Advertisement

Commentary Open Access | 10.1172/JCI198351

Genomic adjusted radiation dose stratifies radiotherapy dosing based on tumor-specific sensitivity in HPV+ oropharyngeal cancer

Sandip K. Rath and David S. Yu

Department of Radiation Oncology and Winship Cancer Institute, Emory University School of Medicine, Atlanta, Georgia, USA.

Address correspondence to: David S. Yu, Department of Radiation Oncology, Emory University School of Medicine, 1365 Clifton Rd. NE, C3008, Atlanta, Georgia, 30322, USA. Phone: 404.778.1758; Email: dsyu@emory.edu.

Find articles by Rath, S. in: PubMed | Google Scholar

Department of Radiation Oncology and Winship Cancer Institute, Emory University School of Medicine, Atlanta, Georgia, USA.

Address correspondence to: David S. Yu, Department of Radiation Oncology, Emory University School of Medicine, 1365 Clifton Rd. NE, C3008, Atlanta, Georgia, 30322, USA. Phone: 404.778.1758; Email: dsyu@emory.edu.

Find articles by Yu, D. in: PubMed | Google Scholar

Published October 1, 2025 - More info

Published in Volume 135, Issue 19 on October 1, 2025
J Clin Invest. 2025;135(19):e198351. https://doi.org/10.1172/JCI198351.
© 2025 Rath et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Published October 1, 2025 - Version history
View PDF

Related article:

Personalized treatment in HPV+ oropharynx cancer using genomic adjusted radiation dose
Emily Ho, … , Javier F. Torres-Roca, Jacob G. Scott
Emily Ho, … , Javier F. Torres-Roca, Jacob G. Scott
Using a large trial of patients treated with radiation therapy for head and neck cancers we show the utility of a genomic method personalizing radiation dose for each patient to optimize their outcomes and minimize their toxicity.
Research Article Clinical Research Oncology

Personalized treatment in HPV+ oropharynx cancer using genomic adjusted radiation dose

Abstract

BACKGROUND A key objective in managing HPV+ oropharyngeal squamous cell carcinoma (OPSCC) is reducing radiation therapy (RT) doses without compromising cure rates. A recent phase II/III HN005 trial revealed that clinical factors alone are insufficient to guide safe RT dose de-escalation. Our prior research demonstrated that the genomic adjusted radiation dose (GARD) predicts RT benefit and may inform dose selection. We hypothesize that GARD can guide personalized RT de-escalation in HPV+ OPSCC patients.METHODS Gene expression profiles were analyzed in 191 HPV+ OPSCC patients enrolled in an international, multi-institutional observational study (AJCC Eighth Edition, stages I–III). Most patients received 70 Gy in 35 fractions or 69.96 Gy in 33 fractions (median dose: 70 Gy; range: 51.0–74.0 Gy). Overall survival (OS) was 94.1% at 36 months and 87.3% at 60 months. A Cox proportional hazards model assessed association between GARD and OS, and time-dependent receiver operating characteristic analyses compared GARD with traditional clinical predictors.RESULTS Despite uniform RT dosing, GARD showed wide heterogeneity, ranging from 15.4 to 71.7. Higher GARD values were significantly associated with improved OS in univariate (HR = 0.941, P = 0.041) and multivariable analyses (HR = 0.943, P = 0.046), while T and N stages were not. GARD demonstrated superior predictive performance at 36 months (AUC = 78.26) versus clinical variables (AUC = 71.20). Two GARD-based RT de-escalation strategies were identified, offering potential survival benefits while reducing radiation exposure.CONCLUSION GARD predicts OS and outperforms clinical variables, supporting its integration into the diagnostic workflow for personalized RT in HPV+ OPSCC.FUNDING This work was supported by the National Cancer Institute through the Cleveland Clinic/Emory ROBIN center (U54-CA274513, project 2), the European Union Horizon 2020 Framework Programme (grant/award 689715), the Italian Association for Cancer Research (AIRC project ID 23573), and the European Research Area Network ERA PerMed JTC2019/Fondazione Regionale per la Ricerca Biomedica project SuPerTreat (Supporting Personalized Treatment Decisions in Head and Neck Cancer through Big Data).

