and DWI were compared.
Standard of Reference Histology was taken as gold standard. In assessing the diagnostic performance of both imaging modalities based on the per-site analysis, evaluation was limited to anatomical sites with histological proof. For the lesions included in the quantification study, those that had no surgical verification were scrutinized on follow-up FDG-PET/CT or CT and determined by consensus between two radiologists. Lesions that showed complete or partial regression following therapy, or lesions that had progressed were considered true positive for peritoneal metastases (19). Static lesions and lesions that regressed without treatment were taken as false positive lesions (20). Statistical Analysis The correlation between SUV and ADC was evaluated by Pearson correlation coefficient. SUV and ADC values were expressed as mean 6 standard deviation. Difference between the tumor volumes on PET and DWI were compared by Wilcoxon signed-rank test.
102
Yu et al.
Figure 1. Inverse correlations between SUVmax and ADCmin (a), SUVmean and ADCmin (b), SUVmax and ADCmean (c), and SUVmean and ADCmean (d) in peritoneal metastasis (n ¼ 34).
Diagnostic characteristics were defined by sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and area under the receiver operating characteristics curve (AUC) for DWI, MRI, DWI/MRI and FDG-PET/CT. Based on per-site analysis, McNemar test was used to compare the diagnostic performances among these techniques. The interobserver agreement was evaluated by Kappa statistics. The relationship between Kappa value and strength of agreement was as follows: poor ¼ less than 0.2, fair ¼ 0.2–0.4, moderate ¼ 0.4–0.6, good ¼ 0.6–0.8, and very good ¼ 0.8–1. Statistical analyses were performed using SPSS (version 16.0, SPSS, Chicago, IL, USA). All P values were two-sided, and P < 0.05 were considered as statistically significant. RESULTS Patient characteristics and imaging findings were summarized in Table 1. Peritoneal metastases from 5 patients were histological proven through percutaneous biopsy and laparotomy with site-by-site correlation. The FDG-PET/CT to surgery time was 18 6 13 days. The remaining patients were confirmed with radiological follow-ups with mean follow-up duration
of 316 6 141 days. PM_1, PM_2, and PM_5 had lesions that progressed on follow-up CT or FDG-PET/ CT. PM_3 had 5 lesions that showed partial response to chemotherapy. Thirty-four lesions were quantitatively evaluated with mean diameter of 4.8 6 2.4 cm (range: 2.0–14.4 cm). Nineteen lesions had histological proof and 15 lesions were verified by follow-up FDG-PET/CT or CT. No significant difference was found between tumor volumes measured on PET and DWI (Wilcoxon signed-rank test, P ¼ 0.651). The differences between the initial and repeated measurements 6 months later were close to 0 and all within the two standard deviations. These lesions demonstrated ADCmin of 0.54 6 0.14, ADCmean 1.34 6 0.42, SUVmax 5.8 6 2.3, and SUVmean 3.7 6 1.4. ADCmin was significantly and inversely correlated with SUVmax (r ¼ 0.508, P ¼ 0.002, Fig. 1a) and SUVmean (r ¼ 0.513, P ¼ 0.002, Fig. 1b). Similar inverse correlations were also identified between ADCmean and SUVmax (r ¼ 0.528, P ¼ 0.001, Fig. 1c), as well as ADCmean and SUVmean (r ¼ 0.548, P ¼ 0.001, Fig. 1d). For site-based analysis, a total of 69 histologically proven sites were analyzed. Table 2 shows the averaged values of the diagnostic characteristics
SUV and ADC in Peritoneal Metastasis
103
Table 2 Summary of Diagnostic Characteristics (Mean Value in Percentage) of Different Imaging Sequences and Modalities for Per-Site–Based Analysis With Reference to Histology
Sensitivity Specificity Accuracy PPV NPV AUC
DWI
MRI
DWI/MRI
FDG-PET/CT
81 99 95 97 95 90
88 99 97 97 96 93
88 99 97 97 96 93
88 100 97 100 96 94
assessed by radiologist 1 and 2, comparing the diagnostic performances of FDG-PET/CT, DWI, MRI, and DWI/MRI in peritoneal metastases detection (Fig. 2). Although statistically insignificant, DWI/MRI was slightly more sensitive in depicting peritoneal metastases (88%) as compared to DWI (81%). The diagnostic accuracy of DWI/MRI was comparable to that of FDGPET/CT (P > 0.05). In the right subphrenic space, DWI/MRI has one false positive result due to epiphrenic lymph node. There were two false negative sites on both DWI/MRI and FDG-PET/CT due to urinary bladder serosal deposits. The AUC for DWI, MRI, DWI/MRI, FDG-PET/CT fell between 0.9 and 1. The interobserver agreement was excellent for FDG-PET/ CT (k ¼ 1.000), DWI/MRI (k ¼ 0.956), MRI (k ¼ 0.956) and DWI (k ¼ 0.954) for site-based analysis. DISCUSSION Peritoneal metastases can be subtle and still extensively disseminated at the time of presentation (21). The
provision of peritoneal imaging has facilitated clinical decision-making, especially in surgical planning, and has improved patient outcome (2,3) and forms an important part of disease evaluation. The present study verified the inverse correlation between SUV and ADC in peritoneal metastases. SUV, a semi-quantitative parameter of glucose metabolism in tumor, is useful for tumor detection, characterization, prognosis, and treatment response monitoring in different cancers (22,23). SUVmax is the most clinically used parameter due to its ease of use and reproducibility without the dependence on tumor size and shape (24). SUVmean, on the contrary, is dependent on the threshold used for volume approximation, for which there is no consensus agreement. Both fixed and adaptable thresholds have been used in PET/CT to delineate tumor boundaries (23,24). In this study, we found that using fixed 45% threshold provided the best concordance between volumes measured by CT and PET with the lowest mean variance. ADC values are used for tumor grading, nodal metastases detection, and treatment response monitoring (25–29). A patient-specific threshold was applied to exclude poorly fitted voxel, which may lead to underestimation in ADC measurement, for more accurate quantification (30). Although SUV and ADC represent two diverse facets of cell biology, our results reveal a relationship between tumor metabolism and tissue cellularity. Studies showed that tumor metabolism correlated with cell proliferation and cell density (31,32). Tumors that have high cell proliferative indices have higher tissue cellularity, and therefore more restricted in diffusion, giving rise to lower ADC values. Understanding the biologic
Figure 2. A 53-year-old female diagnosed with serous adenocarcinoma of peritoneum showing peritoneal thickening at the right subphrenic space identified by white arrows shown on axial DWI images with b value of 0 (a), constructed ADC map using three b values (0, 400, 800 s/mm2) (b), 6 min post-contrast–enhanced image (c), and fused axial PET/CT image (d).
