| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Investigations |
Research Program Radiological Diagnostics and Therapy, German Cancer Research Center (Deutsches Krebsforschungszentrum) [H-P. S., M. B., P. B., I. Z., M. V. K., G. v. K.], and Departments of Otolaryngology, Section Oncology [A. D., B. V., P. W., H. W.], and Radiation Therapy, Radiological Clinic [V. R., M. W.], University of Heidelberg, 69120 Heidelberg, Germany
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-fluoro-ß-alanine (FBAL) were resolved in the tumor spectra. The median of the 5-FU and FBAL levels was significantly higher (more than 2-fold) at the second compared with the first examination, whereas the level of DHFU did not change. This effect could indicate an increased delivery of 5-FU into the interstitial space of the tumor in the course of the combined treatment, which would result in an enhanced exposure of the tumor cells to the drug. A potential mechanism for synergy between radio- and chemotherapy is discussed, but alternative mechanisms are also being considered. The findings indicate that a method is available to rationally address the design of dosing schedules in concurrent therapy regimens. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fluorine-19 NMR spectroscopy permits noninvasive monitoring of the pharmacokinetics of 5-FU in vivo (10, 11, 12) . The 19F nucleus has 100% natural abundance and a NMR sensitivity comparable to that of 1H. There is no 19F-NMR background signal from the tissue because the physiological concentration of mobile fluorine (less than 10-6 M) is below the detection threshold. The large chemical shift range allows the spectral resolution of NMR signals of 5-FU and different 5-FU metabolites even at the relatively low field strength of 1.5 T commonly used for clinical NMR spectroscopy. Previous 19F-NMR studies with human patients undergoing 5-FU chemotherapy, which focused mainly on the examination of liver metastases, indicate a correlation of intratumoral 5-FU levels and response to treatment (13, 14, 15) .
The aim of our clinical study was to detect possible changes of intratumoral 5-FU pharmacokinetics in patients in the course of simultaneous radiotherapy and chemotherapy. For this purpose, 19F-NMR spectroscopy examinations were performed in patients with advanced head and neck carcinomas. Randomized clinical trials indicate that radiochemotherapy using 5-FU in combination with other drugs, particularly cisplatin, improves therapeutic outcome of head and neck cancer in terms of improved locoregional control, disease-free survival, and overall survival rates (5) . We present what we believe to be the first report on 19F-NMR-monitored intratumoral pharmacokinetics of 5-FU in patients undergoing a simultaneous radiochemotherapy.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 |
500 cm3). Detailed patient data are given in Table 1
.
|
The definitions of the target volumes were based on CT or magnetic resonance imaging scans in all of the patients, and the dose was calculated to midplane using the custom-developed software "Ziviplan Heidelberg." With the use of 6-MeV photons, the primary tumor including a safety margin and the upper and middle neck nodes were irradiated through parallel opposed fields, and the lower neck nodes were irradiated with an anterior oppositional field. Individual blocks were used to spare healthy tissue as far as possible. With a dose of 27 Gy at the isocenter, the spinal cord was outside the photon fields. The uninvolved posterior neck was treated according to CT findings with electrons of selected energy using a dose of 2.5 Gy/day applied five times a week (total dose, 55 Gy). The boost comprising the macroscopic tumor was delivered by opposed lateral fields.
Two cycles of chemotherapy with 5-FU (dose, 600 or 1000 mg/m2 of b.s. x day) and carboplatin (70 mg/m2 of b.s. x day) were applied on days 1–5 (week 1) and days 29–33 (week 5). Tumor response was assessed 6 weeks and 3 months after therapy by determining the tumor size from CT or ultrasonographic images. Response was finally classified according to the WHO criteria as complete remission (CR, complete disappearance of the tumor), partial remission (PR, reduction of tumor size by 50% or more), stable disease (SD, reduction of tumor size by less than 50% or increase of tumor size by less than 25%), and progressive disease (PD, increase of tumor size by more than 25%).
