Predictive factors of severe radiation-induced lymphopenia in proton-treated patients

Abstract

Purpose or Objective Radiation-induced lymphopenia (RIL) has been reported in solid tumours and is associated with worse survival outcomes. There is little known thus far of potential predictive factors for developing severe RIL. This single institution study is aimed at establishing potential predictive factors of severe RIL in a cohort of patients treated with proton beam therapy (PBT). Materials and Methods One hundred and twenty patients treated with pencil beam scanning PBT without synchronous chemotherapy, from December 2018 to March 2020 at The Christie PBT centre were included in the study. All patients had pre-treatment lymphocyte count and on-treatment serial lymphocyte counts. Treatment sites were divided into three categories: central nervous system (CNS), head and neck (H&N), which included base of skull, and spine (for paraspinal and spinal treatment of the cervical, thoracic, lumbar or sacro-coccygeal region). Severe RIL was defined as grade 3 or higher based on nadir lymphocyte count < 0.5x109/L during PBT according to CTCAE v5.0. Clinical target volume (CTV) was obtained from planning images. Vx was calculated from RT images as the volume receiving x Gy or higher normalized to body weight to account for patients’ size. For patients without weight data (n=4), it is imputed using a 3rd power fit of weight vs age from the available data. To investigate the association between the incidence of severe RIL and dosimetric factors, multivariate logistic regression was used. To avoid multi-collinearity, dosimetric parameters were included in the multivariate regression one at a time together with other factors known to affect lymphocyte count. Results Fifty-four (45%) of the patients were female. The median age was 15 years (range 1 – 82). Treatment sites were CNS 69 (58%), H&N 26 (22%) and spine 25 (21%). Diagnoses are shown in Table 1. Prescribed PBT dose ranged from 50.4 Gy (RBE) to 75.6Gy (RBE). The median absolute lymphocyte count prior to treatment was 1.9X109/L (range 0.3 – 6.4) and during treatment 1.2X109/L (range 0.1-2.8). Seventeen patients (14%) developed grade 3 or 4 lymphopenia. In univariate analysis, severe RIL was associated with pre-treatment lymphocyte count (Odds Ratio (OR) 0.29, p=0.004), age (OR 0.93, p=0.018), spinal treatment site (OR 8.53, p<0.001), CTV (Ln) (OR 2.42, p<0.001), and V2-V30Gy. On multivariate analysis, severe RIL was associated with pre-treatment lymphocyte count, and V2-20Gy. Conclusion Severe RIL is seen most commonly in patients having treatment involving their spine. Baseline lymphocytes and low to intermediate dose volumes, V2 through V20 Gy, are predictive factors of developing severe RIL in patients treated with PBT. Further work in progress will better define dose volume limits to predict patients likely to develop severe RIL who may therefore require prophylactic antibiotics.