104
interactions, the relationship between SUV and ADC has been demonstrated in several other tumors. Ho et al (24) was the first to report a significant inverse correlation between relative SUVmax and relative ADCmin in mixed types of primary cervical cancer. The inverse correlation between SUV and ADC was also found in primary rectal, non-small cell lung cancer and gastrointestinal stromal lesions (22,23,33). In concordance, our study presented significant inverse correlations between ADC and SUV in 34 peritoneal metastases, regardless of their primaries with highly reproducible quantitative measurements. This robust correlation between ADC and SUV demonstrated the feasibility of using ADC as a diagnostic adjunct in peritoneal metastasis, similar to SUV. As cytotoxic treatment breaks up cell membrane, restores normal water molecular diffusion and extracellular matrix, the ADC value increases or becomes normalized, DWI has the capability of early treatment response assessment before tumor size changes (34). Further studies are encouraged to evaluate the potential roles of SUV and ADC in treatment response assessment in peritoneal metastases. In our study, DWI and MRI had inferior sensitivity to DWI/MRI and FDG-PET/CT, but the difference was not statistically significant. The inferior diagnostic performance of DWI may be caused by poor anatomic localization due to limited spatial resolution, low signal-to-noise ratio, and susceptibility artifacts. DWI viewed in conjunction with MRI improved the overall diagnostic performance in peritoneal metastasis detection, although modest. DWI had better background and bowel suppression while conventional MRI improved anatomic landmark depiction. This underlines the complementary value of DWI and conventional MRI. Even with this modest increment, the authors believe that DWI should be a routine sequence added to abdominal imaging with a small trade-off of an extra 5–6 min of scanning time. Our results demonstrated that DWI/MRI and FDGPET/CT had comparable diagnostic performance in peritoneal metastases detection. This finding is in concordance with previous study (5) concluding that DWI/MRI can be used as the alternative imaging technique for peritoneal dissemination detection when FDG-PET/CT is not available. In Satoh et al (5), two different cohorts of patients were recruited. Majority of the FDG-PET/CT patients (15/27) had gastrointestinal malignancies while patients who underwent DWI/MRI had genitourinary malignancies (16/23). Therefore, no direct comparison could be made in these two cohorts of patients with different malignancies. Soussan et al (3) reported that DWI/MRI had higher sensitivity and specificity than FDG-PET/CT, but the results were statistically insignificant. This finding could be related to the large number of subcentimeter lesions included in this study (12/27), which could underestimate the glucose uptake on PET due to the partial volume effect, therefore reducing the diagnostic performance of FDG-PET/CT. The two false negative lesions on both MRI and FDG-PET/CT were micro-metastases with thin layers of cells over the urinary bladder serosal surface, only
Yu et al.
confirmed by routine biopsy of the region and not visible on gross inspection during surgery. FDG excretion in the urinary bladder will reduce the contrast between tumor and the high background uptake, thus reducing the detection rate. There are several limitations in this study. First, the number of patients recruited was small, preventing meaningful patient-based analysis. Second, patients enrolled in this study had various histological types, which may have confounded the correlation between SUV and ADC, in that different histology may have different biologic behaviors. Nevertheless, we found a significant correlation between SUV and ADC. Finally, recruiting patients with suspected peritoneal metastasis on FDG-PET/CT may lead to selection bias, in that patients with small lesions undetected on FDGPET/CT were excluded from the study. In conclusion, the negative correlation between SUV and ADC provides complimentary information on tumor biology, with similar diagnostic accuracy between FDG-PET/CT and DWI/MRI. These techniques can be used in synergistic manner, or as an alternative imaging when access to FDG-PET/CT is limited, in peritoneal dissemination evaluation and add values to our current understanding of the biologic behaviors of these tumors. REFERENCES 1. Healy JC, Reznek RH. The peritoneum, mesenteries and omenta: normal anatomy and pathological processes. Eur Radiol 1998;8: 886–900. 2. Low RN. MR imaging of the peritoneal spread of malignancy. Abdom Imaging 2007;32:267–283. 3. Soussan M, Des Guetz G, Barrau V, et al. Comparison of FDGPET/CT and MR with diffusion-weighted imaging for assessing peritoneal carcinomatosis from gastrointestinal malignancy. Eur Radiol 2012;22:1479–1487. 4. Low RN, Sebrechts CP, Barone RM, Muller W. Diffusion-weighted MRI of peritoneal tumors: comparison with conventional MRI and surgical and histopathologic findings--a feasibility study. AJR Am J Roentgenol 2009;193:461–470. 5. Satoh Y, Ichikawa T, Motosugi U, et al. Diagnosis of peritoneal dissemination: comparison of 18F-FDG PET/CT, diffusionweighted MRI, and contrast-enhanced MDCT. AJR Am J Roentgenol 2011;196:447–453. 6. Kyriazi S, Kaye SB, deSouza NM. Imaging ovarian cancer and peritoneal metastases-current and emerging techniques. Nat Rev Clin Oncol 2010;7:381–393. 7. Berthelot C, Morel O, Girault S, et al. Use of FDG-PET/CT for peritoneal carcinomatosis before hyperthermic intraperitoneal chemotherapy. Nucl Med Commun 2011;32:23–29. 