Efficacy of Palliative Radiation Therapy (RT) for Diffuse Large B-Cell Lymphoma: A Population-Based Retrospective Review

Abstract

To investigate outcomes of palliative RT in patients with diffuse large B-cell lymphoma (DLBCL) and clinical factors that impact treatment efficacy. We hypothesize that factors can be identified which are associated with improved response. All patients with DLBCL who received palliative RT from 2001-2015 in the province were reviewed for patient characteristics, treatment details, and outcomes. Two hundred and seventeen patients who received 370 courses of palliative RT were identified. Median age at RT was 76 years, and 57% of courses were in male patients (n=211). Median equivalent dose in 2 Gy fractions (EQD2) was 19 Gy. Size of treated lesion was documented in 240 courses; median largest dimension was 5.8 cm. Irradiated sites were 22% skin, 18% head and neck, 14% spine, 13% abdomen, 10% pelvis, 7% bone, 6% thorax, 5% axilla, and 5% other. Indications for palliative RT were 42% pain, 22% enlarging mass, 20% obstruction or peripheral nerve compression, 9% spinal cord compression, 4% bleeding, 2% non-healing wound, and 1% pruritis. Clinical response assessment was available for 274 courses (74%), of which 42 courses had radiologic follow up; 2 courses had only radiologic response assessments. Symptom resolution was achieved in 42% of courses (114/274), symptom improvement in 40% (110/274), and stability in 13% (36/274); there was symptom progression in 5% (14/274). For courses given for pain relief, complete pain relief was achieved in 51% of courses (60/118) and partial pain relief in 36% (43/118). Using a logistic generalized estimating equation method, factors associated with symptomatic benefit and/or radiological response were initial stage I/II vs III/IV (94% vs 76%, p<0.0005), response vs no response to initial chemotherapy (84% vs 69%, p=0.03), concurrent vs no concurrent steroid use (71% vs 88%, p=0.001), skin vs non-skin disease (96% vs 79%, p<0.0005), and spinal vs non-spinal disease (61% vs 85%, p=0.008). Local control (LC) at 6 months was 65% (95% confidence interval [CI] 54-74%). Six month LC was 78% vs 50% when lesion size was <5.8 vs ≥5.8 (p=0.04); 74% vs 61% when initial stage was I/II vs III/IV (p=0.04); 51% vs 85% when the time from diagnosis to first relapse or progression was <1 year vs ≥1 year (p=0.02); and 75% vs 27% when initial response to chemotherapy was partial or complete vs none (p<0.0005). LC was not significantly associated with age, skin disease, spinal disease, EQD2, concurrent steroids use, or number of previous chemotherapy lines. On multivariable Cox regression analysis, response to initial chemotherapy (HR 4.3, 95% CI 2.1-8.5) was associated with improved LC. Palliative RT for DLBCL is effective for symptom improvement and local disease control. In deciding upon RT for incurable DLBCL, factors to consider include lesion size, response to initial chemotherapy, and anatomical site.