A single‑center, open‑label phase II trial tested combined ipilimumab and nivolumab in patients with advanced, progressive well‑differentiated neuroendocrine tumors (WD‑NET). The protocol used Simon’s two‑stage design with the primary endpoint defined as objective response rate (ORR) by RECIST v1.1 and a predefined activity threshold of greater than 15%. Investigators administered ipilimumab 1 mg/kg every six weeks and nivolumab 240 mg every two weeks, with treatment allowed for up to two years. The trial planned to accrue 56 patients overall but closed early; the source reported that Stage 1 enrollment was incomplete at the time of closure.
Nine patients were enrolled before the trial stopped accrual. The cohort included six patients with lung NET, two with pancreatic NET, and one with small‑bowel NET. The mean age across enrolled patients was 63.7 years. The source specifies that inclusion required measurable disease per RECIST 1.1, documented progression within 12 months of enrollment, and prior lines of therapy consistent with the protocol (lung NET after at least one line; GI/pancreas after at least two lines). The article indicates standard laboratory and performance status eligibility criteria were applied, and that archival or new core biopsies were required for correlative studies.
Of the nine enrolled patients, eight were evaluable for response according to RECIST v1.1. Four of these eight evaluable patients achieved a confirmed objective response, yielding an observed ORR of 50% (95% CI: 16–84) in the evaluable group. The source notes that confirmation of responses followed RECIST v1.1 rules (confirmation on a subsequent scan at least four weeks after initial response). The investigators explicitly stated that, because accrual stopped early, no formal statistical conclusion could be applied to efficacy outcomes. The article does not present median progression‑free survival or overall survival estimates for this prematurely closed cohort within the provided text.
Five of nine patients experienced immune‑related adverse events while on study. The source lists immune‑related toxicity as an important secondary endpoint and indicates that toxicity grading used NCI CTCAE v4.0. Treatment discontinuation rules, dose‑delay criteria, and monitoring procedures were described in the protocol. The excerpted text refers to tables that enumerate immune‑related adverse events and serious adverse events; however, detailed per‑patient grade distributions and management outcomes are provided in those article tables rather than in the passage quoted here.
Correlative analyses were performed on archival or newly obtained FFPE tumor tissue. The study team used immunohistochemistry to evaluate immune infiltrates (CD3, CD8, CD20, PD‑1) and PD‑L1 expression. Digital image analysis quantified cell density per mm2 and percent‑stained surface area in annotated tumor and invasive front regions. Investigators reported a higher CD3 infiltration in pretreatment tumors from patients who achieved radiologic responses compared with nonresponders. The source specifies automated PD‑L1 staining protocols and dual CD20/CD3 staining methods; the article includes technical IHC parameters and blocking/detection workflows. The text does not provide a full numeric breakdown of PD‑L1 positivity or other biomarker thresholds in the narrative excerpt, and those results are presented within the article figures and tables.
The protocol defined ORR as confirmed complete response or partial response divided by the number of treated subjects with at least one RECIST scan, with exact binomial 95% confidence intervals calculated for ORR. The two‑stage Simon design planned 20 patients in Stage 1 and 56 total patients overall. The trial closed prior to completing Stage 1 accrual (9 of 20 planned for Stage 1), and the source explicitly states that no formal statistical conclusion could be drawn because of incomplete accrual. The observed 50% ORR therefore represents an observed signal in a small number of evaluable patients and is described by the authors as hypothesis‑generating rather than definitive proof of activity. The article includes tables listing baseline characteristics, response data, and adverse events; the narrative excerpt focuses on primary results and correlative findings without providing additional time‑to‑event endpoints in the quoted section.
The clinical trial is registered on ClinicalTrials.gov under identifier NCT03420521. The source article refers readers to figures and multiple tables that contain the full baseline demographics, granular response data, immune‑related adverse event details, serious adverse event listings, and IHC quantification methods. These supplementary materials are necessary to review per‑patient outcomes, toxicity grades, and full correlative biomarker data that the narrative excerpt does not enumerate.
Limitations and unresolved questions documented in the source
The principal limitation reported by the authors is premature trial closure with incomplete Stage 1 accrual, which prevents formal statistical inference. Sample size was small (nine enrolled; eight evaluable), and the authors emphasize that the observed ORR of 50% should be considered hypothesis‑generating. The source did not report comprehensive time‑to‑event outcomes (median PFS or OS) for this cohort within the provided excerpt. Additional questions remain about the durability of responses, the full toxicity profile by grade for the cohort, and the reproducibility of the CD3 infiltration association with response in larger, fully accrued cohorts.