8. Kim CK, Park BK, Choi JY, Kim BG, Han H. Detection of recurrent ovarian cancer at MRI: comparison with integrated PET/CT. J Comput Assist Tomogr 2007;31:868–875. 9. Dirisamer A, Schima W, Heinisch M, et al. Detection of histologically proven peritoneal carcinomatosis with fused 18F-FDG-PET/ MDCT. Eur J Radiol 2009;69:536–541. 10. Kitajima K, Yamasaki E, Kaji Y, Murakami K, Sugimura K. Comparison of DWI and PET/CT in evaluation of lymph node metastasis in uterine cancer. World J Radiol 2012;4:207–214. 11. Herneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol 2003;45:208–213. 12. Heusner TA, Kuemmel S, Koeninger A, et al. Diagnostic value of diffusion-weighted magnetic resonance imaging (DWI) compared to FDG PET/CT for whole-body breast cancer staging. Eur J Nucl Med Mol Imaging 2010;37:1077–1086. 13. Song I, Kim SH, Lee SJ, Choi JY, Kim MJ, Rhim H. Value of diffusion-weighted imaging in the detection of viable tumour after neoadjuvant chemoradiation therapy in patients with locally
SUV and ADC in Peritoneal Metastasis
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
advanced rectal cancer: comparison with T2 weighted and PET/ CT imaging. Br J Radiol 2012;85:577–586. Nagamachi S, Wakamatsu H, Kiyohara S, et al. Comparison of diagnostic and prognostic capabilities of F-18-FDG-PET/CT, I-131-scintigraphy, and diffusion-weighted magnetic resonance imaging for postoperative thyroid cancer. Jpn J Radiol 2011;29: 413–422. De Paepe K, De Keyzer F, Wolter P, Oyen R, Verhoef G, Vandecaveye V. 5.3 Tesla whole-body diffusion-weighted imaging for staging lymphoma: comparison with PET/CT. Cancer Imaging 2011;11 Spec No A:S178. Takenaka D, Ohno Y, Matsumoto K, et al. Detection of bone metastases in non-small cell lung cancer patients: comparison of whole-body diffusion-weighted imaging (DWI), whole-body MR imaging without and with DWI, whole-body FDG-PET/CT, and bone scintigraphy. J Magn Reson Imaging 2009;30:298–308. Lagana M, Rovaris M, Ceccarelli A, Venturelli C, Marini S, Baselli G. DTI parameter optimisation for acquisition at 1.5T: SNR analysis and clinical application. Comput Intell Neurosci 2010: 254032. Gong NJ, Wong CS, Chu YC, Gu J. Treatment response monitoring in patients with gastrointestinal stromal tumor using diffusion-weighted imaging: preliminary results in comparison with positron emission tomography/computed tomography. NMR Biomed 2013;26:185–192. Boellaard R, O’Doherty MJ, Weber WA, et al. FDG PET and PET/ CT: EANM procedure guidelines for tumour PET imaging: version 1.0. Eur J Nucl Med Mol Imaging 2010;37:181–200. Lee KR, Tavassoli FA, Prat J, et al. Tumours of the ovary and peritoneum. In: Tavassoli FA, Devilee P, editors. Pathology and genetics tumours of the breast and female genital organs. 3rd edition. Volume 4. Lyon: IARC Press; 2003. p 113–202. Klaver YL, Lemmens VE, de Hingh IH. Outcome of surgery for colorectal cancer in the presence of peritoneal carcinomatosis. Eur J Surg Oncol 2013;39:734–741. Regier M, Derlin T, Schwarz D, et al. Diffusion weighted MRI and 18F-FDG PET/CT in non-small cell lung cancer (NSCLC): does the apparent diffusion coefficient (ADC) correlate with tracer uptake (SUV)? Eur J Radiol 2012;81:2913–2918. Gu J, Khong PL, Wang S, Chan Q, Law W, Zhang J. Quantitative assessment of diffusion-weighted MR imaging in patients with
105
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
primary rectal cancer: correlation with FDG-PET/CT. Mol Imaging Biol 2011;13:1020–1028. Ho KC, Lin G, Wang JJ, Lai CH, Chang CJ, Yen TC. Correlation of apparent diffusion coefficients measured by 3T diffusionweighted MRI and SUV from FDG PET/CT in primary cervical cancer. Eur J Nucl Med Mol Imaging 2009;36:200–208. Matoba M, Tonami H, Kondou T, et al. Lung carcinoma: diffusion-weighted mr imaging--preliminary evaluation with apparent diffusion coefficient. Radiology 2007;243:570–577. Gourtsoyianni S, Papanikolaou N, Yarmenitis S, Maris T, Karantanas A, Gourtsoyiannis N. Respiratory gated diffusion-weighted imaging of the liver: value of apparent diffusion coefficient measurements in the differentiation between most commonly encountered benign and malignant focal liver lesions. Eur Radiol 2008;18:486–492. Nakajo M, Kajiya Y, Kaneko T, et al. FDG PET/CT and diffusionweighted imaging for breast cancer: prognostic value of maximum standardized uptake values and apparent diffusion coefficient values of the primary lesion. Eur J Nucl Med Mol Imaging 2010; 37:2011–2020. Pauls S, Schmidt SA, Juchems MS, et al. Diffusion-weighted MR imaging in comparison to integrated [(1)(8)F]-FDG PET/CT for N-staging in patients with lung cancer. Eur J Radiol 2012;81:178–182. Dudeck O, Zeile M, Pink D, et al. Diffusion-weighted magnetic resonance imaging allows monitoring of anticancer treatment effects in patients with soft-tissue sarcomas. J Magn Reson Imaging 2008;27:1109–1113. Jones DK, Basser PJ. “Squashing peanuts and smashing pumpkins”: how noise distorts diffusion-weighted MR data. Magn Reson Med 2004;52:979–993. Herholz K, Pietrzyk U, Voges J, et al. Correlation of glucose consumption and tumor cell density in astrocytomas. A stereotactic PET study. J Neurosurg 1993;79:853–858. Higashi K, Ueda Y, Yagishita M, et al. FDG PET measurement of the proliferative potential of non-small cell lung cancer. J Nucl Med 2000;41:85–92. Wong CS, Gong N, Chu YC, et al. Correlation of measurements from diffusion weighted MR imaging and FDG PET/CT in GIST patients: ADC versus SUV. Eur J Radiol 2012;81:2122–2126. Morse DL, Galons JP, Payne CM, et al. MRI-measured water mobility increases in response to chemotherapy via multiple celldeath mechanisms. NMR Biomed 2007;20:602–614.
