Abstract
Planning hepatic 90Y radioembolization activity requires balancing toxicity with efficacy. We developed a dual-tracer SPECT fusion imaging protocol that merges data on radioactivity distribution with physiologic liver mapping. Methods: Twenty-five patients with colorectal carcinoma and bilobar liver metastases received whole-liver radioembolization with resin microspheres prescribed as per convention (mean administered activity, 1.69 GBq). As part of standard treatment planning, all patients underwent SPECT imaging after intraarterial injection of 37 MBq of 99mTc-macroaggregated albumin (99mTc-MAA) to simulate subsequent 90Y distribution. Immediately afterward, patients received 185 MBq of labeled sulfur colloid (99mTc-SC) intravenously as a biomarker for normal hepatic reticuloendothelial function and SPECT was repeated. The SPECT images were coregistered and fused. A region-based method was used to predict the 90Y radiation absorbed dose to functional liver tissue (DFL) by calculation of 99mTc-MAA activity in regions with 99mTc-SC uptake. Similarly, the absorbed dose to tumor (DT) was predicted by calculation of 99mTc-MAA activity in voxels without 99mTc-SC uptake. Laboratory data and radiographic response were measured for 3 mo, and the survival of patients was recorded. SPECT-based DT and DFL were correlated with parameters of toxicity and efficacy. Results: Toxicity, as measured by increase in serum liver enzymes, correlated significantly with SPECT-based calculation of DFL at all time points (P < 0.05) (mean DFL, 27.9 Gy). Broad biochemical toxicity (>50% increase in all liver enzymes) occurred at a DFL of 24.5 Gy and above. In addition, in uni- and multivariate analysis, SPECT-based calculation of DT (mean DT, 44.2 Gy) correlated with radiographic response (P < 0.001), decrease in serum carcinoembryonic antigen (P < 0.05), and overall survival (P < 0.01). The cutoff value of DT for prediction of 1-y survival was 55 Gy (area under the receiver-operating-characteristic curve = 0.86; P < 0.01). Patients who received a DT of more than 55 Gy had a median survival of 32.8 mo, compared with 7.2 mo in patients who received less (P < 0.05). Conclusion: Dual-tracer 99mTc-MAA–99mTc-SC fusion SPECT offers a physiology-based imaging tool with significant prognostic power that may lead to improved personalized activity planning.
Yttrium-90 radioembolization is a rapidly emerging radionuclide treatment modality for hepatic malignancy that improves progression-free and overall survival in appropriate patients (1–4). Activity planning aims to maximize the effect of treatment while keeping toxicity acceptably low. However, the current activity calculation methods are based on empiric data with regard to both efficacy and safety, without established dose–response relationships (5). Resin microsphere activity is calculated on the basis of body surface area (BSA) and fractional liver involvement, whereas glass microsphere activity is calculated on the basis of a whole-liver partition model derived from the MIRD equations for dose calculation (6). Neither accounts for the heterogeneous intrahepatic microsphere distribution and the resultant differential radiation absorbed dose in tumors and normal liver tissue.
A potential approach to better customization of activity is anatomic partition modeling (7–9). The involved liver may be segmented into 2 compartments, and the desired activity is calculated on the basis of intercompartmental volumes and activity ratios. Because compartmental boundaries are drawn on the basis of anatomic images, these methods are time-consuming and subject to considerable error in tumor delineation. Partition modeling in its present capacity is therefore advised only for patients with a limited number of large, hypervascular tumors (10). In most patients with liver metastases, however, disease is diffusely distributed throughout the liver, includes multiple tumors that are difficult to delineate, and therefore is not amenable to anatomic partition modeling (5,10).
We introduce a physiology-based segmentation tool that uses a dual-tracer SPECT technique combining 99mTc-macroaggregated albumin (99mTc-MAA) SPECT for simulation of 90Y activity distribution and 99mTc-sulfur colloid (99mTc-SC) SPECT for evaluation of functional liver parenchyma (reticuloendothelial function). We aimed to validate the utility of this method by correlating its results with efficacy and toxicity in patients treated with whole-liver radioembolization.
