| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Radiation Oncology |
1 From the Departments of Radiotherapy and Radio-oncology (A.d.V., J.G., P.L.), Magnetic Resonance (C.K., W.J., T.K.), Biostatistics (K.P.P.), and Radiodiagnostic 1 (W.B.) and the Institute of Histology and Embryology (P.D.), Leopold-Franzens-Universität Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria; and the GSF Research Center for Environment and Health, Institute for Radiation Biology, Neuherberg, Germany (T.G.). From the 1998 RSNA scientific assembly. Received November 29, 1999; revision requested January 28, 2000; revision received March 20; accepted April 4. Supported in part by grants from Schering Germany and Schering Austria. Address correspondence to A.d.V. (e-mail: alexander.devries@uibk.ac.at).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Perfusion data in 11 patients with cT3 (clinical staging, tumor invaded the perirectal tissue) rectal carcinoma who underwent preoperative chemoirradiation were analyzed. Perfusion data were acquired by using a T1 mapping sequence with a whole-body magnetic resonance (MR) imager. After contrast medium was intravenously infused at a constant rate, concentration-and-time curves were evaluated for arterial blood and tumor. All patients underwent MR imaging before and at constant intervals during chemoirradiation. Clinical stages before therapy were compared with surgical stages after therapy.
RESULTS: Spatial and temporal resolution on dynamic T1 maps were sufficient to reveal changes in contrast medium accumulation in the tumor. Comparison of PI values and radiation dose showed a significant increase in the 1st (P = .003) and 2nd weeks (P = .01) of treatment; values subsequently returned to pretreatment levels or showed a renewed increase. High initial PI values correlated with greater lymph node downstaging (P = .042).
CONCLUSION: Dynamic T1 mapping provides a suitable tool for monitoring tumor microcirculation during chemoirradiation and offers the potential for individual optimization of therapeutic procedures. Furthermore, these results indicate that the PI map may serve as a prognostic factor.
Index terms: Magnetic resonance (MR), contrast enhancement, 757.12143 • Magnetic resonance (MR), inversion recovery, 757.121413 • Magnetic resonance (MR), perfusion study • Pelvic organs, MR, 757.121411, 757.121413, 757.12143 • Pelvic organs, therapeutic radiology, 757.1269 • Rectum, neoplasms, 757.321 • Rectum, therapeutic radiology, 757.1266, 757.1269
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
At present, the intervals for temporal administration of additional chemotherapy are empiric based (1,2,7,8). In a mouse model, Kallman (9) and Kallman et al (10) demonstrated a clear dependence of the effect of combined chemoirradiation on the timing of chemotherapy. Ideally, the planning of chemotherapy should encompass the monitoring of any vascular changes in tumor tissue during fractionated radiation therapy. Thus, a step toward individualized tumor therapy is to quantitatively assess the parameters, such as spatially differentiated measurement of perfusion, capillary permeability, and extracellular volume, that characterize the microcirculation of a tumor. For practical clinical application, it is essential to measure these parameters before and during therapy with the use of minimally invasive or noninvasive methods.
Contrast medium–enhanced dynamic magnetic resonance (MR) imaging has been used in a few studies (11–17) to evaluate tumor microcirculation by using a standard MR imaging contrast medium such as gadopentetate dimeglumine. Feldmann et al (14) and Hawighorst et al (16) evaluated blood supply at the beginning of treatment (radiation therapy alone or with hyperthermia) in primary tumors with a wide range of entities, stages, and recurrence. Mayr et al (17) measured blood perfusion in cervical carcinoma at different stages and correlated the MR imaging data with radiation therapy outcome. All of these studies lacked measurements obtained at periodic intervals covering the entire course of treatment, especially during the initial stages of therapy.