Original Research
Correlation Between Tissue Metabolism and Cellularity Assessed by Standardized Uptake Value and Apparent Diffusion Coefficient in Peritoneal Metastasis Xue Yu, BS,1 Elaine Yuen Phin Lee, FRCR,1* Vincent Lai, FRCR,1 and Queenie Chan, PhD2 Key Words: FDG-PET/CT; diffusion-weighted imaging; SUV; ADC; peritoneal metastases J. Magn. Reson. Imaging 2014;40:99–105. C 2013 Wiley Periodicals, Inc. V
Purpose: To evaluate the correlation between standardized uptake value (SUV) (tissue metabolism) and apparent diffusion coefficient (ADC) (water diffusivity) in peritoneal metastases. Materials and Methods: Patients with peritoneal dissemination detected on 18F-fluorodeoxyglucose positron emission tomography combined with computed tomography (FDG-PET/CT) were prospectively recruited for MRI examinations with informed consent and the study was approved by the local Institutional Review Board. FDGPET/CT, diffusion-weighted imaging (DWI), MRI, and DWI/MRI images were independently reviewed by two radiologists based on visual analysis. SUVmax/SUVmean and ADCmin/ADCmean were obtained manually by drawing ROIs over the peritoneal metastases on FDG-PET/CT and DWI, respectively. Diagnostic characteristics of each technique were evaluated. Pearson’s coefficient and McNemar and Kappa tests were used for statistical analysis.
METASTATIC PERITONEAL MALIGNANCIES, commonly arise from stomach, colon, pancreas, lungs, breasts, and gynecological cancers (1), are challenging to image, because of the extensive surface area and sometimes small volume of tumor (2). As new curative treatments emerge, it is crucial to detect and characterize peritoneal metastases at an early stage for precise patient selection (3), diagnosis and therapy outcome monitoring (4–6). 18 F-fluorodeoxyglucose positron emission tomography combined with computed tomography (FDG-PET/CT) is a functional technique based on glucose metabolism that has been used in peritoneal imaging (3,5–7) with reported sensitivity and specificity of 43–100% (3,7–9) and 90–100% (3,9), respectively. The standardized uptake value (SUV) derived from FDG-PET/CT is a diagnostic and prognostic parameter in assessing malignancy. MRI also plays an important role (2–5,8,10), especially with the use of gadolinium-enhanced MRI in depicting peritoneal metastases of small volume (4). The addition of diffusion-weighted imaging (DWI) has improved sensitivity, 74–90% as compared to 52–73% in conventional MRI (3,4,8). Apparent diffusion coefficient (ADC) measured by DWI has been reported to be inversely correlated with tissue cellularity (11). As ADC varies according to the changes of tissue microstructure and pathophysiological state, it can provide information on water diffusivity and tumor aggressiveness. The diagnostic performances of FDG-PET/CT and combined DWI/conventional MRI (DWI/MRI) have been compared in several types of carcinoma (10,12– 16) with mixed results. FDG-PET/CT performed better in breast and thyroid carcinoma (12,14), while DWI/ MRI was more superior in rectal cancer (13). Only two studies (3,5) have compared the utility of DWI/MRI and FDG-PET/CT in assessing peritoneal metastases.
Results: Eight patients were recruited for this prospective study and 34 peritoneal metastases were evaluated. ADCmean was significantly and negatively correlated with SUVmax (r ¼ 0.528, P ¼ 0.001) and SUVmean (r ¼ 0.548, P ¼ 0.001). ADCmin had similar correlation with SUVmax (r ¼ 0.508, P ¼ 0.002) and SUVmean (r ¼ 0.513, P ¼ 0.002). DWI/MRI had high diagnostic performance (accuracy ¼ 98%) comparable to FDG-PET/ CT, in peritoneal metastasis detection. Kappa values were excellent for all techniques. Conclusion: There was a significant inverse correlation between SUV and ADC.
1 Department of Diagnostic Radiology, University of Hong Kong, Hong Kong, China. 2 Philips Healthcare, Hong Kong, China. Contract grant sponsor: University of Hong Kong. *Address reprint requests to: E.Y.P.L., Room 406, Block K, Queen Mary Hospital, Pokfulam Road, Hong Kong. E-mail: [email protected] Received March 11, 2013; Accepted July 12, 2013. DOI 10.1002/jmri.24361 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
99
100
The aims of this study were to assess the correlation between SUV measured by FDG-PET/CT and ADC from DWI in peritoneal metastases and to evaluate the diagnostic performances of FDG-PET/CT and DWI/MRI in peritoneal metastases detection.
MATERIALS AND METHODS This study was approved by the local Institutional Review Board. Patients From September 2011 to June 2012, 8 patients (mean age, 56 6 6 years old; range, 45–63 years), with peritoneal metastases found on FDG-PET/CT were prospectively recruited for MRI. Informed consent was obtained from each patient. Patients who had treatment interventions between the two examinations were excluded. The temporal difference between FDGPET/CT and MRI was 9 6 8 days. FDG-PET/CT FDG-PET/CT was performed using an integrated PET/CT scanner (Discovery VCT, GE Healthcare, BioScience Corp., NJ). All patients were required to fast for at least 6 hours. After verifying that the patient’s glucose level was below 8 mmol/L, FDG (4.8 MBq/kg body weight) was injected intravenously. One hour after administration of FDG, CT was obtained from skull base to proximal thigh (imaging parameters: 80– 100 mAs, 120 kV, 0.5 s/rotation, 2.5 mm slice thickness, pitch of 0.984:1, field of view 50 cm), followed by PET acquisition. Each bed position took 2 min 30 s with approximately 6 bed positions per patient. CT was used for attenuation correction. PET images were reconstructed using an ordered-subset expectation maximization algorithm with 14 subsets and 2 iterations. MRI All patients were required to fast for at least 6 hours before the MRI examination to minimize bowel peristalsis. Oral contrast (Metamucil) was administrated to distend the bowel loops to facilitate serosal metastases detection. Metamucil (The Proctor and Gamble Co., Cincinnati, OH) was mixed with water (50 g per liter), and patients were requested to drink 1.5–2 liters of the mixture 2 h before scanning. MRI was performed on a 3 Tesla (T) Achieva scanner (Philips Healthcare, Best, the Netherlands) with a torso coil for both abdominal and pelvic imaging. Conventional MRI included coronal T2 weighted turbo spin echo (repetition time/echo time [TR/TE], 3000/ 80 ms; FOV, 230 322 mm2; matrix, 152 188; turbo spin echo factor, 65), and 3D e-THRIVE dynamic contrast-enhanced imaging (TR/TE, 3/1.38 ms; FOV, 350249 mm2; matrix, 204 145; 4 sets of axial images obtained at 0, 25 s, 90 s, and 3 min after intravenous injection of 0.15 mmol/kg of gadolinium).
