Highlight
– In a pooled secondary analysis of 3 randomized trials (NRG/RTOG 0129, 0522, 1016), radiotherapy (RT) interruptions were common (28% of patients) and, when quantified by duration, were associated with clinically meaningful increases in locoregional failure (LRF) and worse overall survival (OS).
– Each additional 7 days of RT interruption corresponded to a hazard ratio of ~1.45 for LRF and ~1.41 for OS; absolute 3‑year LRF increases were larger for p16‑negative or non‑oropharyngeal cancers and for advanced T/N stage.
Background
Head and neck squamous cell carcinoma (HNSCC) is curable in many patients when treated with combined‑modality therapy that commonly includes definitive radiotherapy (RT) with or without systemic therapy. Evidence accumulated over decades has linked prolonged overall treatment time and unplanned RT interruptions to worse tumor control, a biologic finding attributed to accelerated tumor cell repopulation during gaps in treatment. During the COVID‑19 pandemic, disruptions to cancer care raised concerns that missed or delayed RT fractions might translate into higher recurrence and mortality. Assessing the magnitude of harm associated with RT interruptions, and whether harms differ by HPV (p16) status and disease stage, has direct implications for triage and mitigation strategies in routine and crisis settings.
Study design
This article summarizes and interprets a secondary pooled analysis of individual patient data from three contemporary randomized trials: NRG/RTOG 0129, 0522, and 1016 (ClinicalTrials.gov identifiers NCT00047008; NCT00265941; NCT01302834). The parent trials enrolled patients with locoregionally advanced HNSCC treated with definitive RT plus systemic therapy (various concurrent chemotherapy or targeted agents depending on the trial). For this secondary analysis, the authors included patients treated with RT and classified disease into two groups: (1) p16‑positive oropharyngeal squamous cell carcinoma (p16+ OPSCC), and (2) p16‑negative oropharynx plus all other head and neck subsites regardless of p16 status (termed locally advanced HNSCC, LAHNSCC).
Primary exposures were: (a) occurrence of any RT interruption (binary yes/no), and (b) duration of RT interruption (continuous, modeled per 7‑day increment). Outcomes were locoregional failure (LRF) and overall survival (OS). Multivariable Cox proportional hazards models adjusted for relevant clinical and tumor characteristics; interaction testing evaluated effect modification by p16 status.
Key findings
Population and exposure:
- 1549 patients were included in the binary interruption analysis (mean age 57 years, 12.9% female); 1048 (67.7%) had p16+ OPSCC and 501 (32.3%) were LAHNSCC.
- Overall, 439 patients (28.3%) experienced at least one RT interruption.
- Length of interruption was available for 1083 patients (69.9%).
Binary interruption results (any gap vs none):
- Any RT interruption was associated with an adjusted hazard ratio (HR) for LRF of 1.04 (95% CI, 0.90–1.36) — not statistically significant.
- The adjusted HR for OS with any interruption was 1.22 (95% CI, 0.99–1.50) — a point estimate suggesting worse survival that narrowly missed conventional statistical significance.
Duration‑based results (continuous predictor):
- Each additional 7‑day RT interruption was associated with substantially greater risk: HR for LRF 1.45 (95% CI, 1.12–1.89); HR for OS 1.41 (95% CI, 1.07–1.86).
- These associations were consistent across p16 groups (no evidence of effect modification), but absolute harms differed by p16 status and stage.
Predicted absolute impacts:
- Using covariate‑adjusted model predictions, a mean 7‑day interruption was associated with a 3‑year absolute increase in LRF of 4.1% for patients with p16+ OPSCC and 9.1% for patients with LAHNSCC.
- Predicted 3‑year LRF detriment ranged from about 2.0% for a patient with non‑T4, non‑N3 p16+ OPSCC to 11.2% for a patient with LAHNSCC with T4N3, p16‑negative disease.
Interpretation of effect sizes
The dichotomous analysis (any interruption) diluted signal seen when interruption length is modeled continuously. This suggests that small, brief, isolated disruptions may have limited impact, while longer cumulative gaps substantially increase the risk of locoregional relapse and mortality. The magnitude of risk per 7‑day increment is clinically meaningful, particularly for patients with p16‑negative tumors and high T/N stage, who already have higher baseline recurrence risk.
