Mapping Hypoxia in Renal Carcinoma with Oxygen-enhanced MRI: Comparison with Intrinsic Susceptibility MRI and Pathology

Purpose To cross-validate T1-weighted oxygen-enhanced (OE) MRI measurements of tumor hypoxia with intrinsic susceptibility MRI measurements and to demonstrate the feasibility of translation of the technique for patients. Materials and Methods Preclinical studies in nine 786–0-R renal cell carcinoma (RCC) xenografts and prospective clinical studies in eight patients with RCC were performed. Longitudinal relaxation rate changes (∆R1) after 100% oxygen inhalation were quantified, reflecting the paramagnetic effect on tissue protons because of the presence of molecular oxygen. Native transverse relaxation rate (R2*) and oxygen-induced R2* change (∆R2*) were measured, reflecting presence of deoxygenated hemoglobin molecules. Median and voxel-wise values of ∆R1 were compared with values of R2* and ∆R2*. Tumor regions with dynamic contrast agent–enhanced MRI perfusion, refractory to signal change at OE MRI (referred to as perfused Oxy-R), were distinguished from perfused oxygen-enhancing (perfused Oxy-E) and nonperfused regions. R2* and ∆R2* values in each tumor subregion were compared by using one-way analysis of variance. Results Tumor-wise and voxel-wise ∆R1 and ∆R2* comparisons did not show correlative relationships. In xenografts, parcellation analysis revealed that perfused Oxy-R regions had faster native R2* (102.4 sec–1 vs 81.7 sec–1) and greater negative ∆R2* (−22.9 sec–1 vs −5.4 sec–1), compared with perfused Oxy-E and nonperfused subregions (all P < .001), respectively. Similar findings were present in human tumors (P < .001). Further, perfused Oxy-R helped identify tumor hypoxia, measured at pathologic analysis, in both xenografts (P = .002) and human tumors (P = .003). Conclusion Intrinsic susceptibility biomarkers provide cross validation of the OE MRI biomarker perfused Oxy-R. Consistent relationship to pathologic analyses was found in xenografts and human tumors, demonstrating biomarker translation. Published under a CC BY 4.0 license. Online supplemental material is available for this article.

Animals were housed in specific pathogen-free rooms in autoclaved, aseptic microisolator cages with a maximum of four animals per cage. Food and water were provided ad libitum. The mice were routinely monitored for appearance of palpable tumors. Animals were from a pilot study so no formal power calculation was performed. All animals that were entered into the protocol survived and gave analyzable data. Since all animals had one scan and no treatment, blinding and randomization was not required. No adverse effects were experienced by any animal.

Preclinical Anesthetic Method
Anesthesia was induced with a 10 mL/kg intraperitoneal injection of fentanyl citrate (0.315 mg/mL) plus fluanisone (10 mg/mL) (Hypnorm; Janssen Pharmaceutical Ltd, High Wycombe, UK), midazolam (5 mg/mL) (Hypnovel; Roche, Welwyn Garden City, UK) and sterile water (at 1:1:2 ratio) (2). Mice were positioned in a 3 cm birdcage coil on a custom built platform to isolate the tumor which was surrounded by dental paste (3M; Bracknell, UK) to minimize motion and susceptibility artifacts. Gas delivery (medical air or 100% oxygen) was continuous at 2 l/min through a nose piece. Warm air maintained animal core temperature at 37C. Lateral tail vein cannulation was performed with a heparinized 27G butterfly catheter (Venisystems, Hospira, Royal Leamington Spa, UK) for intravenous administration of gadolinium contrast agent.

Clinical Tumor Pathology Analysis
Whole nephrectomy was performed and specimens were transferred to the pathology laboratory within 30 minutes. Tissue sections (4 µm) were obtained from formalin fixed paraffin embedded tumor material. Immunohistochemistry for the hypoxia-regulated gene glucose transporter 1 (GLUT1) staining was performed using a rabbit polyclonal antihuman GLUT1 (GT-12A; Alpha Diagnostics International, San Antonio, USA) at 10 µg/mL concentration, detected using rabbit Envision Plus HRP (Dako K4003) and visualized (brown stain) using DAB. A hypoxia score (H score) was generated by an experienced consultant clinical pathologist (GNB). All pathology images were scanned using a Leica SCN400 slide scanner microscope (Leica Microsystems, Milton Keynes, UK) at 40× magnification.

Image Analysis
Comparable analyses were performed for both preclinical and clinical data. R2* values were calculated using an exponential fit to the data, for both preclinical and clinical tumors.
In preclinical data analysis, the R1(air) was calculated as the average of the first two R1 maps. The R1(O2) was derived from the R1 map acquired during oxygen breathing. Voxels were defined as oxygen enhancing if the ∆R1 > than (2 × voxel native R1 x cohort coefficient of variation (CoV) for the R1; where the CoV was determined from two air-breathing R1 acquisitions. In clinical data analysis used a least squares fit to the multiple inversion time data to define voxel R1. The R1(air) was derived from the average of the first nine R1 maps. The R1(O2) was derived from the average of the last nine time points at the end of the oxygen breathing phase. Voxels were defined as oxygen enhancing by performing a t test on the R1 values from the air and oxygen phases.
Preclinical analysis derived the model free parameter initial area under the curve was calculated from 0 to 60 seconds (IAUC60; units mmol.kg 1 ) (3) and voxels with IAUC60 > 0 were considered enhancing. Clinical analysis compared signal intensity before (first 14 time points) and after (time points 100-155) contrast agent injection, with significant enhancement for each voxel determined by a t test.