In this retrospective study the toxicity, dosimetry and efficacy of salvage PRRT using
177Lu-DOTATATE was analyzed in 35 metastasized and multimodally pretreated NET patients. Previously published data on salvage PRRT reported a tolerable toxicity with promising treatment efficacy [
12,
14]. Nonetheless, hematological and renal toxicity are still regarded as the main side effects and dose limiting factors for PRRT [
11] and are of particular concern in the salvage setting. Although we observed a significant decrease in WBC, erythrocytes and platelet count both after initial and salvage PRRT, blood values recovered and were still within the limits of current PRRT guidelines. Recently Löser et al. presented similar results with significantly decreasing platelets and WBC in 30 patients undergoing
177Lu-DOTATATE therapy with 17 and 13 patients receiving a cumulative therapy activity of ≤29.6 GBq (conforming to ≤4 PRRT cycles with 7.4 GBq) or > 29.6 GBq (conforming to salvage PRRT), respectively [
13]. Interestingly, platelet counts of ≥399,000 cells/μl was associated with worse survival, which was not observed in our cohort. In our analysis one patient (2.9%) showed reversible grade 3 hematotoxicity after salvage PRRT (anemia with reduced hemoglobin to 6.8 g/dl). Studies with larger patient numbers describe similar rates of grade 3/4 hematotoxicity. Recently, Garske-Román et al. presented data on 200 patients receiving
177Lu-DOTATATE therapy according to a dosimetry based study protocol with the goal to not exceed a cumulative kidney dose of 23 Gy [
25]. Grade 3 or 4 hematotoxicity was observed in 30 patients (15%). Bergsma et al. observed grade 3/4 hematotoxicity in 34 of 320 patients (11%) receiving up to 29.6 GBq of
177Lu-DOTATATE. Data on patients receiving ≥8
177Lu-DOTATATE therapy cycles have been published recently by Yordanova et al. in 15 patients [
15] and by van der Zwan et al. in 13 patients within a larger cohort of 181 salvage PRRT patients receiving ≥5 cycles [
14]. Yordanova et al. observed grade 3/4 hematotoxicity in 4 patients (23%) between PRRT cycles 1 and 4, in 3 patients (13%) between cycles 5 and 8 and in none of 9 patients between cycles 9 to 13. Similarly van der Zwan et al. observed grade 3 /4 hematotoxicity in 13 patients (7.2%) after salvage therapy (cylces 5 and 6) and in 1 patient (7.7%) after re-salvage therapy (cycle 7 and 8). Regarding nephrotoxicity, no grade 3 / 4 toxicity was obversved in our cohort. The median annual decrease in TER of 0.03 ± 0.07 (2.25 ± 0.48%) normed to the lower limit is comparable to data by Werner et al. [
23]. Contrary to their findings we could not detect a significant correlation between high initial median TER and a decrease in TER. In our cohort the decrease of absolute TER before initial and after salvage PRRT was higher than reported by Werner et al. indicating a preserved kidney function in their cohort. This might be explained by the larger number of therapy cycles in our salvage PRRT cohort. Kidney dosimetry revealed a mean cumulative absorbed dose of 23.8 ± 6.5 Gy in the salvage PRRT setting without correlation with loss of TER. This indicates that salvage PRRT with a median of 6 cycles results in an absorbed kidney dose which is within the limit of 23 Gy, established as the maximum tolerated dose by means of external beam radiation. In a large group of 323 NET patients Bergsma et al. observed no grade 3 or 4 nephrotoxicity or annual decrease of renal function > 20% and calculated a mean kidney dose of 20.1 ± 4.9 Gy in 228 patients by planar scintigraphy [
26]. They concluded that the threshold of 23 Gy seems to be too low. Low rates of Grade 3 and 4 nephrotoxicity even in the salvage setting support this conclusion [
14,
15]. However, as the kidneys are still considered as the dose limiting organs, SPECT based dosimetry protocols as presented by Garske-Román et al. represent interesting approaches, particularly as tumor based dosimetry is not feasible in NET [
25]. In contrast to other authors we found a significant difference in cumulative absorbed doses of different locations of metastases with higher doses in liver and lymph node metastases compared to bone and peritoneal lesions [
13]. Supporting the thesis of a tumor sink effect in NET [
27], we found a significant inverse correlation of increasing mean kidney doses and decreasing mean tumor doses over the course of initial and salvage PRRT. This supports the thesis that lower mean absorbed tumor doses lead to higher mean absorbed kidney doses. However, the high number of other therapy options including liver targeted therapies might also have an impact on SSTR expression and absorbed dose in tumors. Further trials with higher patient numbers and stratification according to previous therapies are needed to answer this question.
