Streptozotocin plus 5-fluorouracil followed by everolimus or the reverse sequence in patients with advanced pancreatic neuroendocrine tumors (SEQTOR-GETNE phase III study): a randomized clinical trial

Abstract

Background

Everolimus or streptozotocin plus 5-fluorouracil (STZ/5-FU) are approved treatments for patients with pancreatic neuroendocrine tumors (panNETs). The SEQTOR trial aimed to assess the optimal treatment sequence.

Patients and methods

SEQTOR was an international, open-label, randomized, crossover, phase III trial that recruited adults with unresectable or metastatic, advanced, well-differentiated panNET. Patients received 10 mg/day of everolimus followed upon progression by STZ/5-FU; or the reverse sequence. The primary endpoint was the 35-month progression-free survival (PFS) rate after first- and second-line treatment; however, due to slow accrual and longer survival, it was changed to the 12-month PFS rate following first-line treatment (12-mPFS₁).

Results

Patients were randomized to everolimus (n = 72) or STZ/5-FU (n = 69) first. The 12-mPFS₁ was 71.4% [95% confidence interval (CI) 59.4% to 81.6%] and 61.8% (95% CI 49.2% to 73.3%) (odds ratio 0.65, 95% CI 0.32-1.32) with a median PFS₁ of 19.4 versus 22.7 months for everolimus and STZ/5-FU, respectively. STZ/5-FU achieved a significantly higher overall response rate in first-line (11.6% versus 30.3%, P = 0.012) and second-line (30.6% versus 9.1%, P = 0.072) treatments. No differences were shown in overall survival (median 61.7 versus 50.6 months in everolimus first and STZ/5-FU first, respectively; hazard ratio 1.43, 95% CI 0.86-2.37). Discontinuations of everolimus were more frequent.

Conclusion

STZ/5-FU and everolimus were not statistically different in PFS rates, but STZ/5-FU achieved higher response rates.

Highlights

• •The SEQTOR trial provides evidence supporting that everolimus and STZ/5-FU are both valid treatments for panNETs.
• •Sequencing of both drugs should be tailored to the patient’s specific clinical characteristics and disease presentations.
• •Comorbidities and safety profile of each drug should also be considered for optimal treatment decision-making.

Introduction

Pancreatic neuroendocrine tumors (panNETs) are rare neoplasms, accounting for ∼2% of pancreatic tumors, with an increasing global incidence.^1,2 Advanced progressive panNETs present diverse treatment options; however, the optimal sequence remains debated, and predictive factors for therapy selection are lacking.

Clinical guidelines recommend the use of somatostatin analogues (SSA) as the first-line treatment of grade 1 tumors with a low Ki-67 index or slow growth, based on the results of the phase III trials PROMID and CLARINET.2, 3, 4, 5, 6, 7 The mammalian target of rapamycin (mTOR) inhibitor everolimus, tyrosine kinase inhibitors (TKIs) such as sunitinib, along with chemotherapy schemes such as streptozotocin (STZ) and 5-fluorouracil (5-FU), capecitabine and temozolomide (CAPTEM), or peptide receptor radionuclide therapy (PRRT) are viable options, primarily recommended for patients with grade 2 tumors with Ki-67 index >10%.8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 For instance, STZ and 5-FU have reported objective response rates (ORR) of up to 42.7% and a median progression-free survival (PFS) of ∼12 months in most recent reports.11, 12, 13, 14, 15 Everolimus showed lower ORR rates of ∼5% and a median PFS of 11 months in the phase III Radiant-3 study.^8,9 The decision to administer such antiproliferative treatments is predominantly based on patient comorbidities and tumor growth behavior, despite clinical trials reporting similar efficacy outcomes for these treatments in slow-growing or grade 1 tumors as well. Additionally, long-term control remains elusive, and most therapies do not prevent the need to switch treatments. Determining the optimal treatment sequence is paramount in this scenario, as the lack of direct comparative efficacy data and validated predictive markers complicate further decisions regarding treatment prioritization. Several randomized trials are currently evaluating PRRT against everolimus, sunitinib, or chemotherapy in advanced grade 1-3 gastroenteropancreatic NETs (GEP-NETs).^17,21,22 The COMPETE and OCLURANDOM studies recently showed improvement in PFS versus everolimus and sunitinib, respectively. The median PFS was 23.9 months with ¹⁷⁷Lu-edotreotide versus 14.1 months with everolimus in COMPETE, and 20.7 months with PRRT versus 11.0 months with sunitinib in OCLURANDOM.²⁰ The BETTER2 study (NCT03351296) compares STZ/5-FU versus CAPTEM, with or without bevacizumab in pre-treated or treatment-naive, well-differentiated grade 1-3 panNETs.²³

