Daily Adaptive Radiotherapy Delivers Improved Efficacy and Safety Outcomes in HNSCC

David J. Sher, MD, MPH

Daily adaptive radiotherapy (DART) with 1-mm planning target volume (PTV) margins is feasible in patients with head and neck squamous cell carcinoma and was associated with improved dosimetric parameters compared with standard radiotherapy, according to preliminary outcomes from the phase 2 DARTBOARD trial (NCT04883281) that were presented at the 2023 ASTRO Annual Meeting.

Patients in the DART arm received a lower mean ipsilateral parotid dose, at 11.5 Gy (standard error [SE], 1.0) vs 16 Gy (SE, 1.5) in the image-guided radiation therapy (IGRT) arm (P = .02). Patients in the DART arm also received a lower mean contralateral parotid dose, at 8.9 Gy (SE, 1.1) vs 11.8 Gy (SE, 1.9) in the IGRT arm (P = .20). Additionally, the mean ipsilateral and contralateral submandibular gland doses were lower in the DART arm, at 42.2 Gy (SE, 3.3) and 28.2 Gy (SE, 2.7), respectively, vs 56.3 Gy (SE, 2.4) and 36.5 Gy (SE, 2.9), respectively, in the IGRT arm (P < .01; P = .04).

“Adaptive radiotherapy holds the promise of updating the treatment plan with anatomic changes to both the patient and the disease,” David J. Sher, MD, MPH, professor of radiation oncology, chief of Head and Neck Radiation Oncology Service and vice-chair and medical director of Clinical Operations and Quality, at UT Southwestern Medical Center in Dallas, Texas, said during the presentation. “DART allows not only more frequent adaptation but also margin reduction. With that margin reduction comes clear in silico benefits to dosimetric organs at risk, particularly the salivary glands.”

The DARTBOARD investigators hypothesized that DART with 1-mm PTV margins would decrease the incidence of patient-reported xerostomia 1 year after treatment.

Eligible patients included those with stage I to IVB squamous cell carcinoma of the oropharynx, larynx, or hypopharynx. Those with stage I glottic larynx squamous cell carcinoma were excluded.

This trial used the Varian Ethos system to deliver cone beam computed tomography (CBCT)–guided adaptive radiotherapy. For each patient, the physician contoured on the CBCT, calculated a treatment plan on synthetic computed tomography (CT), performed calculation-based quality assurance (QA), then delivered the radiotherapy. The plan optimization was based on the original CT simulation directives, the original volume of the primary gross tumor volume (GTV) was preserved, except for baseline electron return effect interface changes, and the nodal GTV was reduced, except for in nodes with pre-existing extranodal extension.

Patients received stage-dependent treatment with radiotherapy, accelerated radiotherapy, or chemoradiotherapy alone. Eligible chemotherapy included cisplatin, carboplatin/paclitaxel, or cetuximab (Erbitux). This study used the UT Southwestern involved nodal radiotherapy (INRT) paradigm, and effectively, there was no elective neck irradiation. Investigators identified suspicious nodes by utilizing either radiologic criteria or an artificial intelligence (AI) classification model.

Patients (n = 50) were randomly assigned to receive INRT with or without DART and were stratified by oropharynx squamous cell carcinoma vs other disease sites. The primary end point of this trial was results of the Xerostomia Questionnaire at 1 year.

“There are many ways of doing GTV-to-PTV expansions in head and neck cancer,” Sher noted, explaining that the most common method is GTV plus 5 mm to a clinical target volume (CTV) at 70 Gy, with an additional PTV margin of 3 to 5 mm.

However, patients in the INRT arm of DARTBOARD (n = 26) received a 5-mm GTV-to-PTV margin in the primary and nodal GTV, as well as in the primary CTV. Patients in the DART arm (n = 24) had the same structures contoured daily with a 1-mm GTV-to-PTV margin, except 2 mm sup/inf because of the CBCT thickness. Over 35 fractions, the primary and nodal GTV received 70 Gy of radiotherapy. Lymph nodes in the primary GTV that were at least 17 mm received 63 Gy of radiotherapy in 35 fractions, and the remaining nodes, which were in the same station as the involved nodes or defined by AI, received 56 Gy of radiotherapy in 35 fractions.

“These treatments were extremely tight,” Sher emphasized while sharing an example of a patient treated in the study who had T1/N1, P16-positive oropharynx cancer. He explained that this patient had significant sparing of their bilateral prodded glands and contralateral neck.

The mean ages in the IGRT and DART arms were 61.5 years (SE, 2.1) and 60.7 years (SE, 1.9), respectively (P = .34). Most patients in each arm were male (81%; 92%; P = .27), and 77% and 75% of the patients in each arm had oropharynx disease (P = .57). Regarding American Joint Committee on Cancer v. 7 T stage, in the IGRT arm, 15%, 21%, 19%, and 35% of patients had T1, T2, T3, and T4 disease, respectively; in the DART arm, these rates were 21%, 25%, 42%, and 13%, respectively (P = .17). Regarding AJCC 7 N stage, in the IGRT arm, 19%, 8%, 35%, and 38% of patients had N0, N1, N2a, and N2b disease; these rates were 29%, 4%, 29%, 29%, and 8%, respectively, in the DART arm (P = .5). Most patients in the IGRT (96%) and DART (88%) arms received chemoradiotherapy vs radiotherapy alone (P = .24).