To our knowledge, there is no clinical study in which microcirculatory changes in a specific tumor entity were monitored during the entire course of therapy. In our study, we used a noninvasive MR imaging method to evaluate the perfusion data obtained during the entire course of chemoirradiation in patients with primary rectal carcinomas who were preoperatively treated. We also compared the perfusion data with therapy outcome (N or T downstaging [improvement in N or T stages]), as evaluated after surgery.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
Each patient received combined chemoirradiation with a total radiation dose of between 38.3 and 44.0 Gy before surgery; single doses of 1.1 Gy were administered twice daily. Also, 350 mg/m2 5-fluorouracil was continuously administered through an implanted Port-A-Cath Deltec CADD-1 system (SIMS Deltec; St. Paul, Minn) on each treatment day. Radiation fields in a three-box technique included the rectal canal and adjacent lymph nodes. Each combined chemoirradiation treatment in this study started on Monday, and results were interrupted on weekends. After 4 weeks of treatment, the patients had a recovery interval of 2 weeks and were scheduled for surgery in the following week.
This trial was approved by the local institutional review board. Each patient was fully informed and provided written consent.
Evaluation of Perfusion Parameters
We used a direct approach, namely dynamic T1 mapping with snapshot fast low-angle shot (FLASH) sequences (19), to quantify concentration-and-time curves (20). This method relied on the assumption that the change in bulk tissue relaxation rate, or R1 equals 1/T1, is a linear function of contrast medium concentration in the tissue. Relative concentration-and-time curves for the contrast medium were determined from the set of dynamically acquired T1 maps and were calculated on the basis of region-of-interest (ROI) measurements in arterial blood (right external iliac artery). The ROIs were sized according to the cross-sectional area of the vessel (usually 6 mm; range, 5–7 mm) and over the entire tumor tissue (Fig 1). All ROIs were placed by the same author (A.d.V.), who was blinded to the outcome of the patients.
|
|
|
|
Pulse Sequence
For the evaluation of T1 maps, an inversion-recovery snapshot FLASH sequence, as described by Gneiting et al (19) and Deichmann and Haase (22), was performed by using a 1.5-T whole-body imager (Magnetom Vision; Siemens, Erlangen, Germany). The pulse sequence consisted of a preceding nonselective inversion pulse followed by the acquisition of a series of 16 T1-weighted snapshot FLASH images. In the measurement of the T1 time, the nonselective inversion pulse was used to exclude errors, which may arise from the noninverted water spins flowing with the blood into the imaging plane during data acquisition.
To guarantee sufficiently accurate sampling of fast relaxation dynamics as they occur in blood after contrast medium administration, a short repetition time of 3.9 msec, a gradient echo time of 2.0 msec (3.9/2.0), and flip angle of 
= 5° were chosen, with the acquisition of a k-space–reduced scanning matrix of 64 x 128 elements (zero filled to 128 x 128). The use of these parameters resulted in an acquisition time of 250 msec per snapshot FLASH image. A linear off-center phase-encoding scheme was used (16 lines off center) to improve the reliability of the T1 mapping sequence with small T1 values. Sixteen snapshot FLASH images were acquired after the initial inversion pulse, with a total 4-second acquisition time for one T1 map. Complete T1 relaxation prior to the acquisition of the T1 map is crucial for the precision and reliability of the T1 maps. To achieve this relaxation and, thus, to exclude steady-state effects, the interval between consecutive T1 maps was chosen to be at least 14 seconds.
A quadrature phased-array body coil was used without performing intensity-profile correction to obtain unmanipulated image intensities. The accuracy of the obtained T1 values was verified by means of phantom measurements, where spectroscopically determined T1 values were compared with our T1 map data (19). Thereby, a mean deviation from the spectroscopic data of 2.8% and a maximum deviation of 5.6% was determined for all physiologic T1 values down to 150 msec.
T1 maps were calculated by using a least-square–fitting routine on a pixel-by-pixel basis and by taking progressive saturation effects into account. For practical purposes, T1 values were encoded by using a color look-up table, where blue regions denoted areas with a long T1-time, and red denoted regions with a short T1-time. A color change toward red signified a reduction in T1 due to contrast medium uptake (Fig 1).
Patient Measurements
The initial MR imaging investigation was performed before the onset of radiation therapy and was repeated at constant intervals once weekly during the course of treatment. The patients received an intravenous injection of 20 mg Buscopan (Boehringer, Ingelheim, Germany) prior to mapping to minimize peristaltic movement.
