NCI9673 Part B
NCI9673 (Part B): ETCTN Randomized Phase II Study of Nivolumab With or Without Ipilimumab in Refractory, Metastatic Squamous Cell Carcinoma of the Anal Canal
Phase 2N = 100NCT02314169
- Population
- Patients with refractory, incurable anal cancer (at least one prior systemic therapy for incurable disease, no prior immunotherapy); 100 randomised and treated (52 nivolumab, 48 nivolumab plus ipilimumab).
- Design
- Randomised phase II trial of the NCI Experimental Therapeutics Clinical Trials Network (Part B of NCI9673). Patients were randomly assigned to nivolumab 480 mg intravenously every 4 weeks alone or with ipilimumab 1 mg/kg every 8 weeks. Primary endpoint: progression-free survival, compared with a log-rank test at a one-sided alpha of 0.1 and 90% power. Secondary endpoints: radiographic response, overall survival and grade 3 or worse adverse events. Accrual October 2018 to May 2024; data cutoff 18 June 2024; at least one prior systemic therapy for incurable disease and no prior immunotherapy required; no crossover from nivolumab alone to the combination.
- Primary endpoint
- PFS
- Arms
- Nivolumab (control, n = 52): Nivolumab 480 mg intravenously every 4 weeks
- Nivolumab plus ipilimumab (experimental, n = 48): Nivolumab 480 mg intravenously every 4 weeks with ipilimumab 1 mg/kg every 8 weeks
- People
- Morris VK, senior author Eng C
PFSHR 0.86Δ median PFS+0.8 monthsΔ ORR+4.1 pointsOSHR 0.98Δ median OS+4.1 monthsΔ Grade 3 or worse treatment-related adverse events+13 points
Show the numbers
| Endpoint | Analysis | Nivolumab plus ipilimumab | Nivolumab | Effect |
|---|---|---|---|---|
| PFSRandomised patients (n = 100); the full text does not state the analysis population | primaryMorris 2026 | median 3.7 months | median 2.9 months | HR 0.86P=.25The abstract gives 90% CIs for the medians, not 95% CIs: 1.9-3.8 months (nivolumab) and 2.0-5.6 months (nivolumab plus ipilimumab). The hazard ratio of 0.86 has an interval of 0.60-1.23 that the abstract labels 95% and the Results text labels 90%; with a one-sided alpha of 0.1 and 90% CIs for the medians, the interval is not stored as a 95% interval. The paper names no reference arm; the ratio is read as nivolumab plus ipilimumab relative to nivolumab alone (direction inferred from the medians and the Results wording). One-sided log-rank P = .25, alpha 0.1. |
| ORRParticipants evaluable for response (88: 46 nivolumab, 42 nivolumab plus ipilimumab) per the Results text | primaryMorris 2026 | 21.5% | 17.4% | P=.89Among 88 response-evaluable participants (46 on nivolumab alone, 42 on the combination) the Results give 95% CIs of 9.1-31% and 12-36%. 17.4% matches 8 of 46; 21.5% is not an exact fraction of 42 (9 of 42 is 21.4%), so the percentages are kept as printed. |
| OSRandomised patients (n = 100); the full text does not state the analysis population | primaryMorris 2026 | median 20 months | median 15.9 months | HR 0.98The abstract gives a 90% CI for the hazard ratio (0.63-1.51); the Results text prints 0.64-1.51, also a 90% CI. Neither is a 95% interval, so none is stored. The paper names no reference arm (read as nivolumab plus ipilimumab relative to nivolumab alone). |
| Grade 3 or worse treatment-related adverse eventsTreated patients (100: 52 nivolumab, 48 nivolumab plus ipilimumab) | primaryMorris 2026 | 25% | 12% | Six patients (12%) with nivolumab alone and 12 patients (25%) with nivolumab plus ipilimumab. |
| Endpoint | Analysis | Nivolumab plus ipilimumab | Nivolumab | Effect |
|---|---|---|---|---|
| PFSRandomised patients (n = 100); the full text does not state the analysis population | primaryMorris 2026 | median 3.7 months | median 2.9 months | HR 0.86P=.25The abstract gives 90% CIs for the medians, not 95% CIs: 1.9-3.8 months (nivolumab) and 2.0-5.6 months (nivolumab plus ipilimumab). The hazard ratio of 0.86 has an interval of 0.60-1.23 that the abstract labels 95% and the Results text labels 90%; with a one-sided alpha of 0.1 and 90% CIs for the medians, the interval is not stored as a 95% interval. The paper names no reference arm; the ratio is read as nivolumab plus ipilimumab relative to nivolumab alone (direction inferred from the medians and the Results wording). One-sided log-rank P = .25, alpha 0.1. |
| ORRParticipants evaluable for response (88: 46 nivolumab, 42 nivolumab plus ipilimumab) per the Results text | primaryMorris 2026 | 21.5% | 17.4% | P=.89Among 88 response-evaluable participants (46 on nivolumab alone, 42 on the combination) the Results give 95% CIs of 9.1-31% and 12-36%. 17.4% matches 8 of 46; 21.5% is not an exact fraction of 42 (9 of 42 is 21.4%), so the percentages are kept as printed. |
| OSRandomised patients (n = 100); the full text does not state the analysis population | primaryMorris 2026 | median 20 months | median 15.9 months | HR 0.98The abstract gives a 90% CI for the hazard ratio (0.63-1.51); the Results text prints 0.64-1.51, also a 90% CI. Neither is a 95% interval, so none is stored. The paper names no reference arm (read as nivolumab plus ipilimumab relative to nivolumab alone). |
| Grade 3 or worse treatment-related adverse eventsTreated patients (100: 52 nivolumab, 48 nivolumab plus ipilimumab) | primaryMorris 2026 | 25% | 12% | Six patients (12%) with nivolumab alone and 12 patients (25%) with nivolumab plus ipilimumab. |
- Nivolumab (control)
- Nivolumab plus ipilimumab (experimental)
Median-only curves assume a constant hazard (exponential) and fit plateaus, as seen with some immunotherapies, poorly.
Data table (modelled values)
| Months | Nivolumab | Nivolumab plus ipilimumab |
|---|---|---|
| 0 | 100.0% | 100.0% |
| 2 | 62.0% | 68.8% |
| 5 | 30.3% | 39.2% |
| 7 | 18.8% | 26.9% |
| 9 | 11.6% | 18.5% |
- Nivolumab (control)
- Nivolumab plus ipilimumab (experimental)
Median-only curves assume a constant hazard (exponential) and fit plateaus, as seen with some immunotherapies, poorly.
Data table (modelled values)
| Months | Nivolumab | Nivolumab plus ipilimumab |
|---|---|---|
| 0 | 100.0% | 100.0% |
| 13 | 56.7% | 63.7% |
| 25 | 33.6% | 42.0% |
| 38 | 19.1% | 26.8% |
| 50 | 11.3% | 17.7% |