Neoadjuvant chemoimmunotherapy was associated with higher pathologic and radiographic response rates than neoadjuvant immunotherapy alone in patients with resectable head and neck squamous cell carcinoma, according to a systematic review and meta-analysis of early-phase, mostly single-arm prospective trials — findings that support further study but cannot establish superiority over immunotherapy alone or demonstrate a survival benefit.
Researchers reviewed 23 prospective phase 1 and phase 2 trials that included 751 patients with previously untreated, resectable head and neck squamous cell carcinoma (HNSCC). The analysis included 357 patients who received neoadjuvant chemoimmunotherapy, 102 who received dual-agent immunotherapy, and 292 who received single-agent immunotherapy prior to planned definitive surgery.
The most common primary tumor site was the oral cavity, and most patients with reported staging had locally advanced disease. HPV-associated cancer was reported in 14% of patients, limiting interpretation in HPV-positive disease, particularly for physicians treating patients with HPV-positive oropharyngeal HNSCC.
The primary outcomes included major pathologic response (MPR), complete pathologic response (CPR), and complete radiographic response by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. MPR was defined as less than 10% viable tumor remaining after resection, and CPR was defined as no viable tumor after resection. Secondary outcomes included 1-year overall survival and grade 3 to 5 adverse events.
Pooled MPR plus CPR rates were 66% among patients treated with chemoimmunotherapy, compared with 18% among those treated with dual-agent immunotherapy and 6% among those treated with single-agent immunotherapy. CPR rates were 38%, 5%, and approximately 3%, respectively.
No pathologic response was reported in 78% of patients receiving single-agent immunotherapy, compared with 19% of those receiving chemoimmunotherapy.
Radiographic response rates also favored chemoimmunotherapy, although radiographic response appeared to underestimate pathologic response across all treatment groups. Partial or complete radiographic responses occurred in 66% of patients receiving chemoimmunotherapy, 9% receiving dual-agent immunotherapy, and 6% receiving single-agent immunotherapy. Stable disease by RECIST 1.1 criteria was reported in 73% of patients receiving single-agent immunotherapy, 63% receiving dual-agent immunotherapy, and 30% receiving chemoimmunotherapy. Progressive disease occurred in 4% of patients in the chemoimmunotherapy group, compared with 26% in the dual-agent immunotherapy group and 17% in the single-agent immunotherapy group.
Not all patients proceeded to surgery after neoadjuvant therapy. The researchers reported that 2% of patients receiving single-agent immunotherapy, 12% receiving dual-agent immunotherapy, and 15% receiving chemoimmunotherapy did not undergo surgery after neoadjuvant treatment. The most common reason was patient refusal, followed by protocol deviations leading to study exclusion. This is clinically relevant in the curative-intent setting, where delayed or omitted surgery may offset potential benefits of neoadjuvant therapy, though in the chemoimmunotherapy group some attrition may have reflected response-driven treatment modification — one of the goals being explored with neoadjuvant approaches in this setting.
Overall survival was reported infrequently and was not sufficient for comparing regimens. Among studies reporting 1-year overall survival, rates ranged from 94% to 100% in chemoimmunotherapy studies, 95% in dual-agent immunotherapy studies, and 88% to 96% in single-agent immunotherapy studies.
Among studies that reported toxic effects, grade 3 to 5 adverse events occurred in 17% of patients receiving chemoimmunotherapy, 3% receiving dual-agent immunotherapy, and 29% receiving single-agent immunotherapy. No grade 5 adverse events were reported. The researchers cautioned that adverse events were inconsistently reported across studies, and the rates should not be interpreted as evidence that single-agent immunotherapy is more toxic than chemoimmunotherapy.
The study had several limitations. Many included trials were single-group studies, preventing direct comparisons between immunotherapy and chemoimmunotherapy. Heterogeneity was high across chemoimmunotherapy trials, and subgroup analyses by disease subsite, stage, or HPV status could not be performed. Several studies had missing data, and survival outcomes and adverse events were limited by inconsistent reporting.
The findings may be most relevant to patients with HPV-negative, locoregionally advanced disease, but this remains hypothesis-generating. The low proportion of patients with HPV-associated disease in the analysis limits generalizability to HPV-positive HNSCC.
In an invited commentary, Parth Desai, MD, of Fox Chase Cancer Center, and colleagues wrote that the study offered a timely synthesis focused on adding chemotherapy to neoadjuvant immunotherapy. However, they emphasized that pathologic response has not been fully validated as a surrogate for survival in HNSCC — a caveat central to interpreting the findings. Although MPR and CPR rates were higher with chemoimmunotherapy, it remains unclear whether those responses will translate into improved overall survival, durable disease control, or safer treatment de-escalation. The commentary noted that in KEYNOTE-689, only approximately 1 in 10 patients experienced a major pathologic response and overall survival assessment remained ongoing, yet perioperative immunotherapy favorably modified surgical pathology, resulting in fewer high-risk pathologic features and a reduced need for adjuvant chemotherapy in the experimental arm.
The researchers concluded that future phase 3 trials should directly compare neoadjuvant chemoimmunotherapy with immunotherapy alone in patients with resectable HNSCC and should evaluate survival, treatment de-escalation, patterns of failure, toxicity, surgical feasibility, and attrition.
Disclosures: Dr. Van Abel reported consulting for Intuitive Surgical without compensation. Dr. Price reported consulting for PDS Biotechnology. Dr. McGarrah reported consulting for Techspert and Biolumina. Dr. Routman reported nonfinancial support from Naveris, grant support from the Paul Calabresi Program in Clinical/Translational Research at the Mayo Clinic Comprehensive Cancer Center, advisory board participation for Adela with remuneration paid to Mayo Clinic, and a patent licensed to Exact Sciences. Dr. Desai reported personal fees from Abdera, Rigel, Boehringer Ingelheim, and Merck and research support from Ride Hard Breathe Easy. Dr. Gross reported grants from Ascendis and Regeneron and personal fees from Regeneron, Merck, GeoVax, Pyxis, AiCME, Med Learning Group, OncLive, and UpToDate. Dr. Contrera reported research support from the National Institutes of Health/National Center for Advancing Translational Sciences, the American College of Surgeons, Haystack Oncology, Droplet Biosciences, and Naveris.