Although both everolimus and STZ/5-FU are considered suitable first-line treatments for advanced panNETs, their optimal sequence and relative efficacy and safety have not yet been evaluated in head-to-head prospective randomized studies. Given this context, we designed the first randomized trial to assess the optimal sequence of STZ/5-FU and everolimus in patients with advanced panNETs. Unfortunately, given the rise of new treatment options, the study design needed to be modified and failed to obtain statistically powered evidence of sequentiality.

Patients and methods

Trial design and treatment

SEQTOR (EudraCT: 2013-000726-66/clinicaltrials.gov: NCT02246127) is an international, multicenter, open-label, randomized, crossover, investigator-initiated, phase III trial conducted in Spain, Germany, France, Denmark, the Netherlands, Sweden, Italy, and the United Kingdom (Supplementary Tables S1 and S2, available at https://doi.org/10.1016/j.esmoop.2025.105922) and led by the Grupo Español de Tumores Neuroendocrinos y Endocrinos (GETNE). The patients were randomized by an independent web-based application in a 1 : 1 ratio to receive either everolimus followed by STZ/5-FU upon disease progression (arm A) or the reverse sequence (arm B). Randomization was stratified only according to Eastern Cooperative Oncology Group performance status (ECOG-PS).

Everolimus was administered at 10 mg/day once daily, and STZ/5-FU was administered according to one of the following schemes: STZ 500 mg/m² and 5-FU 400 mg/m² on days 1-5 every 6 weeks (Moertel scheme); or STZ 500 mg/m² on days 1-5 and 5-FU 400 mg/m² on days 1-3 followed by a 1-day treatment of STZ 1 g/m² and 5-FU 400 mg/m² every 3 weeks (Uppsala scheme). Patients remained on the study treatment until disease progression, unacceptable toxicity, consent withdrawal, or death. Previous or concurrent SSA for symptom control was allowed but not mandated.

Patients

Eligible patients were adults with an ECOG-PS of 0-2. They had a histologically confirmed diagnosis of unresectable or metastatic, advanced, well-differentiated (World Health Organization grade 1-2) panNET. Disease progression to previous treatments had to be radiologically documented within 12 months before inclusion in the study. Treatment-naive patients (i.e. first-line systemic treatment of advanced/metastatic disease) were also eligible if they were considered potential candidates for either chemotherapy or everolimus. Patients were required to have adequate renal function (serum creatinine <1.5 × upper limit normal) and measurable disease according to the Response Evaluation Criteria in Solid Tumors (RECIST 1.0). Patients should not have received prior chemotherapy, mTOR inhibitors, or TKIs; PRRT within 6 months before study entry; or radiation or immunotherapy within 4 weeks before study entry. See the protocol for further details on eligibility criteria.

Written informed consent was obtained from all patients before study entry. The trial was approved by the independent ethics committees of the participating institutions [first ethics committee (Bellvitge Hospital, Barcelona) approval reference: 20140725] and competent authorities of the participating countries. The study was carried out in accordance with the Declaration of Helsinki and the applicable local regulatory requirements and laws.

