Evaluation of Substantial Reduction in Elective Radiotherapy Dose and Field in Patients With Human Papillomavirus–Associated Oropharyngeal Carcinoma Treated With Definitive Chemoradiotherapy

Key Points

Question  Is substantial reduction in elective radiotherapy dose and volume associated with adequate locoregional tumor control in patients with human papillomavirus–associated oropharyngeal carcinoma treated with definitive chemoradiotherapy?

Findings  In this cohort study, reduced radiotherapy volume and dose of 30 Gy to the elective regions with concurrent chemotherapy in 276 patients with locally advanced human papillomavirus–positive oropharyngeal carcinoma was associated with a 24-month locoregional control rate of 97.0% and overall survival of 95.1%.

Meaning  These findings suggest that major de-escalation of radiotherapy in elective regions was feasible while maintaining locoregional tumor control.

Importance  Several de-escalation strategies for human papillomavirus (HPV)–associated oropharyngeal carcinoma (OPC) have focused on deintensifying gross disease treatment. Reduction of radiotherapy dose and target volume to subclinical regions may achieve good clinical outcomes with favorable patient quality of life (QOL).

Objective  To determine outcomes from a systematic approach of reducing radiotherapy dose and target volume to the elective treatment regions in patients with HPV-associated OPC undergoing concurrent chemoradiotherapy (CCRT).

Design, Setting, and Participants  This retrospective cohort study included 276 consecutive patients with HPV-positive OPC receiving CCRT from March 1, 2017, to July 31, 2019. Data were analyzed from February 23 to September 13, 2021.

Interventions  Elective nodal and subclinical regions received 30 Gy of radiotherapy in 15 fractions, followed by a cone down of 40 Gy in 20 fractions to gross disease for a total dose of 70 Gy. The high retropharyngeal nodal basins in the node-negative neck and bilateral levels IB and V basins were omitted.

Main Outcomes and Measures  Patients were followed up to evaluate locoregional control as the primary outcome and distant metastasis–free survival, progression-free survival, and overall survival as secondary outcomes. Quality-of-life data were obtained at each visit when feasible.

Results  Among the 276 patients included in the analysis, the median age was 61 (range, 36-87) years; 247 (89.5%) were men; and 183 (66.3%) had less than 10 pack-years of smoking history. Most patients (251 [90.9%]) were White. Overall, 87 (31.5%) had cT3-cT4 disease and 65 (23.5%) had cN2-cN3 disease per the 8th edition of the American Joint Committee on Cancer Staging Manual. One hundred seventy-two patients (62.3%) completed 300-mg/m² high-dose cisplatin therapy. During a median follow-up of 26 (range, 21-32) months, 8 patients developed locoregional recurrence, including 7 at the primary site or gross nodes that received a total dose of 70 Gy and 1 with a persistent node not previously identified as gross disease that received a total dose of only 30 Gy. The 24-month locoregional control was 97.0%; progression-free survival, 88.0%; distant metastasis–free survival, 95.2%; and overall survival, 95.1%. During treatment, 17 patients (6.2%) required a feeding tube. At 24 months, most of the QOL composite scores (jaw-related problems, pain, social contact, eating, speech, and swallow) were comparable or superior to baseline measures except for senses, dry mouth, muscular tension, and cognitive functioning, which improved over time but remained marginally worse than baseline.

Conclusions and Relevance  This cohort study found that the evaluated de-escalation strategy for elective regions showed favorable clinical outcomes and QOL profiles. Long-term follow-up data will help affirm the efficacy of this strategy as a care option for treating HPV-associated OPC with primary CCRT.

