Treatment of advanced BP-NETS with lanreotide autogel/depot vs placebo: the phase III SPINET study

Prospective data are lacking on early somatostatin analog (SSA) therapy in bronchopulmonary neuroendocrine tumors (BP-NETs; typical carcinoids and atypical carcinoids (TCs and ACs)). SPINET (EudraCT: 2015-004992-62; NCT02683941) was a phase III, double-blind study of lanreotide autogel/depot (LAN; 120 mg every 28 days) plus best supportive care (BSC) vs placebo plus BSC, with an optional open-label treatment phase (LAN plus BSC). Patients had metastatic/unresectable, somatostatin receptor (SSTR)-positive TCs or ACs. Recruitment was stopped early owing to slow accrual; eligible patients from the double-blind phase transitioned to open-label LAN. The adapted primary endpoint was progression-free survival (PFS) during either phase for patients receiving LAN. Seventy-seven patients were randomized (LAN, n = 51 (TCs, n = 29; ACs, n = 22); placebo, n = 26 (TCs, n = 16; ACs, n = 10)). Median (95% CI) PFS during double-blind and open-label phases in patients receiving LAN was 16.6 (11.3; 21.9) months overall (primary endpoint), 21.9 (12.8, not calculable (NC)) months in TCs, and 13.8 (5.4; 16.6) months in ACs. During double-blind treatment, median (95% CI) PFS was 16.6 (11.3; 21.9) months for LAN vs 13.6 (8.3; NC) months for placebo (not significant); corresponding values were 21.9 (13.8; NC) and 13.9 (13.4; NC) months, respectively, in TCs and 13.8 (5.4; 16.6) and 11.0 (2.8; 16.9) months, respectively, in ACs. Patients’ quality of life did not deteriorate and LAN was well tolerated. Although recruitment stopped early and the predefined sample size was not met, SPINET is the largest prospective study to date of SSA therapy in SSTR-positive TCs and ACs and suggests clinical benefit in TCs.

Many NETs overexpress somatostatin receptors (SSTRs), leading to the development of somatostatin analogs (SSAs) as antisecretory and antiproliferative treatments across different NET primaries (Del Olmo-Garcia et al. 2021).Two SSAs have been shown to significantly delay progression or death in metastatic gastrointestinal NETs: lanreotide autogel/depot (LAN) in patients with slowly progressive well-or moderately differentiated gastroenteropancreatic (GEP) NETs (Caplin et al. 2014), and octreotide long-acting release (OCT) in patients with well-differentiated midgut NETs and a mainly low hepatic tumor burden (Rinke et al. 2009).SSAs also have favorable tolerability profiles and are the initial mainstay of medical treatment for gastrointestinal NETs with a good prognosis.
TCs and ACs express SSTRs, providing a rationale for their treatment with SSAs (Vesterinen et al. 2019).However, when the current study (SPINET) was designed, the data on the use of SSAs in patients with these tumors were limited to those from a small subgroup (n = 11) of the RADIANT-2 study (Pavel et al. 2011) and retrospective analyses (Lopez et al. 2014, Karra et al. 2015, Sullivan et al. 2015), and LAN was not yet recommended as an antiproliferative treatment for TCs and ACs in US and European guidelines (Öberg et al. 2012(Öberg et al. , Kunz et al. 2013)).The SPINET study was designed to address this unmet need by prospectively evaluating the efficacy and safety of LAN in patients with SSTR-positive TCs and ACs.

Materials and methods
Enrollment began in July 2016.In July 2018, recruitment was stopped owing to slow accrual, resulting primarily from changes to US and EU guidelines.The protocol was amended, including the primary objective and primary and secondary endpoints (see below).Eligible patients still receiving treatment during the double-blind phase could enter the open-label treatment phase to receive open-label LAN.