“There was no significant difference in any of the [doses to the] swallowing glands at risk in the oropharynx population, although there was arguably a trend in oral cavity dose,” Sher explained. The mean oral cavity doses in the DART and IGRT arms were 24.7 Gy (SE, 2.1) and 29.4 Gy (SE, 2.3), respectively (P = .14). In the DART arm, the mean doses in the superior constrictor, middle constrictor, inferior constrictor, and esophagus were 38.4 Gy (SE, 3.9), 29.7 Gy (SE, 4.2), 11.1 Gy (SE, 3.3), and 4.7 Gy (SE. 0.7), respectively. In the IGRT arm, the mean doses in these locations were 43.3 Gy (SE, 4.1), 29.1 Gy (SE, 2.8), 9.3 Gy (SE, 1.0), and 5.7 Gy (SE, 1.0), respectively (P = .4, .91, .6, and .47, respectively).

In the DART arm, the time from the CBCT to the end of radiotherapy was approximately 33 minutes, the door-in to door-out time for the patients was an average of 39.4 minutes, the average contour time was approximately 12.6 minutes, and the doctors spent a total of 22 minutes at the console, which Sher noted included recontouring, waiting for the plan to recalculate, QA, signage, and the delivery of the plan.

Patients in the DART arm experienced significantly fewer incidences of grade 2 or higher dermatitis than those in the IGRT arm (8% vs 31%; P = .048). The rates of grade 2 or higher mucositis were 75% and 92% in the DART and IGRT arms, respectively (P = .10). Additionally, there were no significant differences in low-, intermediate-, or high-risk dysphagia between the 2 arms. Grade 2 or higher dysphagia occurred in 75% and 81% of patients in the DART and IGRT arms, respectively (P = .63), and grade 3 or higher dysphagia occurred in 8% and 19% of patients, respectively (P = .27).

In the oropharynx cohort (n = 38), patients in the DART arm experienced a statistically significant improvement in the incidence of grade 2 or higher mucositis, at 72% vs 95% in the IGRT arm (P = .055). Additionally, the rate of grade 2 or higher dermatitis was 6% in the DART arm vs 30% in the IGRT arm (P = .052). Furthermore, in this population, grade 2 or higher dysphagia occurred in 72% and 80% of patients in the DART and IGRT arms, respectively (P = .57), and grade 3 or higher dysphagia occurred in 11% and 15% of patients, respectively (P = .72).

At 3 months, there was no significant difference in PROs between the 2 arms, although there was a slight numerical improvement favoring the DART arm, Sher explained. In the overall population, the Xerostomia Questionnaire scores were 33.3 (SE, 5.5) in the DART arm vs 38.6 (SE, 5.7) in the IGRT arm. In the oropharynx population, the Xerostomia Questionnaire scores were 31.7 (SE, 6.2) in the DART arm vs 40.6 (SE, 6.0) in the IGRT arm.

At 6 months, the Xerostomia Questionnaire scores were 28.5 (SE, 5.4) in the DART arm vs 36.1 (SE, 5.9) in the IGRT arm, although these differences were not statistically significant. In the oropharynx population, the Xerostomia Questionnaire scores were 28.3 (SE, 6.5) in the DART arm vs 39.1 (SE, 6.4) in the IGRT arm. However, the EORTC sticky saliva scores were significant and favored the DART arm in both the overall and oropharynx populations, at 35.4 (SE, 4.8) vs 57.4 (SE, 8.4) and 38.5 (SE, 5.1) vs 61.9 (SE, 10.4) in the DART and IGRT arms, respectively. “We do suspect this is due to the improved submandibular gland sparing,” Sher said.

At a median follow-up from the end of radiotherapy of 9.6 months (interquartile range, 6.1-12.7), no marginal recurrences were observed. One patient in the IGRT arm experienced an in-field local recurrence. Additionally, 2 patients developed distant metastases, 1 of whom died from oropharyngeal hemorrhage, and 1 of whom received salvage local therapy and experienced a subsequent in-field regional recurrence. Moreover, 2 patients experienced synchronous regional recurrence and distant metastasis, both of whom had bulky N2b disease at presentation.

One limitation of this trial is that it was based on an INRT platform, which Sher notes has questionable generalizability. Additionally, this trial has limited follow-up and small patient numbers.

Disclosures: Dr Sher reports research funding from Varian Medical Systems and honoraria from UpToDate and Red Journal.

Reference

Sher DJ, Avkshtol V, Lin MH, et al. Acute toxicity and efficiency outcomes in the DARTBOARD randomized trial of daily adaptive radiotherapy for head and neck squamous cell carcinoma. Presented at: 2023 American Society for Radiation Oncology Annual Meeting (ASTRO); October 1-4, 2023; San Diego, CA. Abstract LBA08.