Correlation among DCE-MRI measurements of bone marrow angiogenesis, microvessel density, and extramedullary disease in patients with multiple myeloma
Journal
American Journal of Hematology
Journal Volume
87
Journal Issue
8
Pages
837-839
Date Issued
2012
Author(s)
Abstract
We investigated correlations among angiogenesis parameters of the lumbar vertebrae measured by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), microvessel density (MVD) in bone marrow (BM), and extramedullary disease (EMD) in patients with multiple myeloma (MM). Forty-nine MM patients were enrolled. Two semiquantitative parameters, Peak and Slope, were obtained from the DCE-MRI signal-intensity curve; three more quantitative parameters, Amp, Kep, and Kel, were generated from bicompartmental modeling. Apart from Kep, all parameters were found to correlate positively with MVD (r range, 0.323–0.594; all P < 0.03). Multivariate analysis indicated that the only factors significantly associated with MVD were Amp and plasma cell percentage in BM. Comparing angiogenesis parameters for patients with EMD at the time of DCE-MRI versus those who did not showed a high Amp (≥0.08) as the only significant factor associated with EMD (odds 6.33; P = 0.045). During follow-up (median, 76 months), 4 more patients developed EMD. Accumulative incidence for developing EMD over time was significantly higher for patients with high Amp than those with low Amp (P = 0.0254). In conclusion, Amp correlated strongly with MVD in BM and also EMD in patients with MM. Amp measurement might be helpful for identifying MM patients at risk for EMD. Multiple myeloma (MM) is a malignant plasma cell proliferation typically found in bone marrow (BM) [1]. Although MM cells (MCs) depend on the BM microenvironment to provide the signals essential for their growth and survival [2], in a fraction of patients MCs acquire the ability to proliferate in sites outside the BM. Such occurrences appear as extramedullary disease (EMD), indicating that MCs have become independent of the BM microenvironment [1]. The exact mechanism underlying the development of EMD in MM patients is not clear. One hypothesis suggests an alteration in the interaction between MCs and the BM microenvironment [2, 3]. Within the BM microenvironment, angiogenesis might play a major role in not only promoting the growth and survival of MCs but also the disease progression itself [2-4]. The interaction between MCs and BM endothelial cells upregulates a number of angiogenic cytokines, such as vascular endothelial growth factor or matrix metalloproteinases. Such cytokines further stimulate BM angiogenesis and myeloma progression [5, 6], as well as possible extramedullary dissemination [4]. To date, dynamic contrast enhancement magnetic resonance imaging (DCE-MRI) is one of the most widely used noninvasive methods of measuring the perfusion and permeability of a biological tissue in the body, such as vertebral BM [7]. In MM patients, the angiogenesis parameters generated from DCE-MRI of vertebral BM reportedly correlate strongly with histological grade of infiltration, osteolytic bone involvement, microvessel density (MVD), and serum markers of disease activity [8, 9]. However, prior to our study no data were available on the correlation between degree of BM angiogenesis and the development of EMD in MM patients. We thus examined the correlation between angiogenesis parameters generated from DCE-MRI of vertebral bodies, together with MVD in BM (obtained from the posterior iliac crest), with the manifestation of EMD in patients with MM. Between September 2004 and April 2006, a total of 49 patients with MM were enrolled. Of this group, 27 (55%) patients had newly diagnosed MM (NDMM), 6 patients (12%) were in a post-treatment plateau (PTRP), and 16 (33%) had progressive disease (PD). This study was approved by our institutional ethics committee, and written informed consent was obtained from all patients in accordance with the Declaration of Helsinki (ClinicalTrials.gov Identifier: NCT00166855). The salient clinical characteristics of the 49 patients at enrollment are listed in Table I. Notably, 19 of the 49 patients (39%) had EMD at the time of their DCE-MRI, among whom 7 patients (37%) had more than one manifestation of EMD. BM samples were obtained from the posterior iliac crest in all 49 patients at approximately the same time as DCE-MRI was performed (median 2 days; range, −3 to 7 days). The mean MVD [vessels/400× high-power field (HPF)] was 16.7 (range 1.5–40.2), with significant differences being found for patients in different subgroups (P = 0.006). Patients with NDMM had a mean MVD of 20.3 (95% confidence interval [CI: 15.3–25.2]), patients in the PTRP group had a mean MVD of 3.2 (95% CI: 1.4–5.0), and patients with PD had a mean MVD of 15.7 (95% CI: 9.9–21.5). The time-signal intensity (SI) curve shown in DCE-MRI correlated strongly with tissue MVD, where a high peak and steep slope were associated with high MVD (Fig. 1A,B). By contrast, a lower peak and gentler slope were associated with lower MVD (Fig. 1C,D). Moderate correlations were found between MVD and the two semiquantitative parameters Peak and Slope (r = 0.540 and 0.502, respectively; both P < 0.001). Amp and Kel, but not Kep, were also moderately correlated with MVD (r = 0.594 and 0.323, respectively; P < 0.001 and P = 0.024, respectively). Other salient characteristics significantly correlated with MVD were beta2-microglobulin (r = 0.305; P = 0.035), C-reactive protein (r = 0.348; P = 0.018), and percentage of MCs in BM (r = 0.637; P < 0.001). Further multiple linear regression analysis showed that only Amp (r = 45.7; 