Expert commentary and biological plausibility
Biologically, the findings are concordant with long‑standing radiobiological evidence that clonogenic tumor cells can undergo accelerated repopulation when treatment is interrupted, reducing the effective tumoricidal dose of RT delivered. Clinically, earlier observational series and systematic reviews have reported worse local control and survival with prolonged overall treatment time in head and neck cancers; this pooled randomized‑trial secondary analysis adds contemporary, patient‑level evidence from trials with standardized RT delivery and central oversight.
From a clinical perspective, these results emphasize several points:
- Avoiding treatment interruptions should remain a high priority. Even modest average delays (e.g., a week) were associated with measurable absolute increases in recurrence risk in higher‑risk patients.
- The absolute harm is heterogeneous: p16+ OPSCC patients—especially those with early T/N categories—have lower baseline LRF, and therefore the absolute increase due to an interruption is smaller; in contrast, p16‑negative, advanced tumors show larger absolute detriment and may warrant prioritized mitigation efforts.
- When interruptions are unavoidable, compensatory measures (e.g., treatment acceleration, weekend treatments, or adjusted dosimetry) are used in practice, but optimal corrective strategies were not evaluated in this analysis and require prospective evaluation or guideline‑based individualized decisions.
Limitations
- Secondary analysis: Although based on randomized trial cohorts, the exposure (treatment interruption) was not randomized; residual confounding is possible (for example, interruptions may correlate with intercurrent illness, poor performance status, or social determinants that themselves influence outcomes).
- Missing data on interruption length for ~30% of patients could introduce bias if data were not missing at random.
- Details on reasons for interruption, compensatory RT strategies, or subsequent systemic therapy modifications were limited; these factors could modify outcome associations.
- Trial populations may differ from contemporary community practice (selection bias), although the multi‑trial pooled design improves generalizability relative to single‑institution series.
Clinical implications and practice considerations
For clinicians and multidisciplinary teams treating patients with HNSCC, the evidence supports the following pragmatic principles:
- Proactively identify and address barriers to on‑schedule RT delivery (transport, nutrition, dental issues, comorbidity optimization, social support, and infection control in pandemics).
- Prioritize uninterrupted RT scheduling for p16‑negative and advanced‑stage patients when resource constraints exist (triage during surges, staffing shortages, or patient illness).
- When interruptions occur, involve the radiation oncology team early to consider evidence‑based compensatory approaches (accelerated fractionation, weekend treatments, or dose adjustments) while balancing toxicity; consult institutional protocols or national guidelines.
- Document reasons for interruptions and ensure robust survivorship and follow‑up planning, as these patients may be at higher risk and benefit from close surveillance.
Conclusion
This secondary analysis of three large randomized trials demonstrates that radiotherapy interruptions in head and neck cancer are associated, in a duration‑dependent manner, with increased locoregional failure and worse overall survival. The absolute magnitude of harm is greater in p16‑negative and high‑stage disease. While brief, isolated scheduling delays may have limited effect for lower‑risk p16+ oropharyngeal cancers, clinicians should endeavor to avoid or mitigate interruptions—particularly for higher‑risk patients. These findings inform prioritization and mitigation strategies during resource limitations and reinforce the radiation oncology principle that treatment continuity matters.
Funding and clinicaltrials.gov
The pooled analysis used data from NRG/RTOG 0129, 0522, and 1016. ClinicalTrials.gov identifiers: NCT00047008; NCT00265941; NCT01302834. Funding sources are reported in the primary article (Gharzai et al., 2025).
References
1. Gharzai LA, Morris E, Schipper MJ, et al. Treatment Interruption and Outcomes in Head and Neck Cancer: A Secondary Analysis of 3 Randomized Clinical Trials. JAMA Otolaryngol Head Neck Surg. 2025 Dec 4:e254203. doi:10.1001/jamaoto.2025.4203.
2. Ang KK, Harris J, Wheeler R, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med. 2010 Jul 1;363(1):24‑35.
3. Bese NS, Hendry J, Jeremic B. Effect of prolongation of overall treatment time due to unplanned interruptions during radiotherapy of head and neck cancer: a systematic review. Radiother Oncol. 2007 Nov;83(1):24‑30.
4. National Comprehensive Cancer Network (NCCN). NCCN Clinical Practice Guidelines in Oncology: Head and Neck Cancers. Accessed 2025. https://www.nccn.org
5. American Society for Radiation Oncology (ASTRO). Statements and guidance for radiation therapy during the COVID‑19 pandemic. 2020. https://www.astro.org
Author note
This article was prepared by a medical science writer to summarize and interpret the findings of the referenced secondary analysis for clinicians and policy‑minded readers. It is not a guideline and does not replace individualized clinical judgment.