Endpoints and assessments

The study was originally designed to evaluate the 35-month PFS1+2 rate as the primary endpoint. PFS1+2 was defined as the time from randomization to the second documented progression of disease (PD) or death from any cause, encompassing both lines of treatment. However, due to slow patient accrual—partly attributable to the approval of PRRT for the treatment of GEP-NETs—and the indolent course of the disease, which led to prolonged progression-free intervals, the protocol was amended. The new primary endpoint was the 12-month PFS rate to the first treatment (12-m PFS₁). The 12-m PFS₁ was defined as the proportion of patients alive without PD 12 months after randomization, as assessed by local investigators. Consistency analyses for the primary endpoint were carried out (see Supplementary Material, Figure S1, Tables S3 and S4, available at https://doi.org/10.1016/j.esmoop.2025.105922). Tumor radiological assessment was carried out using computed tomography (CT) scans, evaluated both locally by investigators, and centrally by an ENETS blinded independent review committee (BIRC), following the RECIST 1.0 criteria (Supplementary Table S5, available at https://doi.org/10.1016/j.esmoop.2025.105922). RECIST v1.1 was not used because it was not validated in panNETs by the time of study conception, and the authors of the protocol did not want to miss information of the evaluation method (RECIST 1.0 measures more target lesions and therefore they thought it may be more sensitive).

The secondary endpoints included ORR to both the first and second treatments, defined as the proportion of patients with a complete response or partial response as their best radiological assessment; duration of response (DoR); PFS to the first treatment (PFS₁), defined as the time from randomization to PD or death from any cause; PFS to the second treatment (PFS₂), defined as the time from initiation of the second treatment to PD or death and combined PFS₁₊₂, calculated as the time from randomization to the second PD or death. Additional secondary endpoints included patient-reported quality of life (QoL) assessed using the European Organisation for Research and Treatment of Cancer QLQ-C30 questionnaire and the NET-specific module, QLQ-GINET21, at baseline, first progression, and 30 days after treatment discontinuation. Overall survival (OS) was defined as the time from randomization to death from any cause.

Safety was assessed at each visit based on the number of adverse events (AEs), which were graded according to the National Cancer Institute Common Terminology Criteria for AEs version 4.0. Treatment feasibility and compliance were evaluated based on the number of dose reductions, dose delays, and total dose administered.

Statistical analysis

The sample size was calculated assuming a 12-month PFS₁ of 50% for everolimus as the null hypothesis.^8,9 A two-sided binomial test with an alpha of 0.05 and 80% power determined that a minimum of 132 patients were required to detect a 25% increase with STZ/5-FU. Assuming a 5% dropout rate, the total sample size was set at 140 patients (70 per arm).²⁴

Efficacy was assessed in the intention-to-treat population and safety was assessed in all patients who received at least one dose. Baseline characteristics and outcomes were analyzed using standard descriptive statistical methods. Missing data were not imputed. Time-to-event endpoints were estimated using the Kaplan–Meier method and Cox regression analysis to obtain hazard ratios (HR) or logistic regression for odds ratios (OR) in the subgroup analysis. Cohen’s kappa test was used to compare the distribution of best objective responses and to measure the agreement between assessment criteria and central versus local review. The Wilcoxon test, adjusted using Holm’s method, was used to assess changes in QoL over time. All statistical tests were two-tailed with a significance level set at P < 0.05. All statistical analyses were carried out using R²⁵ and SPSS.²⁶

Results

Patients

From June 2014 to July 2021, 141 patients accrued across eight countries were randomized to receive everolimus followed by STZ/5-FU (arm A; n = 72) or the reverse sequence (arm B; n = 69) and 135 received at least one dose of treatment (Figure 1). Overall, 69 patients (51.1%) transitioned to the second scheduled study treatment. Patient demographics and disease characteristics were balanced between the treatment groups (Table 1). Tumors were grade 2 in 63 (87.5%) and 55 (79.7%) patients treated with upfront everolimus and upfront STZ/5-FU, respectively. Previous treatments were administered to 36 patients (50.0%) in the everolimus-first group and 26 (37.7%) in the STZ/5-FU first group, consisting of SSA in 31 (43.1%) and 24 (34.8%) patients, respectively.

Figure 1.

CONSORT diagram showing patient allocation. 5-FU, 5-fluorouracil; ITT, intention-to-treat; pts, patients; SP, safety population; STZ, streptozotocin.

Table 1.

