Development of Total Toxicity Burden Model in Head and Neck Cancer: Correlation of Clinician and Patient Reported Outcomes

Abstract

<h3>Purpose/Objective(s)</h3> Total Toxicity Burden (TTB) is an aggregate estimate of toxicity that has been piloted in clinical trial design to improve upon conventional univariate hypothesis testing. Head and neck cancer (HNC) treatment causes many acute and late toxicities that may be captured by objective and subjective assessments including physician grading, patient reported outcomes (PRO), and swallow studies. We developed a novel composite toxicity metric, TTB-PRO, that combines PROs and clinician reported toxicities. To determine if TTB-PRO can better represent patient toxicity in HNC clinical trials, we investigated the correlation between PRO and clinician reported toxicities individually and through this composite metric. <h3>Materials/Methods</h3> The TTB-PRO weighting was developed through a modified Delphi method involving 4 HNC radiation oncologists. We retrospectively assessed physician reported toxicity (CTCAE v 4.0), PRO scores (EORTC QLQ-H&N35) and videofluoro swallow studies for patients treated on a prospective institutional trial in locally advanced HNC (NCT02031250). Individual toxicities of HNC RT (dysphagia, xerostomia, mucositis, dermatitis, ORN) were graded and scored based on thresholds in 3 realms: PROs, objective data, and toxicity duration (acute, and 1-, 3-, and 12-months post-RT). TTB (using only clinician-graded toxicity) and TTB-PRO metrics were calculated through a weighted-sum for each patient. Spearman correlation coefficients were calculated to evaluate whether physician reported toxicity correlated with objective data, PROs, and TTB-PRO scores. <h3>Results</h3> A total of 81 patients were available for analysis; 41 patients received standard of care 70 Gy RT and 40 received an RT boost to 80 Gy. Mean age was 63 years, 88.9% were male. Sixty-eight patients were T4, 46 were N2b or higher, and 51 were p16 positive. Grade 3+ dysphagia at 3 months as graded by CTCAE, PRO, and VFSS was 24.6%, 12%, and 20.9%, respectively. Table 1 describes the differences in grade 3+ xerostomia (3 and 12 mo) and dysgeusia (12 mo). Spearman correlation coefficients (ρ) for physician vs. patient graded dysphagia at 1-, 3-, and 12-months post-RT were 0.06, 0.30, and 0.05, respectively. The ρ values for physician vs. patient reported xerostomia were 0.18 and 0.31 at 3 and 12 months, respectively. At 12 months, the dysgeusia ρ was -0.13. The ρ value for 12-month TTB vs. TTB-PRO scores was 0.63. <h3>Conclusion</h3> Physician scored toxicity had weak correlation with PROs in a prospective clinical trial of locally advanced HNC. TTB-PRO may be a more sensitive means to capture aggregate toxicity representing the impact of treatment on patient quality of life. This novel toxicity metric may be a useful tool to assess toxicity in future prospective HNC clinical trials.