Patients with human papillomavirus (HPV)–associated oropharyngeal cancer (OPC) constitute most patients with OPC in the US.¹ Preclinical and clinical studies^2,3 have demonstrated enhanced sensitivity to chemoradiotherapy in these patients. The standard 7-week chemoradiotherapy course, although effective, leads to considerable treatment-related toxic effects. Standard radiotherapy dose and field for head and neck cancers often included treatment of large elective volumes to cumulative doses of approximately 50 to 60 Gy, whereas grossly evident disease on imaging received a total dose of 70 Gy. These doses were derived from data generated in the era of 3-dimensional conformal radiation,^4-7 and the elective volumes typically targeted regional nodal basins and tissues adjacent to the gross tumor suspected of harboring microscopic disease based on surgical patterns of spread.⁸

Several clinical trials^9-15 have shown that in patients with select HPV-positive OPC, radiation dose to gross disease can be safely reduced, typically by 10 to 16 Gy. However, many trials did not evaluate radiation dose reduction in elective regions systematically, and a wide range of elective doses were used. Even with the contemporary intensity-modulated radiotherapy technique, elective nodal radiotherapy still contributes significantly to acute and long-term toxic effects to critical organs.^16-18 Studies of patterns of treatment failure in patients with OPC have shown that most locoregional recurrences occur in the high-dose radiation fields, indicating that it might not be necessary to treat the elective regions with doses of 50 to 60 Gy.^19-21 Because the effectiveness of elective treatment is measured by failure rate in the treatment field, instead of a strict threshold effect for the treatment of gross disease, past radiation dose range for microscopic disease might not be applicable in the context of modern imaging technology.^22,23 Results from 2 recent prospective trials of 40-Gy elective neck volumes^24,25 showed uncompromised locoregional control, regardless of anatomical site of disease and HPV status.

Beginning in 2017, we systematically started treating the subclinical regions that have historically been included in the elective treatment volume to a dose of 30 Gy and spared selected negative neck, retropharyngeal, level 1b, and level V nodal basins.^26-29 Herein, we summarize the 2-year follow-up outcomes of patients treated using this regimen.

Consecutive patients with newly diagnosed, previously untreated HPV-positive OPC receiving definitive concurrent chemoradiotherapy (CCRT) with cisplatin or other chemotherapy from March 1, 2017, to July 31, 2019, were included in this cohort study. The new radiotherapy regimen was not intended for patients treated with concurrent cetuximab. The HPV diagnosis was confirmed by both p16 immunohistochemistry staining and HPV-RNA in situ hybridization. Patients who underwent excisional biopsy of the primary site or lymph nodes with residual gross tumor burden after diagnostic biopsy were included. The institutional review board at Memorial Sloan Kettering Cancer Center approved this study with a waiver of informed consent for a retrospective analysis of patient data. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

There were only 2 dose levels for most of the radiotherapy plans: 30 Gy to the elective treatment regions (15 fractions of 2 Gy), followed by a cone down of 40 Gy for a total of 70 Gy to all sites of gross disease, as described previously.²⁹ We used positron emission tomography and computed tomography (CT), CT with contrast if no contraindication, and contrast magnetic resonance imaging to delineate the gross primary tumor and nodal volumes treated with a dose of 70 Gy (GTV70_primary and GTV70_nodes, respectively). We then expanded via anatomical routes of spread of GTV70_primary, typically around 5-mm margins, to derive the planning target volume (PTV70_primary). The PTV70 of gross nodal disease (PTV70_nodes) was obtained by a concentric expansion of 3 to 5 mm from GTV70_nodes. There was no clinical target volume expansion around GTV70_primary or GTV70_nodes. The clinical target volume of 30 Gy (CTV30) encompassed all regions of potential subclinical disease spread (retropharyngeal, retrostyloid, and levels II to IV for node-positive neck; levels II to IV for node-negative neck if receiving radiotherapy). With daily kilovoltage or cone-beam CT imaging, our planning target volume of 30-Gy (PTV30) margins of the elective regions were reduced to 2- to 3-mm expansion from CTV30. We also ensured that PTV30 was at least 1.3 cm from the GTV70_primary. Patients were considered for ipsilateral nodal radiotherapy if they had a well-lateralized T1 tonsil primary tumor with 1 to 2 ipsilateral involved lymph nodes with a diameter of less than 3 cm each. In some instances, indeterminate nodes or areas immediately adjacent to the gross primary disease were treated with a dose of 50 Gy per physician discretion. Treatment techniques were intensity-modulated radiotherapy (256 patients) or intensity-modulated proton therapy (20 patients). Select examples of target volume delineation are shown in the eFigure in the Supplement.

