OtherClinical Investigations

Comparative Analysis of Striatal FDOPA Uptake in Parkinson’s Disease: Ratio Method Versus Graphical Approach

Vijay Dhawan, Yilong Ma, Vandhana Pillai, Phoebe Spetsieris, Thomas Chaly, Abdelfatihe Belakhlef, Claude Margouleff and David Eidelberg
Journal of Nuclear Medicine October 2002, 43 (10) 1324-1330;

Abstract

Striatal-to-occipital ratio (SOR) and influx constant Kiocc are commonly used as analytic parameters in l-3,4-dihydroxy-6-18F-fluorophenylalanine (FDOPA) PET studies. Both have been shown to be useful in discriminating Parkinson’s disease (PD) patients from healthy subjects. We evaluated the relative performance of SOR and influx constant (Kiocc) in the clinical assessment of nigrostriatal dopaminergic function in PD. Methods: Twenty-one parkinsonian patients (Hoehn and Yahr scale I–IV; mean age ± SD, 56 ± 9.2 y) and 11 healthy subjects (mean age, 60 ± 16 y) underwent 3-dimensional dynamic FDOPA scanning from 0 to 100 min. After spatial realignment, PET images at each frame were integrated by summing 4 central striatal slices, and time-activity curves (TACs) were generated after placing a standard set of elliptic regions of interest over striatal and occipital structures. SOR and Kiocc values for each subject were then computed from TACs at different times using an input function from the occipital cortex. Results: Both SOR and Kiocc showed significant bilateral decreases in striatal dopamine uptake in the PD group compared with the control group. SOR values estimated for 10-min frames between 65 and 95 min are statistically equivalent in group discrimination. In addition, SOR values in the caudate and putamen correlated strongly with Kiocc, especially toward the end of the scanning epoch. Both parameters correlated significantly and comparably with Unified Parkinson’s Disease Rating Scale motor scores. Conclusion: These results suggest that SOR determined from a single 10-min scan at 95 min is as accurate as Kiocc in separating PD patients from healthy subjects and in predicting clinical measures of disease severity.

Imaging with PET and l-3,4-dihydroxy-6-18F-fluorophenylalanine (FDOPA) has long been an established procedure to quantify the integrity of dopamine function in the human brain. Both striatal-to-occipital ratio (SOR) and influx constant (Kiocc) have previously been measured noninvasively in dynamic mode using region-of-interest (ROI) approaches (1–4). With the use of reference tissue, both parameters computed for putaminal ROIs have been found to separate Parkinson’s disease (PD) patients from healthy subjects, but with varying degrees of accuracy. We have consistently shown that SOR is a reliable imaging indicator of disease severity in parkinsonism (5–7). Similarly, Kiocc measurements have been used extensively to quantify regional dopamine metabolism (8,9) and to estimate the rate of disease progression (3).

Both parameters have become popular in recent years because their measurement is simple and does not require taking blood samples. SOR may offer a practical advantage because it can be determined by static data acquisition whereas Kiocc requires dynamic scans over a longer time. A long study in patients with advanced PD not only poses a serious compliance issue but also increases potential bias from subject movement. Thus, the application of a ratio index such as SOR may be useful in quantifying nigrostriatal dopamine function in parkinsonism and related disorders. Two important issues, however, need to be addressed before the broad implementation of ratio methods with FDOPA PET: The first is the dependence of SOR on time after the tracer reaches equilibrium, and the second is the relative merits of SOR and Kiocc as descriptors of nigrostriatal dopaminergic degeneration. In this study, we examined these questions by comparing FDOPA PET data from a set of PD patients and a set of healthy volunteers. Compared with our previous study that compared SOR and Kiocc using 2-dimensional data acquisition (1), we have this time used 3-dimensional data acquisition on a more sensitive PET scanner with shorter time frames.

MATERIALS AND METHODS

Subject Characteristics

We investigated 21 patients with moderate PD (13 men, 8 women; mean age ± SD, 56 ± 9.2 y; Hoehn and Yahr stage I–IV; n = 5, 6, 7, and 5 in stages I, II, III, and IV, respectively) (10). All patients stopped dopaminergic medications 12 h before FDOPA PET. We also selected 11 age-matched healthy volunteers as control subjects (8 men, 3 women; mean age, 60 ± 16 y). All patients and healthy volunteers signed a consent form after receiving a detailed explanation of the clinical protocol, which was approved by the Institutional Review Board at North Shore University Hospital.