Yu et al.
Free-breathing DWI with background body signal suppression was acquired in the axial plane using single short spin echo sequence with 3 b-factors of 0, 400, and 800 s/mm2 before gadolinium injection. The scanning parameters for DWI were: TR/TE, 7600/60 ms; field of view (FOV), 350 318 mm2; matrix, 164 150; slice thickness, 5–7 mm; gap, 0; number of excitations (NEX), 2; acquisition time, 5–6 min. The ADC map was generated automatically by manufacturer’s software (Achieva, Philips Healthcare, Best, the Netherlands). Image Evaluation and Data Analysis Two radiologists (E.L. and V.L.) with 5-year and 6year radiology experience independently and qualitatively reviewed the FDG-PET/CT and MRI, on four separate sessions in the following order: DWI, MRI, DWI/MRI, and FDG-PET/CT with temporal difference of 2 weeks. Both radiologists were blinded to the patients’ information and results of previous examinations, and were unaware of each other’s readings. In each review session, the sequence of patients was randomized to avoid recall bias from prior review session. Previously described 16 anatomic sites were scrutinized for the presence and absence of peritoneal metastasis: left and right subphrenic spaces, right subhepatic space, left and right paracolic gutters, greater and lesser omentum, lesser sac, stomach, small bowel, small bowel mesenteries, colon, pelvis, ovaries, uterine serosa, and bladder (4). Given that only patients with positive FDG-PET/CT were selected as an inclusion criterion, diagnostic performance based on patient-based analysis will be biased by selection criterion and therefore only lesion-based analysis was performed. FDG-PET/CT was analyzed on dedicated workstation (Advantage Workstation, 4.3; GE healthcare, NJ). Peritoneal metastases were depicted as focal abnormal hypermetabolic lesions, which were above background activity and physiological uptake. The locations of peritoneal metastases were coregistered on CT images. All peritoneal metastases were documented according to the prior 16 anatomical sites. For measurable lesions (lesions more than 1.0 cm in the shortest axis), maximum SUV (SUVmax) was derived using a rectangular three-dimensional (3D) region of interest (ROI) drawn on PET images to cover the entire tumor, but carefully avoiding surrounding tissue outside the lesion. SUVmean was calculated as the average value of SUV within threshold-defined ROI. The threshold of 40%, 45%, and 50% of the SUVmax were tested, with the threshold of 45% giving the smallest variance between the tumor volumes measured on CT and threshold-defined PET, and, therefore, was chosen. MRI images were analyzed on dedicated console (Achieva, Philips Healthcare, Best, the Netherlands). Quantitative analysis was performed by in-house program written in MatLab (Mathworks). The ADC map was reviewed side by side with DWI. Hyperintense lesion on DWI with suppressed signal intensity on the ADC map, in these 16 anatomical sites, was
SUV and ADC in Peritoneal Metastasis
101
Table 1 Patient Characteristics, SUV, and ADC Values Lesion
Patient
Primary cancer
1
PM_1
2 3 4 5 6 7 8 9 10 11 12 13 14
PM_2
Squamous cell carcinoma of cervix Endometrioid carcinoma of uterine corpus Adenocarcinoma of lung
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
PM_5
PM_3
PM_4
PM_6 PM_7
PM_8
G3 endometrioid serous adenocarcinoma of uterine corpus Adenocarcinoma of colon
Clear cell carcinoma of ovaries High-grade serous adenocarcinoma of ovaries Serous adenocarcinoma of peritoneum
Lesion Location
Confirmation
SUVmean (g/ml)
ADCmin (103 mm2/s)
ADCmean (103 mm2/s)
5.4
3.6
0.6
1.1
SUVmax (g/ml)
Pelvis
CT
Right subphrenic space Right subhepatic space Pelvis Right paracolic gutter Right paracolic gutter Right paracolic gutter Pelvis Pelvis Right paracolic gutter Pelvis Pelvis Pelvis Pelvis
PET/CT PET/CT PET/CT CT CT CT CT CT Histology Histology Histology Histology Histology
7.1 8.4 6.4 3.5 5.0 8.0 10.3 2.4 7.7 3.3 10 7.9 4.5
4.6 5.3 4.2 1.9 3.2 5.3 6.4 1.5 5 2.2 6.2 5.1 2.6
0.4 0.4 0.4 0.5 0.9 0.5 0.5 0.5 0.5 0.5 0.5 0.3 0.3
1.4 0.9 0.7 2.1 1.8 1.2 1.4 1.5 1.4 1.4 1.4 0.9 1.1
Pelvis Right subphrenic space Right subphrenic space Left paracolic gutter Right subphrenic space Right subphrenic space Pelvis Ovaries Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Left subhepatic space Pelvis Pelvis
CT CT CT CT CT CT Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology Histology
3.6 3.4 2.9 3.0 3.3 2.8 8.0 6.5 8.6 5.9 5.9 7.1 3.4 4.4 3.9 8.4 7.8 6.1 6.3 7.4
2.4 2 1.8 1.8 2 1.9 5 4.3 5.9 3.7 3.7 4.5 2.1 3.2 2.3 5.2 4.8 3.6 3.9 4.5
0.7 0.7 0.7 0.9 0.8 0.7 0.5 0.5 0.5 0.5 0.4 0.5 0.5 0.5 0.6 0.5 0.5 0.5 0.5 0.5
1.8 1.9 1.9 2.0 2.1 1.8 1.3 1.4 0.8 1.0 0.7 0.9 1.2 0.9 1.2 0.8 1.0 0.9 0.8 1.8
documented as suspicious peritoneal metastasis. Measurable tumor was contoured manually using irregular shaped ROIs on every slice of the DWI (b ¼ 0 s/mm2) images containing the lesion, and then automatically transferred to the ADC map by the in-house program. The mean ADC (ADCmean) was measured by averaging the ADC for each voxel within the ROI. A patient-specific threshold was applied to remove poorly fitted voxels induced by noise for the calculation of minimum ADC (ADCmin) and ADCmean. Noise was taken by averaging the mean signal values of 5 ROIs (1010 pixels) placed outside the body on 5 different slices (17,18). Conventional MRI sequences were scrutinized for presence of peritoneal metastases based on the previously mentioned 16 anatomical sites. In the quantitative part of the study, only measurable lesions were included for analysis with their respective SUV and ADC. All ROI were placed by X.Y., and verified by E.L.. The same exercise was repeated 6 months later in 10 randomly selected measurable lesions to assess for reproducibility of SUV and ADC using Bland-Altman plot through visual analysis. Tumor volumes measured on PET and DWI were compared.