MATERIALS AND METHODS
Patient Selection
The study included 25 patients from January 2007 to January 2011 who had multiple unresectable liver metastases from colorectal carcinoma and received whole-liver treatment with resin microspheres in a single session. There were 10 men and 15 women, of mean and median age 58 y (range, 25–80 y). All patients had liver-dominant disease for which hepatic metastases were considered to be the most longevity-threatening component. Exclusion criteria included mismatch between 99mTc-MAA and subsequent 90Y-microsphere injection site, imaging failure (SPECT not performed, injection failure), and staged treatment in 2 separate sessions. Clinical data of the studied population are summarized in Table 1. All patients underwent combined 99mTc-MAA–99mTc-SC SPECT imaging as part of their diagnostic work-up. Data were handled in accordance with the Health Insurance Portability and Accountability Act. The institutional review board approved this study, and the requirement to obtain informed consent was waived.
Demographics, Baseline Characteristics, and Oncologic Histories of Cohort
Radioembolization
Activity calculations and treatments were performed according to international consensus guidelines (11–13). All patients were treated with resin microspheres (SIR-Spheres; SirTex Inc.). The prescribed activity was calculated on the basis of BSA and tumor liver involvement (LI), defined as a fraction of the total liver parenchyma (14):
Eq. 1.
Significant hepatopulmonary shunting was compensated for by the recommended activity adjustment (13).
Pretreatment V-Vial (Wheaton Industries, Inc.) activity and posttreatment V-Vial, tubing, and catheter activity were measured in a leak-proof Nalgene (Thermo Fisher Scientific) container using a MicroRem meter (Thermo Scientific/Bicron) at a set standard geometry. Measurements were processed by calibrated conversion algorithms (SirTex Inc.) to calculate the percentage of residual activity. The hepatic administered activity (A) was determined by correcting the prepared activity for residual activity and for the fractional lung shunt (LS).
Eq. 2
Clinical and laboratory follow-up was performed 2, 4, 8, and 12 wk after treatment and at intervals prescribed by the medical oncologist thereafter. Follow-up contrast-enhanced CT at 3 mo was used to analyze objective response according to Response Evaluation Criteria in Solid Tumors (RECIST 1.1). PET scans were not generally available for review. One masked author performed over-reads of all clinical film interpretations.
Imaging Procedures
At the time of the preparatory angiography session and after endovascular skeletonization of the hepatic artery, 99mTc-MAA (37 MBq) was administered intraarterially to simulate the planned treatment with 90Y-microspheres. Activity lower than the standard 150 MBq was used to facilitate combined 99mTc-MAA–99mTc-SC SPECT imaging.
Whole-body planar scintigraphy and hepatic SPECT imaging were performed within 1 h after tracer administration to calculate the lung shunt fraction, to exclude any extrahepatic deposition, and to depict intrahepatic radiopharmaceutical distribution. Immediately after the first SPECT scan, without moving the patient in the scanner to optimize image fusion, an excess of 99mTc-SC (185 MBq) was injected intravenously and SPECT imaging was repeated after a 5-min delay.
SPECT data were acquired on a dual-head Infinia Hawkeye 4 γ camera (GE Healthcare). SPECT images were acquired on a 64 × 64 matrix (voxel size, 0.884 mm3) using a 130- to 150-keV energy window and a low-energy high-resolution collimator. All SPECT data were acquired in 120 projections (15 s per projection) over a 360° full circular orbit.