Different response assessment criteria have been applied for therapy monitoring in NET patients after PRRT. RECIST 1.1 has been validated for NET patients within the NETTER-1 trial [
9]. Nonetheless, response data specifically for the salvage PRRT setting are reported relatively rarely. Compared to other studies our response rates seem considerably lower. A comparative presentation is given in Table
5. This might be explained by the multimodal treatment of NET patients in our center, which is based on an interdisciplinary tumor board decision. Salvage PRRT patients included in the cohorts of Sabet et al. and van der Zwan et al. did not receive additional therapies prior to salvage PRRT. For instance, in our cohort a considerably high number of patients received local ablative liver therapies not only prior to the first PRRT (34.4%) but additionally prior to salvage PRRT (25.7%). Other intermediate therapies included surgery, bone targeted therapy, everolimus and chemotherapy. Thus salvage PRRT was not the first therapy option at the time of progression after initial PRRT but third or fourth line in many cases. Median PFS after salvage PRRT at a median follow-up of 25 months was only 6 months with 8 patients being censored at the time of analysis. Median PFS after four initial PRRT cycles was 33 months and within the range of 27–40 months as described in other studies [
7,
8,
25]. This could again be explained by the relatively late sequence of salvage PRRT in a considerably high number of patients in our cohort. Other groups report median PFS after salvage PRRT ranging from 13 to 18.9 months [
12,
14,
15]. The only difference compared to these studies is the multimodal treatment between initial PRRT and salvage PRRT in our patient cohort. Garske-Román et al. present a similar patient cohort to ours [
25]. However the therapy sequence is not further specified and local ablative liver therapies were not included. Additionally the PFS has been described for the whole patient cohort with respect to the cumulative kidney dose and not the number of therapy cycles. Thus the PFS of salvage PRRT patients is not described. However, in 39 patients receiving 5 or more cycles of
177Lu-DOTATATE, a cumulative dose of 23 Gy to the kidneys was reached only in 26 patients and 19 had died. However, they defined the common threshold of 23 Gy to be associated with a higher survival regardless of the number of therapy cycles [
25]. In our cohort univariate Cox Hazard analysis revealed a tendency towards higher overall survival in patients with a mean cumulative absorbed kidney dose higher than 19.5 Gy (
p = 0.052) compared to patients who received lower doses. As PRRT is mainly performed with a standardized dose of 7.4 GBq per cycle the added value of individualized dosimetry is still a matter of debate. Our study and the data by Garske-Román et al. show that higher absorbed kidney dose are associated with higher survival and remission rates. Personalized dosimetry enables to aim for the maximum of the tolerable dose to receive the best possible outcome.
Table 5
Best response after salvage PRRT
This study | 35 | 5–8 | 25 | 1 (3.1) | 26 (81.6) | 5 (15.3) |
| 33 | 5–8 | 23 | 7 (21.2) | 14 (42.4) | 11 (33.3) |
| 168 | 5–8 | 30.4 | 26 (15.5) | 100 (59.5) | 33 (19.6) |
One limitation of this study is the retrospective, monocentric study design. The patient cohort is relatively small and although the heterogeneity of different treatment options during the course of disease appears to have a major impact on response and survival, the interpretation is limited. Furthermore patients in our cohort present with various NET primary tumors with different tumor characteristics in terms of agressiveness and prognosis. This particularly applies for lung NET with typical and atypical histological variants, which have high impact on therapy response and surivival. Due to the small sample size, a subgroup analysis was performed only in small intestinal NET and demonstrated longer PFS compared to other primary tumors (data not shown). However, such data needs validation in larger cohorts, such as recently described by van der Zwan et al. [
14]. Nonetheless it could be demonstrated that the majority of our heterogeneous patient population shows disease stabilization in terms of PR and SD after multimodal therapy and progression.