Purpose: The aim of this study was to investigate the relationship between the in vivo derived kinetic parameters of 3′-deoxy-3′-18F-fluorothymidine (18F-FLT) and the proliferation rate measured in vitro by Ki-67 staining in patients with newly diagnosed high-grade gliomas.

Experimental Design: Thirteen patients with newly diagnosed high-grade gliomas were investigated with 18F-FLT and methyl-11C- l-methionine (11C-MET) positron emission tomography (PET) and T1-, Gd-T1–, and T2-weighted magnetic resonance imaging on consecutive days. Tracer kinetic parameters of 18F-FLT as well as the standardized uptake value and the tumor-to-background (T/B) ratio of 18F-FLT and 11C-MET were determined. Data of kinetic modeling, standardized uptake value, and T/B values derived from 18F-FLT-PET were compared with T/B values derived from 11C-MET-PET and to the in vitro proliferation marker Ki-67.

Results: A significant correlation was observed between the metabolic rate constant Ki and the proliferation index as measured by Ki-67 immunostaining [Ki, r = 0.79 (P = 0.004)]. Also, the phosphorylation rate constant k3 correlated with Ki-67 [k3, r = 0.76 (P = 0.006)], whereas the rate constant for transport through the blood brain barrier K1 showed a weaker correlation with Ki-67 [K1, r = 0.62 (P = 0.044)]. No significant correlation between 11C-MET and 18F-FLT uptake ratios and Ki-67 was observed.

Conclusions: This study shows that kinetic analysis of 18F-FLT tracer uptake is essential for the in vivo assessment of tumor proliferation in high-grade gliomas, whereas uptake ratios of 11C-MET and 18F-FLT failed to correlate with the in vitro determined proliferation marker. Thus, kinetic analysis of 18F-FLT might provide an accurate method for the assessment of early response to glioma treatment in the future.

Positron emission tomography (PET) with radiolabeled amino and nucleic acids allows metabolic imaging of tumor activity in vivo. Methyl-11C- l-methionine (11C-MET) and 2-18F-fluoro-deoxy-d-glucose have been established as markers in the diagnosis of gliomas. 2-18F-fluoro-deoxy-d-glucose enables to detect brain tumors because of their increased glucose consumption. However, the high cortical background level of glucose limits the capacity of 2-18F-fluoro-deoxy-d-glucose to distinguish tumoral tissue from normal brain tissue. 11C-MET enables to determine with high sensitivity the delineation of the extent of the tumor, the effect of treatment, and the differentiation of recurrent tumor from radiation necrosis (1). Methionine uptake correlates to microvessel density (2), to the proliferative cell nuclear antigen index indicating the malignancy of brain tumors (3), and to the expression of the LAT1 amino acid transporter (4). However, kinetic analysis remains limited because of the short half-life of 11C and the fast metabolism of 11C-MET.

Shields et al. (5) have developed 3′-deoxy-3′-18F-fluoro- l-thymidine (18F-FLT) as a tracer to image proliferation in vivo. They established 18F-FLT as an analogue substrate of thymidine, which is intracellulary phosphorylated by the thymidine kinase 1 (TK1). TK1 is a cytosolic enzyme that is expressed with the onset of the S phase during DNA synthesis and that is decreased in nondividing cells. Activity of TK1 increases up to 10-fold during the S phase, after which it is directly degraded (6). Compared with normal proliferating tissue, the increase of TK1 activity is even higher in proliferating tumor cells (7) and can be imaged by 18F-FLT as selective substrate for TK1 that converts 18F-FLT to its nucleotide monophosphate. This 18F-FLT monophosphate is not further metabolized and accumulates in the cell.

In several clinical studies, 18F-FLT has been validated to assess proliferation of different types of tumors in vivo (811). Correlations between the standard uptake value (SUV) of 18F-FLT and the Ki-67 expression in vitro have been shown in lung cancer, malignant lymphoma, colorectal cancer, and recently, in brain tumors (1215). In a mixed population of patients with newly diagnosed and recurrent gliomas, our group recently showed that 18F-FLT uptake (a) enables to differentiate between low-grade and high-grade tumors; (b) is mainly due to increased transport and to a lower extent to phosphorylation by TK1; and, finally, (c) that 18F-FLT-, 11C-MET-PET, as well as Gd-enhanced magnetic resonance imaging (MRI) yield complementary information on the activity and the extent of gliomas (16).

The long half-life of 18F allows for kinetic analysis, providing the differentiation between metabolized and nonmetabolized 18F-FLT in the tissue. Especially in brain tumors, kinetic modeling enables to distinguish between increased 18F-FLT uptake due to increased transport through the blood brain barrier (BBB) and increased uptake due to an increased TK1 activity (16, 17).

The aim of this study was to measure tumor proliferation noninvasively in vivo in patients with nontreated newly diagnosed high-grade gliomas using kinetic modeling and to investigate the correlation between the kinetic parameters of 18F-FLT and the immunohistochemical proliferation marker Ki-67.