Transverse sections (field of view, 30 cm; thickness, 5 mm) through the pelvis of each patient were selected so that both the tumor and arterial vessels (external iliac arteries) could be clearly identified on the images (Fig 1). During the pretreatment examination, a section position for the T1 maps was chosen, which allowed the simultaneous delineation of arterial blood and tumor tissue. With this position, suitable anatomic markers such as bone structures in the pelvis were defined and were used for navigation in the following treatment measurements.
For dynamic T1 mapping, a contrast medium dose of 0.05 mmol of gadopentetate dimeglumine per kilogram of body weight (Magnevist; Schering, Erlangen, Germany) was infused at a constant rate into the right brachial vein by using a syringe pump over 4 minutes (flow rate, 90–105 mL/h). The constant-rate infusion was started after two precontrast T1 maps had been obtained. After the onset of the prolonged bolus injection, contrast medium uptake was recorded by obtaining 31 T1 maps at intervals of 14 seconds. The relatively slow contrast medium washout period was monitored by obtaining 15 T1 maps at intervals of 2 minutes. The resultant total time required for an examination, including the acquisition of transverse T1-weighted turbo spin-echo images (800/12; turbo factor, 3; 350 mm; matrix, 256) before and after contrast enhancement, was about 50 minutes.
Statistical Analysis
All statistical analyses were performed by using SPSS version 8.0 (SPSS, Cary, NC). Because of some deviations from a normal distribution, nonparametric procedures were applied. Paired comparisons were performed by using the Wilcoxon test. Since this study aimed to identify trends and substantial changes, no adjustments in P values for multiple testing were performed. The correlation between succeeding values were evaluated by using the Spearman rank correlation coefficient.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
With correct section positioning, both the tumor and arterial blood in the external iliac arteries were depicted on the T1 maps. During contrast medium application, T1 values reached a minimum of 0.40 second in arterial blood, 0.65 second in tumor, and 0.80 second in muscle. All of these values were well within the spectroscopically verified sensitivity range of the sequence. The spatial resolution of the T1 maps proved to be sufficient to reveal regions of different uptake kinetics in individual tumors. During constant-rate infusion, a steady and nearly linear increase in the concentration of contrast medium in arterial blood was observed. After termination of the infusion, a sharp peak was seen, followed by a multiexponential decay. The concentration in the tumor tissue peaked a few minutes after the concentration peaked in arterial blood (Fig 2).
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Dynamic MR imaging with prior administration of gadopentetate dimeglumine has been used in a few studies to evaluate tumor microcirculation (12,14–17). At MR imaging with gadopentetate dimeglumine, T1-weighted signal intensities are widely used in the calculation of concentration-and-time curves. However, relative changes in signal intensity are not ideally suitable for use in quantitative assessment of contrast medium concentration and, thus, of tumor perfusion due to a variety of factors (24,25). To avoid this source of error, we used dynamic T1 mapping with snapshot FLASH sequences as a direct approach to quantification (20,22).
This method has the potential for online monitoring of the individual microcirculatory response of a tumor to therapy. From the series of T1 maps, reliable concentration-time curves and, thus, the uptake and washout of contrast medium can be measured with excellent spatial and temporal resolution in both the tumor and its feeding arteries. The concentration-and-time curves obtained reflect the well-known pharmacokinetic behavior of an extracellular contrast medium such as gadopentetate dimeglumine and are not falsified by nonlinear dose-effects of the contrast medium or other errors typically associated with dynamic studies of signal intensity images (24,25).
We administered the contrast medium as a prolonged bolus over 4 minutes with a syringe pump; we used a constant rate of infusion to ensure a steady state between intravascular and interstitial contrast medium concentrations. An instantaneous bolus injection technique, as reported in most previous publications (12,14,16,17), does not necessarily represent a steady-state single-bulk situation during the first pass of the contrast medium and, thus, might lead to errors in the deduced concentration-and-time curves (26,27).
Although our method has proved to be feasible in a clinical setting and yields consistent data on tumor perfusion (13), there are limitations concerning data acquisition time and sample size. As with other sampling methods (eg, polarographic oxygen measurements), it is questionable whether the sample measured in only a restricted area or a single plane is representative enough to support estimations about the behavior of the whole tumor. Meanwhile, our group succeeded in developing a multisection T1 mapping technique that has already been validated in tests (28). In future studies, the sampling rate can be raised substantially, and a wider coverage of tumor tissue can be achieved.