All patients were followed up per our institutional standard of care, typically every 3 months in the first year after completion of CCRT, every 4 to 6 months in the second to fifth year, and yearly thereafter. At these visits, the following assessments were performed: evaluations of toxic effects and history and physical examinations. Evaluations with positron emission tomography–CT, CT, and/or magnetic resonance imaging were typically performed at 3 to 4 months after CCRT and repeated yearly thereafter to year 2 or year 3.

All patients were asked to complete the Gothenburg Trismus Questionnaire (GTQ) at each visit per our departmental routine practice. The GTQ is a comprehensive, self-administered, trismus-specific questionnaire with different aspects of objective measures of 13 items and the following 3 domains: jaw-related problems, eating limitation, and muscular tension. There are an additional 8 single items.³⁰ The GTQ was validated against the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Head and Neck Module (EORTC QLQ H&N35) as well as the EORTC QLQ Core Questionnaire 30 (EORTC QLQ-C30).^30,31 In addition to the GTQ, 54 of the 276 patients also enrolled in a prospective trial of adaptive radiotherapy planning³² during their treatment course to account for disease response and temporal variations in anatomy, with quality of life (QOL) evaluated at each visit using the EORTC QLQ H&N35 and EORTC QLQ-C30.^33-35

Data were analyzed from February 23 to September 13, 2021. The primary outcome of interest was 24-month locoregional control (LRC). Secondary outcomes included progression-free survival, distant metastasis–free survival, overall survival, and QOL. We used the Kaplan-Meier method to calculate the cumulative incidence of each event and survival rates with death as a competing risk. We evaluated factors associated with LRC using multivariable Cox proportional hazards regression models. All analyses were performed with SAS, version 9.4 (SAS Institute Inc) and SPSS, version 26 (IBM Corporation).

The GTQ was scored by calculating a mean for each domain, which was then transformed to a scale ranging from 0 to 100, where a higher score indicated greater perceived dysfunction.³⁰ For ease of comparison and analysis, the 8 single items were grouped into 4 categories: facial pain, facial pain impact, jaw limitation, and jaw limitation impact. Composite scores of each category were calculated accordingly. Similar to GTQ, composite scores of EORTC QLQ-HN 35 and EORTC QLQ-C30 were calculated according to the EORTC scoring manuals.^33-36

Table 1 shows the baseline patient and clinical characteristics for the 276 participants. The median age of the cohort was 61 (range, 36-87) years. Two hundred forty-seven patients were men (89.5%) and 29 were women (10.5%). We included self-reported patient race and ethnicity in the analysis to show the distribution of different racial and ethnic groups in our cohort. Eight patients (2.9%) were Asian, 10 (3.6%) were Black, 1 (0.4%) was Hispanic, 251 (90.9%) were White, and 6 (2.2%) were of unknown race and ethnicity. Half of the patients (135 [48.9%]) were never smokers; 183 (66.3%) had a smoking history of less than 10 pack-years. There was a wide range of disease-stage distribution in the cohort; 87 patients (31.5%) had cT3-cT4 disease and 65 (23.5%) had cN2-cN3 disease per the 8th edition of the American Joint Committee on Cancer Staging Manual. Most patients (259 [93.8%]) had a primary site at either the base of the tongue or the tonsil. Most patients (213 [77.2%]) received high-dose cisplatin, and 172 of 213 patients (80.7%; 62.3% overall) were able to complete a cumulative dose of 300 mg/m². Other commonly used chemotherapy regimens included weekly carboplatin with paclitaxel (Taxol) or fluorouracil. The new radiotherapy regimen was not intended for patients treated with concurrent cetuximab; however, 5 patients who received concurrent cetuximab were incidentally treated with this regimen and were included in the analysis.

All 276 patients were prescribed a radiation dose of 30 Gy to the elective and subclinical regions and cone down to 70 Gy to the gross disease. In 87 patients (31.5%), indeterminate nodes or areas immediately adjacent to the gross disease were treated with a dose of 50 Gy. Most patients who received an intermediate dose of 50 Gy were treated within the first year after implementing the new dosing regimen before physicians gradually became more comfortable treating the entire elective regions with 30 Gy.