Patients
Full inclusion and exclusion criteria are summarized in Supplementary Table S1 (see section on supplementary materials given at the end of this article).Patients, who were enrolled by study investigators at specialist centers, were eligible if they were aged ≥ 18 years and had metastatic and/or unresectable, pathologically confirmed TCs or ACs with positive SSTR imaging.They also needed at least one measurable lesion of the disease on CT or MRI (Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 (Eisenhauer et al. 2009)) and an Eastern Cooperative Oncology Group (ECOG) performance status score of 0 or 1. Patients were excluded if they had received previous SSA treatment, more than two lines of chemotherapy (cytotoxic, molecular targeted therapy, or interferonalpha) for TCs/ACs, or chemotherapy in the 4 weeks before randomization, or if they had functional disease requiring SSA treatment for symptom management.No assessment of disease progression was performed before study entry (not an inclusion criterion).

Study design and interventions
SPINET (EudraCT: 2015-004992-62; NCT02683941) was a phase III, prospective, multicenter, randomized, double-blind, placebo-controlled study with an optional open-label LAN treatment phase and an open-label follow-up phase.Patients were randomly allocated (2:1) to receive subcutaneous injections every 28 days with LAN (120 mg) plus best supportive care (BSC) or placebo plus BSC, with randomization stratified by TC vs AC and prior vs no prior chemotherapy (see Supplementary Material for information about randomization and blinding).
The plan was to enroll 216 patients across Canada, Europe, and the USA, but recruitment was stopped early owing to slow accrual.Before recruitment was stopped, patients receiving a placebo with centrally confirmed progression (RECIST v1.1 (Eisenhauer et al. 2009)) during the double-blind phase could enter the open-label treatment phase to receive LAN plus BSC.After recruitment was stopped, eligible patients remaining in the double-blind phase (i.e.patients receiving LAN without centrally confirmed disease progression and patients receiving placebo) could transition to open-label LAN + BSC.
The study ended 18 months after the last patient was randomized, in February 2020.

Assessments and endpoints
To assess disease progression, patients underwent a CT or MRI scan at baseline and every 12 weeks during the study.Procedures for local and central review are summarized in Supplementary Fig. S1.If locally assessed disease progression was not confirmed centrally, the local decision prevailed, in accordance with US Food and Drug Administration requirements.Other assessments during the study are summarized in Supplementary Tables S2 and S3.Some endpoints were amended after recruitment was stopped.The primary endpoint was changed from centrally confirmed progression-free survival (PFS) during the double-blind phase to centrally confirmed PFS during the double-blind or open-label treatment phases for patients randomized to LAN only; PFS endpoints for LAN vs placebo became secondary endpoints.Key secondary endpoints, assessed for the LAN and placebo groups during the double-blind phase, were centrally and locally assessed PFS, objective response rate (ORR), and time to treatment failure (TTF).Other secondary endpoints, assessed during the double-blind and open-label treatment phases, included changes in the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Core 30 (QLQ-C30), proportions of patients with a deterioration in QLQ-C30 scores (≥ 10-point decrease), and changes in biomarker levels (serum chromogranin (CgA) and urinary 5-hydroxyindoleacetic acid (5-HIAA); see Supplementary Material).Exploratory endpoints included clinical benefit rate (CBR), defined as the proportion of patients with complete response, partial response, or stable disease, and tumor growth rate (TGR) during the double-blind and open-label treatment phases.Safety endpoints were treatment-emergent adverse events (TEAEs), laboratory parameters (hematology and biochemistry), vital signs, and ECG results.Patients had gallbladder echography if biological abnormalities and/or clinical inflammatory symptoms occurred during the study.

Statistical analysis
The planned sample size was 216 patients (LAN, n = 144; placebo, n = 72; i.e. 2:1 ratio), based on 90% power to detect a clinically meaningful treatment difference of 4 months for PFS (10 months for LAN and 6 months for placebo; i.e. an expected hazard ratio (HR) of 0.6), with a type 1 error rate of 0.05.
The intention-to-treat (ITT) population comprised all patients randomly assigned to study treatment.The open-label ITT population comprised all ITT patients who entered the open-label treatment phase and received at least one LAN injection during this phase.
The safety population included all patients who received at least one injection of study medication.
PFS and TTF were analyzed using the Kaplan-Meier product limit method.Associated HRs were estimated using a stratified Cox proportional hazards model, and 95% CIs were calculated using the Brookmeyer-Crowley method.For ORR, CBR, and the proportions of patients with decreases in biomarker CgA and QLQ-C30 scores, 95% CIs were calculated using the Clopper-Pearson method.Statistical significance was assessed using a stratified log-rank test.
TGR was calculated from radiologic scans in the 12 months before the study, at baseline, and every 12 weeks during the study, and estimated using the sum of diameters of target lesions (100 (exp(3 log (Dt/D0)/t) − 1)) (D0 and Dt = tumor sizes at times 0 and t).Given that the provision of historical scans was optional, TGR at baseline could only be computed for the subgroup of patients for whom historical scans were available.