Baseline patient characteristics

Characteristic	Everolimus/STZ-5FU (n = 72)	STZ-5FU/Everolimus (n = 69)	SEQTOR trial (n = 141)
Median age, years (range)	59 (33-83)	58 (33-80)	58 (33-83)
Sex, n (%)
Male	41 (56.9)	44 (63.8)	85 (60.3)
Female	31 (43.1)	25 (36.2)	56 (39.7)
ECOG-PS, n (%)
0	50 (69.4)	47 (68.1)	97 (68.8)
1	20 (27.8)	22 (31.9)	42 (29.8)
2	2 (2.8)	0 (0)	2 (1.4)
Tumor grade WHO, n (%)
G1	9 (12.5)	12 (17.4)	21 (14.2)
G2	63 (87.5)	55 (79.7)	118 (83.7)
Unknown	0 (0)	2 (2.9)	2 (1.4)
Metastases at inclusion, n (%)
Mx	1 (1.4)	1 (1.4)	2 (1.4)
M0	2 (2.8)	6 (8.7)	8 (5.7)
M1	69 (95.8)	62 (89.9)	131 (92.9)
Metastases sites, n (%)
Liver	61 (84.7)	56 (81.2)	117 (83.0)
Bone	7 (9.7)	10 (14.5)	17 (12.1)
Lymph nodes	4 (5.6)	6 (8.6)	10 (7.1)
Lung	2 (2.8)	3 (4.3)	5 (3.5)
Ki-67 index, n (%)
≤2	9 (12.5)	14 (20.3)	23 (16.3)
3-20	62 (86.1)	51 (73.9)	113 (80.1)
Unknown	1 (1.4)	4 (5.8)	5 (3.5)
Previous lines, n (%)
0	36 (50.0)	43 (62.3)	79 (56.0)
1	32 (44.4)	22 (31.9)	54 (38.3)
2	4 (5.6)	4 (5.8)	8 (5.7)
Type of previous lines, n (%)
SSA	31 (43.1)	24 (34.8)	55 (39.0)
Radiopharmaceuticals	4 (5.6)	3 (4.3)	7 (5.0)
Others	1 (1.4)	1 (1.4)	2 (1.4)

ECOG-PS, Eastern Cooperative Oncology Group performance status; SSA, somatostatin analogues.

Treatment compliance

The median duration of the first-line treatment was 11.2 months [95% confidence interval (CI) 7.3-19.8 months] for everolimus and 7.7 months (95% CI 5.1-13.9 months) for STZ/5-FU; for the second-line treatment, it was 7.4 months (95% CI 3.4-12.4 months) for everolimus and 4.7 months (95% CI 3.5-7.5 months) for STZ/5-FU. The main reasons for discontinuation were PD (55.1% for everolimus versus 53.8% for STZ/5-FU) and AEs (26.5% versus 19.2%, respectively). Renal toxicity caused the discontinuation of STZ/5-FU in one patient (Supplementary Table S6, available at https://doi.org/10.1016/j.esmoop.2025.105922). Dose delays or reductions were required in 59.4% (n = 41) on everolimus and 13.6% (n = 9) on STZ/5-FU during the first-line treatment (P < 0.001); and 54.5% (n = 18) on everolimus and 13.9% (n = 5) on STZ/5-FU during the second-line treatment (P = 0.001).

Efficacy

At data cut-off, 45 (62.5%) patients treated with everolimus and 48 (69.6%) treated with STZ/5-FU experienced PFS events, as assessed by the investigators. No differences were found between the arms for the primary endpoint; the 12-month PFS₁ rates were 71.4% (95% CI 59.4% to 81.6%) and 61.8% (95% CI 49.2% to 73.3%) for everolimus and STZ/5-FU, respectively (OR 0.65, 95% CI 0.32-1.32, P = 0.229). The BIRC 12-month PFS₁ rate was 69.5% (95% CI 58.9% to 82%) versus 63.6% (95% CI 52.7% to 76.6%) for everolimus and STZ/5-FU, respectively (OR 1.05, 95% CI 0.53-2.08, P = 0.677) (Supplementary Figure S2, available at https://doi.org/10.1016/j.esmoop.2025.105922). The Kaplan–Meier estimated investigator-assessed median PFS₁ was outstanding in both groups, 19.4 months (95% CI 16.8-27.6 months) versus 22.7 months (95% CI 13.3-28.6 months) for everolimus and STZ/5-FU, respectively (HR 1.16, 95% CI 0.77-1.75, P = 0.474) (Figure 2A). SSAs were administered concurrently with everolimus in 10 patients (14.5%), reporting a median PFS₁ of 27.2 months [95% CI 9 months to not reached (NR)], which was also not statistically different from that observed with STZ/5-FU in the post hoc analysis (HR 1.4, 95% CI 0.6-3.1, P = 0.428) (Supplementary Figure S3, available at https://doi.org/10.1016/j.esmoop.2025.105922).