Image Processing and Analysis
Data were reconstructed by applying filtered backprojection and a Butterworth postreconstruction filter (Fc, 0.23; order, 6), using Segami software (Segami Corp.). For each patient, corrected 99mTc-SC images were constructed by subtracting 99mTc-MAA images (first SPECT) from the combined 99mTc-MAA–99mTc-SC images (second SPECT). Voxels positive for 99mTc-SC were assigned to the functional-liver compartment, whereas voxels negative for 99mTc-SC but positive for 99mTc-MAA were assigned to the tumor compartment. Next, a map of functional liver was produced by applying a threshold that identified all voxels with 10% or more of the maximum 99mTc-SC uptake per voxel, using software programmed in IDL 6.1 (Research Systems, Inc.). Similarly, a tumor map was generated from the 99mTc-MAA images by applying the 10% threshold to the 99mTc-MAA images and excluding all voxels with 99mTc-SC uptake (Figs. 1 and 2). On the basis of these maps, tumor and functional-liver regions of interest (ROIT and ROIFL) were defined as all voxels negative for 99mTc-SC but positive for 99mTc-MAA (ROIT) and as all voxels positive for 99mTc-SC (ROIFL). Partition was based solely on these physiologic parameters with no need for anatomic delineation and its inherent error. For comparison, alternative thresholds of 5% and 20% were also evaluated.
Schematic representation of physiologic partition model of liver. (A and B) Maximum-intensity projection of 99mTc-MAA SPECT (A) and 99mTc-SC SPECT MIP (B), showing concentrated uptake of 99mTc-MAA in and around 2 tumors and relative photopenia in these areas on 99mTc-SC image. (C) Tumor map contains only those voxels positive for 99mTc-MAA uptake and negative for 99mTc-SC uptake. (D) Functional liver map contains all voxels positive for 99mTc-SC uptake. Because of spill-over and partial-volume effects, threshold function was used to define voxels as positive for 99mTc-MAA or 99mTc-SC uptake.
Physiologic partition model of liver. (A and B) Transaxial slices (thick slab) of 99mTc-MAA SPECT (A) and 99mTc-SC SPECT (B), showing concentrated uptake of 99mTc-MAA in and around physiologic tumor tissue in right lobe and relative photopenia in these areas on 99mTc-SC image. (C and D) Coronal reformats of 99mTc-MAA SPECT (C) and 99mTc-SC SPECT (D) confirm this yin-yang phenomenon.
ROI volume and 99mTc-MAA activity in each ROI were translated to a radiation absorbed dose using A (GBq), the ROI volume in liters, and an estimated density of all hepatic tissue of 1.029 kg/L (based on the specific density of soft tissue in general). The mean radiation absorbed dose to tumor (DT) was thus calculated, derived from the MIRD equations for dose calculation (6):
Eq. 3where MAA activity ROIT is the 99mTc-MAA activity in ROIT (i.e., tumor tissue as defined by 99mTc-MAA uptake threshold), total MAA activity is the administered 99mTc-MAA activity corrected for the lung shunt, A is the administered 90Y activity in GBq, 1.029−1 is the density conversion factor for hepatic tissue, volume ROIT is the volume of ROIT in liters, and 50 is the conversion factor for 90Y from GBq/kg to Gy derived from MIRD (6).
Similarly, the mean DFL was calculated as
Eq. 4where MAA activity ROIFL is the 99mTc-MAA activity in ROIFL (i.e., normal liver as defined by 99mTc-SC uptake threshold), total MAA activity is the administered 99mTc-MAA activity corrected for the lung shunt, and volume ROIFL is the volume of ROIFL in liters.
For comparison, the mean whole-liver radiation absorbed dose (DWL) was calculated using conventional MIRD formulae, assuming homogeneous distribution and absorption of all the administered activity and energy in the liver:
Eq. 5where volumeWL is the volume of the whole liver in liters, calculated from preprocedural CT scans.
