Comparison of hypoxia-inducible factor-1alpha expression before and after transcatheter arterial embolization in rabbit VX2 liver tumors.

PURPOSE: To test the hypothesis that transcatheter arterial embolization (TAE) induces expression of hypoxia-inducible factor-1alpha (HIF-1alpha) within the same rabbit VX2 liver tumor. MATERIALS AND METHODS: Seven VX2 tumors were grown in the livers of five New Zealand white rabbits. Ultrasonography-guided biopsy was performed before and 10 minutes after TAE in all tumors. Pre- and post-TAE tumor biopsy specimens along with post-TAE whole liver tumor sections were stained with HIF-1alpha antibody and analyzed for percentage of HIF-1alpha-positive nuclei by using a spectral unmixing system mounted on a high-powered microscope. Statistical data comparisons were performed with the Wilcoxon signed-rank test (alpha = 0.05). RESULTS: TAE of liver tumors resulted in a statistically significant increase in the mean percentage of HIF-1alpha expression. The mean percentage of HIF-1alpha-positive stained nuclei increased from 23% +/- 3.5 in pre-TAE biopsy specimens to 41% +/- 8.7 in post-TAE biopsy specimens (P < .02). The increase was even more significant when the mean percentage of HIF-1alpha-positive stained nuclei from the same pre-TAE biopsy specimens was compared with sections from post-TAE whole tumor specimens (60% +/- 8.9, P < .02). CONCLUSIONS: The results of this study revealed that hypoxia caused by TAE of VX2 liver tumors activates HIF-1alpha, a transcription factor that in turn regulates other pro-angiogenic factors.

Tumor response after yttrium-90 radioembolization for hepatocellular carcinoma: comparison of diffusion-weighted functional MR imaging with anatomic MR imaging.

PURPOSE: Anatomic magnetic resonance (MR) imaging assessment of hepatocellular carcinoma (HCC) response to yttrium-90 ((90)Y) radioembolization may require 3 months before therapeutic effectiveness can be determined. The relationship between anatomic MR and diffusion-weighted imaging (DWI) changes after (90)Y therapy is unclear. The present study tested the hypothesis that apparent diffusion coefficient (ADC) values on DWI at 1 month precede anatomic tumor size change at 3 months after (90)Y radioembolization. MATERIALS AND METHODS: In this prospective study, 20 patients with HCC (16 men) enrolled between April 2005 and July 2006 underwent lobar (90)Y therapy with mean doses of 141 Gy (right lobe) and 98 Gy (left lobe). Anatomic 1.5-T MR imaging (gadolinium-enhanced T1-weighted gradient-recalled echo) and DWI (single-shot spin-echo echo-planar imaging; b value of 0, 500 sec/mm(2)) were performed at baseline (0-3 weeks before (90)Y therapy) and at 1 and 3 months after (90)Y therapy. Tumor size and ADC values were measured and compared, and the percentage change in ADC was compared to the change in tumor size (minimum >5% change in size), with use of a paired t test (alpha = .05). RESULTS: Yttrium-90 therapy was successfully delivered in all patients. The mean baseline ADC of 1.64 x 10(-3) mm(2)/sec +/- 0.30 significantly increased to 1.81 x 10(-3) mm(2)/sec +/- 0.37 at 1 month (P = .02), and to 1.82 x 10(-3) mm(2)/sec +/- 0.23 at 3 months (P = .02). The mean baseline tumor size of 83.0 cm(2) +/- 63.7 did not change statistically at 1 month (84.1 cm(2) +/- 62.1; P = .75) or 3 months (74.0 cm(2) +/- 57.0; P = .10). The overall mean ADC percentage change at 1 month of 10.5% +/- 23.1% preceded an overall mean tumor size percentage change at 3 months of -18.5% +/- 31.5% (P = .03). CONCLUSIONS: HCC tumor response assessed with DWI at 1 month preceded anatomic size changes at 3 months after (90)Y therapy. DWI may assist in early determination of the response or failure of (90)Y therapy for HCC.

90Y Radioembolization for metastatic neuroendocrine liver tumors: preliminary results from a multi-institutional experience.