The median follow-up for surviving patients was 26 (range, 1-47) months. The overall 24-month LRC was 97.0% for the entire cohort, 88.0% for progression-free survival, 95.2% for distant metastasis–free survival, and 95.1% for overall survival (Figure 1). Eight patients (2.9%) developed locoregional recurrence; of these, 7 (87.5%) had disease recurrence in the primary site and/or gross disease in nodes that received a full 70-Gy dose, and 1 patient had persistent gross disease in a node that was not previously identified and received only 30 Gy. One patient with locoregional recurrence also developed a synchronous distant recurrence. Five patients had neck-only recurrence; all these patients underwent successful salvage treatment by neck dissection and were still alive without evidence of disease at the time of last follow-up. There was otherwise no recurrence in the subclinical fields receiving a dose of 30 Gy or in the omitted elective nodal regions. Nineteen patients (6.9%) developed distant-only recurrence.

We then performed multivariable Cox proportional hazards regression analyses to evaluate factors associated with LRC. Of all variables included in the proportional hazards regression model (age, disease site, sex, race and ethnicity, smoking history, T category, N category, and chemotherapy), only advanced N category was associated with worse LRC (hazard ratio, 3.4 [95% CI, 1.1-10.6]; P = .03). There was no difference in LRC between patients receiving high-dose cisplatin and weekly cisplatin or other chemotherapy combinations (hazard ratio, 0.8; 95% CI, 0.2-3.5; P = .71). The only patient with treatment failure in the 30-Gy subclinical region received a cumulative cisplatin dose of a 300-mg/m² bolus.

During treatment, 17 patients (6.2%) had a feeding tube placement, with a median duration of feeding tube dependance of 5 (range, 1-23) months. Four patients had a feeding tube in place at the time of death; 2 of these had locoregional disease recurrence within 6 months after CCRT with feeding tubes still in place and died shortly after, 1 died of complications from treatment, and 1 died of unknown cause. No feeding tube was placed prophylactically in the study cohort.

We analyzed GTQ data collected at baseline and at 3 to 6 months, 12 (±3) months, and 24 (±3) months after radiotherapy completion. Figure 2 shows the mean GTQ composite scores of the study participants by symptom domain. Compared with baseline, all domain scores worsened at 3 to 6 months after radiotherapy but improved steadily over time. At 24 months, facial pain scores were notably better than baseline (mean, 7 [95% CI, 5-9] vs 12 [95% CI, 10-14]). All other scores were comparable to baseline, except for the muscular tension scores, which remained marginally worse (mean, 13 [95% CI, 10-16] vs 9 [95% CI, 7-11]).

In addition to GTQ scores, 54 of the 276 patients were also enrolled in a prospective adaptive radiotherapy trial³² with QOL metrics evaluated at baseline and follow-up visits by EORTC QLQ H&N35 and EORTC QLQ-C30 scores, as shown in Figure 3. Dry mouth and global function scores exhibited the most variations after treatment. At 24 months, most of the QOL composite scores (pain, social contact, social eating, speech, and swallow) were comparable or superior to baseline measures except for senses (taste), dry mouth, and cognitive functioning.

In this retrospective cohort study, we demonstrated uncompromised disease control and favorable toxic effect profiles using a systematic approach to reduce radiation dose and volume to elective nodal regions in patients with HPV-positive OPC undergoing CCRT. Despite 31.5% of patients having cT3-cT4 disease, the 2-year LRC, progression-free survival, and overall survival rates remained high. Only 1 patient had a locoregional failure in the elective fields treated with a 30-Gy dose. Only 17 patients (6.2%) required feeding tube placement during and after completion of chemoradiotherapy. Most patients were able to receive a full 3 cycles of high-dose cisplatin. The QOL analysis of the study cohort showed that most of the composite scores were comparable or even superior to baseline measures.

