A longitudinal prospective mixed model analysis of patient reported outcomes after head and neck conformal reirradiation

Presentation: AHNS-064
Topic: Radiation Therapy
Type: Oral
Date: Thursday, May 2, 2019
Session: 10:15 AM - 11:00 AM Scientific Session 8 - Outcomes & Miscellaneous
Authors: Courtney Pollard, III, MD, PhD1, Theresa Nguyen, BS1, Sweet P Ng, MBBS1, Gary B Gunn, MD1, Adam S Garden, MD1, Steven J Frank, MD1, Jay P Reddy, MD, PhD1, William H Morrison, MD1, Clifton D Fuller, MD, PhD1, David I Rosenthal, MD1, Charles S Cleeland, PhD2, Tito R Mendoza, PhD2, Jack Phan, MD, PhD1
Institution(s): 1Department of Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, USA, 2Department of Symptom Research, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Purpose or Objective: Conformal radiotherapy can potentially lower patient symptom burden associated with head and neck (HN) reirradiation (reXRT).  Here, we prospectively evaluate patient reported outcomes (PRO) using the MD Anderson Symptom Inventory – Head and Neck module (MDASI-HN) in patients with HN malignancies reirradiated with intensity modulated radiation therapy (IMRT), stereotactic body radiation therapy (SBRT) or proton beam therapy (PBT). 

Material and Methods: Between 2014 -2018, 185 patients with previously irradiated HN malignancies received either IMRT (30-35 fractions over 6-7 weeks), SBRT (5 fractions over 2 weeks) or PBT (30-35 fractions over 6-7 weeks) reirradiation (reXRT) and enrolled in our prospective reXRT protocol with longitudinal PRO assessment.  Protocol eligibility include biopsy-confirmed HN cancer, documented prior HN radiation to ≥40 Gy, and curative intent reXRT.  Pre- and post-treatment (2 weeks, 3 months, 6 months) MDASI-HN symptom (22 items) and interference (composite of 6 items) scores were assessed for patients that completed the planned reirradiation.Symptom and interference severity (0-10 numeric scale) were based on mean scores of individual items and compared by reXRT modality (IMRT v. SBRT v. PBT).  Additional potential covariates included sex, age, performance status (ECOG PS), reXRT site (skull base v. mucosal v. non-mucosal), reXRT volume, chemotherapy and surgery.   Univariate and multivariate mixed models were used to determine covariates effects. 

Results: Of the 109 patients eligible for assessment, 51 (47%) received IMRT, 33 (30%) received SBRT, 25 (23%) received PBT, 37% were mucosal reXRT, and 40% were skull base reXRT.  Of the 23 MDASI-HN items evaluated and stratified by reXRT modality, a significant difference in mean score was found for fatigue (P=0.007), mucus production (P=0.044), and difficulty swallowing/chewing (P=0.043), all favoring SBRT (figure below).

Mixed model analysis demonstrated an independent effect of ECOG PS (P=0.001) and reXRT modality (P=0.006) on fatigue, with patients receiving SBRT reporting less fatigue (mean score 2.73; 95% CI 1.85-3.61) than those receiving IMRT (3.90; 95% CI 3.10-4.70; P=0.027) or PBT (4.71; 95% CI 3.68-5.73; P=0.002).

ReXRT modality did not independently impact mucous production on mixed model analysis, which was independently associated with ECOG PS (P=0.002) and reXRT site (P=0.006).  Patients receiving mucosal reXRT (mean score 3.96; 95% CI 3.09-4.82) had higher mucous production symptom burden compared to skull base (2.28; 95% CI 0.86-3.69; P=0.025) and non-mucosal reXRT (2.66; 95% CI 1.84-3.47; P=0.004).

Similarly, reXRT modality did not impact difficulty swallowing/chewing on mixed model analysis, which was associated with ECOG PS (P=0.006) and reXRT site (P=0.029).  Patients with PS 0 and non-mucosal reXRT (4.20; 95% CI 3.18-5.22; P=0.009) had lower difficulty swallowing/chewing scores compared to PS 1-2 and mucosal reXRT (2.76; 95% CI 1.80-3.73; P=0.009).

Conclusion: In a longitudinal prospective analysis, patient reported symptom burden, as assessed by MDASI-HN, within 6 months following definitive HN reXRT with IMRT, SBRT and PBT was low.  ECOG PS and site of retreatment were the most significant factors affecting symptom severity. These results suggest that the effect of radiation modality on patient reported symptom burden is minimal, with lower acute symptom burden associated with a shorter treatment course.