PURPOSE: Minimally invasive therapies such as transarterial chemoembolization and radiofrequency ablation are used for hepatic metastatic neuroendocrine tumor (NET) therapy. Results from another minimally invasive therapy, radioembolization, remain unknown. The purpose of this multicenter open label phase II study was to assess the efficacy and safety of yttrium-90 (Y) radioembolization for treating hepatic metastatic NET using a primary outcome of tumor response and secondary outcomes of serologic toxicities and survival. MATERIAL/METHODS: In this multicenter study, all patients underwent lobar radioembolization using glass or resin Y radioembolic agents. Patients were assessed serologically and radiographically at 2 to 4 weeks and then at 1 to 3 month intervals after treatment. We 1) compared liver volumes, radiation doses, and serologic liver function tests (unpaired t test, P = 0.05) and 2) assessed tumor response, serologic toxicity, and median survival from first Y therapy. The clinicaltrials.gov identifier was NCT00532740. RESULTS: Forty-two patients underwent radioembolization using glass (mean age 58 +/- 12 years) or resin (mean age 61 +/- 11 years) microspheres. A statistically significant greater median radiation dose was delivered to each lobe using glass (right lobe 117 Gy; left lobe 108 Gy) than using resin (right 50.8 Gy; left 44.5 Gy) (P < 0.01). Using Response Criteria in Solid Tumors, 92% of glass and 94% of resin patients were classified as partial response or stable disease at 6 months after treatment. Six patients experienced grade 3/4 toxicities during the follow-up period. Median survival was 22 months (glass) and 28 months (resin) (P = 0.82). CONCLUSION: Y radioembolization of metastatic NET is a viable therapy with acceptable toxicity. Further investigation is warranted.

Use of a combined MR imaging and interventional radiology suite for intraprocedural monitoring of uterine artery embolization.

PURPOSE: The optimal embolic endpoint for uterine artery embolization (UAE) is unknown. It is difficult to quantify substasis endpoints with the use of conventional x-ray interventional radiology (IR) guidance. Although magnetic resonance (MR) imaging can detect perfusion changes, intraprocedural perfusion changes within targeted uterine leiomyomas and the remaining uterine wall remain unknown. A hybrid MR/IR unit was used to test the hypothesis that MR imaging can detect changes in uterine perfusion-dependent signal enhancement immediately after UAE. MATERIALS AND METHODS: In this prospective study, UAE was performed in a hybrid MR/IR unit in women with symptomatic uterine leiomyomas. This MR/IR unit contains a wide-bore 1.5-T MR scanner connected by a sliding table to an adjacent x-ray digital subtraction angiography unit. Gadolinium-enhanced MR imaging was performed before and after UAE. Relative signal-to-noise ratio (SNR) was measured within each tumor and the adjacent uterine wall, and mean relative SNR changes were compared before and after UAE with the paired t test (alpha=0.05). RESULTS: UAE was technically successful in all six women, in whom 10 tumors were assessed (seven intramural, two submucosal, one subserosal). Mean relative SNR of the tumors before UAE was 62.2+/-25.0 and was reduced to 41.1+/-17.7 after UAE (P<.01). Mean relative SNR of the adjacent uterine wall was 64.2+/-14.3 before UAE and decreased to 28.8+/-14.9 after UAE (P<.01). CONCLUSIONS: Immediate reductions in perfusion-dependent enhancement in targeted uterine leiomyomas and the adjacent uterine wall can be detected during UAE with the use of a hybrid MR/IR unit. Further studies are now warranted to compare long-term clinical outcomes versus immediate changes in perfusion at the time of UAE.

Radiation dose limits and liver toxicities resulting from multiple yttrium-90 radioembolization treatments for hepatocellular carcinoma.

