Exploring quantitative MRI biomarkers of head and neck post-radiation lymphedema and fibrosis: Post hoc analysis of a prospective trial

Abstract

Importance Quantifying Head and Neck Lymphedema and Fibrosis (HN-LEF) is crucial in the investigation and management of this highly prevalent treatment sequelae in head and neck cancer (HNC). The HN-LEF grading system classifies physically palpable soft-tissue injury categorically. Imaging biomarkers from MRI may serve to complement or validate physical HN-LEF grading when assessing the effectiveness of therapeutic interventions or toxicity profiles of patients. Objective To explore the relationship between 1) physical HN-LEF classification in submental and oral regions of interest (ROI) and the MRI T1- and T2-weighted signal intensity (SI) in close proximity regions, and 2) a novel HN-LEF score and MRI T1 and T2 structural volumes. Design: Post hoc analysis of pilot single-arm MANTLE trial (NCT03612531). Setting: Single institution, NCI-designated comprehensive cancer center. Participants A total of 16 individuals (mean [SD] age, 68.28 [7.0] years; 3[19%] female) enrolled in the MANTLE trial underwent MRI. All participants were disease-free at least two years post radiotherapy with grade ≥2 fibrosis (in any cervical ROI) and grade ≥2 dysphagia (per DIGEST). Over a 12-week period, participants engaged in manual therapy sessions accompanied by concurrent standardized multiparametric, serial MRI examinations and palpation-based HN-LEF evaluations at 3-time points: baseline, post-manual therapy, and post-washout. Exposures: The independent variable HN-LEF included its categorical classification (No-LEF, A-B = edema, C= edema + fibrosis, D=fibrosis) and a novel metric (10-point scale) derived from the HN-LEF categories (considering both type and severity classification). Main Outcomes and Measures: The T1- and T2-weighted MRI SI was examined by Kruskal-Wallis tests in relation to HN-LEF categories and the novel HN-LEF score. We hypothesized higher T2 SI in edema states, higher T1 SI in fibrotic states, and decreasing structural volume as the HN-LEF score increased. Results: We identified differences in mean ranks among HN-LEF categories in relation to the MRI SI (A-B and C are higher than D and No-LEF for T2 SI, and A-B is the highest for T1). Furthermore, six pairs of FOM volumes on MRI demonstrated a strong negative correlation (p<0.05) with the HN-LEF score at adjacent palpable sites: digastric vs. submental left (ρ = -0.421; 95% CI, -0.65~ -0.10, T1), mylohyoid vs. submental left (ρ = -0.36; 95% CI, -0. 62~ 0.03, T1), digastric vs. submental left (ρ = --0.45; 95% CI, -0. 72~ -0.06, T2), genioglossus vs. Intraoral left (ρ = -0.47; 95% CI, -0. 74~ -0.07, T2), mylohyoid vs. Intraoral left (ρ = -0.48; 95% CI, -0. 75~ -0.09, T2), tongue base vs. Intraoral left (r = -0.42; 95% CI, -0. 71~ -0.01, T2). Conclusions and Relevance This exploratory analysis provides hypothesis-generating data supporting further study of MRI SI as an imaging biomarker of edematous soft tissue states after RT in HNC, but does not support the hypothesized T2 SI relationship with fibrotic tissue states. The inverse correlation between the novel HN-LEF scores and structural volumes points to the potential validity of this novel metric assuming structural volume diminishes as patients move from edema to fibrotic states. This study highlights the potential for enhancing the LEF quantification using imaging metrics, which might further aid in the early detection and precise measurement of lymphedema and fibrosis severity in post-radiation HNC patients.