| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Investigations |
Medicity Research Laboratory and Radiopharmaceutical Chemistry Laboratory, Turku PET Centre; Department of Oncology and Radiotherapy; and Department of Otorhinolaryngology and Head and Neck Surgery, Turku University Central Hospital, Turku, Finland
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Key Words: [18F]fluoroerythronitroimidazole • hypoxia • PET • head and neck cancer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
None of the methods to measure the hypoxic fraction in human tumors is currently used routinely (5). Recent reviews (5,6) have presented histomorphometric methods, a polarographic microelectrode technique, cytologic analysis of DNA strand breaks, and hypoxia markers that can be detected by external imaging either with magnetic spectroscopy or with radionuclide labeling. Of these, a commercially available microelectrode system (Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany) has been designated as the gold standard. In this system, a thin-needle electrode is inserted into the tumor and gives signals that are proportional to the oxygen partial pressure in the tissue. Disadvantages include difficulties in distinguishing between measurements made in necrotic and viable regions and poor access to deep-seated tumor sites such as primaries in the esophagus, larynx, and lungs (7).
Recently, great interest has arisen in developing radiolabeled markers of the hypoxic fraction for noninvasive imaging of solid tumors (6). The advantage of PET is the ability to image tumor metabolism quantitatively in any part of the body. For example, the primary tumor and the lymph node metastases can be imaged simultaneously in head and neck cancer and in lung cancer (8,9) with PET and [18F]FDG, a glucose analog commonly used in oncologic PET studies. In vitro experiments show that hypoxia increases uptake of [3H]FDG in cell lines established from patients with melanoma, ovarian carcinoma, and head and neck carcinoma because glucose is the preferred fuel for energy metabolism under low-oxygen concentrations (10,11). The use of [18F]FDG for detection of hypoxia is, however, impeded by many other factors—such as cellular viability and growth—that contribute to [18F]FDG uptake in tumors.
Hypoxic cells can be specifically imaged with PET using radiolabeled nitroimidazole compounds (12–14), of which 18F-labeled fluoromisonidazole ([18F]FMISO) has been introduced in the clinical setting (12,15,16). So far, [18F]FMISO has been used to study patients with head and neck, lung, prostate, and brain tumors. Although clearly having potential as a probe for detection of hypoxia, [18F]FMISO may have a low target-to-background ratio because of the lipophilicity that causes neurotoxicity in its parent drug when used as a radiosensitizer in pharmacologic doses (17). 18F-labeled fluoroerythronitroimidazole (FETNIM), originally developed by Yang et al. (13), was found to be more hydrophilic than [18F]FMISO and is eliminated rapidly from well-oxygenated tissues, resulting in higher tumor-to-liver and tumor-to-blood ratios in rat mammary cancer than are obtained with [18F]FMISO. These original observations prompted us to study the uptake and metabolism of [18F]FETNIM in tumor-bearing mammals (18) and, finally, in cancer patients. Our goal was to study the time course of [18F]FETNIM uptake in relation to blood flow and glucose metabolism in head and neck cancer and to validate a noninvasive approach to the assessment of hypoxic retention of [18F]FETNIM, which appears to show predominantly reversible binding in tumors in the first 2–3 h after injection.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
[18F]FDG.
[18F]FDG was synthesized with a computer-controlled apparatus by a previously described nucleophilic substitution procedure (23). The radiochemical purity of [18F]FDG always exceeded 98%, and the specific radioactivity at the end of radiosynthesis was >40 GBq/µmol.
[18F]FETNIM.
[18F]FETNIM was synthesized from 1-(2'-nitro-1'-imidazolyl)-2,3-O-isopropylidene-4-tosyloxybutane by nucleophilic displacement of the tosyloxy-group with [18F]F-. The method of radiochemical synthesis was a modification of the method of Yang et al. (13) and was reported separately (18). The radiochemical purity of [18F]FETNIM exceeded 95%, and the specific radioactivity exceeded 300 GBq/µmol.
PET Acquisition
PET was performed using an Advance scanner (General Electric Medical Systems, Milwaukee, WI) operated in 2-dimensional mode. The scanner consists of 18 rings of bismuth germanate detectors yielding 35 transverse slices spaced by 4.25 mm. The imaging field of view is 55 cm in diameter and 15.2 cm in axial length (24).
