Dissimilar DNA Damage to Blood Lymphocytes After ¹⁷⁷Lu-Labeled DOTATOC or Prostate-Specific Membrane Antigen Therapy

DISCUSSION

In this study, the largest to date on this topic, we have shown the feasibility of quantifying RIFs in the blood lymphocytes of patients undergoing either ¹⁷⁷Lu-DOTATOC or ¹⁷⁷Lu-PSMA therapy. The literature on RIFs abounds because of radiographic or teletherapeutic procedures, but we know of just 3 prior studies investigating RIFs due to ¹⁷⁷Lu-labeled radionuclide therapies (14–16), contributing 43 patients in total. Two reports (14,16) have addressed ¹⁷⁷Lu-DOTATOC/DOTATE therapy in a pool of 27 patients, and another (15) examined 16 recipients of ¹⁷⁷Lu-PSMA. Our cohort of 48 patients is thus the largest population consecutively enrolled in a single-center study of RIFs after ¹⁷⁷Lu-labeled radioligand therapy. To facilitate comparison of results, we measured absorbed dose and dose rate as in the largest related studies heretofore conducted, by Eberlein et al. (19) and Schumann et al. (20).

Once ¹⁷⁷Lu-labeled radioligand was administered, the RIF count increased rapidly from DNA damage, peaking at 5 min (nominally) after full injection (Fig. 1). This finding is corroborated by studies involving x-ray imaging, indicating that RIF counts maximize within several minutes (5,8,12–14). However, other reports of ¹⁷⁷Lu-labeled pharmaceuticals seem contradictory in terms of RIFs. Denoyer et al. (21) have recorded maximum RIF counts for ¹⁷⁷Lu-DOTATATE at 2 h after injection and sometimes at 30 min after injection, whereas Eberlein et al. (19) and Schumann et al. (20) cite RIF maximums occurring at several hours after injection. The reason for such discrepancies is unclear.

One possible explanation is variation in the delivery procedures used. Similar to steps taken in other studies, we infused agents directly from their vials over periods of about 20 min. There was thus a gradual dilution, responsible for tapering of activity concentration after an initial peak. Because RIFs also correlate with dose rate and thus with activity concentration, a RIF peak was reached earlier by us than by other groups. In studies of other groups, large syringes of radiopharmaceuticals and syringe pumps were used, resulting in more even dispersion of activity concentration and dose rate to blood. Also, our first blood sampling (5 min after injection) was earlier than in other studies, although this 5-min time point was only nominal, since the exact time of blood sampling was not available for the 5- and 30-min-postinjection time points. Although accuracy was the aim, the actual time of blood sampling could have been delayed in many patients because of workload and task complexity at the start of infusion. In actuality, a sampling time point of between 5 and 30 min after injection was likely. Together with our slower infusion method, the 5-min time point might be comparable to a time point of about 15 min after injection in a protocol with a faster infusion rate. On the other hand, in literature comparisons of activity concentration as derived from the blood time–activity curve fit at 10 min after injection, the mean of 236 ± 108 kBq/mL in our study is similar to representative patients reported by Delker et al. (22) for ¹⁷⁷Lu-PSMA and Sandström et al. (23) for ¹⁷⁷Lu-DOTATATE. This finding could indicate that we did not considerably underestimate time-integrated activity and dose contributions from the early phase after infusion.

RIF counts gradually declined from peak levels over time (Fig. 1). However, they were significantly higher than at baseline levels at most time points tested. Only a few very late time points failed in this regard, further disadvantaged by low patient numbers. This general pattern of RIFs is affirmed in other studies (19–21), where RIF counts dwindle and approach pretherapeutic levels. Any disputes over whether or when baseline and late RIF counts become similar may be the fault of insufficient sampling, given the considerable interpatient variability displayed.

When comparing subgroups of patients treated with ¹⁷⁷Lu-DOTATOC and ¹⁷⁷Lu-PSMA, we found the RIF count at 5 min after injection to be significantly higher for the ¹⁷⁷Lu-PSMA group, whereas at 72 h after injection, the count was significantly higher for the ¹⁷⁷Lu-DOTATOC group. Average injected activities were similar (¹⁷⁷Lu-PSMA: 6,300 ± 1,377 MBq; ¹⁷⁷Lu-DOTATOC: 6,420 ± 1,263 MBq). The higher early RIF count for ¹⁷⁷Lu-PSMA is attributable to a higher activity concentration in the blood (reflecting increased dose rate), which our data support (Fig. 4). Unfortunately, the higher RIF count for ¹⁷⁷Lu-DOTATOC at 72 h after injection is not similarly explained. In fact, the dose rate for the ¹⁷⁷Lu-DOTATOC group was lower than that for the ¹⁷⁷Lu-PSMA group at every time point. Moreover, the dose rate for ¹⁷⁷Lu-PSMA (vs. ¹⁷⁷Lu-DOTATOC) was about 60%–90% higher until 4 h after injection, whereas RIF count was only marginally higher by comparison. These inconsistencies raise the issue of a potential lapse in the dosimetry model when calculating blood dose and dose rate. For example, the current model applied here by us and regularly by others does not take into account the relatively high splenic uptake of ¹⁷⁷Lu-DOTATOC, which may add to irradiation of blood cells and raise RIF counts at later time points. Like others before us, we see that the γ-H2AX determination could help in improving and validating existing dosimetry methods and dose limits, which in turn has the potential to increase clinical acceptance of dosimetry. There is an indication that personalized, dosimetry-based dosing could lead to increased tumor doses (24) and better response rates (25). Ultimately, a 3-dimensional determination of γ-H2AX RIF by PET or SPECT, as O’Neill et al. demonstrated in a preclinical setting recently (26), has the potential to improve dosimetry models for many tissue types.