Ethical approval
The study was conducted in compliance with the Declaration of Helsinki, the International Conference on Harmonization Good Clinical Practice guidelines, and institutional review board/independent ethics committee requirements.All patients had to provide written informed consent before enrollment.

Patients
Overall, 77 patients were enrolled (51 randomized to LAN, 26 to placebo) (Fig. 1); 21 in the LAN group and 19 in the placebo group received LAN during the open-label treatment phase (Fig. 1).Given that recruitment was stopped prematurely, the planned sample size was not met, and statistical analyses are considered descriptive.The median (range) on-study duration was 13.1 (1-34) months (double-blind phase), 8.1 (0-21) months (open-label treatment phase), and 22.8 (2-36) months (overall).Five patients (9.8%) in the LAN group and four patients (15.4%) in the placebo group had received prior systemic therapy (Table 1).
Overall, the mean (s.d.) age was 66.2 (12.5) years, and 54.5% of patients were male.Among them, 58.4% had TCs (29 patients in the LAN arm and 16 patients in the placebo arm) and 41.6% had ACs (22 patients in the LAN arm and 10 patients in the placebo arm (Table 1)).There were some imbalances in baseline characteristics between the two groups, including the median time since diagnosis and the proportion of patients with ECOG performance status 0 or 1 (Table 1 and Supplementary Table S4).
During the double-blind phase, the centrally assessed ORR was 14.0% for LAN and 0% for placebo, but the difference was not statistically significant (Table 2).CBR was similar (90.0%vs 92.0%) between the two groups (Table 2).TTF (median (95% CI)) was 13.3 (5.6; 14.1) months for LAN vs 9.8 (5.4;13.6)months for placebo (HR 0.86 (95% CI, 0.50; 1.50)), but this difference was not statistically significant (P = 0.582) (Fig. 3); see Supplementary Table S7 for reasons for treatment failure.The ORR for LAN was 6.0% on local assessment and 14.0% on central assessment.The ORR and CBR based on local assessments are summarized in Supplementary Table S8 and further information provided in the Supplementary Material.

Safety
During the double-blind phase, TEAEs were reported for similar proportions of patients in the LAN and placebo groups (96.1% and 96.2%, respectively); in most cases, the events were of grade 1 or 2 severity (Table 3).TEAEs leading to treatment discontinuation were reported in two patients (3.9%) receiving LAN and three patients (11.5%) receiving placebo (  (E) During the double-blind phase in patients randomly assigned to LAN or placebo (in patients with atypical carcinoids).Data from patients who did not die and did not have confirmed disease progression were censored on the date of the last radiological assessment at which the target lesions were evaluated by central review.a One patient was excluded from the analysis because their data were censored at baseline (because the baseline assessment was prior to randomization, this would have yielded a negative PFS).PFS in months was calculated as follows: (date of event -date of randomization) / 30.4375.AC, atypical carcinoid; HR, hazard ratio; ITT, intention-to-treat; LAN, lanreotide autogel/depot; NC, not calculable; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors; TC, typical carcinoid.