Analysis of our data on the basis of PI values involved the use of data from only the first 10 minutes in the concentration-and-time curves. Therefore, in future studies, the total observation time of contrast medium kinetics can be shortened substantially. Shortening the acquisition time for T1 maps from approximately 50 to 10 minutes should allow the use of this method as a adjunct to any diagnostic MR examination, with only a slight increase in examination time.
Another point of concern is the influence of the chemotherapeutic agent on the microcirculation of normal and tumor tissue. Only a few articles reported data for high-dose administration of 5-fluorouracil. Kakinuma and Ohwada (29) described no substantial change in gastric mucosal blood flow in rats after intravenous administration of low-dose (50 mg/kg) 5-fluorouracil but did report a substantial and dose-dependent decrease after the administration of high doses of 5-fluorouracil (100 or 200 mg/kg). In RIF-1 murine tumors, Li et al (30) described a substantial increase of blood perfusion, which was estimated by uptake of 86 Rb+, after treatment with 5-fluorouracil (100 or 200 mg/kg, intraperitoneal injection); 86 Rb+ uptake in normal tissues (skin, muscle) was unaffected. At present, no data are available concerning the effects of continuous administration of 5-fluorouracil with its inherent low chemotherapeutic concentration levels in the bloodstream. Current investigations aim to clarify whether an indirect effect due to the reduction of vital cells in tumor exists (31).
In 21 patients with different tumors (colorectal, gastric, hepatocellular carcinomas), Harte et al (32) used positron emission tomography to show a close correlation between uptake and retention of radioactively labeled 5-fluorouracil and blood flow. Zanelli et al (6) found the following three peaks of 3H-vincristine sulfate uptake in tumors in CBA/He mice: at 14 (1st week), 28 (2nd week), and 40 (3rd week) Gy. Radiobiologic observations have proved that the capillary system becomes more permeable after irradiation (33). Our results indicate that the increased microcirculation can be explained as a combination of increased perfusion and extraction fraction.
PI, our main parameter for assessing tumor microcirculation, combines two important parameters: tissue perfusion and extraction fraction. It can be speculated that the accumulation of nutrients and therapeutic agents is mainly dominated by these two parameters (34). Therefore, PI measurements may be used to reliably monitor microcirculatory changes that are highly relevant to tumor therapy. We use both effects, tumor perfusion and the extraction fraction, for assessing tumor microcirculation with T1 mapping. The contrast agent used, gadopentetate dimeglumine, has a molecular weight similar to that of most chemotherapeutic agents. Thus, the MR imaging data should resemble chemotherapeutic extraction kinetics.
The increase in PI values during the 2nd to 3rd weeks of chemoirradiation suggests a notable increase in blood supply to the tumor. The degree and time course varies between individual patients. This variability can be explained by the complex effects of the combined therapy, which directly affect the tumor vasculature and which act on the tumor cells to cause secondary changes to the blood supply.
Monitoring the tumor perfusion data could guide the optimal timing for the administration of chemotherapeutic agents during fractionated irradiation. Monitoring would allow us to reduce the total dose of chemotherapeutic agent used while maintaining or even improving the therapeutic result. As a consequence, this reduction could lead to a reduction of side effect–induced interruption of chemoirradiation, which has been shown to have a substantial negative effect on therapeutic success (35,36). In any case, our results may lead to a better understanding of tumor behavior during fractionated radiation therapy.
Measurement of PI values before therapy seems to have value in the prognosis of tumor treatment outcome. Our data showed a significant correlation between higher PI values before treatment and N downstaging (P = .042), which, in turn, has a substantial influence on therapy outcome, as Janjan et al (37) demonstrated. This finding can be explained by the combination of two previous observations: Hawighorst et al (16) showed a positive correlation between microvascular density in tumor tissue and their microcirculatory data. Vascular density in tumor tissue itself relates significantly with therapy outcome, as Weidner et al (38) and Kohno et al (39) showed.