### Patients

Thirteen patients suffering from newly diagnosed primary central nervous system tumors (9 male and 4 female; median age, 64.0 y; range, 35-71 y) were included in this prospective study after giving their written informed consent on multimodal PET and MRI. The study protocol differed from our previous study (16) only with respect to patients with newly diagnosed tumors. After the PET scan, except for one, all gliomas were confirmed by histology and classified according to WHO grade (stereotactic biopsy, n = 7; resection, n = 5). Five patients were classified as astrocytoma grade III, one as oligoastrocytoma grade III, and six as glioblastoma (grade IV; Table 1). All patients underwent Gd-diethylenetriaminepentaacetic acid–enhanced MRI within 6 d before the PET investigation. PET investigation was done when there were suspicious findings for tumor proliferation in the MRI.

Table 1.

Clinical data, PET uptake values, kinetic rate constants k3 and Ki, and percentage expression of the in vitro proliferation marker Ki-67

Pat. no.Age (y)SexLocationHistology11C-MET uptake ratio18F-FLT uptake ratio18F-FLT SUVk3 (tumor; 1/min)Ki (tumor; 1/mL/g/min)Ki-67 (%)
57 R/BG — 1.6 3.83 0.8 0.0055 0.0016 —
58 L/T Astrocytroma grade III 3.73 10.8 2.42 0.1578 0.0396 70
70 R/PT Glioblastoma 3.2 5.67 1.87 0.055 0.0173 40
71 R/O Astrocytoma grade III 1.8 5.4 1.61 0.0267 0.0086 20
56 R/T Astrocytoma grade III 2.46 2.29 0.44 25
66 L/P Glioblastoma 3.6 7.79 1.88 0.044 0.0132 30
35 L/PO Glioblastoma 3.73 1.32 0.0223 0.012 —
69 R/BG Astrocytoma grade III 2.4 2.16 0.41
54 BS Astrocytoma grade III 1.9 2.34 0.63 30
10 63 L/P Glioblastoma 4.6 6.99 1.7 0.03 0.0143 20
11 67 L/P Glioblastoma 2.8 6.45 1.75 0.237 0.029 55
12 64 R/F Glioblastoma 3.2 6.77 1.3 0.0472 0.0161 45
13 66 R/F Oligoastrocytoma III 4.5 4.03 1.24 0.0833 0.0122 60
Pat. no.Age (y)SexLocationHistology11C-MET uptake ratio18F-FLT uptake ratio18F-FLT SUVk3 (tumor; 1/min)Ki (tumor; 1/mL/g/min)Ki-67 (%)
57 R/BG — 1.6 3.83 0.8 0.0055 0.0016 —
58 L/T Astrocytroma grade III 3.73 10.8 2.42 0.1578 0.0396 70
70 R/PT Glioblastoma 3.2 5.67 1.87 0.055 0.0173 40
71 R/O Astrocytoma grade III 1.8 5.4 1.61 0.0267 0.0086 20
56 R/T Astrocytoma grade III 2.46 2.29 0.44 25
66 L/P Glioblastoma 3.6 7.79 1.88 0.044 0.0132 30
35 L/PO Glioblastoma 3.73 1.32 0.0223 0.012 —
69 R/BG Astrocytoma grade III 2.4 2.16 0.41
54 BS Astrocytoma grade III 1.9 2.34 0.63 30
10 63 L/P Glioblastoma 4.6 6.99 1.7 0.03 0.0143 20
11 67 L/P Glioblastoma 2.8 6.45 1.75 0.237 0.029 55
12 64 R/F Glioblastoma 3.2 6.77 1.3 0.0472 0.0161 45
13 66 R/F Oligoastrocytoma III 4.5 4.03 1.24 0.0833 0.0122 60

Abbreviations: Pat. no., patent number; M, male; F, female; R, right; L, left; BG, basal ganglia; BS, brain stem; T, temporal; PT, parietotemporal; PO, parietooccipital; O, occipital; P, parietal; F, frontal.

### PET

Data acquisition. PET imaging was done on an ECAT EXACT (CTI/Siemens; in-plane full-width at half maximum, 6 mm; slice thickness, 3.375 mm; axial field of view, 162 mm) and an ECAT EXACT HR (CTI/Siemens; in-plane full-width at half maximum, 3.6 mm; slice thickness, 3.125 mm; axial field of view, 150 mm). Ten-minute transmission scans with rotating germanium-68/gallium-68 sources were done for attenuation correction of the PET data.

11C-MET and 18F-FLT syntheses were done as described previously (16). The radiolabeling yield of 18F-FLT was 10% ± 1.5%, and the radiochemical purity of 18F-FLT was >98%.

The mean injected dose of 18F-FLT was 321.9 ± 85.1 MBq (range, 111-370 MBq). From all patients, arterialized blood samples were taken by a peripheral i.v. catheter. 18F-FLT-PET images were acquired in the following dynamic sequence: 6 × 10, 3 × 20, 2 × 30, 2 × 60, 2 × 150, and 16 × 300 s.

Tracer accumulation was recorded in a three dimensional mode >60 min in 47 transaxial slices of the entire brain as described previously (2, 16).

For coregistration to the anatomic data, T1, T2, and contrast enhanced MRI scans were done in all patients on a 1.5 T system (Gyroscan Intera; Philips Medical Systems). Magnetic resonance images were coregistrated to summed PET images with an accuracy of 2 mm or better (18).