We found a strong and not unexpected correlation between RIF count and dose rate in all patients, as well as in ¹⁷⁷Lu-DOTATOC and ¹⁷⁷Lu-PSMA treatment subgroups. Indeed, past publications have shown that newly induced foci are repaired within reasonably short time-frames. According to Horn et al. (18), the decay in RIFs is biexponential, with a half-life of 1.4–1.7 h for the predominantly fast component. This time is considerably shorter than the interval between most of our sampling points. In effect, RIF totals rely heavily on numbers of continuously created foci and, thus, on dose rates. The only other study describing this effect is one by Schumann et al. (20), claiming such a correlation for later (≥48 h) time points only. We are unaware of any other research on the correlation of dose rate and RIF counts in the setting of ¹⁷⁷Lu-labeled radiotherapeutics.

In examining absorbed dose and RIF count, we found a correlation until 4 h after injection (no later), as is explained by ongoing reparative processes in which detectable RIFs decline as rates of repair exceed rates of injury. The absorbed dose is a monotonously rising parameter and thus will not linearly correlate with RIF counts over all time points. These tenets are supported by Eberlein et al. (19) and Schumann et al. (20), who documented linear correlations until 5 and 2.6 h after injection, respectively. A potential enhancement for the modeling of interplay between absorbed dose and RIFs might be the incorporation of repair rates, as instituted by Eberlein et al. (19) for ¹⁷⁷Lu-DOTATOC and Mariotti et al. (27) for x-ray irradiation.

It is notable that for ¹⁷⁷Lu-PSMA, absorbed dose to the blood was significantly higher in our study (0.030 ± 0.011 mGy/MBq) than in an effort by Schumann et al. (20) (extrapolated as ∼0.0136 mGy/MBq) at comparable time points (48 h after injection). On the other hand, the absorbed dose of ¹⁷⁷Lu-DOTATOC reported here (0.017 ± 0.009 mGy/MBq) was similar to that (0.011 mGy/MBq) documented by Eberlein et al. (19). Because the dosimetry methods were comparable, we can only speculate about the basis for this departure in ¹⁷⁷Lu-PSMA values. Bias in determining blood activity concentration, an otherwise pivotal factor in absorbed dose to blood, is a possibility. Also, Schuman et al. used a PSMA variant (PSMA I&T; Scintomics GmbH), as opposed to our use of ¹⁷⁷Lu-PSMA-617 (ABX Advanced Biochemical Compounds), which may not share blood clearance properties. Differing patient populations may be responsible for these deviations as well. For instance, a substantially higher tumor burden in one study group would likely reduce circulating time and associated residual activity in the blood.

Ionizing radiation is a well-known cause of dose-dependent deteriorations in all hematopoietic cell lines (28). The time course after exposure depends on cell type. As a rule, it is quickest for lymphocytes (hours), followed by granulocytes (days), and erythrocytes (weeks). Thrombocytes and platelets decline within days after exposure (29). The extent to which RIF counts indicate deterministic and stochastic radiation damage is still debated (30). In the context of ¹⁷⁷Lu-labeled therapies, RIFs in blood lymphocytes are presumed to positively correlate with increased DNA damage, potential DNA disrepair, and functionally impaired cells. We have also shown that the dose rate in blood, which is chiefly driven by blood activity concentration, correlates with RIFs. One may therefore assume that RIF increments in blood lymphocytes are indicative of damage to other hematopoietic cells and their derivatives. In our patient population, we did find a significant negative correlation (P < 0.05) between short-term decline (3 d of follow-up) in thrombocyte counts of peripheral blood and RIF counts of lymphocytes, suggesting that damage on a microscopic level (number of DSBs) may eventually impact a macroscopic system (blood parameters). Only one other study has seemingly corroborated this observation. Denoyer et al. describe a relation between peak RIF counts in the first 4 h after injection and changing numbers of peripheral blood lymphocytes (21). Unfortunately, we could not test for therapy-related blood lymphocyte changes, because complete blood cell counts with differentials were unavailable in most of our patients.