Statistical Analysis
A commercial software package was used for statistical analysis (SPSS for Windows, version 19.0; SPSS Inc.). All continuous variables were tested for normal distribution probability using Kolmogorov–Smirnov tests, including normality plots. Median and range were reported for non–normally distributed variables, mean and range for normally distributed variables. For individual correlation of 2 continuous variables, the Pearson or Spearman correlation coefficient was used, depending on normality. Survival analysis was performed using Kaplan–Meier curves with the log rank test for comparison. Receiver-operating-characteristic analysis was used to determine cutoff values for clinical use of the presented parameters. A P value of less than 0.05 was considered statistically significant.
RESULTS
The patients selected were heavily pretreated, with almost all (92%) having received systemic chemotherapy and bevacizumab (Table 1). Half the patients (52%) also received epidermal growth factor receptor antagonists, and almost half (48%) had undergone liver-directed treatments including resection, ablation, embolization, and external-beam radiotherapy. All patients maintained a good performance status (Eastern Cooperative Oncology Group scores of 0–1) and baseline laboratory values within acceptable ranges. All patients had multiple metastases in both liver lobes, with a median estimated tumor involvement of the liver of 25% (range, 5%–60%). All were treated with resin microspheres with whole-liver treatment in a single session. Mean A was 1.69 GBq (range, 0.97–2.33 GBq), and mean DWL was 48.8 Gy (range, 29.8–68.5 Gy). Only 3 patients restarted systemic treatment within our 3-mo follow-up period after radioembolization (irinotecan plus bevacizumab after 2 mo, cetuximab after 2 mo, and oxaliplatin plus 5-fluorouracil after 1.5 wk). At least 4 other patients restarted some form of systemic treatment beyond 3 mo.
With a 10% threshold, where any voxel with at least 10% of the maximum 99mTc-MAA or 99mTc-SC activity was considered positive for that marker, the mean DFL was 27.9 Gy (range, 11.9–41.6). The changes in liver function tests (serum bilirubin, aspartate aminotransferase [AST], alanine transaminase [ALT], alkaline phosphatase, and albumin) and blood counts (white blood cell count, hemoglobin, and platelets) during the first 3 mo after radioembolization are depicted in Figure 3. Significant correlations were found between DFL and changes in serum liver enzymes (AST, ALT, and alkaline phosphatase) at virtually all time points (r = 0.38–0.69; P < 0.01). Only weaker trends were found between DFL and blood counts, bilirubin, and albumin. Comparing laboratory values with DWL instead of DFL yielded a similar but weaker pattern of correlations, with only a few reaching statistical significance (AST in weeks 4–8 and ALT in week 2–8).
Hematologic and serum chemistry changes from baseline and 2, 4, 8, and 12 wk after radioembolization, expressed as mean percentage change. WBC = white blood cell count; bilirubin = total serum bilirubin; alk phos = alkaline phosphatase.
According to the National Cancer Institute’s Common Terminology Criteria for Adverse Events, version 4.02, these changes in laboratory values were classifiable as mild toxicity in most patients (Table 2). A significant correlation was found between the cumulative toxicity grades and DFL (r = 0.48; P < 0.01) (Fig. 4) and DWL (r = 0.44; P < 0.05). Toxicity grades were most pronounced for the liver enzymes AST, ALT, and alkaline phosphatase. Total bilirubin levels increased but reached thresholds of toxicity in only a few patients, whereas minor decreases in albumin levels were classified as toxicity in most patients because of low pretreatment levels already near the toxicity threshold (Table 2; Fig. 3). Broad biochemical toxicity, defined as a 50% increase in each of the 3 measured liver enzymes, occurred in 9 of 25 patients (36%). Receiver-operating-curve analysis revealed a good predictive value of DFL to predict this composite toxicity, with an area under the curve of 0.88 (P < 0.01; 95% confidence interval, 0.66–0.99). All cases of a 50% increase in all 3 liver enzymes occurred at a DFL of 24.5 Gy and above.