Figure 2.

Survival outcomes in arm A (everolimus followed by STZ/5-FU) and arm B (STZ/5-FU followed by everolimus). (A) Kaplan–Meier analysis of progression-free survival (PFS) to first treatment (PFS₁). (B) Multivariable hazard analysis of PFS₁ to first treatment assigned in selected subgroups. (C) Kaplan–Meier analysis of PFS to second treatment (PFS₂). (D) Multivariable hazard analysis of PFS₂ to first treatment assigned in selected subgroups. (E) Kaplan–Meier analysis of overall survival. (F) Multivariable hazard analysis of OS in selected subgroups. The gold dots and error bars represent the HR and 95% CIs. 5-FU, 5-fluorouracil; CI, confidence interval; ECOG-PS, Eastern Cooperative Oncology Group performance status; Eve, everolimus; HR, hazard ratio; OS, overall survival; PFS, progression-free survival; STZ, streptozotocin.

Investigator-assessed median PFS₂ was 8.8 months (95% CI 5.2-30.2 months) versus 9.5 months (95% CI 6.6-19.3 months) (HR 1.6, 95% CI 0.94-2.73, P = 0.082) (Figure 2C and D) and BIRC median PFS₂ was 8.8 months (95% CI 5.2-27.7 months) versus 10.7 months (95% CI 6.6-19.3 months) (HR 1.5, 95% CI 0.88-2.61, P = 0.132) (Supplementary Figure S2, available at https://doi.org/10.1016/j.esmoop.2025.105922) both for STZ/5-FU and everolimus, respectively. The median PFS₁₊₂ was 37.5 months (95% CI 27.1-53.7 months) versus 32.6 months (95% CI 23.7-41.1 months) for everolimus first and STZ/5-FU first, respectively (HR 1.4, 95% CI 0.90-2.19, P = 0.135) (Supplementary Figure S4, available at https://doi.org/10.1016/j.esmoop.2025.105922).

Significant differences in the ORR were observed between everolimus and STZ/5-FU as the first treatment (11.6% versus 30.3%; Fisher´s exact P = 0.012) (Figure 3A). The BIRC assessment reported an ORR of 10.3% (95% CI 4.2% to 20.1%) versus 30.2% (95% CI 19.2% to 43%) for everolimus and STZ/5-FU, respectively (P = 0.004) (Supplementary Table S5, available at https://doi.org/10.1016/j.esmoop.2025.105922). Median DoR was 4.5 months (95% CI 0-35.2 months) for everolimus and 25.2 months (95% CI 22.3-35.1 months) for STZ/5-FU. In the second treatment, STZ/5-FU achieved a higher ORR than everolimus by investigator (30.6% versus 9.1%, Fisher´s exact P = 0.072) and BIRC (30.3% versus 9.7%, Fisher´s exact P = 0.062) (Figure 3C). Median DoR to second treatment was 15 months (95% CI 2.8-32 months) for STZ/5-FU. Only three patients responded to everolimus as second-line treatment with a DoR of 2, 21, and 27 months.

Figure 3.

