Abstract
BACKGROUND: Few studies concentrate on spleen dosimetry of radiotherapy for gastric cancer (GC). Although there is no consensus on the spleen dose-volume threshold for lymphopenia, several studies indicated that the higher the spleen dose, the higher the risk of lymphopenia. This study aimed to identify the appropriate spleen dosimetric parameters for predicting grade 4 + lymphopenia in patients with locally advanced GC. MATERIAL AND METHODS: A total of 295 patients treated with nCRT and nChT from June 2013 to December 2021 at two major centers were included, of whom 220 were assigned to the training cohort and 75 to the external validation cohort. RESULTS: Grade 4 + lymphopenia was more common in the nCRT than in the nChT group (49.5% vs. 0, P < 0.001 in the training cohort; 25.0% vs. 0, P = 0.001 in the external validation cohort). Age ≥ 60 years (P = 0.006), lower pretreatment absolute lymphocyte count (P = 0.001), higher spleen volume (SPV) (P = 0.001), and higher V(20) (P = 0.003) were significant risk factors of grade 4 + lymphopenia for patients treated with nCRT. Patients with grade 4 + lymphopenia had significantly worse PFS (P = 0.043) and showed a negative correlation trend with OS (P = 0.07). Limiting V(20) to < 84.5% could decrease the incidence of grade 4 + lymphopenia by 35.7%. The predictive effectiveness of the multivariable model in the training and external validation cohorts was 0.880 and 0.737, respectively. CONCLUSION: Grade 4 + lymphopenia during nCRT was more common than nChT, and was associated with a worse PFS in GC patients. Constraining the spleen V(20) to < 84.5% may indirectly improve outcomes through lymphocyte preservation.