Laboratory, Clinical, and Survival Outcomes Associated With Peptide Receptor Radionuclide Therapy in Patients With Gastroenteropancreatic Neuroendocrine Tumors

This cohort study examines laboratory, clinical, and survival outcomes among patients with gastroenteropancreatic neuroendocrine tumors (NETs) treated with peptide receptor radionuclide therapy.


Introduction
Neuroendocrine tumors (NETs) are a heterogeneous group of neoplasms, with primary tumors commonly originating in the pancreas, intestines, and lungs. [1][2][3] These tumors often present with distant metastatic disease, and given the multitude of treatment options available, treatment of metastatic disease is growing increasingly complex. 4 Somatostatin analogues are widely used to treat metastatic NETs and have been shown to slow progression and improve disease-related symptoms. 5,6 Other potential therapies for metastatic NETs include systemic chemotherapy, liverdirected therapy, surgical treatment, and peptide receptor radionuclide therapy (PRRT). 4,[7][8][9] In PRRT, a radionuclide linked to a somatostatin analogue is used, allowing for targeted delivery of radiotherapy to somatostatin receptor-expressing NETs, which represent most NETs. 10,11 The first randomized clinical trial assessing the efficacy of PRRT for treatment of NETs was the Neuroendocrine Tumors Therapy (NETTER-1) trial, which included patients with metastatic, welldifferentiated (ie, World Health Organization [WHO] grade 1 or 2) midgut NETs who had progressed on somatostatin analogue therapy. 12 In this landmark phase 3 multicenter trial, patients were randomized to receive lutetium-177-dotatate plus octreotide long-acting repeatable or high-dose octreotide long-acting repeatable alone. Progression-free survival (PFS) was significantly higher in the lutetium-177-dotatate group than in the control group. After results of NETTER-1 supported the effectiveness of PRRT in midgut NETs, the Food and Drug Administration (FDA) approved lutetium-177-dotatate for treatment of somatostatin receptor-positive gastroenteropancreatic NETs in 2018.
Since FDA approval of PRRT, efforts have been targeted at implementation of PRRT within NET clinics across the United States. 13,14 Results from a 2020 study 15 by our group suggested that PRRT could be successfully administered among a diverse US-based group of patients with NETs, and we found that patients with abnormal liver function tests before PRRT had an increased likelihood of discontinuing PRRT prior to completion of the PRRT treatment course.
Prior to FDA approval of PRRT, and apart from NETTER-1, multiple studies [16][17][18][19] reported outcomes of PRRT in patients with NETs. However, to our knowledge, there have been no studies after PRRT was approved describing PRRT outcomes for patients with NETs who underwent treatment in the United States. This unique population of patients received a treatment course that may have differed from that in other parts of the world owing to the delayed PRRT approval in the US and represent a more heterogeneous population than patients enrolled in NETTER-1. Given the increased accessibility to PRRT and increased use of the treatment within the US, it is critical to understand clinical outcomes to establish expectations for patients and clinicians. In this study, we examined the first 2 years of PRRT implementation within a tertiary US NET referral center; we describe the patient group selected for PRRT and examine their laboratory-measured toxic effects and therapeutic response.

Methods
This cohort study was approved through the University of Pennsylvania Institutional Review Board, which granted a waiver of informed consent for participation because this study was deemed to be of minimal risk. This study is reported following the Strengthening the Reporting of Observational

Statistical Analysis
Statistical analysis was performed using Stata/IC statistical software version 15.1 (StataCorp).
Comparative statistical analyses were performed, and PFS and overall survival were analyzed using Cox proportional hazards models. Kaplan-Meier plots were produced. P values were calculated using log-rank test for survival curves. For PFS, events were defined by progression or death. Patients were censored at end of follow-up. Overall survival included death by any cause. Univariable analysis was performed to identify factors associated with survival. All statistical tests were 2-tailed, and significance was defined as P < .05. at first PRRT; 39 (50.0%) patients were men ( Table 1)

