Simplified dosimetry for kidneys and tumors in 177Lu-labeled peptide receptor radionuclide therapy

The results of this study confirmed previous findings that kidney-absorbed doses are relatively constant between treatment cycles, whereas tumor doses slowly decrease with every treatment cycle for 177Lu-DOTATATE therapy [18, 20]. The spread in absorbed doses was found to be much larger for tumors: one normalized standard deviation of the absorbed doses was 28% for kidneys and 75% for tumors. It was also found that absorbed doses, effective half-lives and uptake at 1 d were higher for tumors in patients diagnosed with small intestine NET G1 in comparison to G2. This is in line with the results of Roth et al. [20] who reported higher absorbed doses to G1 tumors when observing a dataset of tumors with different origins combined (pancreas, small intestine, lung and unknown). Roth et al. reported a larger decrease in absorbed dose to G2 tumors for every treatment cycle in comparison to G1 tumors which was not observed for the small intestine NET in the present study. No difference in dosimetric accuracy due to tumor grading was found for any of the approaches investigated in the present study.

As the kidney-absorbed doses were rather constant and the tumor-absorbed doses decreased with every treatment cycle, the dose-per-activity approach performed better for kidneys than for tumors, for which the absorbed dose was increasingly overestimated for each treatment cycle. For any of the investigated approaches in this study to be clinically implemented, the data points with the largest errors must be few and the errors must not be too large as this could result in serious adverse effects for such “outlier” patients [14]. Relatively large errors may be acceptable for single treatment cycles, but not when considering the total dose from all cycles. Thus, the dose-per-activity approach appear unfeasible also for kidneys as it resulted in 2 outliers out of 77 data points for which the total dose corresponded to only 66% and 70% of the “true” total dose. While the biological effects of such inaccuracies are difficult to quantify, such calculations cannot be considered reliable and are therefore not suitable for clinical implementation.

The most accurate single-point dosimetry protocol for kidneys was the 1 d SPECT images together with the fixed effective half-life of 52 h for which the maximum difference in total dose was − 18% in relation to the “true” dose. Interestingly, both approaches using the early SPECT images outperformed the approaches using the late SPECT images for the kidneys. A possible benefit of performing single-point dosimetry at 1 d post-injection is that the patient can undergo both treatment and imaging before leaving the hospital. Promising results for single-point dosimetry using early SPECT images have been reported before. Willowson et al. [16] reported an average deviation of 2% and a maximum difference of 45% when calculating doses using 1 d SPECT images with effective half-lives from the first cycle (corresponding average and maximum difference was 7% and 35% for all kidney data points in the present study). Also, Sundlöv et al. [15] reported average differences of 1 ± 17% (2 standard deviations) and − 2 ± 25% when using 1 d SPECT images and effective half-life from the first treatment cycle or a fixed effective half-life of 51.6 h. The corresponding values for all kidney data points in the present study were 7 ± 13% and 5 ± 10%, respectively.

For tumors, single-point dosimetry was most accurately performed using the late SPECT images. This is in agreement with Hänscheid et al. [11] who reported an increased accuracy in single-point dosimetry for tumors with increasing time post-treatment (Pearson correlation coefficient between approximated time integral and actual time integral of 0.99 at 144 h). It should be noted that the tumor-absorbed doses in the present study were significantly lower in comparison to the “true” doses when calculated using the late SPECT image together with the fixed effective half-life (Fig. 4). This indicates that, unlike for kidneys, single-point dosimetry for tumors is more accurate when using a tumor-specific effective half-life determined from the initial treatment cycle.

Single-point dosimetry for kidneys proved more accurate when using a fixed effective half-life of 52 h instead of a patient-specific value from the first treatment cycle or a dose-per-activity value determined from the first cycle. This could imply that the uncertainty in determining the effective half-life or dose-per-activity values for one treatment cycle is higher than the difference in kinetics between one individual patient and the population for which the fixed effective half-life was calculated. One drawback associated with the present study is that the “true” absorbed doses were calculated from only two data points (1 d and 7 d). Two data points are not sufficient to fully reproduce the associated kinetics of tumor volumes or kidneys, that are likely to behave more closely to a bi-exponential function. However, a third data point at, e.g., 4 d does not provide more information regarding the initial uptake but rather serves to improve the curve fit and reduce the impact from possible errors associated with the 1 d or 7 d image acquisitions (e.g., poor statistics, motion artifacts, etc.).

It should be noted that only SPECT/CT images were included in the present study removing all uncertainties related to the corrections needed when calculating absorbed doses from planar images. Also, scatter correction and CT-based attenuation correction were used on all images reducing potential errors associated with patient geometries. Regarding the small VOI approach used in the present study, it has been shown that absorbed doses can be systematically higher when calculated using maximum activity concentrations in the kidneys in comparison to using activities within a volume delineated on CT images [12]. This difference should however be smaller when using several VOIs throughout the kidneys as in the present study. The fixed effective half-lives of the kidneys and tumors used in the present study were chosen as the median values of all patients which may have limited generalization to other patient cohorts. However, the value of 52 h used for kidneys in the present study is in excellent agreement with published data [8, 14, 16, 18]. For tumors, less data are available for comparison and the variation in effective half-lives is expected to be larger. Roth et al. [20] reported average effective half-lives of 81 h and 103 h for G1 and G2 tumors, respectively. To assess the robustness of our single-point calculations, fixed effective half-lives were varied between 81 and 137 h. This resulted in median values of the total normalized absorbed doses (7 d images) of 1.06 (IQR: 1.04–1.07), 0.98 (IQR: 0.96–1.00) and 0.99 (IQR: 0.97–1.00) for fixed effective half-lives of 81 h, 109 h and 137 h, respectively. This indicates that the calculations are not sensitive to moderate variations in tumor effective half-lives.

In agreement with other studies [12, 14], our previous in-house evaluation showed that when reducing the number of SPECT images from three (1 d, 3 d and 7 d) to two acquisitions, the best accuracy was obtained when using the 1 d and 7 d images. Previous work on single-point dosimetry by Hänscheid et al. [11] suggests that absorbed doses can be determined with satisfactory accuracy for tumors and kidneys when using an image acquired 4 d post therapy. An optimal timepoint for kidneys at 4 d post therapy was also suggested by Sundlöv et al. [15]. When considering these findings together with the results of the present study it could be argued that if single-point dosimetry is to be performed, the images should be acquired at 1 d or 4 d if kidney doses are of interest, 4 d or 7 d if tumor doses are of interest, and at 4 d if both kidney and tumor doses are of interest. This suggests that the imaging schedule can be adjusted to the logistical reality of the clinic within this timeframe without compromising much accuracy in dosimetry.

In conclusion, our results show that single-point dosimetry is feasible for both kidneys and tumors. However, the 1 d SPECT image proved to be optimal for kidneys and the 7 d SPECT proved to be optimal for tumors (independent of tumor grading).

留言 (0)

沒有登入
gif