Authors

Emily Ho, Loris De Cecco, Steven A. Eschrich, Stefano Cavalieri, Geoffrey Sedor, Frank Hoebers, Ruud H. Brakenhoff, Kathrin Scheckenbach, Tito Poli, Kailin Yang, Jessica A. Scarborough, Shivani Nellore, Shauna Campbell, Neil Woody, Tim Chan, Jacob Miller, Natalie Silver, Shlomo Koyfman, James Bates, Jimmy J. Caudell, Michael W. Kattan, Lisa Licitra, Javier F. Torres-Roca, Jacob G. Scott

×
Abstract

Uniform radiation therapy (RT) de-escalation in HPV+ oropharyngeal squamous cell carcinoma (OPSCC) has underperformed in clinical trials, likely due to underlying genomic heterogeneity. In this issue of the JCI, Ho et al. evaluated genomic adjusted radiation dose (GARD), which integrates tumor gene expression with RT dose to estimate biological effect. In 191 locoregionally advanced HPV+ OPSCC patients treated with definitive RT with or without chemotherapy, GARD values varied widely, despite uniform dose delivery, and independently predicted overall survival. These data support a genomically informed framework specific for HPV+ OPSCC patients via GARD for guiding radiation dose de-escalation strategies.

Uniform radiation therapy underperforms in lower-risk HPV+ patients

HPV+ oropharyngeal squamous cell carcinoma (OPSCC), which represents over 70 percent of OPSCC diagnosed in the US (1–3), is among the most curable head and neck cancers. Standard regimens — typically 70 Gy delivered via volumetric modulated arc therapy or intensity modulated radiation therapy (RT) — achieve three-year overall survival (OS) rates of 93% in the lowest risk groups (4) but can be associated with profound toxicities (5). The favorable prognosis associated with clinical success in HPV+ OPSCC has led to American Joint Committee in Cancer downstaging of OPSCC based on HPV status (6, 7), but, paradoxically, this downstaging has also introduced a therapeutic risk by leading to use of uniform RT dose de-escalation strategies guided solely by conventional clinical factors, which may inadvertently undertreat a subset of patients (8–14). The NRG-HN005 trial illustrated this challenge, as a uniform 60 Gy regimen in low-risk patients failed to meet its phase II endpoint (9), highlighting the need for biologically informed approaches that can individualize dose while preserving disease control. In their recent JCI study (15), Ho et al. proposed application of a biologically informed alternative: the genomic adjusted radiation dose (GARD) (16), a model that quantifies the biological effect of a given radiotherapy dose using tumor-specific radiosensitivity derived from a validated 10-gene radiosensitivity index (RSI) signature (comprising AR, c-Jun, STAT1, PKC, RelA, c-ABL, SUMO1, CDK1, HDAC1, IRF1) (17–19) embedded within the linear quadratic model: GARD = D × (α + βD), where α = −ln(RSI)/D, β is fixed at 0.05 Gy2, D represents the total radiation dose delivered in Gy, α represents the linear component of cell killing due to radiation and is derived from the patient’s individual radiosensitivity using the RSI, and β represents the quadratic component of cell killing, attributed to the accumulation of multiple “single hits” or sublethal damage. Thus, a uniform physical RT dose would be expected to display more effective tumor control in an individual with a higher GARD score.

GARD operationalizes biologic heterogeneity

Clinically similar tumors can differ profoundly in their response to radiation due to variation in DNA damage repair efficacy, cell cycle checkpoint control, and tumor-immune interactions, as radiogenomic studies have shown (20). Such heterogeneity challenges the premise that a uniform radiation dose yields equivalent biological effects across patients. Quantifying this variability is essential for moving beyond one-size-fits-all radiotherapy toward strategies that match dose to the tumor’s intrinsic biology. Ho et al. expanded on the GARD framework, reaffirming the principle that physical dose does not equate to biological dose (15). In a prior study, they assessed a clinically homogeneous cohort of 191 locoregionally advanced HPV+ OPSCC patients from the BD2Decide trial (21), all treated with equivalent dose in 2 Gy fractions (EQD2) values of 69–71 Gy with or without chemotherapy. GARD scores in this cohort varied widely from 15.4 to 71.7 (median, 39.1; IQR, 12.6). In the present study, Ho et al. showed that these differences, which were not apparent from tumor-node-metastasis (TNM) stage or Eastern Cooperative Oncology Group (ECOG) performance score (ECOG, 0 in 83% of patients, indicating a high level of function) were similar to earlier results in multiple cancer types, where GARD, based on the 10-gene RSI (18, 19, 22), consistently outperforms physical RT dose in predicting outcomes. Ho. et al. further demonstrated that across all definitive RT patients, each unit increase in GARD was associated with improved OS (hazard ratio [HR] = 0.941, P = 0.041); in the standard-of-care dose subgroup, the HR was 0.920 (P = 0.019). Notably, Ho. et al. showed that patients with GARD ≥42 achieved a 3-year OS of 100%, compared with 90% for GARD <42 (P = 0.0045). Using multivariable analysis in HPV+ patients, including T stage (primary tumor), N stage (spread to lymph nodes), smoking, and ECOG performance, Ho et al. found that GARD remained the only statistically significant predictor of OS (HR = 0.943, P = 0.046). These results align with broader pan-cancer evidence in a pooled analysis across multiple cancer types (22). Similar results have been reported in breast cancer (16, 22), pancreatic cancer (16, 22, 23), and locally advanced rectal cancer, where GARD ≥17 predicted good pathological response (AUC = 0.75) and revealed that only approximately 17% of patients’ optimal biologically calculated doses matched guideline prescriptions (16, 22, 24).