Laboratory-Measured Toxic Effects
At least 1 new grade 2 or greater laboratory-measured toxic effect was found in 47 patients (60.3%) during their treatment course ( Table 2). The most common new grade 2 or greater laboratorymeasured toxic effect was leukopenia, found in 26 patients (33.3%), followed by anemia, found in 16 patients (20.5%); acute kidney injury (AKI), found in 12 patients (15.4%); liver injury, found in 12 patients; and thrombocytopenia, found in 9 patients (11.5%). Of 12 patients with AKI, 6 patients had non-PRRT-related causes of kidney injury documented in the medical record. Among 55 patients who completed 4 doses of PRRT, the most common toxic effect after the final dose was leukopenia, found in 13 patients (23.6%), followed by thrombocytopenia, found in 7 patients (12.7%); anemia, found in 6 patients (10.9%); AKI, found in 3 patients (5.5%); and liver injury, found in 3 patients.
Grade 3 or 4 laboratory-measured toxic effects were observed in 25 patients (32.1%) throughout treatment. Among all patients in the study, the most common grade 3 or 4 laboratory-measured toxic effects were anemia, found in 10 patients (12.8%), and leukopenia, found in 10 patients. Grade 3 or 4 thrombocytopenia was found in 7 patients (9.0%). While 1 patient developed transfusiondependent anemia, this patient had baseline anemia associated with prior treatment with everolimus. No one in the study group developed myelodysplastic syndrome, dialysis-dependent kidney injury, or liver failure. Dosage was reduced for 2 patients owing to laboratory-measured toxic effects.

PFS
During the study period, 30 patients (38.5%) had progression or died, 25 patients (32.1%) had evidence of progression on imaging, and 5 patients (6%) who did not have progression on imaging died during the treatment course. The only significantly different characteristic between patients who progressed and those who did not was NET primary location (Table 1). Median PFS for the entire   study group was 21.6 months (Figure 2, part A). Univariable analysis of factors associated with progression is shown in  Figure 2, part B). Age, sex, tumor grade, prior systemic therapy, prior liver-directed therapy, and prior resection of primary tumor were not associated with statistically significant differences in progression.

Overall Survival
In this study group, 10 patients (12.8%) died after completion of at least 1 dose of PRRT; however, given the short follow-up period, median overall survival was not reached (eFigure in the Supplement). Additionally, 2 patients (2.6%) died after completion of 1 cycle of treatment, 4 patients (5.1%) died after completion of 2 cycles, 2 patients died after completion of three cycles; and 2 patients died after completion of 4 cycles. Univariable analysis of factors associated with survival was significant for age at NET diagnosis (HR, 1.08; 95% CI, 1.01-1.16; P = .02) (eTable 2 in the Supplement).