Standard of Reference Histology was taken as gold standard. In assessing the diagnostic performance of both imaging modalities based on the per-site analysis, evaluation was limited to anatomical sites with histological proof. For the lesions included in the quantification study, those that had no surgical verification were scrutinized on follow-up FDG-PET/CT or CT and determined by consensus between two radiologists. Lesions that showed complete or partial regression following therapy, or lesions that had progressed were considered true positive for peritoneal metastases (19). Static lesions and lesions that regressed without treatment were taken as false positive lesions (20). Statistical Analysis The correlation between SUV and ADC was evaluated by Pearson correlation coefficient. SUV and ADC values were expressed as mean 6 standard deviation. Difference between the tumor volumes on PET and DWI were compared by Wilcoxon signed-rank test.
102
Yu et al.
Figure 1. Inverse correlations between SUVmax and ADCmin (a), SUVmean and ADCmin (b), SUVmax and ADCmean (c), and SUVmean and ADCmean (d) in peritoneal metastasis (n ¼ 34).
Diagnostic characteristics were defined by sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and area under the receiver operating characteristics curve (AUC) for DWI, MRI, DWI/MRI and FDG-PET/CT. Based on per-site analysis, McNemar test was used to compare the diagnostic performances among these techniques. The interobserver agreement was evaluated by Kappa statistics. The relationship between Kappa value and strength of agreement was as follows: poor ¼ less than 0.2, fair ¼ 0.2–0.4, moderate ¼ 0.4–0.6, good ¼ 0.6–0.8, and very good ¼ 0.8–1. Statistical analyses were performed using SPSS (version 16.0, SPSS, Chicago, IL, USA). All P values were two-sided, and P < 0.05 were considered as statistically significant. RESULTS Patient characteristics and imaging findings were summarized in Table 1. Peritoneal metastases from 5 patients were histological proven through percutaneous biopsy and laparotomy with site-by-site correlation. The FDG-PET/CT to surgery time was 18 6 13 days. The remaining patients were confirmed with radiological follow-ups with mean follow-up duration
of 316 6 141 days. PM_1, PM_2, and PM_5 had lesions that progressed on follow-up CT or FDG-PET/ CT. PM_3 had 5 lesions that showed partial response to chemotherapy. Thirty-four lesions were quantitatively evaluated with mean diameter of 4.8 6 2.4 cm (range: 2.0–14.4 cm). Nineteen lesions had histological proof and 15 lesions were verified by follow-up FDG-PET/CT or CT. No significant difference was found between tumor volumes measured on PET and DWI (Wilcoxon signed-rank test, P ¼ 0.651). The differences between the initial and repeated measurements 6 months later were close to 0 and all within the two standard deviations. These lesions demonstrated ADCmin of 0.54 6 0.14, ADCmean 1.34 6 0.42, SUVmax 5.8 6 2.3, and SUVmean 3.7 6 1.4. ADCmin was significantly and inversely correlated with SUVmax (r ¼ 0.508, P ¼ 0.002, Fig. 1a) and SUVmean (r ¼ 0.513, P ¼ 0.002, Fig. 1b). Similar inverse correlations were also identified between ADCmean and SUVmax (r ¼ 0.528, P ¼ 0.001, Fig. 1c), as well as ADCmean and SUVmean (r ¼ 0.548, P ¼ 0.001, Fig. 1d). For site-based analysis, a total of 69 histologically proven sites were analyzed. Table 2 shows the averaged values of the diagnostic characteristics
SUV and ADC in Peritoneal Metastasis
103
Table 2 Summary of Diagnostic Characteristics (Mean Value in Percentage) of Different Imaging Sequences and Modalities for Per-Site–Based Analysis With Reference to Histology
Sensitivity Specificity Accuracy PPV NPV AUC
DWI
MRI
DWI/MRI
FDG-PET/CT
81 99 95 97 95 90
88 99 97 97 96 93
88 99 97 97 96 93
88 100 97 100 96 94
assessed by radiologist 1 and 2, comparing the diagnostic performances of FDG-PET/CT, DWI, MRI, and DWI/MRI in peritoneal metastases detection (Fig. 2). Although statistically insignificant, DWI/MRI was slightly more sensitive in depicting peritoneal metastases (88%) as compared to DWI (81%). The diagnostic accuracy of DWI/MRI was comparable to that of FDGPET/CT (P > 0.05). In the right subphrenic space, DWI/MRI has one false positive result due to epiphrenic lymph node. There were two false negative sites on both DWI/MRI and FDG-PET/CT due to urinary bladder serosal deposits. The AUC for DWI, MRI, DWI/MRI, FDG-PET/CT fell between 0.9 and 1. The interobserver agreement was excellent for FDG-PET/ CT (k ¼ 1.000), DWI/MRI (k ¼ 0.956), MRI (k ¼ 0.956) and DWI (k ¼ 0.954) for site-based analysis. DISCUSSION Peritoneal metastases can be subtle and still extensively disseminated at the time of presentation (21). The
provision of peritoneal imaging has facilitated clinical decision-making, especially in surgical planning, and has improved patient outcome (2,3) and forms an important part of disease evaluation. The present study verified the inverse correlation between SUV and ADC in peritoneal metastases. SUV, a semi-quantitative parameter of glucose metabolism in tumor, is useful for tumor detection, characterization, prognosis, and treatment response monitoring in different cancers (22,23). SUVmax is the most clinically used parameter due to its ease of use and reproducibility without the dependence on tumor size and shape (24). SUVmean, on the contrary, is dependent on the threshold used for volume approximation, for which there is no consensus agreement. Both fixed and adaptable thresholds have been used in PET/CT to delineate tumor boundaries (23,24). In this study, we found that using fixed 45% threshold provided the best concordance between volumes measured by CT and PET with the lowest mean variance. ADC values are used for tumor grading, nodal metastases detection, and treatment response monitoring (25–29). A patient-specific threshold was applied to exclude poorly fitted voxel, which may lead to underestimation in ADC measurement, for more accurate quantification (30). Although SUV and ADC represent two diverse facets of cell biology, our results reveal a relationship between tumor metabolism and tissue cellularity. Studies showed that tumor metabolism correlated with cell proliferation and cell density (31,32). Tumors that have high cell proliferative indices have higher tissue cellularity, and therefore more restricted in diffusion, giving rise to lower ADC values. Understanding the biologic
Figure 2. A 53-year-old female diagnosed with serous adenocarcinoma of peritoneum showing peritoneal thickening at the right subphrenic space identified by white arrows shown on axial DWI images with b value of 0 (a), constructed ADC map using three b values (0, 400, 800 s/mm2) (b), 6 min post-contrast–enhanced image (c), and fused axial PET/CT image (d).