The study was approved by the local Human Investigations Committee according to the principles of the Declaration of Helsinki, and informed consent was obtained from each patient before the NMR examination.
Fluorine NMR.
19F-NMR spectroscopy was performed in a 1.5-T whole-body scanner (Magnetom 63/84 SP; Siemens, Erlangen, Germany) on days 1 and 29 of treatment (i.e., on day 1 of the first and second chemotherapy cycle). A 5-cm-diameter surface coil (built by H-J. Zabel, Deutsches Krebsforschungs-zentrum) tunable to 1H (for shim on the tissue water 1H resonance) and 19F Larmor frequency was positioned centrally on a palpable neck metastasis (Fig. 1)
. To ensure reproducibility in repeat measurements, the positioning of the coil on the tumor as well as the coil’s orientation relative to the direction of the magnetic field were carefully recorded for each patient. The individual 5-FU dose was diluted in a 250-ml solution of 0.9% sodium chloride and administered i.v. at an infusion rate of 5 ml/min (total infusion time, 50 ± 2 min). A series of 19F-NMR spectra was acquired during 5-FU administration. In the case of good cooperation of the patient, additional spectra were recorded for a period up to 30 min after the end of the infusion. Repeat examinations were performed only on patients exhibiting fluorine signals on the first examination (except patient no. 14).
|
Data Evaluation and Statistical Analysis.
For a comparison of 5-FU uptake and metabolism at the various stages of treatment, 19F-NMR spectra were analyzed in two different ways: (a) for the kinetic analysis, FIDs nos. 1, 2, 3, and so forth, obtained during and after 5-FU administration were added according to: 1, 1+2, 1+2+3, and so forth (cumulative sum); and (b) the sum of FIDs acquired during the infusion period (50 min) was used to estimate the amount of 5-FU and of fluorine-containing metabolites present in the tumor tissue during 5-FU administration. The resulting time-domain signals were zero-filled to 8 K data points and apodized by an exponential (
-1 = 10 Hz) and a Gaussian function (-1 = 0.01 Hz) before Fourier transformation and phase correction. The combination of exponential and Gaussian function provides optimum S:N on condition of minimum line-broadening. After baseline correction, Fourier spectra were analyzed using a least-squares-fit routine (VNMR; Varian, Palo Alto, CA).
NMR signal intensities and chemical shifts were referenced to the signal of an external sample (volume, 1 ml; fixed at the opposite side of the surface coil) of TFA (CF3COOH; chemical shift set at 
= 0 ppm). Normalization of signal intensities using the TFA signal allowed the quantitative comparison of the data from different examinations. Normalized signal intensities of the detected fluorine-containing compounds [i.e., the cytostatic drug (5-FU) and its catabolites, DHFU and FBAL] in sum spectra over the drug infusion period were designated by Iinf5-FU, IinfDHFU, and IinfFBAL. Iinf is interpreted as an approximate measure of the accumulated amount of the corresponding compound in the examined metastasis within the infusion period.
Changes of intratumoral drug and metabolite levels in the course of radio/[5-FU+carboplatin]chemotherapy were quantified by means of relative differences of Iinf5-FU, IinfDHFU, and IinfFBAL between the first and the second examinations. These parameters were also tested for a possible correlation to clinical response determined 6 weeks and 3 months after treatment. For the statistical analysis, a nonparametric, paired Wilcoxon test was applied (SAS for Windows; SAS, Cary, NC, USA).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Broad 19F-NMR resonances of 5-FU [chemical shift position () = -93.7 ppm; linewidth (
1/2) 
60 Hz], FBAL ( = -112.7 ppm, 1/2 110 Hz), and DHFU ( = -126.3 ppm, 1/2 85 Hz) were found in the tumor spectra (Table 2
and Fig. 2
). The signal of 5-FU was observed in the spectra of 11 of 17 patients and in 21 of 30 examinations. A comparison of tumor volumes in Table 1
and of Iinf5-FU in Table 3
shows that 5-FU was undetectable in the metastases with volumes larger than 40 cm3 (except for patient no. 12). In most of these examinations, the 5-FU resonance could be observed first at 6–18 min after the beginning of drug infusion. The 5-FU signal increased steadily during the infusion period and decreased rapidly afterwards.
|
|
|
4 ppm downfield from the 5-FU peak) were seen in the spectra obtained in this study.
