Why This Study Matters
Adding immune checkpoint inhibitors to neoadjuvant chemotherapy has already improved outcomes in high-risk, early-stage HER2-negative breast cancer, but a meaningful share of patients still fail to reach a pathologic complete response (pCR) — the surrogate endpoint most closely tied to long-term outcomes in this setting. The I-SPY2 platform trial, an ongoing adaptive Phase 2 study that has enrolled more than 2,500 patients and tested more than 25 investigational regimens since 2010, set out to test whether dual checkpoint blockade — simultaneously blocking PD-1 and LAG-3 — could push response rates higher than single-agent immunotherapy has achieved to date.
Study Design
I-SPY2 is a randomized, open-label, adaptively randomized platform trial that has continuously enrolled patients since 2010, assigning them to experimental or control arms based on receptor subtype (hormone receptor and ERBB2/HER2 status) and MammaPrint molecular risk category (high [MP1] or ultrahigh [MP2]). For this analysis, patients were adaptively randomized between February 2, 2020, and December 9, 2021, with data analyzed from January 1, 2022, through August 5, 2025.
The trial enrolled 78 patients in the PCF arm, compared against 350 historical controls. Both arms received weekly paclitaxel for 12 weeks followed by doxorubicin and cyclophosphamide, then surgery; the intervention arm additionally received four doses of cemiplimab (anti-PD-1) and fianlimab (anti-LAG-3), administered every three weeks — the PCF regimen.
Patient Population
Enrolled patients had stage II or III ERBB2 (HER2)-negative breast cancer classified as high recurrence risk. The PCF arm had a mean age of 47 years (range 39-54); the historical control population had a mean age of 48 years (range 39-57).
Primary Endpoint Results
The primary endpoint was pathologic complete response (pCR). PCF nearly doubled pCR rates relative to control across the overall ERBB2-negative population and both major receptor-defined subtypes:
| Population | PCF (95% CI) | Control (95% CI) |
|---|---|---|
| All ERBB2-negative | 44% (34%-53%) | 21% (17%-25%) |
| Triple-negative | 53% (39%-67%) | 29% (22%-36%) |
| HR-positive / ERBB2-negative | 36% (23%-49%) | 14% (9%-19%) |
Treatment arms in the I-SPY2 platform "graduate" once they reach an 85% Bayesian predictive probability of success in a hypothetical subtype-specific Phase 3 confirmatory trial. PCF graduated across all clinical signatures evaluated.
Subgroup Analyses
Beyond the receptor-defined subgroups above, drawn directly from the published abstract, the study evaluated a predictive immune biomarker called ImPrint. The abstract states qualitatively that PCF "was found to be highly effective" in ImPrint-positive patients, without providing exact figures. Independent institutional and news coverage of this same publication reports substantially higher pCR rates in ImPrint-positive patients than the subtype averages above:
| ImPrint-Positive Subgroup | Reported pCR | Source |
|---|---|---|
| Triple-negative, ImPrint-positive | 83% | External press coverage |
| HR-positive/ERBB2-negative, ImPrint-positive | 91% | External press coverage |
These two ImPrint-positive figures are not stated as exact numbers in the published abstract itself. They are reported consistently by an institutional press release and independent health-news coverage of this same JAMA Oncology publication (matching DOI and publication date), but should be treated as externally sourced rather than abstract-verified until the full peer-reviewed text is available.
Safety Profile
Among the 78 PCF-treated patients, 16 (21%) experienced adrenal insufficiency, including hypophysitis, with 11% classified as grade 3 or 4 in severity. These endocrine adverse events mostly occurred after immunotherapy had already been completed — a pattern the abstract specifically calls out.
Additional detail from external coverage of this publication indicates that diabetes occurred in 4% of PCF patients. Reporting citing the study authors indicates that, because of this toxicity signal, the PCF regimen is not moving directly into a dedicated Phase 3 confirmatory trial, and the investigators are evaluating lower-dose strategies and biomarker-based patient selection instead.
| Adverse Event | Rate in PCF Arm | Source |
|---|---|---|
| Adrenal insufficiency (incl. hypophysitis) | 21% (16/78); 11% grade 3-4 | Abstract |
| Diabetes (new-onset) | 4% | External press coverage |
The bottom line is that we want to individually treat people with the drugs they need and avoid toxicities.
— Claudine Isaacs, MD, lead author, Georgetown Lombardi Comprehensive Cancer Center (quoted in institutional and news coverage of the publication)
Interpretation and Broader Context
These results add to a growing body of evidence that dual checkpoint blockade — pairing a PD-1 inhibitor with a LAG-3 inhibitor — can meaningfully increase pathologic complete response over chemotherapy alone in high-risk, ERBB2-negative early breast cancer, with the largest absolute gains seen in the triple-negative and HR-positive/ERBB2-negative subtypes. At the same time, the substantial rate of adrenal insufficiency and other endocrine immune-related adverse events, some emerging weeks after treatment ended, means this specific four-dose PCF regimen is not yet positioned as a standard-of-care option. The findings reinforce the importance of biomarker-guided patient selection — using tools like the ImPrint immune signature — to concentrate the benefit of intensified immunotherapy in patients most likely to respond, while sparing others from added toxicity.
Future Directions
The published abstract states that these results "warrant further definitive trials." External coverage of the study indicates the investigators are now evaluating dose modifications and ImPrint-based patient selection before pursuing a dedicated Phase 3 confirmatory trial of this regimen.
In this Phase 2 arm of the I-SPY2 platform trial, adding dual PD-1/LAG-3 blockade (cemiplimab plus fianlimab) to neoadjuvant chemotherapy nearly doubled pathologic complete response in high-risk, ERBB2-negative breast cancer — 44% vs 21% overall, with the largest gains in triple-negative (53% vs 29%) and HR-positive/ERBB2-negative (36% vs 14%) disease. The regimen met the platform's graduation threshold across all clinical signatures, but a 21% rate of adrenal insufficiency — often emerging after treatment ended — means the four-dose PCF regimen is headed toward biomarker-guided refinement rather than directly into a Phase 3 confirmatory trial.