PT - JOURNAL ARTICLE AU - Anne Segolene Cottereau AU - Michel Meignan AU - Christophe Nioche AU - Jerome Clerc AU - LAETITIA VERCELLINO AU - Olivier Casasnovas AU - Catherine Thieblemont AU - Irene Buvat TI - A radiomic biomarker characterizing tumor dissemination in patients has complementary predictive value to metabolic tumor volume in a large cohort of diffuse large B-cell lymphoma patients responding to R-CHOP treatment DP - 2020 May 01 TA - Journal of Nuclear Medicine PG - 270--270 VI - 61 IP - supplement 1 4099 - http://jnm.snmjournals.org/content/61/supplement_1/270.short 4100 - http://jnm.snmjournals.org/content/61/supplement_1/270.full SO - J Nucl Med2020 May 01; 61 AB - 270Objectives: The total metabolic tumor volume (MTV) calculated from baseline 18F-FDG PET images is an early prognostic factor in lymphoma. Yet, MTV does not precisely reflect the dissemination of the disease throughout the body. We recently introduced a new radiomic feature measuring the dissemination of the tumor mass within the whole body [1]. Here, we investigate its prognostic value when measured at baseline in a large cohort of Diffuse Large B-Cell Lymphoma (DLBCL) patients responding to R-CHOP treatment. Methods: A total of 301 patients (60 to 80 years old) from an ancillary study of the REMARC trial (NCT01122472) were retrospectively analyzed. All patients had a baseline 18F-FDG PET/CT before R-CHOP and baseline MTV was found to be a strong prognosticator of outcome [2]. Based on the regions delineated to calculate the MTV, the LIFEx software [3] was used to derive the Dmax feature defined as the distance between the 2 lesions that were the furthest apart in the 3D whole body PET scan. The prognostic value of Dmax was investigated using ROC analysis to identify the cut-off values for both Progression Free Survival (PFS) and Overall Survival (OS). Survival functions were obtained using Kaplan Meier (KM) analysis. Multivariate analyses combining Dmax and MTV were conducted to assess the complementarity of the two biomarkers. Results: 291/301 patients had multifocal lesions from which Dmax could be calculated, with median Dmax of 42 cm and median MTV of 242 mL. Among these patients, 240 were stage 3 or 4, with median Dmax of 46 cm and median MTV of 277 mL. The median follow up was 5 years. In univariate analysis of the 291 patients, the Dmax Areas Under the ROC curve (AUC) were 0.628 for PFS (p<0.05) and 0.618 for OS (p<0.05). The predictive value of Dmax for PFS was also observed in the 240 stage 3-4 patients, with AUC of 0.625 for PFS. In multivariate analysis, Dmax>47 cm and MTV>220mL remained independent prognosticators for both PFS (p=0.0006; p=0.0016) and OS (p=0.023; p=0.0007). In the 291 patients, combining MTV (cut-off 220mL) and Dmax(cut-off 47 cm) yielded 3 risk groups for PFS (log-rank test p<0.0001) and OS (p<0.0001): high risk with 2 adverse factors (4y-PFS and OS of 46% and 67%, n=82), intermediate risk with 1 adverse factor (64% and 78%, n=114), and low risk with no adverse factor (87% and 95%, n=95). This stratification remained highly significant in the 240 stage 3-4 patients (p<0.001 for PFS and p<0.003 for OS) with 4y-PFS and OS of 46% and 66% for the high risk group (n=82), 4y-PFS and OS of 60% and 77% for the intermediate risk group (n=100), and 4y-PFS and OS of 79% and 97% for the low risk group (n=58). Conclusions: The predictive value of a radiomic tumor dissemination biomarker measured from baseline 18F-FDG PET is confirmed in a large cohort of DLBCL patients and further improves patient stratification compared to MTV alone.