Data analysis. We used a region-of-interest approach to determine the maximal tracer uptake in 18F-FLT- and 11C-MET-PET of the summed images. The circular region-of-interest with a diameter of 8 mm was placed in the tumor region with the highest tracer uptake. To calculate the uptake ratio, a reference region-of-interest was placed on the contralateral unaffected tissue. Uptake and SUV were then calculated as previously described (2, 16). Coregistration and region-of-interest calculation was done by using the VINCI software (19).

Mankoff et al. (20, 21) established a kinetic model to quantify the incorporation of thymidine into DNA. In analogy to this kinetic model for thymidine, Muzi et al. (22) validated a compartment model for 18F-FLT in patients with lung cancer to quantify TK1 activity. They used a four-parameter two-compartment model with strong constant estimates that correlate to the in vitro proliferation marker Ki-67. In this model, the rate constant K1 determines the transport across the BBB into the tissue and k2 determines the return from the tissue to the blood. The rate constant k3 represents the intracellular phosphorylation of 18F-FLT, and the small proportion that is dephosphorylated back to 18F-FLT is represented by k4 (23). Also, blood volume was included as a parameter in the model.

We also used a four-parameter (+blood volume) two-compartment model including k4. Although Akaike's Information Criterion values between the four-parameter model and the three-parameter model were not significantly different, we decided to use the four-parameter two-compartment model because ignoring k4 as kinetic constant rate in the kinetic model leads to an overestimation of the influx of 18F-FLT as reported previously (24, 25).

18F-FLT is the radiolabeled analogue of the endogeneous thymidine. Thymidine is incorporated into DNA during production and is therefore strongly related to proliferation. The main purpose of 18F-FLT analysis consists in obtaining information about the endogenous processes of thymidine and thereby measuring cellular proliferation. The pathway of thymidine is determined by transport from blood into tissue (KTh1), out of the tissue into blood (kTh2), its phosphorylation (kTh3), and its dephosphorylation (kTh4). In contrast to fluorothymidine, thymidine is incorporated into DNA, i.e., the thymidine pathway has an additional rate constant describing the incorporation of thymidine into DNA (kDNA). This constant cannot be determined from the fluorothymidine analysis. However, the rate of cell proliferation (P), which is proportional to the rate of thymidine incorporation into DNA, can be expressed in terms of the thymidine rate constants and the thymidine concentration in blood (ThB) by

$P{\sim}\frac{K_{\mathrm{Th1}}{\cdot}k_{\mathrm{Th3}}}{\left(1+k_{\mathrm{Th4}}/k_{\mathrm{DNA}}\right){\cdot}k_{\mathrm{Th2}}+k_{\mathrm{Th3}}}{\cdot}Th_{\mathrm{B}}$

If kTh4 is small compared with kDNA, which is a reasonable assumption, because the dephosphorylation rate is low, this expression is independent of kDNA. Because the other reactions are also present in the fluorothymidine pathway, we replace the thymidine rate constants by their fluorothymidine analogues leading to

$P{\sim}\frac{K1{\cdot}k3}{k2+k3}{\cdot}Th_{\mathrm{B}}=Ki{\cdot}Th_{\mathrm{B}}$

Thus, the rate of cell proliferation is proportional to the influx constant rate constant Ki and the thymidine level in the blood plasma.

The time activity curves were determined in the part of the tumor with the highest uptake. Three consecutive brain slices were included in the calculation. Kinetic analysis was done by using PMOD biomedical image quantification and kinetic modeling software (PMOD Technologies Ltd.).

### Histologic assessment of proliferation by immunostaining Ki-67

A representative formalin-fixed, paraffin-embedded section from each specimen was immunohistochemically stained with MIB-1 (Ki-67) antibody by use of the Avidin-Biotin-Peroxidase-Complex method and the DCS Detection kit with 3,3′-diaminobenzidine and H2O2 (DCS). All cells with nuclear staining of any intensity were regarded as positive. Proliferative activity was defined as the percentage of nuclei stained with MIB-1 per total number of nuclei in the biopsy. The fraction of labeled tumor cells, defined as the Ki-67 labeling index, was assessed over four microscopic high power fields (0.16 mm2) that contained the highest average fraction of labeled cells. The Ki-67 labeling index was determined by scoring the fraction of cells stained with the MIB-1 antibody in 5% intervals.

### Statistical analysis

Parametric statistical tests were used to determine significant correlations between the parameters (Pearson correlation analysis). Correlations were considered significant at a level of P value <0.05. Statistical analysis were done by SPSS software (Release 11.0.1.; SPSS, Inc.).

Distribution of 18F-FLT and 11C-MET uptake. The uptake ratios and SUV were calculated by placing a circular region-of-interest in the area of the tumor with highest tracer uptake (Table 1). 18F-FLT uptake ratios varied from 2.2 to 10.8 (mean, 5.3; SD, 2.5), SUV from 0.4 to 2.4 (mean, 1.3; SD 0.6), and 11C-MET uptake ratios ranged from 1.6 to 4.6 (mean, 3.0; SD, 0.9). Restricted to the histologic grading mean values of SUV, 18F-FLT and 11C-MET uptake were higher in patients with WHO grade IV gliomas (n = 6) than in patients with WHO grade III gliomas [n = 6; fluorothymidine-SUV, 1.6 ± 0.3 (WHO IV) versus 1.1 ± 0.8 (WHO III); fluorothymidine uptake ratio, 6.2 ± 1.4 (WHO IV) versus 4.5 ± 3.3 (WHO III); methionine uptake ratio, 3.4 ± 0.6 (WHO IV) versus 2.8 ± 1.1 (WHO III); Fig. 1]. 18F-FLT uptake ratio correlated weakly to 11C-MET uptake ratio (r = 0.65; P = 0.016), indicating the relation between uptake of amino acids and nucleosides in tumor regions with a disrupted BBB.