Similar results exist from previous MR studies (16,17) in which the initial uptake of contrast medium in tumor tissue was shown to be indicative of the effectiveness of radiation therapy. In tumor cells, microcirculatory data might well reflect the oxygen status, a major factor that influences radiosensitivity. However, the numerous interactions between microcirculation, oxygen uptake, and consumption in tumor tissue necessitate further study.
Studies involving larger patient populations and the monitoring of different types of tumor entities with any combination of radiation therapy and chemotherapy might lead to a better understanding of tissue behavior. The future use of more sophisticated tissue compartment models for obtaining our T1 map data will allow a more differentiated observation of physiologic parameters (15).
In summary, our noninvasive method for measuring perfusion data has proved to be a robust and practical tool for monitoring tumor microcirculation during the entire course of fractionated chemoirradiation and for studying its influence on therapy.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, P.L.; study concepts, J.G.; study design, A.d.V., W.J.; definition of intellectual content, A.d.V., J.G.; literature research, A.d.V.; clinical studies, P.L., A.d.V.; experimental studies, C.K., T.K., W.B.; data acquisition, A.d.V., T.G.; data analysis, A.d.V.; statistical analysis, K.P.P.; manuscript preparation, A.d.V.; manuscript editing, W.B., W.J.; manuscript review, P.D.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Bellomi, G. Petralia, A. Sonzogni, M. G. Zampino, and A. Rocca CT Perfusion for the Monitoring of Neoadjuvant Chemotherapy and Radiation Therapy in Rectal Carcinoma: Initial Experience Radiology, August 1, 2007; 244(2): 486 - 493. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V S Khoo and D L Joon New developments in MRI for target volume delineation in radiotherapy Br. J. Radiol., September 1, 2006; 79(Special_Issue_1): S2 - S15. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Atri New Technologies and Directed Agents for Applications of Cancer Imaging J. Clin. Oncol., July 10, 2006; 24(20): 3299 - 3308. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Liu, H. S. Rugo, G. Wilding, T. M. McShane, J. L. Evelhoch, C. Ng, E. Jackson, F. Kelcz, B. M. Yeh, F. T. Lee Jr, et al. Dynamic Contrast-Enhanced Magnetic Resonance Imaging As a Pharmacodynamic Measure of Response After Acute Dosing of AG-013736, an Oral Angiogenesis Inhibitor, in Patients With Advanced Solid Tumors: Results From a Phase I Study J. Clin. Oncol., August 20, 2005; 23(24): 5464 - 5473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Miller, H. H. Pien, D. Sahani, A. G. Sorensen, and J. H. Thrall Imaging Angiogenesis: Applications and Potential for Drug Development J Natl Cancer Inst, February 2, 2005; 97(3): 172 - 187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Dzik-Jurasz The development and application of functional nuclear magnetic resonance to in vivo therapeutic anticancer research: 2002 Sir Godfrey Hounsfield lecture delivered at the President's Day, Manchester Br. J. Radiol., April 1, 2004; 77(916): 296 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A R Padhani MRI for assessing antivascular cancer treatments Br. J. Radiol., December 1, 2003; 76(suppl_1): S60 - S80. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Boiselle, K. F. Reynolds, and A. Ernst Multiplanar and Three-Dimensional Imaging of the Central Airways with Multidetector CT Am. J. Roentgenol., August 1, 2002; 179(2): 301 - 308. [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Boiselle and A. Ernst Recent Advances in Central Airway Imaging* Chest, May 1, 2002; 121(5): 1651 - 1660. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. DeVries, J. Griebel, C. Kremser, W. Judmaier, T. Gneiting, A. Kreczy, D. Öfner, K.-P. Pfeiffer, G. Brix, and P. Lukas Tumor Microcirculation Evaluated by Dynamic Magnetic Resonance Imaging Predicts Therapy Outcome for Primary Rectal Carcinoma Cancer Res., March 1, 2001; 61(6): 2513 - 2516. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
R1), versus time in the arterial blood, muscle, and tumor of patient 6 before treatment. Change in R1 is a linear function of contrast medium concentration in tissue. s = seconds.