Clinical and toxic effects outcomes from this analysis are equivalent or superior to those of other similar trials,^9-15,24,25 many of which used a lower dose to gross disease but a higher dose to elective regions, as summarized in Table 2. Many of the comparable trials also had wider radiotherapy elective nodal coverage and target volume margins. For example, the NRG HN002 trial for p16-positive OPC⁹ used a reduced dose of 60 Gy to gross disease but used a dose of 54 Gy to high-risk subclinical sites and 44 to 48 Gy to low-risk nodal levels, with mandatory coverage of ipsilateral neck nodal levels Ib and V and a total of 0.8- to 1.0-cm expansion from GTV to PTV. Similarly, the dose de-escalation trial for p16-positive OPC conducted by Chera et al¹⁴ also used a dose of 60 Gy to gross disease and 54 Gy to subclinical nodal volumes with coverage of ipsilateral nodal levels Ib and V. The high dose and volume to the elective nodal regions might have in part contributed to the elevated feeding tube rates of 21.7% and 34%, respectively, in these studies. Notably, our cohort had the lowest rate of feeding tube use among patients treated with definitive CCRT.

Other studies have investigated modest reduction of the radiation dose or omission of elective microscopic regions. A pooled analysis of patients with HPV-positive and HPV-negative cancer found that elective nodal volumes of 40 Gy did not result in any treatment failures within the elective targets among cohorts with HPV-positive disease.³⁷ In the postoperative setting, doses of 30 to 36 Gy delivered twice daily with concurrent chemotherapy were shown to be adequate in the surgical bed as well as elective regions for HPV-positive tumors.¹² Two prospective trials using an elective radiation dose of 40 Gy^24,25 also showed uncompromised LRC, regardless of the anatomical site of disease and HPV status. The study by Deschuymer et al²⁵ included 200 patients with nonmetastatic squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, and larynx who received primary or adjuvant radiotherapy with or without concurrent systemic therapy. All patients received a dose of 70 Gy to gross disease and were randomized to elective nodal volumes of either 40 Gy or 50 Gy. With a median follow-up of 7.6 years, only 2 regional recurrences occurred in the patients receiving 40 Gy, the same as that in the group receiving 50 Gy. Of note, 32% of the study population did not receive systemic therapy, and it was unclear whether those who had regional recurrence in the 40-Gy group received concurrent systemic therapy. Another phase 2 single-arm study by Sher et al²⁴ investigated an elective nodal dose of 40 Gy in 72 patients with OPC or laryngeal cancer (90% received CCRT). With a median follow-up of 25 months, there were no solitary elective nodal recurrences. The study had comparable QOL trends in EORTC QLQ H&N35 and QLQ-C30 scores akin to those of our present cohort, although 33% their patients required feeding tube placement during treatment owing to weight loss or dysphagia. The results of our study also showed similar trends but overall more favorable EORTC QLQ H&N35 swallowing function scores compared with that of the RTOG (Radiation Therapy Oncology Group) 1016 study.¹³

Our institution has used a systematic approach of radiotherapy de-escalation since March 2017 for all patients with HPV-positive OPC undergoing CCRT, with further reduction of the elective radiation dose to 30 Gy, based on historical data on HPV-related cancers and modern HPV-specific OPC trials.^38,39 The classic report by Nigro et al³⁸ showed that most patients with squamous cell carcinoma of the anal canal achieved a complete clinical or pathologic response after preoperative chemoradiotherapy to 30 Gy, suggesting that even some gross disease could be eradicated with this dose. In a pilot study conducted at our institution using functional positron emission tomography imaging response with a hypoxia tracer,³⁹ patients were selected to receive a de-escalated radiation dose of 30 Gy to gross disease. All 14 patients treated per protocol had either a complete or partial pathologic response and maintained LRC at 2 years.