95% CI: 13.0–78.3; P = 0.007) and percentage of MCs in BM (r = 0.18; 95% CI: 0.06–0.29; P = 0.003) were independently correlated with MVD. Table II shows the results of our comparison of salient features and angiogenesis parameters between patients with and without EMD at the time of their DCE-MRI. When compared with patients without EMD, patients with EMD displayed significantly greater infiltration of MCs in BM and higher levels of the angiogenesis parameter Amp. Further multivariate analysis using a multiple logistic regression model showed that Amp was the only significant factor associated with EMD (OR 6.33; P = 0.045). The median follow-up period was 76 months (95% CI: 67.9–84.1) from the date of DCE-MRI. During this time, 4 of the 30 patients who did not have EMD at the time of their DCE-MRI will develop EMD thereafter. The accumulative incidence for development of EMD over time was significantly higher for patients with high Amp (≥0.08) than for patients with low Amp (<0.08) (P = 0.0254) (Fig. 2). Additional covariates were identified by univariate analysis as being significantly associated with the development of EMD (after DCE-MRI), including light chain isotype, high Kel (≥0.10), and high calcium (≥2.4 μmol/l). However, multivariate analysis using the Cox regression model did not confirm that any of these factors was independently significant. Correlation between MVD of BM and the time-SI curve from the DCE-MRI. (A) Immunohistochemical staining of BM using anti-CD34 antibody showed a high MVD with numerous thin, winding, and sprouting vessels (200× magnification). (B) The corresponding time-signal intensity (SI) curve was characterized by a high peak value and steep slope, followed by a rapid wash-out phase (49–68 s), with a subsequent slow wash-out equilibrium phase (69–600 s). (C) In contrast, a lower MVD (200× magnification) is shown together with (D) the corresponding time-SI curve with a flatter, lower peak and gentler slope. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] Accumulative incidence for development of EMD. A significant difference was found for the accumulative incidence of EMD developing over time, for patients with high Amp (solid line) versus those with low Amp (shaded line). To the best of our knowledge, this was the first study to demonstrate a possible correlation between the angiogenesis parameter Amp (generated from DCE-MRI of vertebral BM) and EMD in patients with MM. Our study identified high Amp values (≥0.08) as a possible risk factor associated with the development of EMD in MM patients; high Amp values indicate high tissue vascularity and permeability. This finding partly supports the hypothesis that BM angiogenesis may play role in the development of EMD in MM. Our finding that the angiogenesis parameter Amp correlated strongly with MVD and the extent of MC infiltration in BM was consistent with previous research [9]. However, in our study tissue MVD did not emerge as a significant variable correlated with EMD, despite its reputation as a “gold standard” for the detection and quantification of angiogenesis. One possible explanation is that, compared with tissue-based methods, image-based (DCE-MRI) methods of assessing angiogenesis reflect not only vascular density but also permeability. Tissue-based assessments are derived from a limited sampling area and do not reflect blood vessel permeability. Furthermore, image-based assessments advantageously provide an opportunity for subsequent examinations including a much larger area of investigation. Some studies have shown either a broad or no association between tissue MVD and DCE-MRI angiogenesis parameters [10]. Thus, for MM patients, quantitative analysis is necessary using DCE-MRI data and angiogenesis and blood-vessel permeability based microcirculation variables. However, a fuller understanding of what these various measurements imply is necessary before such assessments can be incorporated into routine clinical practice. Several limitations of this study should be noted. First, the design was cross-sectional and the number of study patients was limited; in addition, the disease status of MM patients at the time of their DCE-MRI was heterogeneous. Some, but not all patients had received anti-MM treatment, which could alter tumor angiogenesis, microvessel permeability, and behavior [10]. Nonetheless, our subgroup analysis using the data from NDMM patients only indicated that Amp remained strongly correlated with MVD (data not shown). Another recent study concluded that presence of EMD at diagnosis, rather than any treatment modalities ever used, was the only significant predictor of extramedullary recurrence. This finding suggests that the development of EMD, even during treatment, may possibly reflect different tumor biology [11]. The second main limitation of our study was that the histological specimen was obtained from the posterior iliac crest and not the vertebral column. This is of some concern, because MM may grow in a patchy rather than diffuse pattern; therefore, the degree of infiltration of the BM by MCs cannot be expected to be equal throughout the skeleton. Third, cytogenetic data were available only for a minority of our study patients, and were thus not included in the analysis. Prior research using DCE-MRI has reported an association between gain of 1q21, deletion of 13q14, and deletion of 17p13 on the one hand, and, on the other, distinct patterns of increased microcirculation [12]. However, controversy still surrounds the possible correlation between parameters of DCE-MRI and angiogenesis genes in MM patients [6]. We concluded that among our MM patients, the