A key advance of Ho et al. is the operationalization of GARD for biologically guided dose adaptation specifically for HPV+ OPSCC (15). In silico modeling showed that uniform de-escalation to 60 Gy — as performed in the NRG-HN005 trial — would reduce three-year OS from 94.6% to 92.7%, primarily because 42 additional patients would become biologically underdosed (GARD <42). By contrast, a selective GARD-guided strategy, i.e., reducing dose only in patients predicted to continue to have high GARD values at 60 Gy, preserved the 94.6% OS rate while enabling safe dose reductions in approximately 16% of patients. An even more personalized approach, ensuring a minimum GARD ≥ 32 for each patient, was projected to allow dose reduction in 77.7% of the cohort, while identifying the 22.3% who would require ≥60–70 Gy to achieve optimal tumor control. This model predicted that some patients could be adequately treated with as little as 30 Gy, while others might require doses beyond the conventional 70 Gy.

Predictive performance and clinical implications

Time-dependent receiver operating characteristic analysis in the present study (15) further underscored GARD’s predictive strength: clinical variables alone yielded an AUC of 71.20 at 3 years, GARD alone achieved an AUC of 78.26, and the combination reached an AUC of 83.81 — an absolute gain of more than 12 points over clinical factors alone. However, adding previously defined gene expression clusters conferred no additional benefit. These findings reinforce that physical dose is not synonymous with biological dose and that failing to account for radiosensitivity risks suboptimal patient management.

Limitations of the present study

An important consideration in interpreting the results of this study (15) is that the analyzed BD2Decide dataset was restricted to OS and not local recurrence or disease-free survival, which are important outcome measures in HPV+ OPSCC, and was conducted over a wide range of years of treatment (2008–2017), during which delivery approaches may have advanced; moreover, over 90% of the HPV+ OPSCC patients received concurrent chemoradiotherapy. As such, the independent predictive contribution of GARD should be used with caution in other contexts. Furthermore, immune checkpoint inhibitors (ICIs) were not included in the treatment landscape of the analyzed cohort. The exclusion of ICI-treated patients from the BD2Decide dataset represents a growing gap, as HPV+ OPSCC is immunologically active and increasingly managed with ICI (25). ICIs, including nivolumab and pembrolizumab, are approved in the recurrent/metastatic setting for head and neck cancers (26, 27) and are increasing being incorporated in the latest HPV+ OPSCC clinical trials (9). Whether GARD remains predictive in the presence of immune modulation remains unknown. In addition, radiosensitivity in HPV+ OPSCCs is also likely to be dependent on additional factors, including the tumor immune microenvironment, hypoxia, and epigenetic landscape. Incorporation of these factors, such as hypoxia using 18F-fluoromisonidazole positron emission tomography (28), or incorporation of additional genomic, transcriptomic, and immune profiling data currently being generated from HPV+ OPSCCs patients treated with RT by the National Cancer Institute (NCI) Radiation Oncology Biology Integration Network are likely to further improve guidance of RT dose de-escalation strategies.

