Dosimetric Analysis for Predicting Secondary Lymphedema After Radiation for Head and Neck Cancer

• Published in: International journal of radiation oncology, biology, physics Vol. 111; no. 3; p. e371
• Main Authors: Jaworski, E., Sun, Y., Vineberg, K.A., Mierzwa, M.L., David, C., Morales Rivera, K.A., Schonewolf, C.A., Casper, K., Rosko, A., Swiecicki, P.L., Worden, F., Shah, J.L.
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.11.2021
• ISSN: 0360-3016; 1879-355X
• DOI: 10.1016/j.ijrobp.2021.07.1096

Secondary lymphedema (SL) following radiation (RT) for head and neck (H&N) cancer occurs commonly with multiple adverse consequences1. Primary location, RT dose, months since treatment, and increasing treatment modalities are associated with SL1. We sought to define structures within anatomically relevant compartments and correlate dosimetric parameters of these structures with SL. We hypothesized that SL patients would have received different doses to these structures as compared to patients without SL. This single-institution retrospective study was IRB approved. We identified 23 H&N patients referred to Occupational Therapy (OT) for SL management from 2017 – 2019 after completing H&N RT (SL patients). These patients were matched 1:2 to patients having received H&N RT without OT referral (non-SL patients). Matching was based on primary site, treatment paradigm, RT dose, unilateral vs bilateral neck RT, and RT year. For all 69 patients, we contoured 3 structures: anterior neck soft tissue superficial to larynx from the caudal mandible or cranial hyoid inferiorly to C7, soft tissue superficial to the spinous processes from C1-C7, and subcutaneous tissue superficial to bilateral trapezius muscles. Wilcoxon rank-sum tests were used to compare the following dosimetric parameters of Anterior Neck, Spinous Process, and Trapezius structures between SL (n = 23) vs non-SL (n = 46) patients: mean dose, D0.1 cc [Gy], D100% [Gy], and V15 Gy [%] to V50 Gy [%] in 5 Gy increments. Given this exploratory analysis, we did not adjust for multiple comparisons. SL and non-SL patients were balanced for age, BMI, smoking, steroid use during RT, primary site, disease stage, treatment paradigm, bilateral vs unilateral neck dissection, radiation dose and unilateral vs bilateral neck RT (all P > 0.2). Vascular comorbidities were imbalanced with SL patients more commonly having both diabetes and hypertension (17%) vs non-SL patients (2%), and non-SL patients more commonly having HTN (41% vs 17%, P = 0.03). Oropharynx was the most common site (61%) followed by oral cavity (20%), larynx (13%), salivary (4%), and cutaneous (1%). Anterior Neck, mean dose (P = 0.03), V15 Gy [%] (P = 0.01), V20 Gy [%] (P < 0.01), V25 Gy [%] (P < 0.01), V30 Gy [%] (P = 0.02), V35 Gy [%] (P = 0.02), and V40 Gy [%] (P = 0.04) were significantly higher in patients with SL. Anterior Neck D100% [Gy] (P = 0.07), Spinous Process D100% [Gy] (P = 0.052), and Trapezius V15 Gy [%] (P = 0.06) trended towards significance. To our knowledge, this is the first investigation of dosimetric parameters associated with SL following H&N RT. Our findings suggest that higher volume of low-mid RT dose within the anterior neck is associated with developing SL. This may be related to circumferential neck dose and/or the role of superficial lymphatics within level VI after H&N RT. These findings warrant prospective validation in a larger cohort, but may be considered in the context of RT treatment planning to minimize SL after RT. (HUM00176639).