PURPOSE: To assess the relationship between cumulative hepatic lobar radiation dose and liver toxicities in patients with hepatocellular carcinoma (HCC) treated with multiple sessions of yttrium-90 radioembolization. MATERIALS AND METHODS: Forty-one patients with HCC (age range, 46-82 years) underwent radioembolization with 90Y. Patients were classified according to the Okuda scoring system. All patients received single liver lobar treatments on two or more occasions according to standard clinical 90Y embolization protocol. Cumulative radiation dose to each liver lobe was measured and patients were followed to assess liver toxicities. Statistical analysis was performed with the Student t test and Kaplan-Meier analysis. RESULTS: Patients with Okuda stage I disease received more treatments than those with Okuda stage II disease (mean, 2.65 vs 2.24; P<.05). For average cumulative radiation dose, patients with Okuda stage I disease received 247 Gy (range, 88-482 Gy) and those with Okuda stage II disease received 198 Gy (range, 51-361 Gy; P<.05). A total of 13 toxicities occurred in seven patients (16%). Patients with Okuda stage I disease were given a greater cumulative dose than patients with Okuda stage II disease before worsening of liver function: 390 Gy versus 196 Gy (P<.005). For patients with Okuda stage I disease, a higher cumulative radiation dose was associated with occurrence of one or more toxicities: 222 Gy (no toxicities) versus 390 Gy (>or=1 toxicity; P<.005). No correlation between cumulative radiation dose and liver toxicities existed in patients with Okuda stage II disease. The maximum tolerated dose was between 222 and 390 Gy. Median survival times were 660 and 431 days for patients with Okuda stage I and stage II disease, respectively. CONCLUSIONS: Patients with HCC can tolerate high cumulative radiation doses with 90Y therapy. Compared with patients with Okuda stage II disease, patients with Okuda stage I disease tolerate a higher cumulative radiation dose without liver toxicity, but liver toxicities increase with increasing cumulative radiation doses.

MR imaging assessment of changes in renal function with renal artery stent placement in swine.

PURPOSE: To prospectively test the hypothesis that magnetic resonance (MR) imaging can detect changes in renal function at the time of renal artery stent placement in a swine model of renal artery stenosis (RAS). MATERIALS AND METHODS: In this animal care and use committee-approved study, hemodynamically significant (>50%) RAS was surgically induced in six pigs. MR imaging was employed for assessment of the anatomic and physiologic changes induced by fluoroscopically guided stent placement. With MR imaging, we assessed changes in renal blood flow (RBF), extraction fraction (EF), and single-kidney glomerular filtration rate (skGFR) during the procedure. Arterial diameter stenosis before and after stent placement was assessed with x-ray digital subtraction angiography (DSA). Mean changes in functional and anatomic parameters were compared with the Wilcoxon matched-pairs test, with an alpha level of 0.05. RESULTS: There was no significant change in mean RBF after stent deployment (P=.44). Mean EF increased from 0.19+/-0.08 before stent placement to 0.31+/-0.17 after stent placement (P=.16). Mean skGFR measurements were 25 mL/min+/-16 before stent placement and 41 mL/min+/-28 after stent placement (P<.05). According to x-ray DSA measurements, mean stenosis measurements were 60%+/-12% before stent placement and 24%+/-16% after stent placement (P<.02). CONCLUSIONS: In swine, MR imaging can detect immediate changes in renal function after radiographically guided stent placement for unilateral RAS. This functional MR technique may have applications in the setting of hybrid MR/x-ray DSA procedure suites.

Liver tumors: monitoring embolization in rabbits with VX2 tumors--transcatheter intraarterial first-pass perfusion MR imaging.

PURPOSE: To prospectively test the hypothesis that transcatheter intraarterial first-pass perfusion (TRIP) magnetic resonance (MR) imaging can depict serial reductions in rabbit liver tumor perfusion during transcatheter arterial embolization (TAE). MATERIALS AND METHODS: All experiments had institutional animal care and use committee approval. In four rabbits implanted with eight VX2 liver tumors, catheters were positioned in the hepatic arteries with conventional angiographic guidance. After transfer to the MR imaging suite, serial TAE was performed, with approximately 0.5 million 40-120-microm embolic particles injected at each embolic stage. TRIP MR imaging was performed at baseline and after each subsequent embolic stage (10 minutes between stages). Serial TAE and TRIP MR imaging were repeated until stasis. The first-pass time course of signal enhancement was measured in both tumors and hepatic arteries. Tumor area under the curve (AUC) and maximum upslope (MUS) values, each normalized by arterial input, were measured to assess iterative perfusion reduction. Perfusion measurements across TAE stages were compared with paired t tests and linear regression. RESULTS: AUC decreased from a pre-TAE baseline of 0.408 (95% confidence interval [CI]: 0.330, 0.486) to 0.065 (95% CI: 0.046, 0.085) (P<.001) after TAE. MUS decreased from a pre-TAE baseline of 0.151 (95% CI: 0.121, 0.181) to 0.027 (95% CI: 0.022, 0.031) (P<.001) after TAE. Reductions to AUC and MUS after each embolic stage were statistically significant (P<.006 for each group of paired comparisons). AUC strongly correlated with MUS (r=0.966, P<.001). CONCLUSION: TRIP MR imaging can depict serial reductions in liver tumor perfusion during TAE. TRIP MR imaging offers the potential to target functional embolic end points during TAE.