Conclusion and future directions: toward biologically precise radiation

Use of the GARD model offers a state-of-the-art paradigm for personalizing RT for patients with HPV+ OPSCC. Without a biologically informed approach, empirical RT dose reduction risks undertreating the subset of patients with intrinsically fewer radiosensitive tumors, eroding the gains achieved in this highly curable population. GARD’s ability to stratify radiosensitivity offers a path to maintain cure rates while reducing treatment-related morbidity — providing a model for how precision radiation could be implemented in other cancers. The bigger picture is clear: biologically adaptive dosing with further validation in a prospective randomized trial could redefine RT from a uniform protocol to a dynamic, patient-specific intervention.

Acknowledgments

This work is the result of NIH funding, in whole or in part, and is subject to the NIH Public Access Policy. Through acceptance of this federal funding, the NIH has been given a right to make the work publicly available in PubMed Central.

• DSY is supported by the NIH/NCI [R01CA178999, R01CA254403, R01CA301614, and U54CA274513] and Department of Defense (DOD) Breast Cancer Research Program BC220744.

• SKR is supported by DOD Lung Cancer Research Program LC240534.

Address correspondence to: David S. Yu, Department of Radiation Oncology, Emory University School of Medicine, 1365 Clifton Rd. NE, C3008, Atlanta, Georgia, 30322, USA. Phone: 404.778.1758; Email: dsyu@emory.edu.

Footnotes

Conflict of interest: The authors have declared that no conflict of interest exists.

Copyright: © 2025, Rath et al. This is an open access article published under the terms of the Creative Commons Attribution 4.0 International License.

Reference information: J Clin Invest. 2025;135(19):e198351. https://doi.org/10.1172/JCI198351.

See the related article at Personalized treatment in HPV+ oropharynx cancer using genomic adjusted radiation dose.