Fig. 1.

Coregistered 18F-FLT, 11C-MET, parametric map of metabolism Ki, and MRI T1 + Gd. A, a 58-y-old patient with an astrocytoma grade III. The 18F-FLT-PET shows an uptake of 10.8-fold to the contralateral tissue with a high metabolic constant Ki (Ki = 0.039 mL/g/min) and a 3.73-fold 11C-MET uptake corresponding to a high Ki-67 expression of 70%. B, an oligoastrocytoma grade III of a 66-y-old patient with relatively high 11C-MET uptake (4.5-fold) and a relatively low 18F-FLT uptake (4.03-fold) but high values of Ki (Ki = 0.0122 mL/g/min) and of Ki-67 expression (60%). A 63-y-old patient with a first diagnosed glioblastoma: the 18F-FLT-PET and 11C-MET image in C show high tracer uptake ratios in 18F-FLT (6.77-fold) and in 11C-MET uptake (3.22-fold) to the contralateral tissue, an increased kinetic metabolic constant Ki (0.0161 mL/g/min) according to a high %–Ki-67 expression of 45%.

Fig. 1.

Coregistered 18F-FLT, 11C-MET, parametric map of metabolism Ki, and MRI T1 + Gd. A, a 58-y-old patient with an astrocytoma grade III. The 18F-FLT-PET shows an uptake of 10.8-fold to the contralateral tissue with a high metabolic constant Ki (Ki = 0.039 mL/g/min) and a 3.73-fold 11C-MET uptake corresponding to a high Ki-67 expression of 70%. B, an oligoastrocytoma grade III of a 66-y-old patient with relatively high 11C-MET uptake (4.5-fold) and a relatively low 18F-FLT uptake (4.03-fold) but high values of Ki (Ki = 0.0122 mL/g/min) and of Ki-67 expression (60%). A 63-y-old patient with a first diagnosed glioblastoma: the 18F-FLT-PET and 11C-MET image in C show high tracer uptake ratios in 18F-FLT (6.77-fold) and in 11C-MET uptake (3.22-fold) to the contralateral tissue, an increased kinetic metabolic constant Ki (0.0161 mL/g/min) according to a high %–Ki-67 expression of 45%.

Close modal

Contribution of the kinetic parameter K1 and k3 to 18F-FLT uptake. Data of kinetic analysis are summarized in Table 2. Kinetic model analysis provides the rate constants for transport (K1), reflux (k2), intracellular phosphorylation (k3), and dephosphorylation (k4) of fluorothymidine and the fraction of blood volume. The metabolic rate Ki of 18F-FLT is calculated from the rate constants as described above. To analyze the contribution of the kinetic constants to 18F-FLT uptake measured in PET, we calculated their correlation to 18F-FLT uptake. The kinetic constant K1 showed a weaker relation to 18F-FLT uptake (r = 0.52; P = 0.064) than the phosphorylation rate constant k3 (r = 0.60; P = 0.029; Fig. 2). These results may indicate that in high-grade newly diagnosed gliomas, tracer uptake is more related to the phosphorylation rate than to the transport through the BBB. The metabolic rate of 18F-FLT Ki showed the strongest correlation to 18F-FLT uptake (r = 0.87; P < 0.001).

Table 2.