For illustration, the data set acquired in the examinations of patient no. 12 is presented in Figs. 2
3
4
. Fig. 2
shows the 50-min sum spectrum obtained during the first cycle of chemotherapy. Fig. 3
and Fig. 4
display the kinetics of 5-FU, DHFU, and FBAL as stack plot and plot of integrated peak areas of cumulative sum spectra. Normalized 19F-NMR signal intensities Iinf5-FU, IinfDHFU, and IinfFBAL of all of the examined patients are given in Table 3
. Fig. 5
shows the averages of these data.
|
|
|
= -133.8 ppm (in the spectra of patients nos. 1, 6, and 8) and = -137.7 ppm (in the spectra of patients nos. 1, 3, and 15) that could not be assigned to any known 5-FU metabolite. Neither peak was observed during the follow-up examinations or in experiments on model solutions using the same equipment.
In all of the patients with detectable 19F-NMR signal (11 of 11), higher levels of 5-FU were measured at the second examination (Fig. 5a)
. Mean value and median of Iinf5-FU were larger by factors of 2.5 and 2.8, respectively, which is highly significant (P = 0.001). In 8 of 10 patients, the signal intensity of FBAL was also significantly larger in the second examination (Fig. 5c)
; mean value and median were enhanced by factors of 2.2 and 2.4, respectively (P = 0.027). In 6 of 11 patients, lower signal intensities of DHFU were detected in the second examination. Mean IinfDHFU decreased by a factor of 0.8, which is statistically insignificant; the median of IinfDHFU did not change (Fig. 5b)
.
Data on clinical response of the examined patients are given in Table 1
. Within 6 weeks after the end of treatment, no increase in tumor volumes was observed. The statistical analysis showed no correlation of clinical response determined 6 weeks and 3 months after therapy and 19F-NMR parameters, i.e., normalized signal intensities as well as intensity changes observed in the course of radio/[5-FU+carboplatin]chemotherapy.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The detoxification of 5-FU takes place mainly in the liver. In the first step, the drug is reduced to DHFU by trans-addition of two hydrogen atoms to the C5
6 double bond under consumption of energy. The degradation further proceeds to FUPA and FBAL. FBAL, the most abundant 5-FU metabolite in the body during 5-FU chemotherapy, is finally excreted via the kidneys.
In vitro and in vivo animal studies indicate that 5-FU acts as a radiosensitizer, but the underlying mechanism is poorly understood (6 , 7) . In experiments on cell cultures, Byfield (18) found three prerequisites for radiosensitization owing to 5-FU: (a) the tumortoxicity of the drug; (b) a sufficient amount of 5-FU the tumor cells are exposed to; and (c) the presence of the drug in the tumor tissue for at least 24 h after irradiation.
Although in vitro models, e.g., isolated tumor cells, permit study of the mechanisms of the interaction of 5-FU and ionizing radiation, important environmental factors that are effective in tumors in situ cannot be assessed in these systems. The amount of 5-FU that reaches the tumor cells depends on different mechanisms that are not directly related to the administered 5-FU dose. The major fraction of the infused drug is rapidly eliminated from the plasma by the catabolism in the liver. Although the plasma concentration of 5-FU could be relevant to the amount of 5-FU that reaches the tumor cells, plasma and urine levels of 5-FU and 5-FU metabolites of individual patients—determined ex vivo by means of high-resolution 19F-NMR spectroscopy—did not correlate with tumor response to treatment (19) . However, the intratumoral accumulation and retention of 5-FU, which was first measured by C. A. Presant et al., was found to be related to response to chemotherapy (14) . This result demonstrates the importance of the in vivo determination of intratumoral 5-FU levels.