Objective response rate (ORR) in arm A (everolimus followed by STZ/5-FU) and arm B (STZ/5-FU followed by everolimus). (A) Waterfall plot for overall response rate (ORR) to everolimus and streptozotocin (STZ) plus 5-fluorouracil (5-FU) as first assigned treatment reporting the percentage change from baseline in the sum of the longest diameters of target lesions. (B) Multivariable analysis of ORR to first treatment assigned in selected subgroups. The gold dots and error bars represent the odds ratio and the 95% confidence interval. (C) Waterfall plot for ORR to everolimus and STZ plus 5-FU as second assigned treatment. (D) Multivariable analysis of ORR to the second treatment assigned in selected subgroups. Arm A (sequence everolimus/STZ and 5-FU) in blue and arm B (sequence STZ and 5-FU/Everolimus) in red. 5-FU, 5-fluorouracil; CBR, clinical benefit rate; CI, confidence interval; CR, complete response; ECOG-PS, Eastern Cooperative Oncology Group performance status; NE, not evaluable; OR, odds ratio; ORR, overall response rate; PD, progressive disease; PR, partial response; SD, stable disease; STZ, streptozotocin.

The median follow-up was 36.6 months (95% CI 30.9-39.9 months). Throughout the study period, 36.1% (n = 26) and 50.7% (n = 35) of the patients died in the everolimus first and STZ/5-FU first groups, respectively. The median OS was 61.7 months (95% CI 49.1 months-NR) versus 50.6 months (95% CI 40.9-64.5 months) in the everolimus first and STZ/5-FU first groups, respectively (HR 1.43, 95% CI 0.86-2.37, P = 0.168; Figure 2E).

In the multivariable subgroup analysis, 12-month PFS₁ in grade 1 tumors was 100% (95% CI 100% to 100%) versus 80.8% (95% CI 60% to 100%) for everolimus and STZ/5-FU, respectively (HR 12.01, 95% CI 1.45-99.5) (Figure 2B and Supplementary Figure S5, available at https://doi.org/10.1016/j.esmoop.2025.105922). Multivariable analysis showed higher ORR for STZ/5-FU across patient subgroups, especially in younger patients (9.1% versus 35.1%, OR 6.01, 95% CI 1.69-26.85), those with an ECOG-PS of 0 (12.2% versus 35%, OR 4.73, 95% CI 1.55-16.49), and grade 2 tumors (10.3% versus 30.6%, OR 4.88, 95% CI 1.67-16.28) (Figure 3B). In the second-line treatment, the ORR for grade 2 tumors was 30.3% and 8% for STZ/5-FU and everolimus, respectively (OR 0.15, 95% CI 0.02-0.84) (Figure 3D). The 18-month OS rate in treatment-naive patients was 93.6% (95% CI 85.4% to 100%) and 77.5% (95% CI 65.6% to 91.6%) in everolimus first and STZ/5-FU groups, respectively (HR 2.5, 95% CI 1.13-5.53) (Figure 2F and Supplementary Figure S6, available at https://doi.org/10.1016/j.esmoop.2025.105922). Females and patients with ECOG-PS 1-2 were associated with worse prognosis, regardless of the treatment arm. Tumor grade was the only independent variable that significantly interacted with the treatment arm for PFS₁ and receiving previous treatment of OS (Supplementary Figure S7, available at https://doi.org/10.1016/j.esmoop.2025.105922).

Safety

The toxicity profile revealed a significant increase in oral mucositis, skin disorders (rash, pruritus, and dry skin), hyperglycemia, and edema in patients treated with everolimus. Conversely, gastrointestinal symptoms such as nausea were significantly more common in patients who received STZ/5-FU (Figure 4 and Supplementary Tables S7-S10, available at https://doi.org/10.1016/j.esmoop.2025.105922). Pneumonitis was reported in 7% (n = 5) of patients treated with everolimus as the first-line therapy and in 21% (n = 7) treated with everolimus as the second-line therapy.

Figure 4.

Safety in arm A (everolimus followed by STZ/5-FU) and arm B (STZ/5-FU followed by everolimus). Treatment-related adverse events reported throughout the study period, including most frequent events (cut-off >5%) and those of special interest such as hematologic toxicities in first line (A) and second line (B). P value on the right of the events that had statistically significant differences in rates between study arms (Fisher’s exact test).