All measurements were corrected for scatter, random counts, and dead time (25). Attenuation correction was performed with 2 rotating rod sources containing 68Ge/68Ge (total, 350 million counts). Image reconstruction for the central 30-cm field of view was performed using a transaxial Hanning filter with a cutoff frequency of 4.6 mm to a 128 x 128 image matrix, which resulted in image pixels of 2.34 x 2.34 mm.
PET Imaging
All PET imaging was performed while the patients were fasting. Before [18F]FETNIM studies, patients were advised to fast at least 4 h, and before [18F]FDG studies, preferably overnight. The plasma glucose level was monitored during the beginning of each study. Seven patients had a normal plasma glucose concentration during both [18F]FDG PET and [18F]FETNIM PET, with a median of 5.3 mmol/L (range, 4.75–6.5 mmol/L). By contrast, 1 patient (patient 3) was found to be hyperglycemic both before the [18F]FDG study (plasma glucose level, 8.8 mmol/L) and before the [18F]FETNIM study (plasma glucose level, 9.8 mmol/L). A line was placed into the antecubital vein for injection of the 18F-labeled tracers and the [15O]H2O. In [18F]FETNIM studies, a catheter was placed in the contralateral radial artery for frequent blood sampling. The tumor was located by inspection; palpation; anatomic landmarks; and, if necessary, CT.
PET studies with [18F]FETNIM and [18F]FDG were performed within 20 d of each other (median, 5 d; range, 1–20 d) in random order during the planning phase of radiation treatment. In [18F]FDG studies, the tracer was injected as a 15-s bolus (median dose, 368 MBq; dose range, 355–379 MBq). A static emission scan consisting of three 5-min frames was acquired 45–60 min after injection and was followed by a 10-min transmission scan. [18F]FETNIM studies were performed sequentially with blood-flow and blood-volume measurements, which were obtained using [15O]H2O and [15O]CO, respectively. Laser-guided landmarks and felt pen markings were used, if feasible, so that the positioning of the patient for the second PET study could be the same as for the first.
The sequence of PET acquisition for [18F]FETNIM and blood-flow and blood-volume studies is shown in Figure 1. The study started with a 10-min transmission scan. Thereafter, perfusion was measured with [15O]H2O (median dose, 1,152 MBq; dose range, 1,014–1,800 MBq). Arterial blood was withdrawn with a pump (Alitea, Stockholm, Sweden) at a speed of 6 mL/min to obtain the input function. Blood radioactivity concentration was measured using a 2-channel online detector (General Electric Medical Systems, Uppsala, Sweden). The procedure was recently depicted in more detail (26). The blood volume was measured using inhalation of [15O]CO (0.14%) mixed with room air through a 3-way inhalation flap-valve (median dose, 2,699 MBq; dose range, 2,496–2,992 MBq), as previously described (27). During the scan, 3 arterial samples were taken at 4, 6, and 8 min from the beginning of the inhalation to measure the radioactivity concentration in the blood. An additional blood sample was withdrawn to measure plasma radioactivity to verify that the [15O]CO was bound to red cells and had not transformed into CO2 (27). When the blood volume study was finished, a slow (15 s) bolus of [18F]FETNIM was injected (median dose, 366 MBq; range, 289–385 MBq). During [18F]FETNIM image acquisition, 29 blood samples were withdrawn through the arterial line to measure the time course of the tracer concentration and metabolites in plasma. [18F]FETNIM data were collected continuously from injection until 120 min. The frame rate for the first scan was 4 x 30 s, 3 x 60 s, 5 x 180 s, 8 x 300 s, and 6 x 600 s. Afterward, the patient was allowed to rest for 40 min before the second scan, which was obtained between 160 and 180 min and at a frame rate of 3 x 400 s (Fig. 1).
|
Definition of ROIs in Blood-Flow, Blood-Volume, and [18F]FETNIM Studies.
Whole-tumor ROIs from the 3 consecutive planes with the highest radioactivity as determined by the [18F]FDG study were carefully matched with the corresponding planes on the flow images. Anatomic landmarks seen on transmission scans, such as vocal cords, mandible, cervical vertebrae, nasal cavities, shoulders, and the tumors themselves, were used to confirm the matching of planes between the [18F]FDG and [15O]H2O images. The size and shape of ROIs were not allowed to change during transfer, and this requirement was met more easily by the tumors’ invariably showing an increased radioactivity signal on flow images. Large elliptic ROIs (size, 11.5–13.5 cm2) were defined on dorsal neck muscles in 3 adjacent planes of the transmission scan. The selected muscle ROIs were remote from known tumors and large neck vessels to avoid contamination of radioactivity in muscles by spillover from tissues with high radioactivity. The same whole-tumor ROIs were then transferred to both the corresponding blood-volume image planes and the corresponding [18F]FETNIM image planes. Appropriate locations for tumor ROIs on [18F]FETNIM images were chosen by correlation with anatomic landmarks because of the possibility of patient movement during scanning.