Toxicity after Radioembolization According to National Cancer Institute’s Common Terminology Criteria for Adverse Events, Version 4.02
DFL in grays on x-axis vs. cumulative toxicity in grades according to National Cancer Institute’s Common Terminology Criteria for Adverse Events, version 4.02. r = 0.48 (P < 0.01).
Using the 10% threshold, the mean DT was 44.2 Gy (range, 11.3–105.7 Gy). Response rate according to RECIST 1.1 was only 26.7%, with no complete responders (stable disease, 60.0%; disease control, 86.7%; progressive disease, 13.3%). Responders received a DT (mean ± SD) of 82.7 ± 23.9 Gy, versus 31.0 ± 10.9 Gy for nonresponders (P < 0.001). Responders versus nonresponders received 1.76 ± 0.26 GBq and 1.68 ± 0.43 GBq, respectively (P = 0.74), and a DWL of 49.1 ± 6.3 Gy and 52.8 ± 11.0 Gy, respectively (P = 0.54), indicating that only differences in intrahepatic distribution of the dose and not the total liver dose impacted efficacy. Median liver tumor involvement was 25% for both the responders and the nonresponders (P = 0.98).
CEA levels showed a median decrease of 28.8% (range, −84.2 to +329.5%) at week 2, 50% (range, −94.7 to +329.5%) at week 4, and 31.3% (range, −94.7 to +430.3%) at week 8. These changes correlated with DT at week 4 (r = −0.45; P < 0.05) and at week 8 (r = −0.43; P < 0.05). Correlation was found for DWL only at week 8 (r = −0.53; P < 0.05).
Median follow-up after radioembolization was 40.8 mo (range, 20–67.7 mo). At the time of writing, 5 patients were still alive and 1 patient had been lost to follow-up after 9.7 mo. Median overall survival was 10.8 mo. DT proved to be the only predictor of survival (P < 0.01). No other correlations were found between survival and clinical, laboratory, or procedural parameters, including DWL (P = 0.23), previous systemic (P = 0.20) and liver-directed treatments (P = 0.57), liver tumor involvement (P = 0.79), extrahepatic disease (P = 0.87), and performance status (P = 0.70).
The 1- and 2-y survival rates were 45.8% and 25.0%, respectively. Receiver-operating-curve analysis of DT versus 1- and 2-y survival led to an area under the curve of 0.86 (P < 0.01; 95% confidence interval, 0.71–1.0) and 0.82 (P < 0.05; 95% confidence interval, 0.58–1.0), respectively (Fig. 5). The 1-y survival for patients who received a DT of more than 55 Gy was 100%, whereas the 1-y survival for patients who received less than 55 Gy was 24%. Similarly, the 2-y survival for patients who received a DT of more than 77 Gy was 100%, whereas the 2-y survival for patients who received less than 77 Gy was 10%. The median survival of patients who received a DT of more than 55 Gy was 32.8 mo, whereas patients who received less than 55 Gy had a median survival of only 7.2 mo (P < 0.05) (Fig. 6).
Receiver-operator-characteristic curve for prediction of 1-y survival after radioembolization by DT. At tumor dose of at least 55 Gy, all patients survived at least 1 y. Area under curve = 0.86 (P < 0.01; 95% confidence interval, 0.71–1.0).
Kaplan–Meier curves for patients treated with DT of more than 55 Gy (upper dashed line) vs. less than 55 Gy (lower dashed line), with median survivals of 32.8 mo vs. 7.2 mo (P < 0.05). Survival of total population is depicted as solid line (median, 10.8 mo).
The impact of applying different thresholds to define positive 99mTc-MAA and 99mTc-SC uptake was studied, using 5%, 10%, and 20% of the maximum activity per voxel as thresholds (Table 3). A higher threshold resulted in higher calculated DT and DFL (P < 0.01). This was the result of smaller ROIs and higher activity per voxel caused by the higher thresholds. This increase was more pronounced for DT. The ratio of tumor uptake (DT) to functional liver uptake (DFL) showed an increase from a mean of 1.2 at a 5% threshold to 1.8 at a 10% threshold to 4.2 at a 20% threshold. The predictive power for toxicity, response, and survival was not significantly influenced by different thresholding. Data and statistics were thus reported at the 10% level to balance prognostic power for efficacy and for toxicity.
