Subacute haematotoxicity after PRRT with ¹⁷⁷Lu-DOTA-octreotate: prognostic factors, incidence and course

Subacute haematological grade 3/4 toxicity was observed in 34 (11 %) of 320 patients receiving ¹⁷⁷Lu-DOTATATE. In half of these patients, toxicity persisted for more than 6 months or blood transfusions were required. This is in accordance with data from other groups [5‐10]. Long-term haematological toxicities, such as myelodysplastic syndrome (MDS) and acute leukaemia (AL), have been found in patients receiving PRRT with ¹⁷⁷Lu-DOTATATE [6, 11]. Also in our patient group, MDS and AL were observed, but since these events have a rare complex stochastic character, they will be reported in a separate study. In a recent study, long-term side effects of PRRT with ⁹⁰Y-octreotide and/or ¹⁷⁷Lu-octreotate were investigated [6]. The authors found more haematological toxicity after PRRT in patients with baseline nephrotoxicity (transient or persistent elevation in creatinine). The prolonged circulation time of ¹⁷⁷Lu-DOTATATE in patients with a poor renal function is probably the most important factor that explains the increased toxicity to the BM, as shown by Svensson et al. [20]. Also anaemia is common in these patients due to a reduction in renal erythropoietin production. In our study, poor renal function was also found to be a predictor of haematological toxicity. We also found that a low baseline WBC count is a predictor of grade 3/4 haematological toxicity. This is in line with the findings of a recent study showing that baseline cytopenia is a predictor of haematological toxicity after PRRT [9].

From a dosimetric point of view, it has been theorized that patients with a large tumour burden and high receptor density have lower amounts of circulating activity [21]. Therefore the radiation to normal tissues could be less than in patients with a low tumour burden. However, in our study, patients with more tumour mass at baseline were significantly more likely to have grade 3/4 haematological toxicity. Furthermore, high tumour burden was more frequently found in patients with persistent grade 3/4 haematological or who required blood transfusion. This suggests that tumour burden plays a role in the development and duration of haematological toxicity, in contrast to the recently described tumour sink effect [22]. That study showed that internalization of ⁶⁸Ga-DOTATATE in the tumour leads to a significant decrease in uptake in healthy tissue, the so-called tumour sink effect [22]. The authors extrapolated their results and speculated on a similar effect for PRRT. However, the main contributing factor for radiation dose to an organ (i.e. the BM) from PRRT is the exposure to radiation over time and the organ dose cannot be based on a distribution with only one time point. In our limited subgroup of patients in whom biodistribution and dosimetry data were available, we were not able to demonstrate the tumour sink effect.

Past chemotherapy was not a clear risk factor in our analyses. This finding can be explained by the limited number of patients who received chemotherapy in our series (38 of 320 patients, 12 %). In other studies more than 25 % of patients have had a history of chemotherapy [6, 8, 9]. In particular, chemotherapy regimens with alkylating agents (e.g. cisplatin, carboplatin, oxaliplatin) or topoisomerase II inhibitors (e.g. etoposide) can damage the DNA of haematopoietic cells [23]. A small number (12 %) of patients in our study had received this type of chemotherapy in the past, therefore limiting the statistical power of this (possible) risk factor.

A potential protective effect of splenectomy on the development of haematological toxicity in patients receiving ¹⁷⁷Lu-DOTATATE has recently been reported [9]. The spleen is a major reservoir of blood cells and uptake of radioactivity is mainly caused by the presence of somatostatin receptors on lymphocytes [14, 24]. Blood cells circulating throughout the spleen may be damaged leading to a reduction in peripheral blood cell counts. However, in another clinical study with ¹⁷⁷Lu-DOTATATE/DOTATOC, no correlation was found between dose to the spleen and haematological toxicity during PRRT [25]. In our study, none of the 12 patients who had had a splenectomy developed grade 3/4 haematological toxicity. This supports the idea that splenectomy has a protective effect in patients receiving PRRT, but statistical analysis could not be performed due to the limited number of patients with splenectomy.

PRRT using ¹⁷⁷Lu-DOTATATE shows similarities to ¹³¹I treatment, because of the comparable half-life (6.7 and 8.0 days, respectively) and similar energies of the emitted β radiation (with average energies of 133 and 182 keV, respectively). Therefore in 2000, we accepted a maximum BM dose for PRRT with ¹⁷⁷Lu-DOTATATE, which was adopted from clinical studies with ¹³¹I treatments. The upper BM dose limit was set to 2 Gy, based on early work in thyroid cancer patients treated with ¹³¹I [12]. In that study, 122 doses of ¹³¹I were administered to 59 patients with metastasized thyroid cancer. In 14 administrations of ¹³¹I, serious radiation complications were observed (Table 3, original table). The authors stated that serious radiation complications per treatment cycle were more frequent when the total dose to the blood exceeded 200 rad (2 Gy) with a significance of P = 0.03 (Table 3). However, only 8 of 14 serious radiation complications were related to the BM; the other complications were pneumonitis or vomiting. When we repeated the analysis, taking into account only serious BM complications (Table 3, modified table), no significantly higher frequency of haematological toxicity for BM radiation doses more than 2 Gy could be demonstrated (Fisher’s exact test, P = 0.68). Also, all eight patients with radiation complications (related to the BM) had metastatic disease to the bone. Bone metastases are a source of radiation after PRRT and could contribute to an increase in BM dose. However, in our multivariate analysis, the presence of bone metastases was not a risk factor for developing grade 3/4 haematological toxicity.