Calculation of Blood Flow.
The methods of measuring blood flow with [15O]H2O are based on the principle of inert gas exchange between blood and tissues (30). The autoradiographic method was chosen to calculate the parametric blood-flow images using a 250-s integration time and an arterial input curve (31). In the autoradiographic method, the tissue curve and integral are calculated from the measured decay-corrected arterial time–radioactivity curve for different flow levels, assuming a constant value of 1.0 for the partition coefficient of water in muscle. The integral values and respective blood-flow values are written to a lookup table. The integral of the measured decay-corrected tissue time–radioactivity curve is calculated for each image pixel, and a corresponding blood-flow value is retrieved from the table and written to the flow image. The measured arterial input curve suffers from dispersion in the sampling system, and there is also a delay between the measured input curve and the real entrance of the tracer into the tissue (26,32). The input function for the tissue was corrected for the measured external dispersion in the tubing (2.5 s) by deconvolution with an exponential function (32). The delay was solved by fitting the input curve into the measured tissue curve (32).
Calculation of Blood Volume.
The average blood volume of the tumor was calculated using the equation BV = (average radioactivity in tumor ROI/average radioactivity in blood) x 1/r, where BV is the average blood volume of the tumor and r is the tissue–to–large-vessel hematocrit ratio. A value of 0.91 was adopted from previous studies (26).
Dynamic [18F]FETNIM Imaging.
SUVs for [18F]FETNIM were measured from the whole-tumor and whole-muscle ROIs. Furthermore, a 3 x 3 pixel ROI with the maximum radioactivity within the previously determined whole-tumor ROIs was used for steady-state and kinetic analyses. The time course of tumor-to-muscle retention ratios was then calculated over different time frames using both maximum and whole-tumor ROIs.
Nitroimidazoles are reduced intracellularly in all viable cells; however, retention occurs only in the tissues with a low oxygen concentration, because reoxidation of the original compound is slowed. This retention leads to the formation of reactive and charged products, which may bind covalently to molecules under hypoxic conditions or diffuse more slowly from the tissue (33). Because the graphic analysis of [18F]FETNIM time–radioactivity data showed no irreversible trapping of tracer in tumor or other tissues during the PET study, the graphic analysis for reversible tracers (34,35) was applied. In this model-independent analysis, the plot of the radioactivity ratio of tissue integral to tissue versus the radioactivity ratio of plasma integral to tissue is linear at equilibrium. The slope of the linear phase of the plot equals the distribution volume (DV) of the tracer (35). In our studies, the plot became linear about 20 min after the tracer injection, and the slope was determined from the data measured after 20 min. For a coarse estimate of the DV during 2 periods, the ratio of tissue and plasma radioactivity curves was calculated at the end of the 2 dynamic acquisition phases (90–120 and 160–180 min after injection). Finally, a pixel-by-pixel analysis was applied to study the relationship between blood flow and [18F]FETNIM uptake within the tumor 90 min after the injection.
Statistical Analysis.
Pearson product moment correlation was used to analyze linear correlations between the variables. The results are expressed using medians, minimums, and maximums.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Assessment of Glucose Metabolism, Perfusion, and Blood Volume in Tumor and Muscle
Pertinent findings for the PET studies are given in Table 2. The PET images and corresponding CT scans of 2 patients are shown in Figure 2. In all cases, the primary tumor was clearly visualized on the blood-flow and [18F]FDG scans, and there was good contrast in perfusion and glucose metabolism between tumor and muscle tissues. Because of high tracer uptake in tumors on [18F]FDG scans, we could define the metabolically active tumor volume using an SUV of 4 as a threshold (29). The median metabolically active tumor volume was 15 mL (volume range, 2–400 mL; Table 1). For patients 2 and 8, however, the true volume may have been underestimated by this approach because of partial-volume effects. The tumor-to-muscle blood-flow ratio was invariably high, with a median of 9.1 and a range of 5.4–28.3.
|
|
Dynamic Uptake of [18F]FETNIM
In early phases (i.e., 5–8 min after injection), the pattern of [18F]FETNIM uptake closely resembled perfusion images. On late images, [18F]FETNIM activity in tumor was more equal to that of background mucosal surfaces and neck muscles. Tumor uptake was difficult to distinguish in 3 patients (38%) (Fig. 2A), whereas the remaining 5 patients showed focally increased and visually detectable uptake in at least some parts of the tumor.