Kinetic analysis including k4 in the tumor

Pat. no.Tumor
KiSDK1SDk2SDk3SDk4SDvBSD
0.0016 0.0667 0.0069 0.0064 0.0185 0.1217 0.0055 0.326 —* 0.0492 0.0105
0.0396 0.0114 0.0733 0.0331 0.1344 0.2895 0.1578 0.2911 0.0241 0.0176 0.1343 0.0225
0.0173 0.0182 0.0411 0.0237 0.0753 0.1816 0.055 0.181 0.0149 0.0408 0.1016 0.0231
0.0086 0.0025 0.0123 0.074 0.0117 208.121 0.0267 6308 3.54460 6151 0.112 0.732
0.0000 0.0000 0.0066 0.006 0.0099 0.204 —*   0.07 0.0169
0.0132 0.0123 0.033 0.0081 0.066 0.0813 0.044 0.1057 0.022 0.0394 0.0452 0.0087
0.0120 0.0489 0.0234 0.0087 0.0212 0.0737 0.0223 0.2448 0.0163 0.2049 0.072 0.0129
0.0000 0.0000 0.0056 0.6064 0.0135 838 —*   0.0435 0.07
0.0000 0.0000 0.0086 0.0136 0.0195 7,454 —*   0.0533 0.0291
10 0.0143 0.0213 0.0353 0.0112 0.0442 0.0728 0.03 0.109 0.012 0.0617 0.11 0.016
11 0.0290 0.0101 0.164 0.058 1.107 0.6781 0.237 0.1851 0.03 0.186 0.125 0.0227
12 0.0161 0.0102 0.041 0.0068 0.0732 0.0645 0.0472 0.0789 0.0191 0.0268 0.0644 0.0081
13 0.0122 0.0067 0.032 0.008 0.135 0.1513 0.0833 0.1338 0.024 0.0255 0.044 0.0069
Pat. no.Tumor
KiSDK1SDk2SDk3SDk4SDvBSD
0.0016 0.0667 0.0069 0.0064 0.0185 0.1217 0.0055 0.326 —* 0.0492 0.0105
0.0396 0.0114 0.0733 0.0331 0.1344 0.2895 0.1578 0.2911 0.0241 0.0176 0.1343 0.0225
0.0173 0.0182 0.0411 0.0237 0.0753 0.1816 0.055 0.181 0.0149 0.0408 0.1016 0.0231
0.0086 0.0025 0.0123 0.074 0.0117 208.121 0.0267 6308 3.54460 6151 0.112 0.732
0.0000 0.0000 0.0066 0.006 0.0099 0.204 —*   0.07 0.0169
0.0132 0.0123 0.033 0.0081 0.066 0.0813 0.044 0.1057 0.022 0.0394 0.0452 0.0087
0.0120 0.0489 0.0234 0.0087 0.0212 0.0737 0.0223 0.2448 0.0163 0.2049 0.072 0.0129
0.0000 0.0000 0.0056 0.6064 0.0135 838 —*   0.0435 0.07
0.0000 0.0000 0.0086 0.0136 0.0195 7,454 —*   0.0533 0.0291
10 0.0143 0.0213 0.0353 0.0112 0.0442 0.0728 0.03 0.109 0.012 0.0617 0.11 0.016
11 0.0290 0.0101 0.164 0.058 1.107 0.6781 0.237 0.1851 0.03 0.186 0.125 0.0227
12 0.0161 0.0102 0.041 0.0068 0.0732 0.0645 0.0472 0.0789 0.0191 0.0268 0.0644 0.0081
13 0.0122 0.0067 0.032 0.008 0.135 0.1513 0.0833 0.1338 0.024 0.0255 0.044 0.0069
*

The SD cannot be calculated when the parameter is equal to 0.

k4 is not defined when k3 = 0.

Fig. 2.

Relationship between 18F-FLT uptake ratio and the kinetic constants.

Fig. 2.

Relationship between 18F-FLT uptake ratio and the kinetic constants.

Close modal

Kinetic parameters for transport and proliferation were ∼2-fold higher in grade IV tumors than in grade III tumors, which did not reach statistical significance [K1, 0.056 ± 0.021 (WHO IV) versus 0.023 ± 0.01 (WHO III); k3, 0.072 ± 0.033 (WHO IV) versus 0.045 ± 0.02 (WHO III); Ki, 0.017 ± 0.002 (WHO IV) versus 0.010 ± 0.006 (WHO III)]. Interestingly, K1 and k3 values of the one oligoastrocytoma were higher than K1 and k3 of all but one astrocytoma (K1, 0.032 versus 0.021 ± 0.029; k3, 0.083 versus 0.037 ± 0.069).

Sensitivity analysis of the parameter estimates. The sensitivity analysis calculates the effect of the variation of a parameter on the model output at a certain time, i.e., on the time activity curve. Figure 3 gives a representative example of the sensitivity analysis for this kinetic model. At the beginning, the time activity curve is completely determined by the blood volume. After the bolus has passed the tissue, the transport process starts. The sensitivity of the model to K1 increases during the first 15 minutes until having reached a steady state. The sensitivity to the phosphorylation rate k3 increases rapidly in the first 15 minutes and then with a lower slope over the entire investigation. The negative sensitivity of the time activity curve to the rate constants k2 and k4 reflects that an increase of these constants decreases the model output. At the end of scan time, sensitivity curves for k2, k3, and k4 have not reached their plateau, i.e., steady state has not been reached.

Fig. 3.

Sensitivity curve of the kinetic parameters in the 18F-FLT model. Sensitivity of each parameter conforms to the extent that the parameter changes the model. vB, blood vessel.

Fig. 3.

Sensitivity curve of the kinetic parameters in the 18F-FLT model. Sensitivity of each parameter conforms to the extent that the parameter changes the model. vB, blood vessel.

Close modal

In all cases, the correlation matrix of the model parameters shows covariances between K1 and k2 and between k3 and k4, which arises from the fact that they describe the same reactions (forward and reverse). All other covariances are negligible. Thus, our choice of the model with two compartments and one blood pool is also justified on a statistical basis.

Ki-67 expression. MIB-1 immunohistochemistry was done in 11 of 13 patient biopsy samples. The tumor probe of one patient was too small to evaluate a representative region for Ki-67 expression; the other patient did not undergo surgery after the PET scan. Mean Ki-67 expression in high grade gliomas ranged between 5% and 70% (mean, 31.4%; SD 19.5%). Differences in Ki-67 expression between the WHO grade III and IV were not found (WHO III: mean, 32.5%; SD, 26.2%; WHO IV: mean, 30%; SD, 9.4%) indicating that in vitro measurement of proliferative activity is not necessarily related to WHO grading.

Analyzing Ki-67 according to the histologic type of astrocytoma, the proliferative fraction of the one oligoastrocytoma was with 60% relatively high compared with the astrocytomas (mean, 27%; SD, 11.2).