In the treatment of head and neck carcinomas, higher response rates were observed when 5-FU was administered via continuous infusion in contrast to bolus injection (20) . The combination of radiotherapy and 5-FU chemotherapy showed positive effects when ionizing radiation and the drug were applied concurrently. No benefit was seen when 5-FU alone or in combination with other drugs was administered before or after radiotherapy, e.g., as induction or adjuvant therapy (5) . This agrees with the results of in vitro experiments, which showed that 5-FU has to be present in the tumor tissue for at least 24 h after irradiation to produce a radiosensitization effect (18) .
After the pioneering 19F-NMR examinations of patients by Wolf et al. (11)
, several clinical studies have been performed to monitor 5-FU pharmacokinetics during chemotherapy. After bolus infusion, the intratumoral half-life (t1/2) of 5-FU was determined (21)
. Findlay et al. (13)
and Schlemmer et al. (15)
estimated drug and metabolite levels after continuous and 10-min i.v. infusion of 5-FU, respectively. These studies showed that t1/2 
20 min of free 5-FU in malignant tumors ("trapping"; 21) as well as enhanced intratumoral 5-FU levels are predictive of a therapeutic response (13, 14, 15)
.
In the present study, spatial localization of 19F-NMR signals in neck metastases was performed by means of a small surface coil, which is adequate for the examination of superficial tumors. The simultaneous measurement of the reference signal allowed the control of the coil’s sensitivity as well as the quantitative comparison of drug and catabolite signals from different examinations. Because in most of the 6-min spectra the available signal was poor, FIDs were added to enhance S:N and to reduce errors in the statistical analysis. Although the resulting sum spectra showed sufficient S:N for the evaluation of relative signal intensities, a quantitative estimation of the 5-FU concentration in the tumors was difficult. Taking into account the sensitive volume of the surface coil (approximately 2
R3/3 33 cm3; R = radius of the coil’s loop) and assuming a homogeneous distribution of the drug in the tissue, the minimum 5-FU concentration (c) that can be detected within a 6-min measurement time with our equipment is cmin 30 µM as estimated from experiments on model solutions. However, absolute quantification by in vivo NMR is generally difficult because tumors are different in size and inhomogeneous, and, technically, fully relaxed spectra cannot be acquired when the measurement time is limited.
The frequent and repeated detection of DHFU in this work was surprising. Previous patient studies indicated that the tissue concentration of this intermediate of 5-FU catabolism is generally below the NMR detection limit (Refs. 11
, 22
; an exceptional case with a detectable level of DHFU in the metastatic liver is reported in Ref. 15
). Because the concentration of free DHFU in plasma is too low for 19F-NMR detection in vivo (19)
, the signal at = -126.3 ppm could originate from DHFU produced by the tumor cells. This interpretation agrees with in vitro experiments, which showed that DHFU is the major 5-FU catabolite in SCCs (23)
.
High steady-state concentrations of DHFU are possible if the transition from DHFU to FUPA/FBAL is rate-limiting. This agrees with the observation that, in patients with detectable FBAL and DHFU levels, the DHFU signal appeared before the FBAL signal (Figs. 3
and 4
). Although DHFU is a catabolite of 5-FU, its detection in tumor tissue could be important (24)
. The intermediate is cytotoxic as shown in vitro. Presumably, DHFU is converted back to 5-FU and acts as a storage form of 5-FU (25)
.