Dynamic uptake of [18F]FETNIM by arterial blood, muscle, and tumor in a representative patient (patient 5) is depicted in Figure 3. The plasma concentration of 18F radioactivity peaked immediately with injection and then rapidly redistributed. We could show radiochromatographically that no major metabolites were in the plasma (18).18F activity remained lower in muscle tissue than in plasma during the whole acquisition for all patients. Whole-tumor 18F activity was always higher than muscle 18F activity but was comparable with plasma activity until 120 min for 4 patients and until 180 min for 2 patients and was lower than plasma activity for the remaining patients (Table 2). Maximum tumor activity was higher than plasma activity for 4 patients and close to plasma activity for 4 patients. In the beginning of the study the tumor curve peaked slightly, whereas the muscle curve rose slowly throughout the study to the final level. Maximum tumor-to-muscle SUV ratios at different times are expressed in Figure 4.
|
|
A strong correlation was found between the DV and [18F]FETNIM uptake as calculated using SUV between 60 and 120 min. The highest correlation was found at 90 min (r2 = 0.935; P < 0.0001; Fig. 5). The correlation did not improve at 180 min (r2 = 0.524; P = 0.066).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
In our study, another nitroimidazole compound, [18F]FETNIM, was used to image patients with head and neck cancer before the start of radiotherapy. [18F]FETNIM may theoretically be a more potent indicator of hypoxia than is [18F]FMISO because of greater hydrophilicity and more favorable pharmacokinetics (13,18). On the basis of these 2 studies, both [18F]FMISO and [18F]FETNIM show late hypoxia-specific accumulation at 120–240 min in experimental mammary tumors. However, [18F]FETNIM had not been evaluated in cancer patients previously. Therefore, we undertook a long, dynamic study to learn the optimal time window for assessment of [18F]FETNIM uptake in tumors and surrounding normal tissues. On the basis of this knowledge, we aim to develop a clinically feasible imaging protocol for future PET studies to evaluate hypoxia-specific binding of [18F]FETNIM in cancer patients.
In addition to the [18F]FETNIM study, concurrent perfusion and blood volume imaging and, on a separate day, [18F]FDG PET were performed on all patients. We studied tumor blood flow using [15O]H2O to assess the interrelationship between flow and [18F]FETNIM uptake, and with the aid of [18F]FDG, the metabolically active tumor volume could be defined and the tumor could be more precisely localized for image correlation with other studies.
We found that blood volume in tumors did not significantly differ from that in muscles and the mucosal surfaces of adjacent neck areas. Neither did tumor blood volume correlate with metabolic characteristics depicted by [18F]FDG and [18F]FETNIM imaging. By contrast, tumor blood flow as assessed with a bolus injection of [15O]H2O was high. Most patients had flow rates between 30 and 65 mL/100 g/min, whereas the flow rate in neck muscle was typically <5 mL/100 g/min. Similarly, high rates of blood flow have been reported for cervical and breast tumors using [15O]H2O autoradiography and a dynamic technique, respectively (38,39). The high apparent flow may be caused by high cellularity and a relatively high water-to-lipid content ratio in tumors (39). In fact, parts of tumors may still remain severely hypoxic because of a low oxygen extraction ratio, as studies on cerebral gliomas have suggested (40).
In tumors with a high regional [18F]FETNIM uptake, the average blood flow tended to be high. This relationship was evident in comparison with not only the early phase of [18F]FETNIM uptake but also, to a somewhat lesser extent, the late retention phase. Our observation has several explanations. First, because [18F]FETNIM is hydrophilic, it is no surprise that initial uptake in tissue is governed by blood flow. Second, in the presence of hypoxia, angiogenesis is enhanced and may result in high average blood flow on the macroscopic level. High blood flow, in turn, is associated with neovascularization, which may foretell an even greater hypoxic fraction of the tumor in regions with high apparent flow but low oxygen extraction. To better understand the relationship, we performed a pixel-by-pixel comparison of whole-tumor radioactivity between the corresponding perfusion and [18F]FETNIM images obtained 90 min after injection. The lack of relationship between regional flow and distribution of 18F activity within tumors suggests that flow and retention of [18F]FETNIM may not be coupled. Furthermore, PET is not sensitive in detecting heterogeneity at the microscopic level because autoradiographic images of experimental tumors usually show an inverse relationship between flow and nitroimidazole labeling. In addition, the relative short imaging time—180 min—may not allow sufficiently high covalent binding of [18F]FETNIM to cellular macromolecules (6).