In the first-line treatment, grade ≥3 toxicities were reported in 55.1% and 43.9% for everolimus and STZ/5-FU, respectively. In the second-line treatment, grade ≥3 toxicities were reported in 27.8% and 30.3% for STZ/5-FU and everolimus, respectively. For everolimus, specific grade ≥3 toxicities included hyperglycemia (5.8%, n = 4), oral mucositis (4.3%, n = 3), and pneumonitis (1.5%, n = 1). STZ/5-FU was associated with one case of grade ≥3 hyperglycemia. Grade ≥3 neutropenia occurred in 1.4% (n = 1) of patients treated with everolimus and 4.5% (n = 3) with STZ/5-FU, whereas grade ≥3 febrile neutropenia occurred in 1.4% (n = 1) and 3% (n = 2), respectively. Grade ≥3 anemia and decreased platelet count were reported in 4.3% (n = 3) of the patients, all of whom were related to everolimus. Grade 1-2 renal failure occurred in 13.4% (n = 9) of the patients who received STZ/5-FU as the first treatment and 2.8% (n = 1) of those who received it as a second treatment (Supplementary Figure S8, available at https://doi.org/10.1016/j.esmoop.2025.105922). No toxic deaths were observed.

Quality of life

At baseline, the everolimus-first group reported significantly lower scores on the global QLQ-C30, global health status, role, emotional and social functioning scales, and appetite (Supplementary Figure S9, available at https://doi.org/10.1016/j.esmoop.2025.105922). No significant changes in global scores were observed between baseline and subsequent time points in either arm. Throughout the study, there was a general trend toward worsening global status, physical, role, and social functioning, as well as symptom scales in both arms. Specifically, the decline in physical function in everolimus first was statistically significant, with median scores decreasing from 93.3% (95% CI 80% to 93.3%) at screening to 80% (95% CI 60% to 86.7%, P = 0.026) at C2D1 from the second-line treatment, and to 66.7% (40% to 86.7%, P = 0.028) by the end of the second treatment.

Discussion

The SEQTOR trial represents the first study comparing the sequence of two standard therapies in panNETs head-to-head: chemotherapy with STZ/5-FU and mTOR inhibition with everolimus. The study failed to address the sequentiality of treatments but provides meaningful data regarding the efficacy and safety of both study drugs. The results revealed no significant differences in PFS (primary endpoint) or OS between STZ/5-FU and everolimus. The multivariable analysis, despite limitations in sample size and statistical power, suggests that younger patients with ECOG-PS 0 and grade 2 tumors seemed to benefit more from STZ/5-FU in terms of ORR. Conversely, older patients, ECOG-PS 1-2, and those with grade 1 tumors showed better survival outcomes with upfront everolimus. This is in line with prior observations in real-world populations, suggesting a better response to everolimus in tumors with Ki-67 ≤12%.²⁷ Therefore, we hypothesized that tumor grade, patient performance status, and age may help establish an optimal therapy sequence for each patient, which will require validation in larger cohorts of patients through real-world studies. The efficacy shown by tumor grade and Ki-67 subgroups reinforce the recommendations of the ENETS consensus guidelines that positioned chemotherapy as preferred to everolimus in fast-growing tumors, whereas everolimus is preferred in slow-growing tumors.

Interestingly, treatment-naive patients, that is, those not exposed to SSA, showed numerically longer OS with upfront everolimus, which was maintained in the multivariable analysis after adjusting for other prognostic factors. The increase in everolimus treatment resistance from prior SSA exposure, given that somatostatin receptors and mTOR pathways are interconnected,28, 29, 30, 31 might explain this effect. Additionally, a potential benefit of reserving chemotherapy in patients treated with everolimus is suggested, with prolonged PFS₂ observed when STZ/5-FU was used in the second-line.