Influence of Threshold on Dosimetry Parameters
DISCUSSION
Activity planning for radioembolization remains inexact and unscientific, likely contributing to the rate of nonresponse, which can be up to 80% (1), and to hepatotoxicity, which can occur in up to 20% of patients (15,16). The liver, despite its qualities of dual blood supply, regeneration, and redundancy, is highly radiosensitive. Absorbed doses high enough to be tumoricidal usually exceed the maximum tolerated absorbed dose for the background liver, limiting the application of radiation therapy to hepatic malignancies (17). The inhomogeneous distribution of intraarterially administered radioembolization microspheres can overcome this limitation by delivering the highest absorbed doses to hypervascular, arterially supplied tumors, while limiting deposition in the portal vein–supplied functional liver. However, a reproducible method to predict or to measure the actual absorbed dose to tumor and to functional liver remains elusive.
By the earliest activity planning method, or empiric method, patients were treated with a predetermined activity that reflected only tumor liver involvement. Tumor involvement that was less than 25%, 25%–50%, or more than 50% of the total liver volume was treated with 2, 2.5, or 3 GBq, respectively (18). It was eventually recognized that this method led to an unacceptable incidence of hepatotoxicity due to radioembolization-induced liver disease (14). The empiric method was supplanted by the current BSA method, which has become adopted as the standard for resin microspheres. This change led to a lower incidence of radioembolization-induced liver disease, not so much because of more accurate dosimetry as because administered activity was significantly reduced (15). The BSA method was proposed without scientific derivation and, although relatively safe, results in potentially subtherapeutic doses in patients with enlarged livers (19).
The standard activity planning method for glass microspheres is a MIRD-based calculation that considers the weight of the treated tissue and a target absorbed dose averaged over the entire treatment volume, sometimes erroneously termed the partition method. This method, as currently practiced, uses a nonpartitioned 1-compartment model of uniform distribution to perform the calculations (target, 80–120 Gy) but assumes a large degree of nonuniform distribution to maintain a background functional liver dose of less than 30 Gy and, thus, tolerable hepatotoxicity. The actual degree of nonuniformity is not typically quantified.
Several groups have proposed improved partition method dosimetry based on quantitative 99mTc-MAA SPECT/CT (7–9). A study on 36 patients with 58 lesions of hepatocellular carcinoma treated with glass microspheres found that quantitative 99mTc-MAA SPECT/CT had a predictive value for response to radioembolization and survival (8). The accuracy in predicting the response of an individual lesion at 3 mo was 91%. Three false-positives were encountered in cases involving large, heterogeneous, partially necrotic lesions. This modified partition method has clear advantages over existing methods with regard to tumor dosimetry but has several important limitations: normal-liver tissue dosimetry and toxicity are not addressed; morphologic imaging–based tumor segmentation is limited by tumor number, heterogeneity, necrosis, and infiltrative spread; and the method is operator-dependent and labor- and time-intensive (19). This partition method is most useful in patients with few and sharply delineated tumors. Unfortunately, such patients are only a small subpopulation of those treated by radioembolization.
Our method of using combined 99mTc-MAA–99mTc-SC SPECT imaging offers a physiologic imaging tool that overcomes the limitations of morphologic imaging–based partition modeling. Absorbed dose to both tumor and functional liver is addressed, yielding good prognostic power with regard to both efficacy and toxicity. Tissue segmentation is based purely on physiology, relying on 2 standard nuclear medicine procedures, post–intraarterial 99mTc-MAA scintigraphy and post–intravenous 99mTc-SC scintigraphy (20), obviating manual morphology-based tumor segmentation. Multiplicity, size, distribution, heterogeneity, necrosis, and infiltrative growth of tumors are inherently accommodated. Our method is fully automated and fast and can be operator-independent.
