This registry study is, to our knowledge, the largest yet published of PSMA-targeted RLT in everyday practice and has the additional strength of relying on prospective observation. Our findings are notable in providing evidence of the encouraging efficacy and safety of 177Lu-PSMA-617 RLT even in “real-world” patients. Unsurprisingly, our patients appear to have had less favorable baseline characteristics, and to have received a less intense RLT regimen, compared to the active treatment groups in the multicenter, randomized controlled VISION and TheraP trials [10, 14].
Compared to VISION’s active treatment group (n = 551), our patients had greater prevalence of visceral metastasis (33% vs < 20%) and of markedly impaired general condition (ECOG performance status ≥ 2 in 37% vs 7% of patients). Our sample also had higher median baseline PSA and ALP levels (Table 4), while the cohorts’ histories of prior therapy appeared to be roughly alike [14]. VISION’s active treatment group received a median (minimum–maximum) cumulative 177Lu-PSMA-617 activity of 37.5 (7.0–48.3) GBq in 5 (1–6) cycles, versus 21.2 (5.1–77.8) GBq in 3 (1–13) administrations for our patients. Despite these differences, our study found a rate of ≥ 50% PSA decrease (52% vs. 46%) and median OS (14.5 vs. 15.3 months) that greatly resembled those of VISION patients given RLT.
Table 4 Comparison of key characteristics and results from the most important studies of 177Lu-PSMA-617 RLT in mCRPC highlighting efficacy in the post-taxane setting Compared to TheraP’s active treatment group (n = 99), our patients had a much greater prevalence of visceral metastasis as well ECOG ≥ 2 status, and had higher median baseline PSA and ALP. However, since TheraP used cabazitaxel as the comparator for RLT, our patients were more heavily pre-treated. The planned regimen in TheraP’s active treatment arm (43.5 GBq total in 6 administrations), envisioned a substantially higher cumulative activity in more cycles than were given to our patients. Nonetheless, the biochemical PR rate in the TheraP active treatment patients did not dramatically differ from ours (66% vs 52%).
Lower rates of grade 3/4 AEs were recorded in our patients than in VISION’s active treatment arm [14]. Additionally, while no toxicity-related 177Lu-PSMA-617 discontinuations or treatment-related deaths were seen in our cohort, there were 11.9% and 3.6% rates of these phenomena, respectively, in VISION’s RLT group. One may speculate that there were two important reasons for these differences. First, the rates reported by the VISION investigators appear to be for treatment-emergent rather than treatment-related toxicity, which is what we report. Secondly, as noted above, the VISION patients typically received more intense and lengthier 177Lu-PSMA-617 treatment than did our patients.
Interestingly, the 177Lu-PSMA-617 safety and tolerability profile in the TheraP active treatment group, appeared more like that in our patients (bearing in mind that again, treatment-emergent toxicity is being compared with treatment-related AEs). Most notably, TheraP had only a 1% rate of toxicity-related 177Lu-PSMA-617 discontinuation and no treatment-related deaths. Regarding the comparison of RLT safety in the two studies, one may speculate that the more intense RLT regimen in TheraP may have been counterbalanced by the less heavily pretreated nature of their active treatment group.
A notable efficacy-related finding of our study was that even after statistical adjustment for patient, disease, and prior treatment characteristics, early biochemical disease control by 177Lu-PSMA-617 RLT was significantly associated with OS; this finding confirmed and extended a similar, earlier, novel observation in a 28-patient subgroup of the present cohort, all of whom had liver metastasis [12]. Indeed, early biochemical progression was the strongest studied determinate related to OS in the 254-patient multivariable analysis, with a HR for mortality nearly double or more than double that of the other 4 significant negative prognostic factors identified (Table 2, Fig. 3B). The relationship of biochemical response to PSMA-targeted RLT with OS may suggest cause and effect regarding 177Lu-PSMA-617 treatment and the relatively lengthy OS in our late-stage/end-stage patients. This hypothesis is supported by the findings regarding comparative efficacy of 177Lu-PSMA-617 versus standard care in VISION [14] or versus cabazitaxel in TheraP [10].
