Comparison of Medium- and High-Energy Collimators for ¹³¹I-Tositumomab Dosimetry

MATERIALS AND METHODS

To determine the half-life of ¹³¹I, serial measurements of phantoms containing Na¹³¹I were obtained with several different γ-cameras and collimators. Three different phantoms were used for this study: an anthropomorphic chest phantom (Data Spectrum Corp.), an abdominal torso phantom (Phantom Laboratory), and a contrast-detail phantom (Rollo phantom). The various chambers of the phantoms were filled from an ¹³¹I stock solution having a concentration of 37 MBq/100 mL (1 mCi/100 mL). Two spheres, 3 and 6 cm in diameter, were used in the abdominal torso phantom to simulate abnormal biodistribution and were filled with activity concentrations of 185 and 74 MBq/mL (5 and 2 mCi/mL), respectively. Tables 1 and 2 show the activity distributions in the various chambers of the phantoms used. The total activity in the chest and torso phantoms was approximately 99 MBq (2.7 mCi). The contrast-detail phantom was filled with water, and 37 MBq (1 mCi) of ¹³¹I were added.

Imaging was performed with 4 different γ-cameras (GE Millennium VG [GE Healthcare], Philips Skylight [n = 2], and Philips Vertex V60) and medium-energy general-purpose (MEGP) or high-energy general-purpose (HEGP) collimators. The γ-camera and collimator combinations used are provided in Table 3. Three sets of measurements were obtained with medium-energy collimators, and 2 sets of measurements were obtained with high-energy collimators. Collimator specifications for some of the systems are provided in Table 4. The GE Millennium VG camera was equipped with 15.9-mm-thick NaI crystals, whereas the Philips cameras were equipped with standard 9.5-mm-thick NaI crystals.

Anterior and posterior whole-body images were acquired with the same acquisition protocol as that described for ¹³¹I-tositumomab by Wahl et al. (5). Imaging was performed at 2, 45, 120, 138, 167, 188, and 285 h after the administration of activity into the phantoms. For each γ-camera and collimator combination, a whole-body acquisition was obtained with all 3 phantoms placed together on the table. Anterior and posterior whole-body images were acquired over a 195-cm scan length at a scan speed of 20 cm/min with 20% energy windows centered around 364 keV. Anterior head height was set to approximately 2 cm above the phantom. This distance was maintained for all subsequent scans. After the scans, the phantoms were removed, and a background scan was acquired with the same parameters.

From each imaging session, the total counts were obtained from the anterior and posterior phantom and background images. The background-corrected geometric mean counts (C) were computed by subtracting the anterior and posterior counts from the background images (B_ant and B_post, respectively) and the anterior and posterior counts from the phantom images (C_ant and C_post, respectively), multiplying the resulting 2 counts, and taking the square root, as follows:Eq. 1The background-corrected geometric mean counts were recorded and plotted against the elapsed time from the initial activity calibration to produce a set of time–activity curves for each γ-camera and collimator combination. These data were used to determine exponential equations of best fit of the formEq. 2with Microsoft Excel (Microsoft Corp.); in this equation, A(t) is activity at time t, A₀ is initial activity (activity at time t = 0), and −λt is product of decay constant and time. From these calculated fit parameters, the half-life of ¹³¹I for each γ-camera and collimator combination was obtained as follows:Eq. 3

A similar analysis was performed for anterior-only and posterior-only images to investigate whether there was any difference in the measured ¹³¹I half-life for single-head γ-cameras. An analysis of the measured half-lives at the 2 NaI crystal thicknesses (15.9 and 9.5 mm) was also performed. Images were also evaluated qualitatively to assess the detection of the high-activity sphere in the abdominal torso phantom and other objects in the phantoms.

RESULTS

The time–activity curves from the whole-body imaging sessions are shown in Figure 1. There was a clear separation between the curves for medium-energy (upper 3 curves) and high-energy (lower 2 curves) collimators, reflecting the different sensitivities of the collimators. Figure 2 shows the data from Figure 1 normalized to the counts at time 0 (A₀ in Eq. 2), derived from the line-of-best-fit equation, for each γ-camera and collimator combination.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Time–activity curves from whole-body imaging.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

¹³¹I decay factors from whole-body measurements. Curves shown are data from Figure 1 normalized to activity at time 0 (A₀ in Eq. 2), derived from line-of-best-fit equation.

The half-lives measured from whole-body scans are shown in Table 5. The average measured half-life was 8.15 d (SD, 0.07 d), which is different by 1.6% from the NuDat database (7) reference value of 8.021 d. The average measured half-life for just the high-energy collimators was 8.12 d (SD, 0.11 d), and that for just the medium-energy collimators was 8.18 d (SD, 0.04 d). The half-life was not found by a Student t test to be statistically significantly different between the medium-energy and high-energy collimators (P = 0.44). The observed half-life difference with crystal thickness (8.09 d for 15.9 mm vs. 8.19 d for 9.5 mm) also was not found to be statistically significant (P = 0.06).