90Y radioembolization of metastatic breast cancer to the liver: toxicity, imaging response, survival.

PURPOSE: To present data from patients with breast cancer liver metastases who underwent radioembolization with yttrium (90Y) microspheres. MATERIALS AND METHODS: Using standard 90Y lobar treatment protocol, 27 female patients with progressing liver metastases on standard of care polychemotherapy were treated under an open-label phase 2 protocol. After treatment, we assessed (a) tumor response using computed tomography and/or positron emission tomography, (b) biochemical toxicity, and (c) survival. RESULTS: The mean age of the patients was 52. Seventeen (63%) patients received 20 left lobe treatments (median radiation dose, 123 Gy; mean, 119 Gy), and 20 (74%) patients received 22 right lobe treatments (median radiation dose, 121 Gy; mean, 109 Gy) to the treatment site. No significant dose-difference was noted between the two lobes (P=.69). Tumor response on 90-day follow-up computed tomography showed (a) complete and partial response in nine (39.1%) patients, (b) stable disease in 12 (52.1%) patients, and (c) progressive disease in 2 (8.8%) patients. Positive tumor response on positron emission tomography was noted in 17 (63%) patients. Three of 27 (11%) patients (Eastern Cooperation Oncology Group 1, 2, or 3) showed bilirubin toxicity of grade 3, all of which were attributed to disease progression. Median survival for Eastern Cooperation Oncology Group 0 versus 1, 2, or 3 patients was 6.8 months and 2.6 months, respectively (P=.24) and for patients with tumor burden<25% versus >25% was 9.4 and 2.0 months, respectively (P=.46). CONCLUSIONS: Radioembolization with 90Y brachytherapy device may be a viable therapeutic option for the treatment of breast cancer liver metastases in patients who have progressed or failed on standard of care polychemotherapy.

Effect of transcatheter arterial embolization on levels of hypoxia-inducible factor-1alpha in rabbit VX2 liver tumors.

PURPOSE: To test the hypothesis that transcatheter arterial embolization (TAE) of VX2 rabbit liver tumors increases the expression of hypoxia-inducible factor-1alpha (HIF-1alpha), a transcription factor that regulates the expression of pro-angiogenic genes. MATERIALS AND METHODS: VX2 tumors were implanted in the livers of eight New Zealand white rabbits. Once tumor growth was seen at T2-weighted turbo spin-echo magnetic resonance (MR) imaging, four of the eight rabbits underwent TAE with 45-150-mum polyvinyl alcohol particles. The remaining four rabbits served as non-TAE controls. The TAE end point was stasis of antegrade blood flow. All rabbits were sacrificed for tumor harvest 2 hours after TAE. Tumor tissue and corresponding normal liver tissue in each rabbit liver were stained with anti-human HIF-1alpha monoclonal antibody and reviewed with light microscopy. Percentages of stained viable tumor and normal liver cells were compared by using the Mann-Whitney U test (alpha=0.05). RESULTS: In eight rabbits with 24 discrete liver tumors, the mean percentage (+/-standard deviation) of positive HIF-1alpha-stained cells in the TAE group was greater than that in the control group (19%+/-7.0 vs 12%+/-8.0, respectively) (P=.05). Normal liver tissue in both the TAE and control groups showed no HIF-1alpha staining. CONCLUSION: Although HIF-1alpha is not expressed in normal rabbit liver parenchyma-even after TAE-HIF-1alpha expression is present in implanted VX2 rabbit liver tumors and significantly increased in lesions that have undergone embolization.

Renal artery stenosis in swine: feasibility of MR assessment of renal function during percutaneous transluminal angioplasty.