References
  1. Chaturvedi AK, et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol. 2011;29(32):4294–4301.
    View this article via: CrossRef PubMed Google Scholar
  2. Lechner M, et al. Gender-neutral HPV vaccination in the UK, rising male oropharyngeal cancer rates, and lack of HPV awareness. Lancet Infect Dis. 2019;19(2):131–132.
    View this article via: CrossRef PubMed Google Scholar
  3. Roman BR, Aragones A. Epidemiology and incidence of HPV-related cancers of the head and neck. J Surg Oncol. 2021;124(6):920–922.
    View this article via: CrossRef PubMed Google Scholar
  4. Ang KK, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med. 2010;363(1):24–35.
    View this article via: CrossRef PubMed Google Scholar
  5. Youssef I, et al. Toxicity profiles and survival outcomes among patients with nonmetastatic oropharyngeal carcinoma treated with intensity-modulated proton therapy vs intensity-modulated radiation therapy. JAMA Netw Open. 2022;5(11):e2241538.
    View this article via: CrossRef PubMed Google Scholar
  6. Ho AS, et al. Derivation and validation of the AJCC9V pathological stage classification for HPV-positive oropharyngeal carcinoma: a multicentre registry analysis. Lancet Oncol. 2025;26(8):1113–1122.
    View this article via: CrossRef PubMed Google Scholar
  7. O’Sullivan B, et al. Development and validation of a staging system for HPV-related oropharyngeal cancer by the International Collaboration on Oropharyngeal cancer Network for Staging (ICON-S): a multicentre cohort study. Lancet Oncol. 2016;17(4):440–451.
    View this article via: CrossRef PubMed Google Scholar
  8. Yom SS, et al. Reduced-dose radiation therapy for HPV-associated oropharyngeal carcinoma (NRG Oncology HN002). J Clin Oncol. 2021;39(9):956–965.
    View this article via: CrossRef PubMed Google Scholar
  9. Yom SS, et al. Interim futility results of NRG-HN005, a randomized, phase II/III non-inferiority trial for non-smoking p16+ oropharyngeal cancer patients. Int J Radiat Oncol Biol Phys. 2024;120(2):S2–S3.
    View this article via: CrossRef Google Scholar
  10. Yacoub I, et al. De-escalated management of HPV-positive oropharyngeal carcinoma: improving outcomes with personalized approaches. Semin Radiat Oncol. 2025;35(2):157–165.
    View this article via: CrossRef PubMed Google Scholar
  11. Tsai CJ, et al. Evaluation of substantial reduction in elective radiotherapy dose and field in patients with human papillomavirus-associated oropharyngeal carcinoma treated with definitive chemoradiotherapy. JAMA Oncol. 2022;8(3):364–372.
    View this article via: CrossRef PubMed Google Scholar
  12. Chera BS, et al. Phase II trial of de-intensified chemoradiotherapy for human papillomavirus-associated oropharyngeal squamous cell carcinoma. J Clin Oncol. 2019;37(29):2661–2669.
    View this article via: CrossRef PubMed Google Scholar
  13. Deschuymer S, et al. Randomized clinical trial on reduction of radiotherapy dose to the elective neck in head and neck squamous cell carcinoma; update of the long-term tumor outcome. Radiother Oncol. 2020;143:24–29.
    View this article via: CrossRef PubMed Google Scholar
  14. Sher DJ, et al. Recurrence and quality-of-life following involved node radiotherapy for head and neck squamous cell carcinoma: initial results from the phase II INRT-air trial. Int J Radiat Oncol Biol Phys. 2021;111(3):e398.
    View this article via: CrossRef Google Scholar
  15. Ho E, et al. Personalized treatment in HPV+ oropharynx cancer using genomic adjusted radiation dose. J Clin Invest. 2025;135(19):e194073.
    View this article via: JCI CrossRef Google Scholar
  16. Scott JG, et al. A genome-based model for adjusting radiotherapy dose (GARD): a retrospective, cohort-based study. Lancet Oncol. 2017;18(2):202–211.
    View this article via: CrossRef PubMed Google Scholar
  17. Eschrich SA, et al. Validation of a radiosensitivity molecular signature in breast cancer. Clin Cancer Res. 2012;18(18):5134–5143.
    View this article via: CrossRef PubMed Google Scholar
  18. Eschrich SA, et al. A gene expression model of intrinsic tumor radiosensitivity: prediction of response and prognosis after chemoradiation. Int J Radiat Oncol Biol Phys. 2009;75(2):489–496.
    View this article via: CrossRef PubMed Google Scholar
  19. Eschrich S, et al. Systems biology modeling of the radiation sensitivity network: a biomarker discovery platform. Int J Radiat Oncol Biol Phys. 2009;75(2):497–505.
    View this article via: CrossRef PubMed Google Scholar
  20. Grimes DR. Limitations of the radiosensitivity index as a direct prognostic marker. Lancet Oncol. 2022;23(11):1352–1353.
    View this article via: CrossRef PubMed Google Scholar
  21. Cavalieri S, et al. Development of a multiomics database for personalized prognostic forecasting in head and neck cancer: The Big Data to Decide EU Project. Head Neck. 2021;43(2):601–612.
    View this article via: CrossRef PubMed Google Scholar
  22. Scott JG, et al. Pan-cancer prediction of radiotherapy benefit using genomic-adjusted radiation dose (GARD): a cohort-based pooled analysis. Lancet Oncol. 2021;22(9):1221–1229.
    View this article via: CrossRef PubMed Google Scholar
  23. Jan IS, Ch’ang HJ. Selection of patients with pancreatic adenocarcinoma who may benefit from radiotherapy. Radiat Oncol. 2023;18(1):137.
    View this article via: CrossRef PubMed Google Scholar
  24. Xia H, et al. Validation of a genome-based model for adjusting radiotherapy dose (GARD) in patients with locally advanced rectal cancer. Sci Rep. 2024;14(1):21572.
    View this article via: CrossRef PubMed Google Scholar
  25. Floudas CS, et al. Novel combination immunotherapy and clinical activity in patients with HPV-associated cancers: a nonrandomized clinical trial. JAMA Oncol. 2025;11(4):394–399.
    View this article via: CrossRef PubMed Google Scholar
  26. Ferris RL, et al. Nivolumab for recurrent squamous-cell carcinoma of the head and neck. N Engl J Med. 2016;375(19):1856–1867.
    View this article via: CrossRef PubMed Google Scholar
  27. Cohen EEW, et al. Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous cell carcinoma (KEYNOTE-040): a randomised, open-label, phase 3 study. Lancet. 2019;393(10167):156–167.
    View this article via: CrossRef PubMed Google Scholar
  28. Lee NY, et al. Hypoxia-directed treatment of human papillomavirus-related oropharyngeal carcinoma. J Clin Oncol. 2024;42(8):940–950.
    View this article via: CrossRef PubMed Google Scholar
Version history
  • Version 1 (October 1, 2025): Electronic publication
Advertisement
Advertisement