To examine whether single-head γ-camera measurements had any impact on the measured ¹³¹I half-life, the same analysis was performed for anterior and posterior whole-body scans separately. Anterior and posterior ¹³¹I half-lives were calculated for anterior and posterior projections for each camera and collimator combination and are summarized in Table 6. For all camera and collimator combinations, the ¹³¹I half-life measured with the anterior head was 8.13 d (SD, 0.13 d), and that measured with the posterior head was 8.18 d (SD, 0.13 d). There was no statistically significant difference between the half-life measured with the γ-camera in the anterior orientation and that measured with the γ-camera in the posterior orientation (P = 0.6 for all camera and collimator combinations, P = 0.52 for MEGP, and P = 0.34 for HEGP).

In ¹³¹I-tositumomab imaging, the administered therapeutic dose is determined by measurement of the TBRT with serial whole-body imaging. To investigate how medium-energy collimators affect the calculated administered therapeutic dose, we considered some data from a patient treated with ¹³¹I-tositumomab. For that particular patient receiving ¹³¹I-tositumomab therapy at our institution, the measured TBRT was 114 h, corresponding to an effective half-life of 79 h. From this finding, the biologic half-life (t_B) of ¹³¹I-tositumomab can be determined as follows:Eq. 4In this equation, t_E is the effective half-life and t_P is the physical half-life. On the basis of a physical half-life of 192.5 h (8.021 d), the biologic half-life is 134 h.

Rearranging Equation 4 to produce Equation 5 and using the measured ¹³¹I half-life for each collimator as the physical half-life, we calculated the effective half-life for each γ-camera and collimator combination (Table 7). Because t_P was larger than t_B, small changes in t_P were not expected to have a significant effect on t_B (1/t_P was small compared with 1/t_B), as shown by the following equation:Eq. 5 Table 7 also shows the prescribed therapeutic dose of ¹³¹I-tositumomab for the calculated TBRT values (range, 2.58–2.6 GBq).

Whole-body images from the first day of phantom imaging with the GE Millennium VG are shown in Figures 3 and 4. On the left of each pair is the anterior image, and on the right is the posterior image. Qualitative comparisons between medium- and high-energy collimator images revealed several expected differences. The background in the medium-energy collimator images was much higher than that in the high-energy collimator images. In the medium-energy collimator images, the counts detected outside the phantom were attributable primarily to septal penetration. In both the medium- and the high-energy collimator images, the 2 spheres used to simulate altered biodistributions were clearly visible in the anterior and posterior projections. The visibility of the 3-cm sphere was slightly degraded in the anterior MEGP image compared with the anterior HEGP image, but the sphere was still easily detected.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3. 

Anterior (A) and posterior (B) whole-body images from GE Millennium VG with MEGP collimator.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4. 

Anterior (A) and posterior (B) whole-body images from GE Millennium VG with HEGP collimator.

DISCUSSION

The simplified whole-body dosimetry method described by Wahl et al. (5) provides a direct means of performing ¹³¹I-tositumomab dosimetry, which may be readily and accurately performed at institutions lacking expert technical staff, who would otherwise be required to perform more involved quantitative dosimetry. Because ¹³¹I is involved, institutions and clinics without high-energy collimators would normally be excluded from performing ¹³¹I-based RIT procedures with the whole-body conjugate-view method. Our findings demonstrate that ¹³¹I residence time measurements for RIT dosimetry can be obtained by use of cameras with medium-energy collimators without a significant loss of accuracy in determining TBRT.

Whole-body images for the clinical use of ¹³¹I-tositumomab were also evaluated qualitatively to detect signs of altered biodistribution (i.e., altered liver or spleen uptake, thyroid uptake indicating free iodine or renal obstruction). Although ¹³¹I imaging with medium-energy collimators does not produce the highest-resolution images, the ¹³¹I-tositumomab imaging task is not one that requires high contrast or spatial resolution. Despite the production of lower-contrast images with medium-energy collimators, any alterations to the ¹³¹I-tositumomab biodistribution significant enough to affect the clinical decision to administer the therapeutic dose should still be readily detected. This scenario is demonstrated by the visualization of the 3- and 6-cm spheres in the abdominal torso phantom. Image quality was better in the posterior image than in the anterior image (typical of whole-body scans, in which the posterior detector is usually much closer to the patient); therefore, for single-detector cameras, posterior whole-body imaging is recommended. The similarity in the measured half-life between anterior (8.13 d) and posterior (8.18 d) detectors also suggests that distance from the patient is not a significant factor in TBRT measurements.

CONCLUSION

Serial whole-body scans of 3 phantoms performed on 4 different γ-cameras with medium- and high-energy collimators yielded measured ¹³¹I half-lives of 8.042–8.214 d. The mean half-life for all γ-camera and collimator combinations was 8.15 d (SD, 0.07 d). In single-head configurations, the results were similar, with measured half-lives of 8.13 d (anterior) and 8.18 d (posterior) for all γ-camera and collimator combinations. The measured half-lives at different crystal thicknesses were also similar (8.09 d for 15.9 mm and 8.19 d for 9.5 mm). These values are quite comparable to the reference value of 8.021 d from the NuDat database. The mean half-life measured with medium-energy collimators was 8.18 d, and that measured with high-energy collimators was 8.12 d; this difference is not statistically significant. The detection of altered biodistribution is not compromised when medium-energy collimators are used. On the basis of these results, for the purposes of measuring patient-specific TBRT and visual assessment of images to detect alterations in biodistribution, scanning patients administered ¹³¹I-based RIT agents (such as ¹³¹I-tositumomab) with medium-energy collimators is as accurate as scanning with high-energy collimators in either single- or dual-head configurations.