Discussion
This cohort study of outcomes associated with PRRT in a non-clinical trial-based US NET group after FDA approval of PRRT found that patients with small bowel primary tumors had significantly lower rates of progression compared with patients with pancreatic primary tumors.
One strength of our study is that our NET group was diverse and heterogeneous, which may provide a realistic representation of clinical NET practices. Our study group was heavily pretreated prior to PRRT initiation, unlike the NETTER-1 population. While pretreatment was not found to be associated with progression after PRRT, studies with larger study groups or longer follow-up periods may find pretreatment associations with treatment outcomes. A 2017 study 16 by our group found that prior systemic chemotherapy was a predictor associated with decreased PFS, although this study had a longer follow-up period after PRRT. Ultimately, given the number of treatment options for NETs, more studies focused on the sequence of therapies may help elucidate where best to incorporate PRRT into the NET treatment algorithm.
Our study also differed from NETTER-1 in primary tumor location and grade. Whereas NETTER-1 was restricted to patients with midgut NETs, 12 our patient population had NETs arising from locations outside of the small bowel, including the pancreas, colon, stomach, and lung, which allowed us to compare PFS among the primary tumor sites. Additionally, NETTER-1 included only patients with grade 1 or 2 NETs, while more than 10% of patients in our group had grade 3 NETs. The NETTER-2 trial will provide critical information that may address efficacy of PRRT in non-small bowel and advanced-grade NETs 22 ; however, results from that study will not be available for several years.
While PRRT is generally accepted as a well-tolerated treatment for NETs, 47 out of 78 patients in our study (60.3%) experienced a laboratory-measured toxic effect at some point during their treatment course. Laboratory-measured toxic effects within our study group were similar to those reported in 3 studies from 2019, 13 suggesting that the rate of AKI associated with PRRT is likely lower than reported in this study.
Additionally, a greater proportion of patients in our study group experienced grade 3 or 4 toxic effects compared with patients in NETTER-1: grade 3 or 4 anemia in 10 patients (12.8%) in our study group vs 0 patients in NETTER-1, grade 3 or 4 leukopenia in 10 patients in our study group vs 1 patient (0.9%) in NETTER-1, and grade 3 or 4 thrombocytopenia in 7 patients (9.0%) in our study group vs 2 patients (1.8%) in NETTER-1. 12 This discrepancy may be associated with the high rate of pretreament prior to PRRT initiation in our study group. While 1 patient in our study required ongoing red blood cell transfusion support, no one in our group went on to develop myelodysplastic syndrome, kidney failure requiring hemodialysis, or liver failure. An appreciation of the frequency of these laboratory-measured toxic effects during therapy is important in clinical practice, and it is reassuring that most laboratory-measured toxic effects were transient within our study period.
Prolonged follow-up may be needed to better capture long-term PRRT-related toxic effects.
Measuring PRRT outcomes using PFS and overall survival may be helpful to set expectations for clinicians and patients. We found that 30 patients (38.5%) had disease progression or died during the study period. The median PFS for our study group was 21.6 months, which is similar to the results from the NETTER-1 group, which had a 65.2% estimated rate of PFS at month 20, with a 95% CI of 50.0% to 76.8%. 12 These results are encouraging for implementation of PRRT in clinical practice, where clinicians may expect a substantially lower response to therapy compared with during the stringent conditions of clinical trials. We also found that having a small bowel primary tumor was associated with a lower rate of progression compared with having a pancreatic primary tumor.
Median PFS for patients with primary pancreatic NETs was 13.3 months, whereas median PFS had not been reached in the small bowel group after 22 months. These findings suggest that in clinical practice, patients with small bowel NETs may have the best PRRT outcomes.
We assessed overall survival in our study group; however, because 68 of 78 patients in our study group (87.2%) were living at the time of analysis, we could not assess median overall survival.
In statistical analysis, we found that patients diagnosed with NET at older ages had a higher risk of dying during the study period; however, the HR for increased risk was 1.08, suggesting a modest association. Studies of PRRT with longer follow-up periods in US-based cohorts may be important in better determining the association of PRRT with overall survival in clinical practice.

Limitations
This study has several limitations. The small sample size may be underpowered to detect statistically significant differences between groups. The heterogeneity of the study group may be associated with confounding variables. Other limitations include the limited length of our follow-up period and the substantial proportion of our study population who had not completed 4 doses of PRRT by the time of analysis. Despite these limitations, we generated PRRT outcome data, which may contribute important information about the understudied clinical outcomes of PRRT in US-based NET populations. Additionally, we included all patient imaging after initiation of PRRT, which may not adequately assess tumor response to PRRT, as response may be slow and could be associated with confounding by pseudoprogression. This is a radiographic pattern that suggests progression based on RECIST criteria that may be associated with transient PRRT-induced tumor inflammation. 25 We believe, however, that discovery of imaging-based disease progression prior to completion of PRRT may be clinically relevant, and therefore, should be included in the analysis. Additionally, our study group represents a heavily pretreated population that may not be representative of other practices around the world; however, this population may be more similar, compared with clinical trial populations, to those in other US-based NET practices.

Conclusions
In this cohort study of patients with metastatic NETs, we found that PRRT may have favorable outcomes when administered among a diverse population of patients with NETs. While laboratorymeasured toxic effects were common, they were seldom associated with serious complications during the study period. Furthermore, we found that patients with small bowel NETs had a significantly longer median PFS after PRRT compared with patients with pancreatic or other NETs.
Larger and longer prospective studies are needed to further support these findings and determine how to optimally incorporate PRRT into the treatment of metastatic NETs.

25.
Karfis I, Marin G, Flamen P. An uncommon "pseudo-progression" pattern of morphological response, as early as after one cycle of PRRT, in siNET patients.