Image Acquisition

Dynamic FDOPA PET data were acquired in 3-dimensional mode on an Advance scanner (General Electric Medical Systems, Milwaukee, WI) during the 100 min after injection (11). This camera covered the whole brain with an intrinsic resolution of 4.2 mm and a slice separation of 4.25 mm. All subjects had a light breakfast 4 h before the study and were given 200 mg of carbidopa 1.5 h before the scan to inhibit decarboxylation. The FDOPA preparation and imaging protocols have been described in detail elsewhere (5). Dynamic scanning started at the time of tracer injection at continued until 100 min after tracer injection. Images were reconstructed with a 6-mm Hanning filter to give a 3-dimensional image resolution of about 8 mm. There were 35 image planes per frame, with a matrix dimension of 128 × 128 and a voxel size of 2.34 × 2.34 mm. Corrections were made for random events, scatter, and electronic dead time, and the photon attenuation effect was corrected using a 10-min transmission scan collected with rotating 68Ge rod sources. These images were then transferred to personal computers running Windows NT (Microsoft, Redmond, WA) and were converted into Analyze format (Mayo Clinic, Rochester, MN).

Image Processing

Dynamic frames were realigned to the image at 55 min using the SPM99 program (Wellcome Department of Cognitive Neurology, London, U.K.). PET images between 40 and 100 min were averaged, and a mean image was created by summing 4 central slices (thickness, 17 mm) covering the striatum. The units of radioactivity for time-activity curves (TACs) were kBq/mL. A set of standard elliptic ROIs (55, 160, and 310 cm2 for the caudate region, putaminal region, and occipital region, respectively) were placed over the right and left caudates, putamen, and occipital cortex on the mean image. To reduce noise in the occipital TAC, the values for the left and right sides were averaged.

Image Analysis

TACs in the caudate, putamen, and occipital cortex were computed from the corresponding single-slice PET images. Striatal-to-occipital ratio (SOR) values were generated for each structure using bilaterally averaged occipital ROI data. SOR was calculated for each 10-min time frame: 65, 75, 85, and 95 min after injection. Kiocc was also calculated by graphic analysis over 40–100 min using the nonspecific uptake value from the occipital region. SOR and Kiocc data from the left and right caudates and the putamen were averaged. The 2 parameters were correlated within each striatal ROI (caudate, putamen) by computing Pearson product moment correlation coefficients. ROI data from the PD and the healthy groups were analyzed separately and in combination. Caudate and putaminal data from the PD group were compared with control data using discriminant function analysis (F test). In 16 of the 21 PD patients (mean age, 55 ± 8.6 y; mean composite Unified Parkinson’s Disease Rating Scale [UPDRS] motor ratings, 33 ± 12), both parameters were correlated with individual motor UPDRS ratings.

RESULTS

We first compared SOR with Kiocc in each structure by combining the data from both the PD group and the healthy group. For the combined group of PD patients and healthy subjects, SOR values at times later than 65 min were strongly correlated with Kiocc in the caudate (r2 = 0.75; P < 0.0001) and putamen (r2 = 0.92; P < 0.0001) (Fig. 1; Table 1). This criterion was also significant (P < 0.01) within each of the 2 groups. This correlation became stronger with time after injection; r2 increased from 0.82 to 0.92 at times from 65 to 95 min (Table 1). Caudate uptake behaved similarly, although the accuracy of discrimination was generally lower.

FIGURE 1.
FIGURE 1.

Parametric correlation between SOR (95 min) and Kiocc in 2 striatal structures in both PD and control groups.

View this table:
TABLE 1

Properties of SOR and Kiocc

Both SOR and Kiocc data in the caudate and putamen were greatly reduced in the PD patients compared with the healthy subjects (Table 2). The magnitudes of reduction in the 2 regions ranged from 21% to 34% for SOR (34%–56% for ratio R, defined as SOR − 1, or the ratio of specific activity in the putamen to occipital activity) and 32%–60% for Kiocc. Figure 2 shows a clear separation in mean values between the 2 groups, with a significance level of P < 0.0001. There were fewer miscategorizations (false-negatives) present with SOR than with Kiocc. Moreover, SOR for the putamen reached a higher significance than did Kiocc in the discrimination analysis (F test, Table 1). The decrease in striatal FDOPA uptake and the relative sensitivity of SOR over Kiocc were in line with our results from a different PET scanner (1). When we compared PD patients with healthy subjects, the FDOPA uptake indices given by both parameters were significantly lower in the putamen than in the caudate. This finding agrees with early studies reported by several groups (2,8). In addition, all SOR values at 65, 75, 85, and 95 min after injection were effective in discriminant analysis (P < 0.0001, Table 1), although an improving trend with time was evident (F values increasing from 182 to 251, Table 1).