The control of subclinical metastasis has a shallow linear dose-response curve, reflecting interpatient heterogeneity in occult metastatic tumor burden.²² The commonly used subclinical dose in the range of 50 to 60 Gy was based on historical radiobiological data from limited imaging capability for detecting occult metastasis.^22,23 Because our ability to discriminate between gross disease and benign lymph nodes has improved with high-quality imaging, the results of our study and others mentioned above highlight that a significantly lower dose for elective volumes at approximately 30 Gy with concurrent chemotherapy can be sufficient for patients with HPV-positive OPC. Therefore, the elective radiotherapy dose of 50 to 60 Gy, or a modestly reduced dose of 40 Gy, likely resulted in overtreatment. In addition, studies from multiple institutions have demonstrated that it is safe to omit the uninvolved ipsilateral level Ib and V nodal regions^{20,26,27,40-42} as well as the contralateral high retropharyngeal regions in uninvolved neck.^28,43 Thus, we are confident that a dose of 30 Gy is sufficient for eradiating microscopic disease in our study population. However, whether it is safe to further reduce the elective radiation dose to below 30 Gy or to omit some elective nodal volumes completely remains unknown. It is also unclear whether the same principle can be applied to both HPV-positive and -negative tumors.

Some limitations of this study are worth mentioning. This was a single-institution, nonrandomized, retrospective cohort study. Results from this study have yet to be validated in multi-institutional, prospective settings to confirm reproducibility and generalizability. One patient had a previously unidentified gross node that did not receive a full 70 Gy of radiotherapy, indicating that with this approach of significantly reduced elective dose, one must be vigilant in target delineation to include small-volume gross disease that might be eradicated by the commonly used elective dose in the range of 50 to 60 Gy. Of note, 31.5% of the patients did receive an intermediate dose of 50 Gy for indeterminate nodes or other small areas of concern per the treating physician’s discretion, and this flexibility in dosing has worked well for treatment planning while still keeping most of the elective regions dosed to 30 Gy.

In this retrospective cohort study with a median follow-up of 26 months, the systematic de-escalation of radiotherapy to 30 Gy with reduction of elective volumes irradiated was associated with minimal toxic effects. With longer term follow-up data to validate the results, this regimen can be safely adapted for treating a wide range of patients with HPV-associated OPC undergoing primary CCRT.

Accepted for Publication: September 30, 2021.

Published Online: January 20, 2022. doi:10.1001/jamaoncol.2021.6416

Corresponding Author: Nancy Y. Lee, MD, Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065 (leen2@mskcc.org).

Author Contributions: Dr Tsai had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Tsai, McBride, Riaz, Chen, Pfister, Lee.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Tsai, McBride, Waldenberg, Lee.

Critical revision of the manuscript for important intellectual content: Tsai, Riaz, Kang, Spielsinger, Gelblum, Yu, Chen, Zakeri, Wong, Dunn, Pfister, Sherman, Lee.

Statistical analysis: Tsai, Riaz, Kang, Yu, Zakeri.

Obtained funding: Tsai, Lee.

Administrative, technical, or material support: Tsai, Kang, Spielsinger, Waldenberg, Lee.

Supervision: Tsai, McBride, Chen, Pfister, Lee.

Other (managed patients from the study cohort): Gelblum.

Conflict of Interest Disclosures: Dr Tsai reported receiving consultation fees from Varian Medical Systems outside the submitted work. Dr Riaz reported receiving research funding from ArcherDx and Repare Therapeutics Inc and consulting fees from Illumina, Inc, and PaigeAI. Dr Dunn reported receiving research funding from Regeneron Pharmaceuticals, Inc, Eisai Co, Ltd, CUE-101, and Replimune Group Inc and serving on the advisory board of Merck & Co, Inc, outside the submitted work. Dr Pfister reported receiving grants from the National Institutes of Health (NIH) and the Philanthropy–Serra Fund during the conduct of the study and research support from Hookipa Pharma Inc outside the submitted work. Dr. Lee reported receiving consultation fees from Merck & Co, Inc, Merck EMD, Mirati Therapeutics, Roche, and Elsie Pharmaceuticals. No other disclosures were reported.

Funding/Support: This study was supported by grant R01CA129182 from the NIH and Cancer Center Support Grant P30 CA008748 from the NIH/National Cancer Institute (NCI).

Role of the Funder/Sponsor: The NIH and NCI played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; nor the decision to submit the manuscript for publication.

Meeting Presentation: The study was presented in part at the 2021 Annual Meeting of the American Society for Radiation Oncology “Head and Neck Cancer in 2021: Updates and New Directions”; October 27, 2021; Chicago, IL.