Hypoxia-Guided Therapy for Human Papillomavirus-Associated Oropharynx Cancer

Human papillomavirus-related (HPV+) oropharynx cancer is a distinct form of head and neck squamous cell cancer with an etiology, biomolecular characteristics, treatment responsiveness, and natural history that differ substantially from other head and neck cancers. An important concern has been the possibility that many patients with HPV+ oropharynx cancer are exposed to excessive treatment, derived from trials conducted in populations with predominantly HPV-negative disease, and will therefore unnecessarily carry chronic toxicities for decades following successful treatment. Dysphagia, aspiration pneumonia, xerostomia, hypothyroidism, chronic pain, and increased noncancer mortality have been reported in head and neck cancer survivors and may increase over time (1,2), with subacute toxicity not substantially different for minimally invasive surgery or definitive chemoradiation (1). Thus, many studies now focus on defining risk within the population with HPV+ disease and exploring strategies to reduce the intensity and/or duration of therapy for the most favorable subsets. Choice of local treatment with minimally invasive surgery or radiation, inclusion or omission of chemotherapy, and dose modification have all been explored, with a tension between increased risk of local and regional vs distant recurrence as local or systemic components of therapy are modified. Three trials have explored response to induction chemotherapy as a means to select patients for radiation dose reduction to 54 Gy or lower. A total of 186 patients have been treated in the seminal trials of this approach—the ECOG-ACRIN Cancer Research Group E1308 trial (3), the 2-institution trial of Chen et al. (4), and the Optima study from the University of Chicago (5)—setting benchmarks for disease control in trials of deintensification. Among low-risk patients across these studies, a remarkable 95%-96% progression-free survival (PFS) is reported at 2 years. Function preservation has also been outstanding (6-8). Reduction in radiation dose to 60 Gy is feasible without prior chemoselection in favorable-risk patients if concurrent chemotherapy is given, with a 2-year PFS of 90.5% (9); however, this is a modest level of radiation dose reduction, and patients treated with 60 Gy will likely continue to experience late swallowing dysfunction and other functional impairments. Omission of chemotherapy as part of definitive therapy, even in this favorable-risk group, does not appear to be a suitable strategy, with a 2-year PFS of less than 90%; with the added information on risk provided by pathologic evaluation of nodal stage, radiation alone following transoral surgery yields an outstanding 2-year PFS at 50 or 60 Gy (10).

Riaz et al. (11), in this issue of the Journal, introduce a highly innovative patient selection strategy. Recognizing the role of hypoxia in radiation resistance in head and neck cancer [albeit this association has been inconsistent in HPV+ disease (12,13)] and building on their prior work demonstrating that fluoromisonidazole-positron emission tomography (F-MISO-PET) dynamically reflects hypoxia across the radiation therapy course, they prospectively studied a radical reduction in radiation dose to 30 Gy for those with no pretreatment hypoxia or in whom hypoxia had resolved within the first 2 weeks of initiating radiation. Sensitizing chemotherapy was not deescalated and consisted of high-dose cisplatin on days 1 and 22 or a carboplatin and 5-fluorouracil combination on days 1-4 and 22-25. This was a small study. Nineteen patients with favorable-stage (T1-2, N1-2b) HPV+ oropharynx cancer were enrolled. Although patients with at least a 10 pack-year smoking history—who are recognized as intermediate risk and often excluded from deintensification studies (3,9,14)—were not excluded, only 2 such patients were accrued. Fifteen of the 19 patients met criteria for and were treated with reduced radiation dose. Resection of the primary tumor was performed at baseline, and neck dissection at 4 months postradiation was employed in all patients. The 2-year PFS in the deescalated radiation group was 92.9% and in the overall group was 89.5%; however, no conclusions can be drawn about the clinical utility of this approach because all patients had disease controlled by surgery at both the primary site and neck nodes. The interest of this work therefore lies in correlating the novel hypoxia-imaging strategy and extensive molecular studies. Most notably, none of the 5 patients with baseline normoxic tumors on F-MISOPET had residual disease at neck dissection compared with 6 of 10 patients with evidence of hypoxia on baseline F-MISO PET. Further, patients who did not achieve a pathologic complete response also had significantly fewer genomic deletions with microhomology; however, given differences in etiologic HPV type, baseline tumor burden, circulating HPV DNA, chemotherapy administration, and the surprising finding of Rb or TP53 mutation in 2 of the 17 patients sequenced, conclusions are again limited by heterogeneity in a small sample size. The genomic sequencing findings were then applied to a distinct cohort of patients with favorable-risk smoking history and no pathologic high-risk features treated in a Mayo Clinic trial with 30 Gy and docetaxel postoperative therapy following definitive surgery of both the primary and neck disease (15). Patients with recurrence in the Mayo trial were matched to patients from the same trial who had not experienced recurrence, again demonstrating fewer deletions with microhomology in the worse outcome group (here termed “nonresponders”). Exploration of immune signatures, tumor mutation burden, and specific genomic alterations such as PIK3Ca mutational status and TRAF3 loss of heterozygosity could not be correlated to hypoxia, response, or DNA damage, likely because of the small sample size and multiple comparisons.

The authors are to be congratulated for the careful annotation of materials in this exploratory study. As they move forward to a larger phase II design without confirmatory neck dissection, the high proportion of patients in this cohort who were found with residual disease at neck dissection may argue against intra-treatment normoxia as a criterion for treatment deintensification. On the other hand, the correlation of baseline hypoxia vs normoxia with pathologic complete response at the least argues that hypoxia has a greater role in predicting radioresponsiveness in HPV+ oropharynx cancer than has been recognized and should be evaluated in future studies of deescalation in concert with promising approaches such as circulating HPV DNA monitoring (16), radiomic features (17, 18), the TRAF3/CYLD mutation signature (19), PIK3Ca mutation (20), and DNA integration status (21).

Funding

Not applicable.

Notes

Role of the funder: Not applicable.

Disclosures: The authors have no conflicts of interest to disclose.

Author contributions: Writing- original draft. B.B.Writing – review and editing. J.C.

Data Availability

Not applicable.

References