FIGURE 2.
FIGURE 2.

Discrimination analysis using SOR (95 min) and Kiocc in 2 striatal structures. SOR and Kiocc give similar separation between PD and control groups. However, SOR is associated with higher discriminant scores (F scores) than is Kiocc (Table 1). Error bars indicate mean and SDs.

View this table:
TABLE 2

Mean SOR and Kiocc Values in Striatal Structures: Comparison of PD and Control Groups

Correlations between FDOPA uptake as measured by SOR or Kiocc and clinical indices of motor disability were evaluated (Table 1; Fig. 3). Decreased uptake in the putamen correlated significantly with composite UPDRS motor scores. Both SOR (r2 = 0.46; P < 0.004) and Kiocc (r2 = 0.44; P < 0.005) had similar correlations with disease severity. No significant correlations were found between UPDRS motor ratings and either SOR or Kiocc in the caudate (P < 0.2).

FIGURE 3.
FIGURE 3.

Clinical correlation between UPDRS motor ratings and dopamine uptake in putamen as measured by SOR (95 min) and Kiocc.

DISCUSSION

The correlations of SOR and Kiocc were of comparable magnitude to those described previously in patients with early-stage PD (5). SOR was based on independent 10-min scans starting at 60 min after injection, whereas Kiocc was based on 40–100 min of dynamic data. The linearity of the Kiocc plot improves with time, and the uptake rate is usually calculated from data between 40 and 100 min. SOR, in contrast, reaches a plateau very slowly, and this parameter is better estimated at later times except for increased noise in the occipital counts after about 100 min. The results in Table 1 show that the SOR-versus-Kiocc correlation varied from 0.82 to 0.92 between 65 and 95 min and was highest at 95 min.

Our results demonstrate that both parameters significantly discriminate PD patients from healthy subjects and correlate with independent measures of motor dysfunction in patients as expected on the basis of theoretic considerations. Both SOR and Kiocc successfully discriminate PD patients from healthy subjects and are also equally sensitive as descriptors of disease severity. Indeed, we are currently using SOR as an index to statistically map 3-dimensional topographic changes in FDOPA uptake after therapeutic treatments (7).

Even though SOR values calculated at 65, 75, 85, and 95 min were statistically equivalent in discriminating healthy subjects from PD patients, the F ratio reflecting the power of discrimination improved with time, from 182 to 251 (Table 1). Moreover, the correlation between clinical severity rating (UPDRS) and SOR improved with time, from 0.29 (at 65 min) to 0.46 (at 95 min, Table 1). This finding is important for longitudinal studies, in which higher sensitivity is required to reduce the number of subjects needed to detect a small change.

The mean percentage differences in Kiocc were equal to or larger than corresponding SOR values; however, the SD in SOR was lower than that in Kiocc, especially for PD patients. We have previously reported differences in the coefficient of variation in SOR and Kiocc using 2-dimensional data from a lower-resolution scanner (1).

We have used the average of the left and right sides to establish a conservative estimate (the more affected body side would better discriminate between healthy and PD subjects, but it is often difficult to assign which side is more affected in some individuals. In healthy subjects, averaging the 2 sides to reduce variance is obviously better. Additionally, the discrimination between healthy and PD subjects will dramatically improve if SOR and Kiocc analysis is performed separately on the anterior and posterior parts of the putamen. FDOPA levels have been shown to reveal a distinct anterior-posterior gradient as the disease progresses. It remains to be seen if a similar improvement in correlation between the 2 parameters (SOR and Kiocc) and UPDRS results if one focuses exclusively on the posterior region of the affected putamen (increased noise in the putaminal signal may, to some extent, counterbalance the expected increased sensitivity and specificity of the regional information).