Comparison of proliferation activity in vitro by Ki-67 expression and in vivo by PET. Linear regression analysis revealed a strong correlation between the metabolic rate constant of 18F-FLT Ki and the in vitro expression of Ki-67 (r = 0.79; P = 0.004; Fig. 4). Also, the phosphorylation rate constant of 18F-FLT k3 correlates with Ki-67 (r = 0.76; P = 0.006; Fig. 4). The transport rate K1 shows a weaker correlation with the proliferation marker Ki-67 (r = 0.62; P = 0.044).

Fig. 4.

Correlation of kinetic constants k3 (A) and Ki (B) to proliferation index Ki-67 (MIB). Pearson rank correlation coefficient is r = 0.88 (P < 0.001) for k3 and r = 0.79 (P = 0.004) for Ki, respectively.

Fig. 4.

Correlation of kinetic constants k3 (A) and Ki (B) to proliferation index Ki-67 (MIB). Pearson rank correlation coefficient is r = 0.88 (P < 0.001) for k3 and r = 0.79 (P = 0.004) for Ki, respectively.

Close modal

The correlation of the 18F-FLT uptake ratio and the proliferation index was not statistically significant (r = 0.57; P = 0.70). Likewise the ratio of 11C-MET uptake did not show any significant correlation with Ki-67 expression (r = 0.43; P = 0.19).

In a patient population with newly diagnosed high grade gliomas WHO III and IV, we show that kinetic modeling of 18F-FLT enables to determine tumor proliferation in vivo via the metabolic constant Ki. Furthermore, this in vivo obtained proliferation marker correlates strongly with the in vitro proliferation rate as determined by Ki-67 immunostaining. These data indicate that kinetic modeling of 18F-FLT uptake in tumor tissue facilitates (a) to assess the proliferation rate in vivo and (b) to distinguish the proportion of 18F-FLT uptake that is due to transport through the BBB from the proportion, which is due to intracellular proliferation. This distinction between tracer transport and phosphorylation enables to further analyze 18F-FLT uptake to obtain a highly sensitive tool for the quantification of tumor proliferation and, thereby, the potential for monitoring antiproliferative treatment.

T2-weighted images and contrast-enhanced MRI is commonly used for the primary diagnosis of brain tumors. In MRI, T2-weighted hyperintense signal and contrast enhancement provide rather indirect signs for tumor growth. Therapeutic procedures as chemotherapy, radiotherapy, and surgery even induce disturbance of the BBB and, hence, contrast enhancement. Therefore, MRI is not suitable neither for monitoring response to therapy nor for the prediction of clinical outcome.

In our recent PET imaging studies, we showed in a heterogeneous patient population with newly diagnosed and recurrent gliomas of various WHO grades that tracers such as 11C-MET and 18F-FLT provide complementary information on the activity and extent of a glioma (16). We showed that 11C-MET detects the extent of gliomas with a high sensitivity and specificity, and that uptake of 11C-MET is correlated to microvessel density (2, 26). 11C-MET might therefore be an accurate tool to reflect tumor angiogenesis. However, for the determination of tumor proliferation as well as for monitoring response to antiproliferative therapy strategies 18F-FLT is the more favorable marker.

In vitro studies have shown that 18F-FLT uptake reflects TK1 activity (27, 28). Cytosolic TK1 is a cell cycle–regulated enzyme that is activated during the salvage DNA synthesis pathway. 18F-FLT serves as a selective substrate for TK1 and reacts only to a very small part with the mitochondrial TK2. TK1 converts 18F-FLT to its monophosphate as a first step to incorporate the nucleoside into the DNA. As soon as the cells have passed the S phase, TK1 will be degraded (6). Unlike thymidine, only <1% of 18F-FLT is incorporated into the DNA (29). The rate-limiting factor for 18F-FLT—in contrast to thymidine—is therefore phosphorylation and not the incorporation into the DNA.

We have shown in this study that the proliferation rate can be obtained from the kinetics of 18F-FLT if the rate constant for dephosphorylation of thymidine (kTh4) is small compared with its rate constant for incorporation into DNA. Because kTh4 is small, this condition is always satisfied if kDNA and, therefore, the proliferation rate is sufficiently high—which is always the case in highly proliferating tissues such as tumors.

The metabolic constant Ki determined in vivo strongly correlates with Ki-67 expression. This relation confirms the assumption mentioned above that the proliferation rate is proportional to the product of Ki and the thymidine level in the blood plasma, which has not been determined in this study. We also observed a strong correlation between k3 and Ki-67 indicating that a high proliferation rate is related to a high phosphorylation rate rather than to an increased transport through the BBB in high-grade gliomas. These findings support the notion that a detailed kinetic analysis enables determination of tumor proliferation in vivo.

It is further interesting to examine the relation of 18F-FLT SUV or uptake ratios—which do not require kinetic modeling and can therefore be much easier obtained from the PET data—to proliferation in high-grade gliomas. Uptake ratios did not correlate with the in vitro proliferation marker Ki-67. The failure of uptake values to detect tumor proliferation is due to the fact that uptake values do not allow to distinguish whether the accumulation of 18F-FLT is mainly caused by transport respectively breakdown of the BBB or to proliferation. This assumption conforms to the patient study recently reported by Wells et al. (30) with 11C-Thymidine PET that behaves in analogy to 18F-FLT. In a patient with a brain tumor, they found a strong divergence between the increasing transport parameter and a declining flux constant (Ki) after several treatment strategies. The declining parameter Ki was finally correlated with the patient's clinical outcome. Thus, uptake ratios and SUV might cause misleading estimates of tumor proliferation.