Discussion
SPINET is the first phase III, randomized, placebocontrolled study to evaluate early SSA therapy in patients with TCs and ACs.It provides clinically important evidence regarding the use of LAN in patients with these tumors, for whom few treatment options are available.Although recruitment was stopped prematurely and the predefined sample size was not met, the combined results of the primary, secondary, and exploratory endpoints suggest that LAN has clinical activity, most obviously in the medically important subgroup with TCs, where the median numerical improvement in centrally assessed PFS vs placebo was 8 months.Indeed, the 21.9-month PFS for LAN in patients with TCs reported in SPINET falls within the range of 14.3 months (time to progression (TTP)) to 38.5 months (PFS) reported for OCT in the PROMID trial (patients with metastatic midgut NETs) or LAN in the CLARINET trial (patients with advanced enteropancreatic NETs), respectively (Rinke et al. 2009, Caplin et al. ).The PFS with placebo in SPINET (13.9 months) is also within the range reported for placebo in the PROMID (TTP, 6.0 months) and CLARINET (PFS, 18.0 months) trials (Rinke et al. 2009, Caplin et al. 2014).In addition to the smaller-than-planned sample size, factors that may have limited the extent of the difference between treatment arms include tumor progression before enrollment not being a prerequisite (as in the PROMID trial (Rinke et al. 2009)), and the potentially higher TGR in the LAN arm.In addition to the numerical improvement in PFS, there were greater reductions in serum CgA and in TGR in the LAN arm.TGR, a predictive marker for progression and survival, is useful for assessing slow-growing tumors like NETs, and it has previously been shown that LAN significantly decreased TGR in patients with GEP-NETs (Dromain et al. 2021).Patients' health-related quality of life did not deteriorate during treatment in the present study, and LAN was well tolerated, having a safety profile consistent with that well established in NETs (Bianchi et al. 2011, Martín-Richard et al. 2013, Caplin et al. 2014, Michael et al. 2017, Godara et al. 2019, Caplin et al. 2021, Ito et al. 2021, Rinke et al. 2021).
The main reason for the slow accrual of patients into SPINET was the increasing use of SSAs for TCs and ACs in clinical practice; as such, patients were reluctant to enroll in this study, with the risk of receiving a placebo, when they could access the active treatment with a prescription.The increase in SSA use probably reflects changes to guidelines (Caplin et al. 2015, Kulke et al. 2015), which were updated to recommend the use of SSAs/LAN as first-line antiproliferative treatment in SSTR-positive TCs and ACs to discourage the early use of platinum-based chemotherapy.At the time the guidelines changed, recommendations were based mainly on the extrapolation of findings from research in GEP-NETs (LAN) (Caplin et al. 2014) or midgut NETs (OCT) studies (Rinke et al. 2009), as well as limited data on the use of SSAs for TCs and ACs from a subgroup analysis of the RADIANT-2 study (OCT vs everolimus plus OCT) (Pavel et al. 2011) and three retrospective studies published as abstracts (Lopez et al. 2014, Karra et al. 2015, Sullivan et al. 2015).
After SPINET was initiated, data from a phase II study (LUNA) of SSA therapy (which included patients with progressive TCs and ACs) were published, providing the first prospective signal that SSAs (as well as everolimus alone and everolimus plus SSA) may have antiproliferative effects in these patients (Ferolla et al. 2017).However, LUNA was conducted in patients receiving mainly post-first-line therapy with progressive tumors according to RECIST v1.1 and included a small number of patients with thymic NETs, so its data are not directly comparable with SPINET data.Since then, the first data on cytotoxic chemotherapy, from the prospective, phase II ATLANT study (LAN plus temozolomide; n = 40), have been published (Ferolla et al. 2023).ATLANT included patients with progressive TCs (20.0%),ACs (52.5%), or unspecified carcinoids (27.5%) who had received ≤1 line of treatment with SSA.The study demonstrated an investigator-assessed disease control rate (DCR) at 9 months (primary endpoint) of 35% (partial response, 2.5%); however, this was not statistically significant at a 30% threshold (P = 0.2968) and the primary endpoint was not met.Median PFS was 37.1 weeks, and 32.5% of patients experienced TEAEs that led to the interruption of study medication.In a sensitivity analysis conducted using investigator assessments between months 7.5 and 10.5, DCR was 45.0% (partial response, 7.5%), which was statistically significant at a 30% threshold (P = 0.0320) (Ferolla et al. 2023).From a clinical perspective, these results do not favor the use of temozolomide as first-line therapy in patients with TCs (Baudin et al. 2021).Another treatment option is the mTOR inhibitor everolimus, which is now licensed to treat unresectable or metastatic progressive TCs and ACs (based on the phase III RADIANT-4 study (Yao et al. 2016) and subgroup analysis (Fazio et al. 2018)).However, similar to temozolomide (Taherifard et al. 2024), the safety and tolerability profile of everolimus (Faggiano et al. 2016) is generally considered less favorable than that of LAN (Faggiano 2024), and some suggest that its risk/benefit profile may be more appropriate for use in patients with clinically significant disease progression (Cives & Strosberg 2018).The latest European Society for Medical Oncology guidelines recommend SSA therapy rather than everolimus as first-line treatment in TC and slowly progressing SSTR-positive lung carcinoids because it is better tolerated (Baudin et al. 2021).