Two earlier studies reported Kiocc to be more powerful than SOR in differentiating PD patients from healthy subjects and in detecting the rate of disease progression (2,3). Vingerhoets et al. (12), in 1994, suggested that SOR may be an appropriate parameter for assessing natural evolution based on its smallest within-subject variation, but Ki (estimated using metabolite-corrected blood data) can permit the use of fewer subjects for drug studies based on larger reliability coefficients. However, in a study dealing with reproducibility and the discriminating ability of FDOPA PET in PD, the same investigators found that SOR and Kiocc were similar in their ability to evaluate progressive changes in nigrostriatal dopaminergic function (2). We attribute this finding to the fact that the SOR measure was computed from images acquired in 2 dimensions and integrated over a longer time than was used in this study. Also, the inclusion of the early phase of the scans (from 30 min after injection onward) decreases the sensitivity of the SOR measure (Table 1). We have not directly compared SOR and Kiocc for longitudinal studies, but the correlation between disease severity (UPDRS scores) and the 2 PET-derived parameters (SOR and Kiocc) was similar (0.46 vs. 0.44) at 95 min after injection (Table 1; Fig. 3). Also, the correlation between SOR and Kiocc was 0.92 at 95 min, suggesting that both SOR and Kiocc may be able to reflect disease progression in longitudinal studies in a similar, quantitative manner. Interestingly, we have directly compared SOR with Ki (plasma) in our fetal transplant study that included moderately advanced PD patients and found that SOR was superior to Ki (plasma) for longitudinal changes (7).

In longitudinal studies, R is the preferred parameter because it reflects changes specifically in striatal uptake rather than in total uptake. SOR and R have similar information and can easily be derived from each other (R is simply SOR − 1).

The present findings have been derived from several patients with mild to advanced PD. It is still necessary to analyze patients with a wider range of motor severity to confirm the superior discrimination capability of SOR over Kiocc. Such an analysis can offer a good opportunity to further validate SOR and Kiocc correlations with components of UPDRS motor scores in PD at different stages, as well as to compare the rate of change of these measures over time at different disease stages.

Traditionally, the multiple-time graphical approach (Patlak plot) has been used to estimate parameters of interest for FDOPA PET studies (13,14). The influx transport rate of FDOPA is given by the slope of the line when the ordinate is Cstriatum(t)/ Cplasma(t) and the abscissa is MathCplasma(τ)dτ/ Cplasma(t). A rigorous derivation of the mathematic relationship has been previously published (Eq. 4 (15)). A simplified treatment to show the relationships between KiFD, Kiocc, and SOR follows: MathEq. 1 for t > t*, where t* is defined as the time at which radiotracer activity in reversible tissue compartments reaches an effective steady state with that in the plasma pool.

Vo is a volume of distribution that includes tracer-exchangeable space and the plasma volume. Multiplying both sides of the equation by Cplasma(t)/Coccipital(t), we get: MathEq. 2 The plasma time-activity curve (TAC) refers to FDOPA only, whereas the occipital TAC includes both FDOPA and its metabolite 3-O-methyl-6-18F-fluoro-l-DOPA.

Multiplying the first term on the right-hand side by MathCoccipital(τ)dτ/ MathCoccipital(τ)dτ and rearranging the terms, we get: MathEq. 3 where MathEq. 4 and SOR is simply defined as the left-hand side of Equation 2 (some investigators prefer to use the ratio R(t), defined as the ratio of specific uptake in the striatum [total striatal activity minus occipital activity] divided by the occipital activity, which is equivalent to SOR − 1). Equation 3 has traditionally been used to estimate Kiocc and has the advantage of not requiring any blood sampling (16). Both FDOPA and its metabolite methyldopa cross the blood-brain barrier, and after a sufficiently long time (approximately 20–30 min after injection of FDOPA), Cplasma(t) and Coccipital(t) follow each other closely (both can be approximated by very slow changing of exponential functions or linear functions). Therefore, their ratio becomes approximately constant. Plots of Cplasma(t) and Coccipital(t), as well as their integrals, are shown in Figure 4 and empirically confirm the theoretic assumptions. Equations 3 and 4 show the relationship of the 3 parameters SOR, Kiocc, and KiFD and emphasize that these parameters are related to each other linearly.

FIGURE 4.
FIGURE 4.

(A) Time-activity curves for plasma FDOPA and occipital radiotracer concentration (combined FDOPA and its metabolites) are shown on same scale. (B) Ratio of plasma to occipital activity as function of time. (C) Ratio of integral of plasma activity to integral of occipital activity. occ = occipital; p = plasma.

CONCLUSION

In this study, we have compared the clinical usefulness of SOR and Kiocc in the accurate diagnosis of PD using PET and FDOPA. SOR is generally more sensitive than Kiocc in distinguishing PD patients from healthy subjects while giving a comparable correlation with objective measures of motor deficit. A single short scan at 95 min after tracer injection is sufficient in a clinical research environment dealing with patients with movement disorders. This simplification offers a powerful and economical alternative for investigating the disease mechanism and clinical correlation in longitudinal experiments.

Acknowledgments

We thank Ralph Mattachieri and David Bjelke at our cyclotron and PET imaging center for their excellent technical assistance. This study was funded in part by grants RO1 NS 35069 and P50 NS 38370 from the National Institutes of Health.