In contrast to our results, Chen et al. (15) recently described a significant correlation between 18F-FLT SUV and the expression of the in vitro marker Ki-67 in brain tumors. However, low-grade gliomas were also included in this study. This might suggest that tracer uptake values, determined as SUV, correlate to in vitro proliferation in a WHO grade–dependent manner but not within a single WHO grade. We suggest that the metabolic rate constant Ki is the more sensitive parameter for imaging proliferation of high-grade gliomas than tracer uptake ratios of 18F-FLT and 11C-MET.

The correlation between k3 and the 18F-FLT uptake ratio indicates that in untreated gliomas with a high in vitro proliferation index, 18F-FLT uptake is more related to phosphorylation than to transport of 18F-FLT through the BBB. Also, the strong correlation between k3, and the metabolic rate constant Ki confirms the effect of the phosphorylation rate k3 on the amount of 18F-FLT uptake in the tumor. These results correspond with our previous findings where we found a strong correlation between the kinetic constants K1 and k3 and 18F-FLT uptake ratio (16). Although in the previously studied heterogeneous patient group with newly diagnosed and recurrent gliomas of various WHO grades, we observed a stronger correlation of K1 than of k3 to 18F-FLT uptake. In treated brain tumors, the major proportion of 18F-FLT uptake is due to transport because of the treatment-induced disturbance of the BBB. In contrast, in the present study in nontreated patients with high-grade gliomas, the contribution of the phosphorylation rate—expressed by k3—to 18F-FLT uptake was higher than of the transport rate constant K1. This underlines the capability and the necessity of kinetic analysis to distinguish between tracer uptake caused by an increased transport through the disturbed BBB and cellular tumor proliferation.

It should be pointed out that a limitation of stereotactic biopsy is its limited size, which sometimes is not representative for the entire tumor. However, in our study, PET data were used for guidance of stereotactic biopsy from the most active tumor part.

A detailed analysis of the heterogeneous tumor compartments of 18F-FLT permits to distinguish high-proliferating from low-proliferating tumor areas. Thus, 18F-FLT analysis allows for the detection of the strongest proliferating part of the tumor. Because the most proliferating part of the tumor is mainly responsible for tumor progression, 18F-FLT analysis enables a more precise estimation of the malignancy and may also serve as an accurate measure of the antiproliferative treatment strategies.

In summary, kinetic analysis of 18F-FLT in patients with newly diagnosed high-grade gliomas provides a useful method for the determination of proliferation rate in vivo. However, with regard to the limited number of patients included in this study, the potential of 18F-FLT kinetic analysis for the early prediction of clinical outcome and response to therapy remains to be further investigated. This study indicates that evaluation of therapy efficiency of individual molecular-targeted approaches may benefit from 18F-FLT kinetic analysis.

Grant support: Deutschen Forschungsgemeinschaft (DFG-Ja98/1-2) and the 6th FW EU grant EMIL (LSHC-CT-2004-503569).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: R. Ullrich and H. Backes contributed equally to this work.