angiogenesis parameter Amp, derived from DCE-MRI of vertebral bodies, was strongly correlated with the tissue MVD of BM obtained from the posterior iliac crest. This finding possibly reflects not only the tissue-specific vascularity, but also the vascular permeability in a more sizable area than the MVD of BM. As a result, Amp rather than MVD may correlate with the development of EMD among MM patients. Thus, high Amp might be a risk factor that could help identify MM patients with the potential to develop EMD. The endothelial cells of microvessels were immunostained as described previously [13]. The MVD was calculated as described previously [14]. Our institution's protocols for the use of DCE-MRI have been described previously [15, 16]. In brief, DCE-MRI was performed at the midsection of vertebral bodies from T11 to sacrum, and the values of SI from L2 to L4 were measured by a radiologist and plotted as a time-SI curve. The time-SI curve was then fitted by the Mathematica (v 6.1) software (Wolfram Research, Champaign, IL) using a nonlinear curve-fitting function. The baseline SI (SIbase) on a time-SI curve was defined as the mean SI for the first five images, and the maximum SI (SImax) was defined as the maximum value of the first rapidly rising part of the curve. We set the total duration of DCE-MRI examination at 600 s to track the uptake kinetics of contrast agents. The contrast enhancement rise time (Trise) was defined as the time between SIbase and SImax. The two semiquantitative parameters, Peak and Slope, were calculated as (SImax − SIbase)/SIbase and (SImax − SIbase)/Trise, respectively. The angiogenesis parameters contained in time-SI courses based on DCE-MRI were described by three model quantitative parameters, Amp, Kep, and Kel, calculated using the bicompartmental model [15, 16]. Data for each patient were represented as the average for the parameters of vertebral bodies from L2 to L4. Among these parameters, Peak indicates the contrast material in the intra- and extra-vascular interstitial spaces, representing tissue perfusion; Slope predominantly indicates the contrast agent in the intravascular space, which is determined by tissue vascularization and perfusion as well as capillary permeability; Amp is similar to Slope but provides better quantification of vascularity. The efflux rate constant is represented as Kep, which indicates the permeability. The Kep parameter is a rate transfer coefficient. In this study, EMD in patients with MM was defined as the presence of MCs outside BM, in one of the following forms: soft-tissue mass spreading from the bone (periosseous plasmacytoma), MCs arising in extraosseous organs (extraosseous plasmacytoma), malignant effusion, or plasma cell leukemia. The presence of EMD was diagnosed in most cases by MRI or computed tomography, which were conducted whenever EMD was suspected from clinical, laboratory, or radiographic findings. Histological confirmation was obtained whenever possible. Chi-square or Fisher's exact tests were used for between-group comparisons of the discrete variables. A two-sample t-test or one-way ANOVA was used for between-group comparison of the means. Pearson's correlational tests were used to analyze the continuous variables, and Spearman correlation was used for the nominal variables. Time to development of EMD (TTE) was calculated from the date on which DCE-MRI was performed to the date of documented EMD thereafter. Kaplan–Meier survival curves were constructed to estimate TTE, and the differences between groups were compared by the Log-rank test. Salient variables for clinical and laboratory data were categorized as described previously [17]. The tertile cutoff values for the DCE-MRI quantitative angiogenesis parameters, Amp (0.08, 0.19), Kep (0.008, 0.01), and Kel (0.10, 0.15), were applied in the univariate analysis. Factors that provided statistically significant predictive power in univariate analysis were further tested by multivariate regression analysis of the linear, logistic, or Cox type. Because of our limited case numbers, for multivariate analysis we pooled the cases into two Amp groups, as follows: high Amp, where Amp was ≥0.08, and low Amp, where Amp was <0.08. All directional P-values were two-tailed, with a P-value of 0.05 or less being considered significant for all tests. All analyses were performed using SPSS 19.0 software (Chicago, IL). The authors would like to thank the staff of the National Translational Medicine Resource Center and the Medical Research Centre of National Taiwan University Hospital for their help with the statistical analysis. Shang-Yi Huang*, Bang-Bin Chen , Hsiao-Yun Lu , Hsiu-Hsia Lin , Shwu-Yuan Wei , Szu-Chun Hsu , Tiffany Ting-Fang Shih , * Department of Medicine, National Taiwan University Hospital, Taipei, Taiwan, Republic of China, Department of Medical Imaging and Radiology, National Taiwan University, Medical College and Hospital, Taipei, Taiwan, Republic of China, Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan, Republic of China
SDGs
Other Subjects
adult; angiogenesis; clinical article; comparative study; compartment model; contrast enhancement; correlation analysis; density; extramedullary disease; female; follow up; human; incidence; letter; male; multiple myeloma; multivariate analysis; nuclear magnetic resonance imaging; plasma cell; plasma cell leukemia; plasmacytoma; priority journal; quantitative analysis; risk assessment; statistical significance; Aged; Bone Marrow; Contrast Media; Female; Follow-Up Studies; Humans; Magnetic Resonance Angiography; Male; Microvessels; Middle Aged; Multiple Myeloma; Neovascularization, Pathologic
Type
letter