Optimization of Ultrasmall Superparamagnetic Iron Oxide (P904)-enhanced Magnetic Resonance Imaging of Lymph Nodes: Initial Experience in a Mouse Model

Materials and Methods

This experiment was approved by the Animal Care Committee at our institution (IACUC 10-0235).

Animal preparation. The experiment was performed using 19 6-week-old BALB/c nude mice (ORIENT, Sungnam, Kyunggi, Korea). The animals were allowed food and water ad libitum. Fourteen normal BALB/c nude mice were used to determine the optimal dose of P904 that produces the maximum T2* effect in normal lymph nodes. Murine B16F10 melanoma cells (American Type Culture Collection [ATCC], Rockville, MD, USA), with high potential to metastasize to regional lymph nodes, were implanted into the bilateral upper brachial areas of the remaining five mice to create an animal model of metastatic lymph nodes, following a previously-reported method (17). From the American Type Culture Collection (ATCC, Rockville, MD, USA), the B16F10 cells were obtained and maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum at 37°C (17). For the metastatic lymph node models, MRI was performed with a 3.0T scanner one week after cell implantation.

MRI contrast material. P904, a new USPIO particle for macrophage imaging, was used in this study. P904 consists of a maghemite core and a coating of a low-molecular-weight amino-alcohol derivative of glucose, which is equivalent to 2000 hydroxyl groups per nanoparticle (4). P904 has a hydrodynamic diameter of approximately 21 nm, which is slightly smaller than that of ferumoxtran-10 (i.e. 30 nm) (1). With regard to T1 and T2 relaxivities, r1 and r2 of P904 measured in water at 1.42 T and 37°C were 14 and 87 mM⁻¹ s⁻¹, respectively (4).

MRI. Prior to MRI, the mice were anesthetized by intraperitoneal injection of a solution containing zolazepam (5 mg/kg, Zoletil®; Virbac, Carros cedex, France) and xylazine (10 mg/kg, Rompun®; Bayer-Schering Pharma, Leverkusen, Germany). All MRI were performed using a 3.0 T scanner (TrioTim; Siemens, Erlangen, Germany) with a 4-channel wrist coil.

Optimization of P904 dose for lymph node imaging in normal mice. At day one, a precontrast coronal T2*-weighted three-dimensional (3D) gradient-echo sequence was acquired after routine localization images. P904 was prepared in the following three different doses to find the optimal dose for lymph node imaging: 300, 150, and 75 μmol Fe/kg. Normal BALB/c nude mice (n=14) were divided into three groups and P904 was administered to the mice of each group via their tail veins: 300 μmol Fe/kg (n=5), 150 μmol Fe/kg (n=4), and 75 μmol Fe/kg (n=5). A postcontrast coronal T2*-weighted 3D gradient-echo sequence was acquired 24 hours after the contrast administration. The scans covered the entire region from the neck to the upper thigh.

The imaging parameters, including repetition time, echo time, number of signals acquired, and section thickness, were modified to find the optimized values for accurate depiction of the brachial and inguinal lymph nodes. The optimized parameters thus selected were as follows: repetition time/echo time, 40 ms/22 ms; flip angle, 12°; field of view, 49×70 mm; matrix, 256×180; slice thickness, 0.6 mm; number of excitations, 6; and pixel resolution, 0.2×0.4×0.6 mm.

Metastatic lymph node model. As for the five mice of the metastatic lymph node model, the optimal dose determined using the normal mice was administered via their tail veins. Pre- and postcontrast coronal T2*-weighted 3D gradient-echo images were acquired using the same parameters.

Isolation of lymph nodes and histopathological examination. After the acquisition of the postcontrast T2*-weighted MR images, all five metastatic lymph node models were sacrificed, and their brachial and inguinal lymph nodes were isolated by one author (H.R.C.). The lymph nodes were fixed with 10% formalin and embedded in paraffin. Afterwards, the lymph nodes were sectioned at 0.5-mm intervals, and from each 0.5-mm section, a 4-μm-thick slide was prepared for hematoxylin-eosin (H&E) staining and subsequently examined by a pathologist (J.K.W.) for the presence of metastasis.