1
Jacobs AH, Dittmar C, Winkeler A, Garlip G, Heiss WD. Molecular imaging of gliomas.
Mol Imaging
2002
;
1
:
309
–35.
2
Kracht LW, Friese M, Herholz K, et al. Methyl-[11C]- l-methionine uptake as measured by positron emission tomography correlates to microvessel density in patients with glioma.
Eur J Nucl Med Mol Imaging
2003
;
30
:
868
–73.
3
Sato N, Suzuki M, Kuwata N, et al. Evaluation of the malignancy of glioma using 11C-methionine positron emission tomography and proliferating cell nuclear antigen staining.
Neurosurg Rev
1999
;
22
:
210
–4.
4
Miyagawa T, Oku T, Uehara H, et al. “Facilitated” amino acid transport is upregulated in brain tumors.
J Cereb Blood Flow Metab
1998
;
18
:
500
–9.
5
Shields AF, Grierson JR, Dohmen BM, et al. Imaging proliferation in vivo with [F-18]FLT and positron emission tomography.
Nat Med
1998
;
4
:
1334
–6.
6
Sherley JL, Kelly TJ. Regulation of human thymidine kinase during the cell cycle.
J Biol Chem
1988
;
263
:
8350
–8.
7
Toyohara J, Waki A, Takamatsu S, Yonekura Y, Magata Y, Fujibayashi Y. Basis of FLT as a cell proliferation marker: comparative uptake studies with [3H]thymidine and [3H]arabinothymidine, and cell-analysis in 22 asynchronously growing tumor cell lines.
Nucl Med Biol
2002
;
29
:
281
–7.
8
Wells P, Gunn RN, Alison M, et al. Assessment of proliferation in vivo using 2-[(11)C]thymidine positron emission tomography in advanced intra-abdominal malignancies.
Cancer Res
2002
;
62
:
5698
–702.
9
Vesselle H, Grierson J, Muzi M, et al. In vivo validation of 3′deoxy-3′-[(18)F]fluorothymidine ([(18)F]FLT) as a proliferation imaging tracer in humans: correlation of [(18)F]FLT uptake by positron emission tomography with Ki-67 immunohistochemistry and flow cytometry in human lung tumors.
Clin Cancer Res
2002
;
8
:
3315
–23.
10
Buck AK, Bommer M, Stilgenbauer S, et al. Molecular imaging of proliferation in malignant lymphoma.
Cancer Res
2006
;
66
:
11055
–61.
11
Buck AK, Halter G, Schirrmeister H, et al. Imaging proliferation in lung tumors with PET:18F-FLT versus 18F-FDG.
J Nucl Med
2003
;
44
:
1426
–31.
12
Francis DL, Visvikis D, Costa DC, et al. Potential impact of [18F]3′-deoxy-3′-fluorothymidine versus [18F]fluoro-2-deoxy-d-glucose in positron emission tomography for colorectal cancer.
Eur J Nucl Med Mol Imaging
2003
;
30
:
988
–94.
13
Buck AK, Schirrmeister H, Hetzel M, et al. 3-deoxy-3-[(18)F]fluorothymidine-positron emission tomography for noninvasive assessment of proliferation in pulmonary nodules.
Cancer Res
2002
;
62
:
3331
–4.
14
Wagner M, Seitz U, Buck A, et al. 3′-[18F]fluoro-3′-deoxythymidine ([18F]-FLT) as positron emission tomography tracer for imaging proliferation in a murine B-Cell lymphoma model and in the human disease.
Cancer Res
2003
;
63
:
2681
–7.
15
Chen W, Cloughesy T, Kamdar N, et al. Imaging proliferation in brain tumors with 18F-FLT PET: comparison with 18F-FDG.
J Nucl Med
2005
;
46
:
945
–52.
16
Jacobs AH, Thomas A, Kracht LW, et al. 18F-fluoro- l-thymidine and 11C-methylmethionine as markers of increased transport and proliferation in brain tumors.
J Nucl Med
2005
;
46
:
1948
–58.
17
Muzi M, Spence AM, O'Sullivan F, et al. Kinetic analysis of 3′-deoxy-3′-18F-fluorothymidine in patients with gliomas.
J Nucl Med
2006
;
47
:
1612
–21.
18
Cizek J, Herholz K, Vollmar S, Schrader R, Klein J, Heiss WD. Fast and robust registration of PET and MR images of human brain.
Neuroimage
2004
;
22
:
434
–42.
19
Vollmar S CJ, Sue M, Klein J, Jacobs AH, Herholz K. VINCI - Volume Imaging in Neurological Research, Co-Registration and ROIs Forschung und wissenschaftliches Rechnen 2003. In: Kremer K, Macho V, editors. Göttingen: Gesellschaft für wissenschaftliche Datenverarbeitung; 2004. p. 115–31.
20
Mankoff DA, Shields AF, Graham MM, Link JM, Eary JF, Krohn KA. Kinetic analysis of 2-[carbon-11]thymidine PET imaging studies: compartmental model and mathematical analysis.
J Nucl Med
1998
;
39
:
1043
–55.
21
Mankoff DA, Shields AF, Link JM, et al. Kinetic analysis of 2-[11C]thymidine PET imaging studies: validation studies.
J Nucl Med
1999
;
40
:
614
–24.
22
Muzi M, Vesselle H, Grierson JR, et al. Kinetic analysis of 3′-deoxy-3′-fluorothymidine PET studies: validation studies in patients with lung cancer.
J Nucl Med
2005
;
46
:
274
–82.
23
Krohn KA, Mankoff DA, Muzi M, Link JM, Spence AM. True tracers: comparing FDG with glucose and FLT with thymidine.
Nucl Med Biol
2005
;
32
:
663
–71.
24
Patlak CS, Blasberg RG. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. Generalizations.
J Cereb Blood Flow Metab
1985
;
5
:
584
–90.
25
Spence AM, Muzi M, Graham MM, et al. Glucose metabolism in human malignant gliomas measured quantitatively with PET, 1-[C-11]glucose and FDG: analysis of the FDG lumped constant.
J Nucl Med
1998
;
39
:
440
–8.
26
Kracht LW, Miletic H, Busch S, et al. Delineation of brain tumor extent with [11C] l-methionine positron emission tomography: local comparison with stereotactic histopathology.
Clin Cancer Res
2004
;
10
:
7163
–70.
27
Rasey JS, Grierson JR, Wiens LW, Kolb PD, Schwartz JL. Validation of FLT uptake as a measure of thymidine kinase-1 activity in A549 carcinoma cells.
J Nucl Med
2002
;
43
:
1210
–7.
28
Barthel H, Perumal M, Latigo J, et al. The uptake of 3′-deoxy-3′-[18F]fluorothymidine into L5178Y tumours in vivo is dependent on thymidine kinase 1 protein levels.
Eur J Nucl Med Mol Imaging
2005
;
32
:
257
–63.
29
Grierson JR, Schwartz JL, Muzi M, Jordan R, Krohn KA. Metabolism of 3′-deoxy-3′-[F-18]fluorothymidine in proliferating A549 cells: validations for positron emission tomography.
Nucl Med Biol
2004
;
31
:
829
–37.
30
Wells JM, Mankoff DA, Eary JF, et al. Kinetic analysis of 2-[11C]thymidine PET imaging studies of malignant brain tumors: preliminary patient results.
Mol Imaging
2002
;
1
:
145
–50.