This study has several limitations, the most important of which is that the predefined sample size was not met.The earlier-than-planned transition to the open-label treatment phase reduced the period of observation during the double-blind phase and further limited treatment comparisons and the clinical interpretation of the data.However, the median PFS in SPINET is consistent with that reported in other retrospective and prospective studies of patients with TCs and ACs.In these studies, median PFS ranged from 8.5 months in patients with progressive disease at study entry to 17.4 months in patients with mixed progression status at baseline (Sullivan et al. 2015, Bongiovanni et al. 2017, Ferolla et al. 2017).In addition, the median PFS of both arms in SPINET is consistent with the TTP and PFS results reported in phase III trials in patients with midgut and enteropancreatic NETs (Caplin et al. 2021, Rinke et al. 2009).The conservative estimate of median PFS (10 months for LAN and 6 months for placebo) for the sample size hypothesis was based on the limited evidence (PFS and TTP) in patients with BP-NETs (Pavel et al. 2011, Lopez et al. 2014, Sullivan et al. 2015, Karra et al. 2015, Fazio et al. 2016 ) and midgut NETs (Rinke et al. 2009) that was available at the time of study design.
As noted earlier, the most important limitation of the study is the reduced sample size, which affected data interpretation and probably explains the lack of significance between LAN and placebo for centrally assessed PFS.In addition, progressive disease was not an entry criterion for the study, which means that it was not possible to formally evaluate disease stabilization; however, the randomized nature of the trial was implemented to overcome this limitation.
Another limitation is the difference in results for central and local assessment of tumor response, which may reflect variability in the assessment of metastases and differences in target lesion selection in slowly progressive NETs, as noted in the RADIANT-2 study (Pavel et al. 2011), and a higher rate of false positive targets in the thoracic area.In SPINET, in two patients where central and local responses differed, the numbers of baseline target lesions were different in central and local assessments.The challenges associated with assessing baseline target lesions in patients with NETs are well known (Abramson et al. 2015), hence the robust central assessment process adopted in SPINET (two readers, with adjudication in the event of differences in opinion).Although some differences between local and central assessment were observed during the study, it was only in a single patient that this difference had a significant impact on the tumor response.This patient had two measurable but small target lesions at baseline (16 mm and 18 mm on central measurement and 19 mm and 18 mm on local assessment); one of these lesions became nonmeasurable on central assessment only.A further limitation includes minor imbalances in baseline characteristics between treatment arms; however, the final clinical relevance of these imbalances is unpredictable.It should also be noted that data on the median time from the onset of metastatic disease to the start of treatment are not available.
Strengths of the study include that it enrolled an international population and was a dedicated, prospective evaluation of early SSA therapy for TCs and ACs, and that it represents the largest prospective study to date of early SSA use in this setting.Based on data from countries permitting the collection of race data, the racial diversity of the study population was limited because most patients were European; nevertheless, disease characteristics were generally as expected of those with TCs and ACs in clinical practice.As per regulatory requirements (Zhang et al. 2018), there was a robust, well-established, first-line central assessment of PFS that, nonetheless, allowed individualized care of patients to continue.Thus, the study provides clinically important data.
The difficulties faced in the SPINET study prompt some important reflections on missed opportunities to provide the best level of evidence achievable through the completion of a phase III trial, especially in rare cancers.At the time SPINET was designed, conducting a dedicated phase III study was considered feasible based on the estimated incidence (~1.5 per 100,000 (Öberg et al. 2012)).More importantly, it was anticipated that it would offer significant benefits to participants (including those initially treated with placebo, who could transition to LAN in the open-label treatment phase), as well as patients whose treatment would be informed by its outcomes.With hindsight, optimal changes to treatment guidelines in these circumstances might have included advocating strongly for SSA treatment via clinical trial participation when possible, as an alternative to off-label prescription.It is also the case, however, that meeting the best interests of patients and increasing recruitment into the study would have been possible with increased awareness among thoracic surgeons and pneumologists, as well as improved standardization in the way in which TCs and ACs are characterized in expert NET centers and in the use of prognostic factors to guide treatment decisions.
In conclusion, despite lower-than-target enrollment, SPINET is the largest prospective study to date of SSA therapy in SSTR-positive TCs and ACs.The study provides clinically important data about the activity and tolerability profile of LAN 120 mg every 28 days in unresectable and/or metastatic BP-NETs.The results of SPINET thus provide much-needed data to support the clinical use of SSAs in BP-NETs, mainly TCs, as recommended by European and US guidelines.