Regarding response rates, STZ/5-FU demonstrated superiority in both the first- and second-line treatments, with BIRC assessment reporting an ORR of 30.2%. This benefit in response rate was consistent across subgroups, particularly in younger patients with good performance status, tumor grade 2, and those who had received prior treatments. The DoR to STZ/5-FU, reaching a median of 25.2 months, was longer than that reported in previous studies.^13,14,32

Survival outcomes were comparable with those reported in recent retrospective studies.^13,14,33 However, possibly influenced by the inclusion of treatment-naive patients with Ki-67 <10%, these outcomes exceed those reported in benchmark studies.^{8,9,11,12,32,34} Both treatments demonstrated efficacy outcomes in line with those of other approved¹⁰ or investigational therapies for panNETs.^16,18,35, 36, 37, 38, 39 CAPTEM may be another suitable alternative (with similar response rate reported in the literature¹⁶) and may be more appealing to some patients because of its oral administration, although data on patient’s preferences are lacking in the literature. The approval of PRRT after the SEQTOR trial initiation changed the therapeutic landscape. Based on our data and those from both clinical trials and observational studies to date,^17,20,40 PRRT should still be considered a second-line treatment option for patients with positive somatostatin receptor positron emission tomography with computed tomography (SST-PET/CT).

The inclusion of treatment-naive patients was justified, as SSA studies have not reported a benefit in OS, which justifies its frontline use, and their efficacy in panNETs is lower than in tumors of intestinal origin. In contrast, cytoreductive therapies, which show efficacy in first-line treatment regardless of Ki-67,^17,21 are potential upfront options for selected patients, as specified in the protocol inclusion criteria (e.g. bulky disease or liver involvement). In our study, only 10 patients (14.5%) received everolimus concurrently with SSA, mostly for symptom control. Although permitted by the protocol, SSA coadministration was not mandatory. The combination with SSA is expected to improve outcomes based on data from the recently reported STARTER-NET trial, which reported a median PFS of 29.7 months.⁴¹ However, this subgroup did not show increased efficacy, nor were significant differences observed when compared with STZ/5-FU. Dedicated prospective studies should explore combination approaches rather than single-agent everolimus.

The SEQTOR study safety profile was consistent with previous trials.^8,9,11, 12, 13, 14, 15^,32,34 The safety profiles of both drugs have particularities that should be considered for treatment assignment. Mucositis, skin disorders, hyperglycemia, and pneumonitis were more frequently reported in patients treated with everolimus, whereas renal and gastrointestinal toxicities were more frequent in those treated with STZ/5FU. Renal toxicity was mild, and most events were grade 1-2. Pneumonitis with everolimus was higher in the second-line treatment, which might be justified by cumulative toxicities and worse ECOG-PS in later treatment lines. Of note, everolimus required a high rate of dose modifications (59%), consistent with previous reports.^8,9,27 Grade 3-4 toxicities were in the range for everolimus and STZ/5-FU (55.1% versus 43.9%) and comparable with other orally available chemotherapeutic schemes.¹⁶

The SEQTOR study had several limitations. Firstly, the study was initially designed with PFS₁₊₂ as the primary endpoint but we were pushed to change it to a shorter-term outcome due to the slow accrual driven by the emergence of new systemic treatments that modified the clinical guidelines and jeopardized the original sequential design. The 12-month PFS₁ was ultimately adopted as the primary endpoint to allow direct comparison between everolimus and STZ/5-FU because it could be evaluated with shorter follow-up and had previously been used by other trials in this setting.¹⁷ Based on previous studies, PFS at 12 months was expected to be sufficiently mature.⁵ Some patients had their CT scans carried out within some weeks’ range from the expected 12-months following randomization but the impact of such a variability range in the timepoint for CT scans was minimized by several sensitivity analyses and was considered almost null. Subgroup analyses were exploratory and involved small sample sizes, limiting the interpretation of results. The arms were unbalanced for baseline QoL, which may have affected the interpretation of QoL outcomes, introducing a selection bias. BIRC assessment increased the validity of the results because it was highly concordant with our main investigator assessment. It served to reinforce results and mitigate some of the limitations of the study.

In conclusion, although the assessment of sequentiality was not feasible, both strategies provide comparable 12-month PFS₁ rates, disease control, and OS, indicating that neither is superior as an initial option. STZ/5-FU achieves higher response rates regardless of treatment sequence, suggesting its selection as the preferred therapy when cytoreduction is necessary. However, differences observed in exploratory subgroup analysis suggest individualizing treatment based on patient characteristics, considering tumor grade, expected treatment effect, safety profile, and comorbidities.