Value Of DCE-MRI Research Articles

SU-D-303-02: Impact of Arterial Input Function Selection and T10 Correction On DCE-MRI Tumour Response Prediction Using Compared to Volumetric DCE CT

Purpose: To evaluate the impact of individualized magnitude and phase signal arterial input function (AIF) measurements as well as voxel-based pre-contrast T10 relaxation on tumour perfusion metrics from DCE-MRI compared to DCE-CT using a common 4D temporal dynamic analysis (TDA) method. Methods: Nine patients with 13 brain metastases underwent volumetric DCE-CT (Toshiba, Aquilion ONE) and DCE-MRI (IMRIS 3T Verio) at baseline then 7 and 21 days post-radiosurgery. Voxel-based whole brain TDA was performed on all data using in-house software producing kinetic parameters AUC, Ktrans, Kep, and Vb (using the Modified Tofts model). AIF susceptibility was investigated in DCE-CT by selecting the AIF (from internal carotid artery) or VIF (from Sagittal Sinus). In DCE-MRI an individual Magnitude or Phase-based AIF (Sagittal Sinus) was compared to population-based AIF together with susceptibility to voxel-based T10 maps instead of a constant of 2400 msec. Absolute DCE-MRI values were compared to DCE-CT by Pearson correlation. Results: No significant difference in median Ktrans (0.048 +/−0.03 s-1) or AUC (2785.5 +/−1143.6 HU.s) was found between individual AIF and VIF-based DCE CT analyses. Using individual Magnitude VIF or Phase-based AIF for DCE-MRI (T10 2400 ms) resulted in higher Ktrans values (0.181 +/−0.11 vs. 0.121 +/−0.099 s-1). This is likely resulting from the smaller AIF peak since the population AIF (which more closely resembles CT) correlates better to DCE-CT metrics. Using voxel-based T10 maps caused further statistically-significant increase in Ktrans and AUC (p<0.0006) that could be contributed to a lower median T10 value (1572 +/−594, n=41). Conclusion: This preliminary data highlights the stability of DCE-CT calculations as well as susceptibility of DCE-MRI Ktrans measurements to various imaging factors, including AIF selection and T10 values used for modeling. Efforts to improve voxel-based T10 map calculations are being explored to further explain discrepancies between analysis methods. Brain Tumor Foundation of Canada and Ontario Institute of Cancer Research

Predictive potential of angiogenic plasma biomarkers (PBs) in phase I trial with NGR-hTNF.

e13612 Background: NGR-hTNF is a vascular targeting agent consisting of TNF fused with the peptide NGR, a ligand of a CD13 overexpressed on tumor blood vessels. Methods: Fifteen refractory patients (pts) received 83 cycles (range, 1-29) of NGR-hTNF (0.2-1.6 μg/m2) given every 3 weeks until progression. Median progression-free (PFS) and overall survival (OS) were 3.4 and 6.9 months, respectively. Ninety PBs were tested at baseline and 2-h after first cycle (post-dosing). Significant post-dosing changes in IL1/IL8/IL10/MCP1/MIP1b were noted. Based on their angiogenic role, an additional four PBs were selected: VEGF, matrix metalloproteinase-9 (MMP9), vascular-cell and inter-cellular adhesion molecule-1 (VCAM1 and ICAM1). The associations between these nine PBs and treatment duration (TD; ≤2 v >2 cycles), OS and maximal changes in tumour vascularity (assessed every cycle by DCE-MRI) were investigated. Results: TD was associated with OS (Spearman's r=0.55, p=.04), baseline DCE-MRI values (r=0.86, p=.004) and moderately with changes in DCE-MRI (r=−0.51, p=.08). Baseline ICAM1 levels inversely correlated with both TD (r=−0.56, p=.03) and OS (r=−0.66, p=.01), as well as post-dosing ICAM1 levels (r=−0.55, p=.03 and r=−0.70, p=.009 for TD and OS, respectively). Post-dosing MMP9 levels correlated with OS (r=−0.57, p=.03) and moderately with TD (r=−0.46, p=.08). Greater decreases in DCE-MRI were associated with higher baseline IL10 levels (r=−0.65, p=.03). All the other PBs failed to show any correlation. By receiver operating characteristic analysis, baseline ICAM1 levels ≤171 ng/ml and post-dosing MMP9 levels ≤47 ng/ml yielded 87% sensitivity and 71% specificity in predicting TD. Pts with ICAM1 levels ≤ and >171 had median PFS of 8.2 and 2.4 months (p=.002) and median OS of 10.3 and 3.9 months (p=.008), respectively. Pts with MMP9 levels ≤ and >47 had median PFS of 5.8 and 2.0 months (p=.05) and median OS of 12.5 and 3.9 months (p=.005), respectively. Univariate Cox analysis confirmed significant associations with OS for both ICAM1 (p=.008) and MMP9 (p=.02). Conclusions: The lower the ICAM1 and MMP9 levels, the longer the clinical benefit. The predictive potential of these PBs for NGR-hTNF activity deserves further studies. Author Disclosure Employment or Leadership Position Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration MolMed