Aligned with observations by others [7], another finding of our study that merits mention is that although patients with taxane failure (n = 188) benefitted from 177Lu-PSMA-617, RLT appeared to be more effective in patients who were taxane-naïve (n = 66) (Supplementary Table S5). For example, median (95%CI) PSA-PFS for the respective subgroups was 4.6 (3.9–5.2) months versus 9.2 (7.0–11.4) months, a difference that was statistically significant; median OS was 13.4 (10.4–16.5) months versus 21.6 (13.6–29.5) months, albeit this difference did not attain statistical significance. The seemingly better survival outcomes in the taxane-naïve subgroup may at least partly stem from their less aggressive disease course (reflected by significantly older age and lower prevalence of visceral metastasis) and more intact red marrow compartment (reflected by significantly higher median baseline hemoglobin concentration). Indeed, in our multivariable analysis, history of taxane therapy was not associated with OS (Table 2), although such history was significantly linked with PSA-PFS (Table 1). Nonetheless, our observations regarding taxane pretreatment, coupled with the seemingly comparable activity but much lesser toxicity of 177Lu-PSMA-617 RLT relative to other recently-introduced systemic agents, suggest that PSMA-targeted RLT merits formal study earlier in the mCRPC disease course. This hypothesis also is supported by the multivariable association in our study of early biochemical disease control and longer OS, and, more concretely, by TheraP’s observations of greater biochemical effect and lesser toxicity of 177Lu-PSMA-617 versus cabazitaxel [10].
We found that older age, better general condition (lower ECOG performance status), and lower baseline serum ALP concentration, also were significant independent determinates of longer PSA-PFS (Table 1). One may speculate that older age and better ECOG performance status may to at least some extent be surrogates for earlier-stage or less-aggressive disease, or less pretreatment. Serum ALP concentration is a marker of bone involvement, and thus when elevated, may reflect advanced-stage or more aggressive disease. One may hypothesize that the negative prognostic significance of taxane pretreatment regarding biochemical response to 177Lu-PSMA-617 RLT relates to the prior therapy selecting more resistant or aggressive clones.
Notable limitations of our study should be acknowledged. The most important include lack of a control arm, as well as the single-center nature of the REALITY Registry. Additionally, no fixed-activity protocol was used for 177Lu-PSMA-617 RLT. Instead,177Lu-PSMA-617 activities, and the number of and intervals between RLT cycles were individually chosen based on total tumor burden, sites of metastases, and the patient’s condition including ECOG performance status and organ function. While this strategy increased individualization of therapy, the variability in RLT regimens arguably renders our overall results harder to interpret. Another limitation of our work is that besides OS, this report does not consider efficacy endpoints unrelated to PSA, such as structural or molecular imaging, pain control, or quality-of-life changes. Regarding structural or molecular imaging, accurate response assessment according to Response Criteria in Solid Tumors [32] was not possible, because CT was performed without contrast enhancement and post-therapy restaging was done at different intervals. Also, in everyday care of typically elderly, frail, late-stage or end-stage patients undergoing an experimental treatment, use of PSA was felt to comprise the least invasive, logistically-easiest, and most rapidly-collected efficacy biomarker. The utility of PSA as such is supported by the significant, strong prognostic power regarding OS of early biochemical disease control in our multivariable analysis.
In conclusion, this, the largest-yet-published report of a prospectively-observed cohort with late–end-stage mCRPC, conventional treatment failure, and care in an everyday setting, provides encouraging evidence that 177Lu-PSMA-617 RLT is effective, safe, and well-tolerated outside of clinical trial conditions, and in patients with frequently more advanced disease and heavier pretreatment than those in VISION or TheraP [10, 14]. Early biochemical disease control (PR/SD) after the initial ≤ 2 cycles of RLT was strongly and significantly associated with better OS, even in the presence of negative prognostic factors, e.g., liver metastases. Prospective study of RLT earlier in the disease course may be warranted.