FBAL was observed less frequently. The peak at = -112.7 ppm was significantly broader than the 5-FU resonance and unresolved in agreement with spectra obtained in previous in vivo 19F-NMR studies (11
, 22)
. There are two reasons for the broadening of this resonance: (a) the scalar spin-spin coupling of 19F and 1H nuclei in FBAL which causes a multiplet splitting into eight resonances (19
, 26)
; and (b) signal overlap owing to the small chemical shift difference of only 1.8 ppm of FBAL and its precursor in the catabolic pathway, FUPA [according to high-resolution experiments in vitro (19)
]. The peak assigned to FBAL may, therefore, contain a signal contribution of FUPA. Because 5-FU catabolism mainly takes place in the liver, the signal at = -112.7 ppm will first be related to FBAL, which is produced in the hepatocytes and thereafter transported to the examined tumor region by the blood. A distinct signal contribution from FBAL formed in SCC is less probable. In particular, experiments with human SCC cell lines performed by Spoelstra et al. (23)
showed that only a very small amount of FBAL is produced by the tumor cells.
The significantly higher intensities of 5-FU at the beginning of the second cycle compared with the first cycle of chemotherapy (after 4 weeks of radiotherapy and a total dose of 43.5 Gy) point to alterations of 5-FU pharmacokinetics in the examined lymph node metastases. The NMR parameters indicate enhanced intratumoral drug levels (tumor volumes did not increase between the two spectroscopic examinations). Elevated intratumoral 5-FU concentrations might be explained by a more efficient transfer of the drug from the vascular system into the interstitial space, in accordance with the observation that NMR-detectable 5-FU is preferentially located in the interstitium (27)
. On the other hand, radiation therapy (28)
as well as treatment with 5-FU (29)
are known to increase tumor blood flow in experimental tumors. This effect is consistent with the data from our follow-up magnetic resonance imaging examinations of these patients, which indicate an increased uptake of Gd-DTPA in SCC after radiochemotherapy.4
An increase in tumor blood flow might also explain the enhanced intratumoral levels of FBAL, which is the dominant 5-FU catabolite in plasma (19)
. It is remarkable that 5-FU and FBAL were most frequently undetectable in large metastases (Tables 1
and 3
). This finding can be explained by decreased vascular perfusion in these tumors, which leads to inadequate drug delivery.
In particular, in the largest tumor examined in this study, only DHFU was seen, i.e., appreciable amounts of this catabolic intermediate were produced during the 50-min infusion period even with 5-FU levels below the in vivo 19F-NMR detection limit. Unchanged concentrations of DHFU during radiochemotherapy (Fig. 5b)
further support the conclusion that DHFU is produced in the tumor cells (the different behavior of the three resonances excludes an instrumental effect).
Another possible mechanism for the increase of 5-FU levels is reduced clearance of the drug from the tumor. This explanation would be in line with animal experiments which showed a prolonged retention of 5-FU after bolus administration in irradiated human colon adenocarcinoma in mice (9)
. Altogether, the mechanism that leads to the observed drug accumulation (Fig. 5a)
in SCC of patients in the course of radiochemotherapy is still being examined in detail and different mechanisms must be considered.
In contrast to previous 19F-NMR studies of metastases in the human liver (13, 14, 15) , the present study showed no correlation between intratumoral 5-FU levels (Iinf5-FU) and the reponse to treatment. One explanation is the different tumor type and therapy. In addition, only short-time follow-up data were available, and these showed in most patients of the study a partial response of the treated metastases. Changes of tumor sizes within a period of 3 months after treatment could be inappropriate for classification of response because examined tumors include not only vital tumor cells but also necrotic, inflammatory, and fibrotic tissue.
