Introduction

Utilizing 3D printed fully-customized oral stents may reduce mucositis in patients with oropharyngeal cancer undergoing RT. Our previous work showed that 3D printed stents are comparable to standard stents in terms of patient-reported outcomes (PROs) and positioning reproducibility (Zaid et al., Oral Oncology, 2020).

We present initial findings from the non-randomized lead-in phase of a multi-institutional trial involving six institutions across the United States (Protocol 2020-1153). The digital design and 3D printing of fully customized mouth-opening tongue-depressing (MOTD) oral stents for enrolled patients is centrally located at MD Anderson Cancer Center in Houston, Texas. Fabricated stents are shipped overnight to each participating institution. We aim to demonstrate the clinical feasibility of using this approach and report PRO data for the non-randomized phase of the trial, which will serve as a benchmark for the randomized phase.

Methods

Attending clinicians acquired 3D intraoral anatomy models of enrolled patients using a commercial IOS device (Trios 3, 3Shape) and underwent five design and post-processing stages to fabricate the oral stent. These stages include: (1) bite registration, (2) stent design, (3) pre-processing, (4) SLA printing, and (5) polishing. We measured fabrication time for each stage and calculated material printing cost for each stent based only on cost per mL of the 3D printing resin material. Participating centers received customized oral stents for patients to be utilized throughout treatment (1-2 days prior to CT simulation). The treating physician assessed mucositis scores for various oral sites of each patient based on the severity of ulceration size and erythema (0 = none to 5 = severe) across all timepoints including baseline (before RT), week 3 of RT, and end of treatment. We compared mucositis scores between timepoints using a matched-pairs analysis and recorded adverse events (AEs).

Results

Eight patients enrolled in the lead-in phase (age 48-80 years, 7 male, 1 female, 4 tonsillar, and 4 base of tongue cancer). Seven patients received volumetric modulated arc therapy (VMAT) to a total dose of 6996 cGy in 33 fractions, and one patient received intensity modulated proton therapy (IMPT) to a total dose of 6000 cGy (RBE, relative biological effectiveness) in 30 fractions. All patients received concurrent chemo using carboplatin or cisplatin. Mean total fabrication time for stages (1) to (5) was 7 hours, with a mean material printing cost of $11 per stent. Baseline and week 3 of RT mean mucositis scores were significantly different (mean difference = 0.39, 95% CI [0.06,0.72], p = 0.03). Similarly, mean mucositis scores between baseline and end of treatment were significantly different (mean difference = 0.49, 95% CI [0.1,0.9], p = 0.02). Week 3 of RT and end of treatment of mean mucositis scores were not significantly different (mean difference = 0.1, 95% CI [-0.5,0.7], p = 0.7). Common AEs for all patients included grade 1-3 oral mucositis and radiation dermatitis.

Conclusion

We successfully demonstrated the in-house, centralized development and successful distribution of 3D printed fully customized oral stent devices to participating institutions across the United States for the non-randomized phase of a multi-institutional trial. We will utilize PRO results from this study as a benchmark for evaluating patients enrolled in the randomized phase of the trial.