For the kinetic analysis of [18F]FETNIM uptake, we applied a graphic analysis for reversible tracers (34,35). The graphic analysis for irreversible tracers showed slopes around zero, implying that the irreversible compartment for labeled [18F]FETNIM metabolites is negligible in the studied tissue types (41). The graphic analysis for reversible tracers has been developed and until now applied in the brain receptor studies, but it is not dependent on a particular model (35). Indeed, it provided excellent linear fits for tumor and muscle data in the current study. The resulting DV is assumed to be relatively independent of blood flow. However, at low rates of blood flow, a longer equilibration time may delay the linear phase of the plot. The lower the oxygen concentration in tissue, the longer the [18F]FETNIM metabolites stay in tissue, leading to higher DV and, thus, a positive image of hypoxia.
The calculated DVs correlated significantly with tumor-to-muscle SUV ratios obtained between 90 and 120 min after injection, suggesting that the SUV approach may be used as a surrogate for kinetic analysis in patients for whom blood sampling is not desirable. However, SUV ratios tended to increase over time in some patients (Fig. 4), and the optimal time for steady-state analysis of [18F]FETNIM uptake is not easy to determine. Recognizing the limitations of the SUV calculations, we consider it too early to recommend this index for general use in [18F]FETNIM PET studies (28). Until more confirmatory data are available, estimation of [18F]FETNIM DVs may be safer with the Logan analysis or with measurement of tissue-to-plasma ratios in the late accumulation phase, which requires only a few blood samples (Table 2).
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
For correspondence or reprints contact: Heikki Minn, MD, PhD, Turku PET Centre, Turku University Central Hospital, P.O. Box 52, FIN-20521 Turku, Finland.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. S. Boss, R. V. Olmos, M. Sinaasappel, J. H. Beijnen, and J. H. M. Schellens Application of PET/CT in the Development of Novel Anticancer Drugs Oncologist, January 1, 2008; 13(1): 25 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. E. Stewart, X. Chen, J. Hadway, and T.-Y. Lee Correlation between Hepatic Tumor Blood Flow and Glucose Utilization in a Rabbit Liver Tumor Model Radiology, June 1, 2006; 239(3): 740 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-M. Eschmann, F. Paulsen, M. Reimold, H. Dittmann, S. Welz, G. Reischl, H.-J. Machulla, and R. Bares Prognostic Impact of Hypoxia Imaging with 18F-Misonidazole PET in Non-Small Cell Lung Cancer and Head and Neck Cancer Before Radiotherapy J. Nucl. Med., February 1, 2005; 46(2): 253 - 260. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Piert, H.-J. Machulla, M. Picchio, G. Reischl, S. Ziegler, P. Kumar, H.-J. Wester, R. Beck, A. J.B. McEwan, L. I. Wiebe, et al. Hypoxia-Specific Tumor Imaging with 18F-Fluoroazomycin Arabinoside J. Nucl. Med., January 1, 2005; 46(1): 106 - 113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Wells, T. Jones, and P. Price Assessment of Inter- and Intrapatient Variability in C15O2 Positron Emission Tomography Measurements of Blood Flow in Patients with Intra-abdominal Cancers Clin. Cancer Res., December 15, 2003; 9(17): 6350 - 6356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G R Laking and P M Price Positron emission tomographic imaging of angiogenesis and vascular function Br. J. Radiol., December 1, 2003; 76(suppl_1): S50 - S59. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G J R Cook Nuclear medicine in imaging head and neck malignancies Imaging, September 1, 2003; 15(3): 141 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Tolvanen, K. Lehtio, J. Kulmala, V. Oikonen, O. Eskola, J. Bergman, and H. Minn 18F-Fluoroerythronitroimidazole Radiation Dosimetry in Cancer Studies J. Nucl. Med., December 1, 2002; 43(12): 1674 - 1680. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Chapman, P. Zanzonico, and C. C. Ling On Measuring Hypoxia in Individual Tumors with Radiolabeled Agents J. Nucl. Med., November 1, 2001; 42(11): 1653 - 1655. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| JOURNAL OF NUCLEAR MEDICINE TECHNOLOGY | THE JOURNAL OF NUCLEAR MEDICINE |
