Baseline Total Metabolic Tumor Volume Measured with Fixed or Different Adaptive Thresholding Methods Equally Predicts Outcome in Peripheral T Cell Lymphoma.

Abstract

The purpose of this study was to compare in a large series of peripheral T cell lymphoma, as a model of diffuse disease, the prognostic value of baseline total metabolic tumor volume (TMTV) measured on ¹⁸F-FDG PET/CT with adaptive thresholding methods with TMTV measured with a fixed 41% SUV_max threshold method. One hundred six patients with peripheral T cell lymphoma, staged with PET/CT, were enrolled from 5 Lymphoma Study Association centers. In this series, TMTV computed with the 41% SUV_max threshold is a strong predictor of outcome. On a dedicated workstation, we measured the TMTV with 4 adaptive thresholding methods based on characteristic image parameters: Daisne (Da) modified, based on signal-to-background ratio; Nestle (Ns), based on tumor and background intensities; Fit, including a 3-dimensional geometric model based on spatial resolution (Fit); and Black (Bl), based on mean SUV_max The TMTV values obtained with each adaptive method were compared with those obtained with the 41% SUV_max method. Their respective prognostic impacts on outcome prediction were compared using receiver-operating-characteristic (ROC) curve analysis and Kaplan-Meier survival curves. The median value of TMTV_41%, TMTV_Da, TMTV_Ns, TMTV_Fit, and TMTV_Bl were, respectively, 231 cm³ (range, 5-3,824), 175 cm³ (range, 8-3,510), 198 cm³ (range, 3-3,934), 175 cm³ (range, 8-3,512), and 333 cm³ (range, 3-5,113). The intraclass correlation coefficients were excellent, from 0.972 to 0.988, for TMTV_Da, TMTV_Fit, and TMTV_Ns, and less good for TMTV_Bl (0.856). The mean differences obtained from the Bland-Altman plots were 48.5, 47.2, 19.5, and -253.3 cm³, respectively. Except for Black, there was no significant difference within the methods between the ROC curves (P > 0.4) for progression-free survival and overall survival. Survival curves with the ROC optimal cutoff for each method separated the same groups of low-risk (volume ≤ cutoff) from high-risk patients (volume > cutoff), with similar 2-y progression-free survival (range, 66%-72% vs. 26%-29%; hazard ratio, 3.7-4.1) and 2-y overall survival (79%-83% vs. 50%-53%; hazard ratio, 3.0-3.5). The prognostic value of TMTV remained quite similar whatever the methods, adaptive or 41% SUV_max, supporting its use as a strong prognosticator in lymphoma. However, for implementation of TMTV in clinical trials 1 single method easily applicable in a multicentric PET review must be selected and kept all along the trial.