Figure 1
Figure 1 Patient disposition.Data from the OLFU phase are not reported in this article; the group comprised 19 patients who transitioned from the DB phase (as noted in the figure above) and one patient from the LAN (NPD)-LAN group who entered OLFU from the OLT phase.a All 22 patients transitioned to the OLT phase after protocol amendment #5.DB, double-blind; LAN, lanreotide autogel/depot; NPD, no progressive disease; OLFU, open-label follow-up; OLT, open-label treatment; PBO, placebo; PD, progressive disease; R, randomization.
Figure 2 Centrally assessed (RECIST v1.1) progression-free survival (ITT population).(A) During double-blind or open-label treatment phase in patients randomly assigned to LAN (overall patient population).(B) During the double-blind or open-label treatment phase in patients biochemical, or ECG parameters, estimated glomerular filtration rate, or vital signs during the double-blind and open-label treatment phases.Only two patients receiving LAN required post-baseline gallbladder echography, and neither developed new abnormalities during the study (Supplementary Material).

Table 1
Baseline demography and clinical characteristics (ITT population).It was not permitted to collect race data in France and Poland; b Data missing for three patients in each group; c According to interactive web response system; d Obtained as part of disease history/diagnosis at screening; e Obtained as part of disease history/diagnosis at screening, if available; f Percentages based on the total number of patients with Octreoscan results; g Percentages based on the total number of patients with positron emission tomography scans; a

Table 2
Secondary efficacy endpoints: objective response rate a and clinical benefit rate b (centrally assessed; ITT population).Note that ORR during the open-label treatment phase is not provided because it was not a pre-specified endpoint.Tumors were assessed according to RECIST v1.1 criteria.
a Complete responses plus partial responses; b Complete responses plus partial responses plus stable disease.CBR, clinical benefit rate; ITT, intention-to-treat; LAN, lanreotide autogel/depot; NC, not calculable; NPD, no progressive disease; ORR, objective response rate; PBO, placebo; PD, progressive disease; RECIST, Response Evaluation Criteria in Solid Tumors.

Table 3
Safety profile (safety and open-label ITT populations).Data from the double-blind phase are from the safety population and those from the open-label treatment phase are from the open-label ITT population.aCarcinoid syndrome (n = 1) and Pneumocystis jirovecii infection in the LAN group, and abdominal pain (n = 1), musculoskeletal chest pain (n = 1), and pneumonia (n = 1) in the placebo group; b Incidence ≥15% in any treatment group; c Preferred terms of abdominal pain or upper abdominal pain have been combined.ITT, intention-to-treat; LAN, lanreotide autogel/depot; NPD, no progressive disease; PBO, placebo; PD, progressive disease; TEAE, treatment-emergent adverse event.