Image analysis. The MRI data were digitally transferred from a picture archiving and communication systems (PACS) workstation to a personal computer and processed with Image J (available at http://rsb.info.nih.gov/ij/) and software developed in-house using Microsoft Visual C++.

Qualitative analysis for differentiation of benign and metastatic lymph nodes in metastatic lymph node models: Two investigators (S.H.C. and J.H.K. with seven and 15 years of experience in the interpretation of MRI studies), who were blinded to the histopathologic information, performed the qualitative analysis of the MRIs presented in random order by one author (R.E.Y.). All the MRIs were reviewed on a PACS workstation. A diagnosis of malignancy was made by consensus, using previously reported criteria (13, 18, 19): a lymph node with an overall high SI, an eccentric or central high SI with darkening along its peripheral rim on the postcontrast T2*-weighted MRI was considered to be malignant. On the other hand, a node with an overall dark SI, an overall dark SI apart from its fatty hilum, an overall dark SI with tiny speckles of high SI, or a central low SI on the postcontrast T2*-weighted MRI was considered to be nonmalignant.

Quantitative analysis of the SI of each lymph node: Quantitative analysis of the SI for each lymph node was performed by one author (R.E.Y.). On each section of the T2*-weighted MRIs, a region of interest (ROI) was drawn to include the entire area of a brachial or an inguinal lymph node using Image-J software. The data acquired from each section were summated to derive voxel-by-voxel SIs for the entire lymph node by using software developed in-house. Given the earlier findings that the muscle tissue remains unchanged by the contrast agent (18, 19), the voxel-by-voxel SI values of the lymph node were normalized with respect to the SI of the shoulder muscle, which was measured using a single ROI of 3-5 mm², to cancel the SI fluctuations related to variations in technical parameters between the pre- and postcontrast images. The following formula was used to calculate the SI ratio from the nSIs: SI ratio=(SI_{lymph node[post-P904]}/SI_{muscle[post-P904]})/(SI_{lymph node[pre-P904]}/SI_{muscle[pre-P904]}). Furthermore, SNR, CNR, and CR were calculated using the following equations (20-23): where SI_{lymph nodes} is the average signal intensity of the brachial or inguinal lymph nodes, σ_noise(air) is the standard deviation of the background noise measured at the air with a ROI of 3-5 mm², and SI_muscle is the average signal intensity of the shoulder muscle.

Statistical analysis. Statistical analyses were performed using commercially available software (MedCalc, version 8.0.0.1; MedCalc software, Mariakerke, Belgium). The data for each parameter was assessed for normality with the Kolmogorov-Smirnov test. In all tests, p-values less than 0.05 were considered statistically significant.

Optimization of P904 dose for lymph node imaging in normal mice. For each group to which a different dose of P904 was administered [300 μmol Fe/kg (n=5), 150 μmol Fe/kg (n=4), and 75 μmol Fe/kg (n=5)], the paired Student t-test was used to assess whether there was a significant difference in the mean nSI, SNR, CNR, and CR values of the brachial and inguinal lymph nodes between the pre- and postcontrast images.

To determine the optimal dose of P904, the pre- and postcontrast mean nSI, SNR, CNR, and CR values of the lymph nodes of the three groups were compared, using repeated measurements analysis of variance (ANOVA) with Tukey–Kramer post hoc comparisons for multiple comparisons.

Metastatic lymph node model. Qualitative analysis for differentiation of benign and metastatic lymph nodes: Sensitivity and specificity for differentiation of the benign and metastatic lymph nodes were calculated for mice of the metastatic lymph node model (n=5) to which the optimal dose of P904 was administered. Data clustering (i.e. more than one lesion per mouse) was accounted for with the method reported by Rao and Scott (24).

Metastatic lymph node model. Quantitative analysis of the SI in each lymph node: With respect to the optimal P904 dose, Mann–Whitney U-test was used to assess whether the mean SI ratio of the benign lymph nodes differed significantly from that of the metastatic lymph nodes.

Receiver operating characteristic (ROC) curve analysis was performed to determine the optimal cutoff value of the mean SI ratio, which provided a balance between sensitivity and specificity for the diagnosis of lymph node metastasis. The Az value (areas under the receiver operating characteristic curves) as well as the corresponding sensitivity and specificity at the optimal cutoff value were calculated, along with 95% confidence intervals (CIs).