[ 18 F-FDG PET/CT Metabolic Parameters and Circulating Tumour DNA Mutation Abundance in Diffuse Large B-Cell Lymphoma: Correlation and Survival Analysis].

OBJECTIVE: To investigate the correlation between 18 Fluoro-deoxyglucose positron emission tomography/computed tomography (18 F-FDG PET/CT) metabolic parameters and peripheral blood circulating tumour DNA (ctDNA) in patients with diffuse large B-cell lymphoma (DLBCL), and the prognostic value of these two types of parameters in predicting progression-free survival (PFS).

METHODS: Clinical, PET/CT and ctDNA data of DLBCL patients who underwent peripheral blood ctDNA testing and corresponding PET/CT scans during the same period were retrospectively analyzed. At the time of ctDNA sampling and PET scan, patients were divided into baseline and relapsed/refractory (R/R) groups according to different disease conditions. CtDNA mutation abundance was expressed as variant allele frequency (VAF), including maximum VAF (maxVAF) and mean VAF (meanVAF). Total metabolic tumour volume (TMTV) and total lesion glycolysis (TLG) were obtained by the 41% maximum normalized uptake value method, and the distance between the two farthest lesions (Dmax) was used to assess the correlation between PET parameters and ctDNA mutation abundance using Spearman correlation analysis. The receiver operating characteristic (ROC) curves were used to obtain the optical cut-off values of those parameters in predicting PFS in the baseline and R/R groups, respectively. Survival curves were outlined using the Kaplan-Meier method and log-rank test was performed to compare survival differences.

RESULTS: A total of 67 DLBCL patients ［28 males and 39 females, median age 56.0(46.0, 67.0) years］ were included and divided into baseline group (29 cases) and R/R group (38 cases). Among these PET parameters, baseline TMTV, TLG, and Dmax were significantly correlated with baseline ctDNA mutation abundance, except for maximum standardized uptake value (SUVmax) (maxVAF vs TMTV: r =0.711; maxVAF vs TLG: r =0.709; maxVAF vs Dmax: r =0.672; meanVAF vs TMTV: r =0.682; meanVAF vs TLG: r =0.677; meanVAF vs Dmax: r =0.646). While in all patients, these correlations became weaker significantly. Among R/R patients, only TMTV had a weak correlation with meanVAF ( r =0.376). ROC analysis showed that, the specificity of TMTV, TLG and Dmax in predicting PFS was better than mutation abundance, while the sensitivity of ctDNA mutation abundance was better. Except R/R patients, TMTV, TLG, Dmax, and VAF were significantly different at normal/elevated lactate dehydrogenase in baseline group and all patients (all P <0.05). Survival curves indicated that high TMTV (>109.5 cm3 ), high TLG (>2 141.3), high Dmax (>33.1 cm) and high VAF (maxVAF>7.74%, meanVAF>4.39%) were risk factors for poor PFS in baseline patients, while only high VAF in R/R patients (both maxVAF and meanVAF >0.61%) was a risk factor for PFS.

CONCLUSION: PET-derived parameters correlate well with ctDNA mutation abundance, especially in baseline patients. VAF of ctDNA predicts PFS more sensitively than PET metabolic parameters, while PET metabolic tumour burden with better specificity. TMTV, TLG and VAF all have good prognostic value for PFS. PET/CT combined with ctDNA has potential for further studies in prognostic assessment and personalized treatment.