Footnotes

  • Received Dec. 31, 2001; revision accepted Jun. 4, 2002.

    For correspondence or reprints contact: Vijay Dhawan, PhD, Center for Neuroscience, North Shore-Long Island Jewish Research Institute, 350 Community Dr., Manhasset, NY 11030.

    E-mail: dhawan{at}nshs.edu

REFERENCES

  1. ↵
    Takikawa S, Dhawan V, Chaly T, et al. Input functions for 6-[fluorine-18]fluorodopa quantitation in parkinsonism: comparative studies and clinical correlations. J Nucl Med. 1994;35:955–963.
    OpenUrlAbstract/FREE Full Text
  2. ↵
    Vingerhoets FJ, Schulzer M, Ruth TJ, Holden JE, Snow BJ. Reproducibility and discriminating ability of fluorine-18-6-fluoro-L-dopa PET in Parkinson’s disease. J Nucl Med. 1996;37:421–426.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    Morrish PK, Rakshi JS, Bailey DL, Sawle GV, Brooks DJ. Measuring the rate of progression and estimating the preclinical period of Parkinson’s disease with [18F]dopa PET. J Neurol Neurosurg Psychiatry. 1998;64:314–319.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    Hoshi H, Kuwabara H, Leger G, et al. 6-[18F]fluoro-L-dopa metabolism in living human brain: a comparison of six analytical methods. J Cereb Blood Flow Metab. 1993;13:57–69.
    OpenUrlPubMed
  5. ↵
    Ishikawa T, Dhawan V, Chaly T, et al. Clinical significance of striatal DOPA decarboxylase activity in Parkinson’s disease. J Nucl Med. 1996;37:216–222.
    OpenUrlAbstract/FREE Full Text
  6. Ishikawa T, Dhawan V, Kazumata K, et al. Comparative nigrostriatal dopaminergic imaging with iodine-123-beta CIT-FP/ SPECT and fluorine-18-FDOPA/PET. J Nucl Med. 1996;37:1760–1765.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    Nakamura T, Dhawan V, Chaly T, et al. Blinded positron emission tomography study of dopamine cell implantation for Parkinson’s disease. Ann Neurol. 2001;50:181–187.
    OpenUrlCrossRefPubMed
  8. ↵
    Morrish PK, Sawle GV, Brooks DJ. Regional changes in [18F]dopa metabolism in the striatum in Parkinson’s disease. Brain. 1996;119:2097–2103.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    DeJesus OT, Endres CJ, Shelton SE, et al. Evaluation of fluorinated m-tyrosine analogs as PET imaging agents of dopamine nerve terminals: comparison with 6-fluoroDOPA. J Nucl Med. 1997;38:630–636.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    Fahn S, Elton R. Unified Parkinson’s disease rating scale. In: Fahn S, ed. Recent Developments in Parkinson’s Disease. Floral Park, NJ: Macmillan; 1987:293–304.
  11. ↵
    Dhawan V, Kazumata K, Robeson W, et al. Quantitative brain PET: comparison of 2D and 3D acquisition on the GE Advance scanner. Clin Positron Imaging. 1998;1:135–144.
    OpenUrlCrossRefPubMed
  12. ↵
    Vingerhoets FJ, Snow BJ, Lee CS, et al. Longitudinal fluorodopa positron emission tomographic studies of the evolution of idiopathic parkinsonism. Ann Neurol. 1994;36:759–764.
    OpenUrlCrossRefPubMed
  13. ↵
    Patlak CS, Blasberg RG. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data: generalizations. J Cereb Blood Flow Metab. 1985;5:584–590.
    OpenUrlCrossRefPubMed
  14. ↵
    Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab. 1983;3:1–7.
    OpenUrlCrossRefPubMed
  15. ↵
    Patlak C, Dhawan V, Takikawa S, et al. Estimation of striatal uptake rate constant of FDOPA using PET: methodological issues. In: Umera K, ed. Quantification of Brain Function: Tracer Kinetics and Image Analysis in Brain PET. Amsterdam, The Netherlands: Elsevier Science Publishers BV; 1993:263–268.
  16. ↵
    Brooks DJ, Salmon EP, Mathias CJ, et al. The relationship between locomotor disability, autonomic dysfunction, and the integrity of the striatal dopaminergic system in patients with multiple system atrophy, pure autonomic failure, and Parkinson’s disease, studied with PET. Brain. 1990;113:1539–1552.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Bookmark this article

Jump to section