PURPOSE: To prospectively test--in a swine model of renal artery stenosis (RAS)--the hypothesis that magnetic resonance (MR) imaging can reveal changes in renal function at the time of percutaneous transluminal angioplasty (PTA). MATERIALS AND METHODS: In this animal care and use committee-approved study, high-grade unilateral RAS was surgically induced in six pigs. MR imaging at 3.0 T was used for intraprocedural assessment of the anatomic and physiologic changes induced by x-ray-guided PTA. With use of MR imaging, changes in single-kidney glomerular filtration rate, extraction fraction, and renal blood flow were assessed during PTA. The arterial diameter of stenosis before and after PTA was assessed by using conventional digital subtraction angiography. Mean changes in functional and anatomic parameters were compared by using the Wilcoxon signed rank test (alpha = .05). RESULTS: At digital subtraction angiography, the mean percentage of stenosis was 69% +/- 10 (standard deviation) before PTA and 26% +/- 10 after PTA (P<.03). Mean pre- and post-PTA extraction fraction values were 0.11 +/- 0.03 and 0.19 +/- 0.06, respectively (P<.03). The mean single-kidney glomerular filtration rate before PTA, 19 mL/min +/- 13, increased to 41 mL/min +/- 33 after PTA (P<.03). There was no significant change in mean renal blood flow after PTA (P=.44). CONCLUSION: In swine, MR imaging can reveal changes in renal function after x-ray-guided PTA for unilateral RAS.

Rabbit VX2 tumors as an animal model of uterine fibroids and for uterine artery embolization.

PURPOSE: To determine the suitability of the rabbit VX2 tumor animal model for uterine fibroids and uterine artery embolization (UAE). MATERIALS AND METHODS: The authors implanted and grew one uterine VX2 tumor per rabbit in six rabbits. UAE was performed by using 100-300 microm embolic particles and confirmed with x-ray digital subtraction angiography, magnetic resonance (MR) imaging, and necropsy. Unenhanced and contrast medium-enhanced MR images of VX2 tumors were obtained before and after UAE. Relative MR signal-to noise-ratio (SNR) was measured in the uterine VX2 tumor and in normal uterine tissue before and after UAE and compared by using a paired t-test (P = .05). RESULTS: VX2 uterine tumors were successfully grown, and both VX2 tumor presence in the uterus and UAE were seen angiographically and confirmed with necropsy in all six rabbits. Statistically significant reductions in relative SNRs were measured in tumors (SNR before UAE, 15.3 +/- 5.15; SNR after UAE, 3.84 +/- 3.94; P < .0001). No statistically significant decrease in SNR was measured in normal uterine tissue before and after UAE (P = .63 for the right uterine horn and P = .93 for the left uterine horn). CONCLUSION: Rabbit VX2 uterine tumors may be a suitable animal model of uterine fibroids and UAE.

Comparison of intraarterial MR angiography at 3.0 T with X-ray digital subtraction angiography for detection of renal artery stenosis in swine.

PURPOSE: To compare the accuracy of catheter-directed intraarterial (IA) magnetic resonance (MR) angiography at 3.0 T with that of x-ray digital subtraction angiography (DSA) for the measurement of renal artery stenosis (RAS) in swine. MATERIALS AND METHODS: Unilateral hemodynamically significant RAS (>50%) was induced surgically in six pigs with use of reverse cable ties. One to two weeks after surgery, each pig underwent x-ray DSA and MR angiography before and after percutaneous transluminal balloon angioplasty (PTA). X-ray DSA was performed before and after PTA of RAS by injection of iodinated contrast agent through a 5-F multiple-side hole angiographic catheter placed in the abdominal aorta under fluoroscopic guidance. MR angiography of RAS was performed before and after PTA of RAS on a 3.0-T clinical MR imager with use of gadolinium-based contrast agent. MR angiography and DSA images were analyzed with the full width at half maximum method. Percent stenosis measurements between x-ray DSA and MR angiography were compared with a paired t test and were correlated with linear regression and Bland Altman analysis (alpha = 0.05). RESULTS: Six cases of RAS were induced and imaged successfully with DSA and MR angiography techniques before and after PTA. On x-ray DSA, median stenoses was 64% (95% CI 57%-80%) before PTA and 20% (95% CI 5%-32%) after PTA. Corresponding MR angiography median stenosis measurement was 69% (95% CI 58%-80%) before PTA and 26% (95% CI 16%-36%) after PTA. A paired t test comparison did not show a difference between DSA and MR angiography (P = .16). RAS measurements on MR angiography correlated closely (P < .01) with DSA measurements (r(2) = 0.92). CONCLUSION: In swine, the accuracy of catheter-directed IA MR angiography with use of a clinical 3.0-T MR imaging unit for the measurement of RAS was similar to that of conventional x-ray DSA.