104
interactions, the relationship between SUV and ADC has been demonstrated in several other tumors. Ho et al (24) was the first to report a significant inverse correlation between relative SUVmax and relative ADCmin in mixed types of primary cervical cancer. The inverse correlation between SUV and ADC was also found in primary rectal, non-small cell lung cancer and gastrointestinal stromal lesions (22,23,33). In concordance, our study presented significant inverse correlations between ADC and SUV in 34 peritoneal metastases, regardless of their primaries with highly reproducible quantitative measurements. This robust correlation between ADC and SUV demonstrated the feasibility of using ADC as a diagnostic adjunct in peritoneal metastasis, similar to SUV. As cytotoxic treatment breaks up cell membrane, restores normal water molecular diffusion and extracellular matrix, the ADC value increases or becomes normalized, DWI has the capability of early treatment response assessment before tumor size changes (34). Further studies are encouraged to evaluate the potential roles of SUV and ADC in treatment response assessment in peritoneal metastases. In our study, DWI and MRI had inferior sensitivity to DWI/MRI and FDG-PET/CT, but the difference was not statistically significant. The inferior diagnostic performance of DWI may be caused by poor anatomic localization due to limited spatial resolution, low signal-to-noise ratio, and susceptibility artifacts. DWI viewed in conjunction with MRI improved the overall diagnostic performance in peritoneal metastasis detection, although modest. DWI had better background and bowel suppression while conventional MRI improved anatomic landmark depiction. This underlines the complementary value of DWI and conventional MRI. Even with this modest increment, the authors believe that DWI should be a routine sequence added to abdominal imaging with a small trade-off of an extra 5–6 min of scanning time. Our results demonstrated that DWI/MRI and FDGPET/CT had comparable diagnostic performance in peritoneal metastases detection. This finding is in concordance with previous study (5) concluding that DWI/MRI can be used as the alternative imaging technique for peritoneal dissemination detection when FDG-PET/CT is not available. In Satoh et al (5), two different cohorts of patients were recruited. Majority of the FDG-PET/CT patients (15/27) had gastrointestinal malignancies while patients who underwent DWI/MRI had genitourinary malignancies (16/23). Therefore, no direct comparison could be made in these two cohorts of patients with different malignancies. Soussan et al (3) reported that DWI/MRI had higher sensitivity and specificity than FDG-PET/CT, but the results were statistically insignificant. This finding could be related to the large number of subcentimeter lesions included in this study (12/27), which could underestimate the glucose uptake on PET due to the partial volume effect, therefore reducing the diagnostic performance of FDG-PET/CT. The two false negative lesions on both MRI and FDG-PET/CT were micro-metastases with thin layers of cells over the urinary bladder serosal surface, only
Yu et al.
confirmed by routine biopsy of the region and not visible on gross inspection during surgery. FDG excretion in the urinary bladder will reduce the contrast between tumor and the high background uptake, thus reducing the detection rate. There are several limitations in this study. First, the number of patients recruited was small, preventing meaningful patient-based analysis. Second, patients enrolled in this study had various histological types, which may have confounded the correlation between SUV and ADC, in that different histology may have different biologic behaviors. Nevertheless, we found a significant correlation between SUV and ADC. Finally, recruiting patients with suspected peritoneal metastasis on FDG-PET/CT may lead to selection bias, in that patients with small lesions undetected on FDGPET/CT were excluded from the study. In conclusion, the negative correlation between SUV and ADC provides complimentary information on tumor biology, with similar diagnostic accuracy between FDG-PET/CT and DWI/MRI. These techniques can be used in synergistic manner, or as an alternative imaging when access to FDG-PET/CT is limited, in peritoneal dissemination evaluation and add values to our current understanding of the biologic behaviors of these tumors. REFERENCES 1. Healy JC, Reznek RH. The peritoneum, mesenteries and omenta: normal anatomy and pathological processes. Eur Radiol 1998;8: 886–900. 2. Low RN. MR imaging of the peritoneal spread of malignancy. Abdom Imaging 2007;32:267–283. 3. Soussan M, Des Guetz G, Barrau V, et al. Comparison of FDGPET/CT and MR with diffusion-weighted imaging for assessing peritoneal carcinomatosis from gastrointestinal malignancy. Eur Radiol 2012;22:1479–1487. 4. Low RN, Sebrechts CP, Barone RM, Muller W. Diffusion-weighted MRI of peritoneal tumors: comparison with conventional MRI and surgical and histopathologic findings--a feasibility study. AJR Am J Roentgenol 2009;193:461–470. 5. Satoh Y, Ichikawa T, Motosugi U, et al. Diagnosis of peritoneal dissemination: comparison of 18F-FDG PET/CT, diffusionweighted MRI, and contrast-enhanced MDCT. AJR Am J Roentgenol 2011;196:447–453. 6. Kyriazi S, Kaye SB, deSouza NM. Imaging ovarian cancer and peritoneal metastases-current and emerging techniques. Nat Rev Clin Oncol 2010;7:381–393. 7. Berthelot C, Morel O, Girault S, et al. Use of FDG-PET/CT for peritoneal carcinomatosis before hyperthermic intraperitoneal chemotherapy. Nucl Med Commun 2011;32:23–29. 