Several limitations of our method remain and are the subject of further refinement. One limitation is that the uptake of 99mTc-SC identifies regions of intact reticuloendothelial function, which may or may not coincide perfectly with regions of intact hepatocellular function. The method currently averages all voxels in the ROIs and does not calculate the absorbed dose for individual lesions, which can vary. The calculated absorbed dose is an estimation influenced by partial-volume and spillover effects. Adjustment of the quantification thresholds can partially overcome this problem but can result in large differences in calculated DT across the range of potential thresholds. It remains unproven which threshold correlates best with anatomic tumor delineation, tumor volumes, and radiation absorbed doses. However, such correlations may be unnecessary, since our physiology-based method proved to have a robust predictive power for outcomes regardless of which threshold was chosen. Further improvements are to be expected by introducing dose-point kernel reconstruction algorithms, iterative reconstruction methods, dual-isotope protocols for optimized quantification and coregistration, and CT-based attenuation correction (using SPECT/CT).
The presented feasibility study included only a homogeneous group of 25 heavily pretreated patients with metastatic colorectal cancer, all of whom received whole-liver salvage radioembolization in a single session. However, interpatient differences still exist, and the benefits of improved dosimetry (and possible confounding factors such as treatment history and disease burden) need to be translated to larger cohorts. These preliminary results will also need to be translated to prospective clinical practice. Even though β irradiation has a biologic coefficient of 1.0 (equivalent to x-rays and γ-rays), the effect of prolonged exposure on a decay curve may be different from that of fractionated doses as given by external-beam radiation, and a DFL of 30 Gy may or may not be a valid limit. A prospective dose-escalation study may be conducted to establish safety limits for DFL and to further refine patient inclusion criteria and optimization of dose planning.
CONCLUSION
Fusion 99mTc-MAA–99mTc-SC SPECT imaging offers a true physiology-based imaging tool for outcome prediction after hepatic radioembolization. It offers a robust, reproducible, automated, and fast method, which demonstrates prognostic value and dose–response relationships for both toxicity and efficacy.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. Daniel Sze is on the medical or scientific advisory boards for Surefire Medical, Inc., Treus Medical, Inc., RadGuard Medical, Inc., and Jennerex Biotherapeutics, Inc.; is on the speaker’s bureau for W.L. Gore, Inc.; and has provided clinical trial consultation for Sirtex, Inc., Nordion, Inc., and Biocompatibles, Inc. No other potential conflict of interest relevant to this article was reported.
Footnotes
Published online Oct. 21, 2013.
- © 2013 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- Received for publication March 16, 2013.
- Accepted for publication June 26, 2013.
Jump to section
Related Articles
Cited By...
- Efficacy of Radioembolization with 166Ho-Microspheres in Salvage Patients with Liver Metastases: A Phase 2 Study
- Reply: High Impact of Preferential Flow on 99mTc-MAA and 90Y-Loaded Microsphere Uptake Correlation
- High Impact of Preferential Flow on 99mTc-MAA and 90Y-Loaded Microsphere Uptake Correlation
- Insights into the Dose-Response Relationship of Radioembolization with Resin 90Y-Microspheres: A Prospective Cohort Study in Patients with Colorectal Cancer Liver Metastases
- Systematic Evaluation of Tumoral 99mTc-MAA Uptake Using SPECT and SPECT/CT in 502 Patients Before 90Y Radioembolization
- Reply: 99mTc-MAA-Based Dosimetry for Liver Cancer Treated Using 90Y-Loaded Microspheres: Known Proof of Effectiveness
- 99mTc-MAA-Based Dosimetry for Liver Cancer Treated Using 90Y-Loaded Microspheres: Known Proof of Effectiveness