Another comment on the study by Benua et al. is that the radiation complications per unit administered dose were analysed and the cumulative BM dose per patient was not considered. This resulted in a double count of radiation complications in one patient (after the first and second dose of ¹³¹I). In our study, haematological toxicity was counted only once per patient since the chance of recurrent haematological toxicity in one patient is relatively high. We also analysed the cumulative BM dose instead of complications per administration, considering that multiple sequential treatments reduce the BM reserve. In a more recent article, the BM limit was set to 3 Gy for ¹³¹I treatment, based on 104 treatments in 83 thyroid cancer patients [26]. No permanent BM suppression was observed, but two patients required PLT and red blood cell transfusion because of pancytopenia.

In the past decade, several studies with BM dose estimates using ¹⁷⁷Lu-DOTATATE have been reported (Table 4). Variations in the reported BM dose estimates can partly be attributed to differences in accuracy of dosimetric methods [27]. In a recent Swedish study, 200 patients who were treated with 7.4 GBq ¹⁷⁷Lu-DOTATATE were analysed and BM doses were calculated based on blood-based and organ-based analysis of the whole-body images. The authors calculated a maximum BM dose of 0.4 Gy per cycle of 7.4 GBq, which would result in a cumulative BM dose of 1.6 Gy for four cycles. Our data showed an estimated mean BM dose of 2.0 Gy (SD 0.2 Gy) in 184 out of 320 patients who received four cycles of 7.4 GBq ¹⁷⁷Lu-DOTATATE. Therefore, half of these patients (92 of 184) received a BM dose of more than 2 Gy. If the true maximum tolerated dose to the BM were 2 Gy, these 92 patients would theoretically be more prone to develop haematological toxicity. However, we found haematological toxicity in only 34 of the 320 patients. This supports the idea that a higher BM dose limit for PRRT with ¹⁷⁷Lu-DOTATATE is appropriate. Another argument for a different BM dose maximum is the success of retreatment with extra cycles of PRRT without serious haematological side effects [28]. In our analysis of this type of retreatment, in which selected patients received a cumulative BM dose of approximately 3 Gy, only 5 (16 %) of 32 patients developed grade 3/4 haematological toxicity after two additional cycles of ¹⁷⁷Lu-DOTATATE. In another study reversible haematotoxicity (grade 3/4) was found in 7 (21.2 %) of 33 patients who underwent salvage PRRT [29]. In our ongoing study in Erasmus MC, we have treated a selected group of patients with multiple additional cycles of ¹⁷⁷Lu-DOTATATE with cumulative doses up to 59.2 GBq resulting in an estimated mean BM dose of more than 3 Gy with limited haematological toxicity (unpublished data).

Several groups have investigated the role of BM dosimetry in radionuclide treatments for predicting haematological toxicity. A weak negative correlation (r _p = −0.47) between neutrophils at nadir and measured whole-body absorbed dose was found in 20 patients treated with ¹³¹I-MIBG [30]. In a PRRT study with ⁹⁰Y-DOTATOC, a correlation (R = 0.58) was found between BM absorbed dose and PLT count reduction at nadir [31, 32]. BM dose was calculated in 12 patients based on plasma samples, assuming that the activity concentration in the BM was equal to that in the plasma. ⁸⁶Y-DOTATOC PET was performed after therapy and showed uptake in the vertebrae. Taking the radioactivity in the spine into the dosimetric calculations, a better correlation (R = 0.82) was found between BM absorbed dose and PLT count reduction at nadir. However, 24 patients did not demonstrate sufficient uptake of ⁸⁶Y-DOTATOC in the spine to provide usable BM dose measurements. In our study, we found similar correlation coefficients between the relative decrease in blood cells and (cumulative) radiation dose to the BM. We found no correlation between BM dose and Hb, which can be explained by minimal effects on circulating erythrocytes after 2 Gy of irradiation [33]. These weak correlations between decreases in blood count and BM dose indicate that current dosimetry cannot fully predict haematological toxicity. Additional clinical factors have to be taken into account to predict haematological toxicity in PRRT.

Further research should focus on reporting BM dose in patients receiving PRRT with ¹⁷⁷Lu-DOTATATE. BM dose limits should be explored at the population level with clinical toxicity grading (e.g. CTCAE) as outcome. However, current BM dosimetry is imprecise and varies due to differences in method of BM dose calculation [34]. Also the absorbed BM dose does not reflect the damage done to the haematopoietic stem cell department. If BM dosimetry can reflect the actual dose to haematopoietic stem cells, it will have a more prominent place in clinical practice during PRRT. In vivo markers might also be an option for assessment of BM status after irradiation with PRRT. In a recent article, γ-H2AX foci in lymphocytes were successfully used for monitoring ionizing radiation-induced DNA double-strand breaks in patients treated with ¹⁷⁷Lu-DOTATATE [35]. However, the response of γ-H2AX foci varied significantly between patients over time, making it less suitable for individual monitoring. In future, BM radiation dose could provide information for decision-making in a clinical setting, but at present BM dosimetry plays a minor role in clinical practice. Clinical parameters and blood cell count recovery are currently the most important criteria for individual PRRT planning.