8. Kim CK, Park BK, Choi JY, Kim BG, Han H. Detection of recurrent ovarian cancer at MRI: comparison with integrated PET/CT. J Comput Assist Tomogr 2007;31:868–875. 9. Dirisamer A, Schima W, Heinisch M, et al. Detection of histologically proven peritoneal carcinomatosis with fused 18F-FDG-PET/ MDCT. Eur J Radiol 2009;69:536–541. 10. Kitajima K, Yamasaki E, Kaji Y, Murakami K, Sugimura K. Comparison of DWI and PET/CT in evaluation of lymph node metastasis in uterine cancer. World J Radiol 2012;4:207–214. 11. Herneth AM, Guccione S, Bednarski M. Apparent diffusion coefficient: a quantitative parameter for in vivo tumor characterization. Eur J Radiol 2003;45:208–213. 12. Heusner TA, Kuemmel S, Koeninger A, et al. Diagnostic value of diffusion-weighted magnetic resonance imaging (DWI) compared to FDG PET/CT for whole-body breast cancer staging. Eur J Nucl Med Mol Imaging 2010;37:1077–1086. 13. Song I, Kim SH, Lee SJ, Choi JY, Kim MJ, Rhim H. Value of diffusion-weighted imaging in the detection of viable tumour after neoadjuvant chemoradiation therapy in patients with locally
SUV and ADC in Peritoneal Metastasis
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
advanced rectal cancer: comparison with T2 weighted and PET/ CT imaging. Br J Radiol 2012;85:577–586. Nagamachi S, Wakamatsu H, Kiyohara S, et al. Comparison of diagnostic and prognostic capabilities of F-18-FDG-PET/CT, I-131-scintigraphy, and diffusion-weighted magnetic resonance imaging for postoperative thyroid cancer. Jpn J Radiol 2011;29: 413–422. De Paepe K, De Keyzer F, Wolter P, Oyen R, Verhoef G, Vandecaveye V. 5.3 Tesla whole-body diffusion-weighted imaging for staging lymphoma: comparison with PET/CT. Cancer Imaging 2011;11 Spec No A:S178. Takenaka D, Ohno Y, Matsumoto K, et al. Detection of bone metastases in non-small cell lung cancer patients: comparison of whole-body diffusion-weighted imaging (DWI), whole-body MR imaging without and with DWI, whole-body FDG-PET/CT, and bone scintigraphy. J Magn Reson Imaging 2009;30:298–308. Lagana M, Rovaris M, Ceccarelli A, Venturelli C, Marini S, Baselli G. DTI parameter optimisation for acquisition at 1.5T: SNR analysis and clinical application. Comput Intell Neurosci 2010: 254032. Gong NJ, Wong CS, Chu YC, Gu J. Treatment response monitoring in patients with gastrointestinal stromal tumor using diffusion-weighted imaging: preliminary results in comparison with positron emission tomography/computed tomography. NMR Biomed 2013;26:185–192. Boellaard R, O’Doherty MJ, Weber WA, et al. FDG PET and PET/ CT: EANM procedure guidelines for tumour PET imaging: version 1.0. Eur J Nucl Med Mol Imaging 2010;37:181–200. Lee KR, Tavassoli FA, Prat J, et al. Tumours of the ovary and peritoneum. In: Tavassoli FA, Devilee P, editors. Pathology and genetics tumours of the breast and female genital organs. 3rd edition. Volume 4. Lyon: IARC Press; 2003. p 113–202. Klaver YL, Lemmens VE, de Hingh IH. Outcome of surgery for colorectal cancer in the presence of peritoneal carcinomatosis. Eur J Surg Oncol 2013;39:734–741. Regier M, Derlin T, Schwarz D, et al. Diffusion weighted MRI and 18F-FDG PET/CT in non-small cell lung cancer (NSCLC): does the apparent diffusion coefficient (ADC) correlate with tracer uptake (SUV)? Eur J Radiol 2012;81:2913–2918. Gu J, Khong PL, Wang S, Chan Q, Law W, Zhang J. Quantitative assessment of diffusion-weighted MR imaging in patients with
105
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
primary rectal cancer: correlation with FDG-PET/CT. Mol Imaging Biol 2011;13:1020–1028. Ho KC, Lin G, Wang JJ, Lai CH, Chang CJ, Yen TC. Correlation of apparent diffusion coefficients measured by 3T diffusionweighted MRI and SUV from FDG PET/CT in primary cervical cancer. Eur J Nucl Med Mol Imaging 2009;36:200–208. Matoba M, Tonami H, Kondou T, et al. Lung carcinoma: diffusion-weighted mr imaging--preliminary evaluation with apparent diffusion coefficient. Radiology 2007;243:570–577. Gourtsoyianni S, Papanikolaou N, Yarmenitis S, Maris T, Karantanas A, Gourtsoyiannis N. Respiratory gated diffusion-weighted imaging of the liver: value of apparent diffusion coefficient measurements in the differentiation between most commonly encountered benign and malignant focal liver lesions. Eur Radiol 2008;18:486–492. Nakajo M, Kajiya Y, Kaneko T, et al. FDG PET/CT and diffusionweighted imaging for breast cancer: prognostic value of maximum standardized uptake values and apparent diffusion coefficient values of the primary lesion. Eur J Nucl Med Mol Imaging 2010; 37:2011–2020. Pauls S, Schmidt SA, Juchems MS, et al. Diffusion-weighted MR imaging in comparison to integrated [(1)(8)F]-FDG PET/CT for N-staging in patients with lung cancer. Eur J Radiol 2012;81:178–182. Dudeck O, Zeile M, Pink D, et al. Diffusion-weighted magnetic resonance imaging allows monitoring of anticancer treatment effects in patients with soft-tissue sarcomas. J Magn Reson Imaging 2008;27:1109–1113. Jones DK, Basser PJ. “Squashing peanuts and smashing pumpkins”: how noise distorts diffusion-weighted MR data. Magn Reson Med 2004;52:979–993. Herholz K, Pietrzyk U, Voges J, et al. Correlation of glucose consumption and tumor cell density in astrocytomas. A stereotactic PET study. J Neurosurg 1993;79:853–858. Higashi K, Ueda Y, Yagishita M, et al. FDG PET measurement of the proliferative potential of non-small cell lung cancer. J Nucl Med 2000;41:85–92. Wong CS, Gong N, Chu YC, et al. Correlation of measurements from diffusion weighted MR imaging and FDG PET/CT in GIST patients: ADC versus SUV. Eur J Radiol 2012;81:2122–2126. Morse DL, Galons JP, Payne CM, et al. MRI-measured water mobility increases in response to chemotherapy via multiple celldeath mechanisms. NMR Biomed 2007;20:602–614.