Diffusion-weighted MR imaging for determination of hepatocellular carcinoma response to yttrium-90 radioembolization.

Early detection of the response of hepatocellular carcinoma (HCC) to yttrium-90 radioembolization therapy may be important to permit repeat radioembolization or alternative treatment options. Water-mobility measurements with use of diffusion-weighted (DW) magnetic resonance (MR) imaging are useful for noninvasive interrogation of microstructural tissue properties. Findings of DW MR imaging may serve as an early biomarker of HCC response. This study tested the hypothesis that DW MR imaging can detect changes in tumor tissue water diffusion in response to (90)Y therapy. In each of six patients with HCC included in the study, tumor water diffusion increased significantly after therapy. DW MR imaging is a promising technique for noninvasive assessment of tumor response to (90)Y radioembolization.

Comparison between intravenous and intraarterial contrast injections for dynamic 3D MRI of liver tumors in the VX2 rabbit model.

PURPOSE: To test the hypothesis that catheter-directed intraarterial (IA) contrast agent injection increases tumor enhancement and conspicuity compared to intravenous (IV) injection. MATERIALS AND METHODS: Eight VX2 liver tumors were grown in five rabbits. After positioning a catheter in the hepatic artery, we performed 3D inversion recovery GRE MRI after IA and IV gadopentetate-dimeglumine contrast injections at doses of 0.04 and 0.1 mmol/kg, respectively. Peak enhancement (signal-to-noise ratio (SNR)) and conspicuity (contrast-to-noise ratio (CNR)) were measured for each acquisition. RESULTS: The peak SNR and CNR were 21.7 +/- 5.8 and 17.0 +/- 4.8 (mean +/- SD) after IA injection, and 16.9 +/- 10.2 and 6.2 +/- 2.6 after IV injection. The IA CNR was significantly greater than the IV CNR (P < 0.05), with a >60% increase in CNR for each tumor. For six of the eight tumors the IA SNR was greater than the IV SNR, but statistical significance was not achieved due to the small sample size of the study (P = 0.07). CONCLUSION: We demonstrated the feasibility of using IA injection techniques to improve tumor conspicuity. This strategy could be employed to enhance the detection of small liver tumors or to conserve contrast agent in future MRI-guided transcatheter liver therapies.

In vivo diffusion-weighted imaging of liver tumor necrosis in the VX2 rabbit model at 1.5 Tesla.

OBJECTIVES: We sought to demonstrate the feasibility of using single-shot spin-echo echo-planar imaging for imaging liver tumor necrosis in the in vivo VX2 rabbit model at 1.5 T. MATERIALS AND METHODS: VX2 liver tumors were grown in 4 rabbits. Diffusion-weighted images (DWIs) were acquired during breath-hold using twice refocused SE-EPI (b = 0, 700, 1400 seconds/mm). Anatomic images for tumor size measurements were acquired using T2W TSE. Rabbits were euthanized for subsequent necropsy. Viable and necrotic tumor tissue ADC measurements were performed with reference to hematoxylin and eosin pathology. RESULTS: A total of 8 tumors were grown with diameters ranging from 1.2 to 5.3 cm. Viable and necrotic tumor compartments were clearly differentiated. Apparent diffusion coefficient in necrotic tumor cores, 1.26 +/- 0.11 x 10 mm/s, were significantly greater than those in surrounding viable tumor tissues, 0.74 +/- 0.06 x 10 mm/s (mean +/- SD, P < 0.05). CONCLUSIONS: In vivo DWI of liver tumor necrosis in the VX2 rabbit model is feasible using a 1.5 T clinical magnetic resonance imaging scanner. DWI may permit longitudinal assessment of liver tumor therapies in both preclinical and clinical studies.

Feasibility of blood oxygenation level-dependent MR imaging to monitor hepatic transcatheter arterial embolization in rabbits.

Blood oxygenation level-dependent (BOLD) magnetic resonance (MR) imaging is a noninvasive method to assess changes in oxygen delivery to tissues. It was hypothesized that BOLD MR imaging can detect changes in rabbit VX2 liver tumor oxygenation after hepatic artery embolization with polyvinyl alcohol particles. In four discrete VX2 liver tumors, a statistically significant mean reduction in apparent transverse relaxation time was shown, from 55 milliseconds before embolization to 41 milliseconds after embolization (P < .01). This reduction corresponded to a decrease in hepatic tumor oxygenation. The use of BOLD MR imaging to monitor changes in hepatic tumor oxygenation after embolization is feasible. These functional MR imaging measurements of hypoxia may be targeted as an endpoint for therapy in future studies.