Hepatocellular Carcinoma in Cirrhotic Liver: Double-Contrast-Enhanced, High-Resolution 3.0T-MR Imaging With Pathologic Correlation

To determine the diagnostic performance of double-contrast-enhanced (DCE) magnetic resonance imaging (MRI) at 3.0T in the detection of hepatocellular carcinoma (HCC) in patients with a cirrhotic liver. This study was approved by our Institutional Review Board and did not require informed consent. A total of 61 patients (52 men, 9 women; range 27-83 years) underwent DCE-MRI at 3.0T. DCE-MRI was composed of baseline MR images, superparamagnetic iron oxide-enhanced T2- and T2*-weighted images, and gadobenate-dimeglumine-enhanced dynamic MR images. The diagnosis of HCC was established at transplantation (n = 12) and hepatic resection (n = 49). Three observers independently analyzed each image in random order. The diagnostic performance, sensitivity, positive predictive value, numbers and causes of false positive and negative findings, and interobserver variability were calculated and analyzed on a per-lesion basis. : Sixty-one patients had 95 HCCs. From alternative free-response receiver operating characteristic analysis, averaged diagnostic performance (Az) was 0.90. The mean sensitivity and positive predictive value of DCE-MRI at 3.0T for detection of HCCs were 89% and 97%, respectively. There was a significant difference in the sensitivity for the detection of HCCs between Child-Pugh class A and C patients (97% and 63%, respectively) (P < 0.01). In addition, the sensitivity according to lesion size was significantly different (P < 0.01). There were 7 false-positive lesions [arterio-portal shunt (n = 1), dysplastic nodule (n = 1), hemangioma (n = 2), and partial volume averaging (n = 3)]. Thirteen false-negatives occurred mainly because of small size (n = 4), partial uptake of superparamagnetic iron oxide (n = 6), faint arterial enhancement (n = 7), and location (n = 2). There was good interobserver variability (kappa: 0.622-0.670). DCE 3.0T MR imaging is a highly sensitive modality for diagnosing HCCs in patients with a cirrhotic liver.

Prediction of chemotherapeutic response of colorectal liver metastases with dynamic gadolinium‐DTPA‐enhanced MRI and localized 19F MRS pharmacokinetic studies of 5‐fluorouracil

Systemic chemotherapy is effective in only a subset of patients with metastasized colorectal cancer. Therefore, early selection of patients who are most likely to benefit from chemotherapy is desirable. Response to treatment may be determined by the delivery of the drug to the tumor, retention of the drug in the tumor and by the amount of intracellular uptake, metabolic activation and catabolism, as well as other factors. The first aim of this study was to investigate the predictive value of DCE-MRI with the contrast agent Gd-DTPA for tumor response to first-line chemotherapy in patients with liver metastases of colorectal cancer. The second aim was to investigate the predictive value of 5-fluorouracil (FU) uptake, retention and catabolism as measured by localized (19)F MRS for tumor response to FU therapy. Since FU uptake, retention and metabolism may depend on tumor vascularization, the relationship between (19)F MRS and the DCE-MRI parameters k(ep), K(trans) and v(e) was also examined (1). In this study, 37 patients were included. The kinetic parameters of DCE-MRI, k(ep), K(trans) and v(e), before start of treatment did not predict tumor response after 2 months, suggesting that the delivery of chemotherapy by tumor vasculature is not a major factor determining response in first-line treatment. No evident correlations between (19)F MRS parameters and tumor response were found. This suggests that in liver metastases that are not selected on the basis of their tumor diameter, FU uptake and catabolism are not limiting factors for response. The transfer constant K(trans), as measured by DCE-MRI before start of treatment, was negatively correlated with FU half-life in the liver metastases, which suggests that, in metastases with a larger tumor blood flow or permeability surface area product, FU is rapidly washed out from the tumor.