The lack of a correlation between 19F-NMR parameters and clinical response could be related to the ignorance of the treatment effect of carboplatin. Presently, this question cannot be resolved because treatment modalities using [5-FU+carboplatin]chemotherapy without irradiation or using 5-FU as single chemotherapeutic agent in combination with irradiation were not performed. It can be speculated that the suggested mechanisms for the alteration of intratumoral 5-FU pharmacokinetics during radiochemotherapy, resulting in an enhanced exposure of the tumor cells to the drug, may also be effective for the administered carboplatin. Recent animal experiments demonstrated the feasibility of platinum-195 NMR in vivo for monitoring local disposition kinetics of carboplatin in tissue (30) . However, the sensitivity of this method is insufficient at present for monitoring platinum-containing compounds at clinically relevant concentrations. More sensitive techniques that could be adequate to this purpose are 1H-NMR (31) , single-photon emission CT, or positron emission tomography; for example, Ginos et al. (32) obtained positron emission tomography scans using 13N-labeled cisplatin in patients with brain tumors undergoing chemotherapy with cisplatin.
In conclusion, the intratumoral 5-FU metabolism including the kinetics of 5-FU, DHFU, and FBAL could be monitored in vivo in this 19F-NMR study of SCC patients. The intermediate DHFU was seen in 90% of the examined patients with detectable metabolite signals and may be attributed to the metabolic activity of the tumor cells. Significant enhancements (more than 2-fold) of 5-FU and FBAL levels but no alteration of DHFU amounts in the metastases were observed during the second cycle of chemotherapy. The observation of increased drug levels within the tumor supports the notion of synergy of radiotherapy and 5-FU chemotherapy. This effect could be explained by an increased tumor blood flow related to radiation, chemotherapy, or both; but thus far, the mechanisms that are responsible for the accumulation of 5-FU in treated SCC are not definitive.
The ability of 19F-NMR spectroscopy to assess the pharmacokinetics of fluorine-containing drugs in tumors in vivo is unique and could be of use to optimize treatment protocols and to evaluate new radiation and drug dose schedules.
|
FOOTNOTES |
|---|
1 Supported by a grant of the Medical Faculty of the University of Heidelberg. 
2 To whom requests for reprints should be addressed, at Research Program Radiological Diagnostics and Therapy, German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany. Phone: 49-6221-422525; Fax: 49-6221-422531; E-mail: h.schlemmer{at}dkfz-heidelberg.de
3 The abbreviations used are: 5-FU, 5-fluorouracil; b.s., body surface; 1/2, full width at half maximum; DHFU, 5,6-dihydro-5-fluorouracil; FBAL, -fluoro-ß-alanine; FID, free-induction decay; FUPA, -fluoro-ß-ureidopropanoic acid; Iinf5-FU, IinfDHFU, and IinfFBAL, normalized NMR signal intensities of 5-FU, DHFU, and FBAL, respectively, in sum spectra over the drug infusion period; NMR, nuclear magnetic resonance; SCC, squamous cell carcinoma; S:N, signal:noise ratio; TFA, trifluoroacetic acid; CT, computed tomography.
4 H-P. Schlemmer et al., manuscript in preparation.
Received 10/28/98. Accepted 3/17/99.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
using in vivo 19F magnetic resonance spectroscopy in patients with colorectal cancer: a pilot study. Ann. Oncol., 4: 597-602, 1993.-fluoro-ß-alanine in both rat liver and tumour. Proc. Soc. Magn. Reson. Med. (SMRM), : 244 1993.
This article has been cited by other articles:
|
|
 |
|
  W. M. Spees, T. P.F. Gade, G. Yang, W. P. Tong, W. G. Bornmann, R. Gorlick, and J. A. Koutcher An 19F Magnetic Resonance-Based In Vivo Assay of Solid Tumor Methotrexate Resistance: Proof of Principle Clin. Cancer Res., February 15, 2005; 11(4): 1454 - 1461. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. M. Seddon, G. S. Payne, L. Simmons, R. Ruddle, R. Grimshaw, S. Tan, A. Turner, F. Raynaud, G. Halbert, M. O. Leach, et al. A Phase I Study of SR-4554 via Intravenous Administration for Noninvasive Investigation of Tumor Hypoxia by Magnetic Resonance Spectroscopy in Patients with Malignancy Clin. Cancer Res., November 1, 2003; 9(14): 5101 - 5112. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