Real-time MRI monitoring of transcatheter hepatic artery contrast agent delivery in rabbits.

RATIONALE AND OBJECTIVES: We sought to test the hypothesis that transcatheter hepatic artery delivery of dilute gadolinium (Gd) in rabbits can be monitored in real-time using magnetic resonance imaging (MRI). MATERIALS AND METHODS: Catheters (2F) were inserted via a femoral access into the hepatic arteries of six New Zealand White rabbits under radiographic guidance. After transfer to a 1.5-T MRI scanner, 26 separate hepatic artery injections of 2 mL of 4% Gd and 14 sham injections were performed. Real-time imaging of all injections was acquired using two-dimensional projection inversion recovery-gradient echo. Films of these 40 injections, as well as 10 random repeats, were independently reviewed in a randomized, blinded fashion by two Certificate of Added Qualification-certified interventional radiologists. Observers reported (i) if Gd injection occurred and (ii) if so, the location of delivery. For each observer, we compared sensitivity/specificity for real-time visualization of contrast injection and accuracy of injection localization. Interobserver and intraobserver variability was assessed using the kappa statistic. X-ray digital subtraction angiography was the gold standard for all MRI studies. RESULTS: Both observers had a sensitivity of 100% and a specificity of 93%. Accuracy for intrahepatic contrast delivery was 77% for both observers. Accuracy for extrahepatic delivery was 92% and 96%, respectively. Both interobserver and intraobserver agreement was outstanding. CONCLUSIONS: In rabbits, MRI allows for accurate real-time monitoring of transcatheter hepatic artery delivery of contrast agent. Localization accuracy is higher outside the liver than within the liver. These results can be used as a baseline reference for comparing the accuracy of delivery of Gd-tagged therapies in the future.

The effect of catheter-directed CT angiography on yttrium-90 radioembolization treatment of hepatocellular carcinoma.

PURPOSE: Yttrium 90 radioembolization is a transcatheter therapy for unresectable hepatocellular carcinoma (HCC) that delivers internal radiation to tumors. In contrast to the usual method of lobar regional delivery, catheter-directed computed tomographic (CT) angiography was investigated as a potentially useful technique to evaluate the administration of segmental 90Y tumor radiation doses superselectively without significantly altering liver function or Child-Pugh classification. MATERIALS AND METHODS: Fourteen patients underwent 90Y therapy for unresectable HCC. After standard angiographic placement of a 3-F microcatheter in a segmental hepatic artery supplying the tumor, each patient underwent CT angiography with use of segmental hepatic artery injection of iodinated contrast agent to confirm segmental perfusion and delineate segmental liver volume. 90Y was later injected into the same segmental artery. Target dose was calculated according to infused 90Y activity and targeted hepatic volume with standard lobar volume (before CT angiography) versus segmental liver volume (after CT angiography). The Wilcoxon signed-rank test (alpha = 0.05) was used to compare the estimated 90Y dose before CT angiography with the actual 90Y dose after CT angiography, as well as changes in serum bilirubin level and Child-Pugh classification as a result of treatment. RESULTS: The mean estimated tumor dose before CT angiography (SD) was 100 Gy +/- 43 (range, 35-169 Gy). The mean actual tumor dose after CT angiography was 348 Gy +/- 204 (range, 105-857 Gy), which was significantly greater (P < .001). The mean bilirubin level before treatment was 1.0 mg/dL +/- 0.97 (range, 0.2-4.0 mg/dL), whereas the mean bilirubin level after treatment was 1.3 mg/dL +/- 0.85 (range, 0.5-3.8 mg/dL). This difference, although statistically significant (P = .03), was not clinically important. Thirteen of 14 patients had no change in Child-Pugh class. CONCLUSION: CT angiography can be used to delineate the blood supply and volume to a targeted hepatic segment, allowing superselective 90Y radioembolization. This approach significantly increases effective 90Y tumor radiation dose without clinically altering liver function or Child-Pugh class.