LAUSR

Dear Sir, Interim FDG-PET/CT imaging performed after the first two cycles of chemotherapy and visually evaluated by Deauville criteria (DS) has proved to be the best prognostic tool in patients with Hodgkin Lymphoma (HL) [1]. Nevertheless, recent studies indicate that DS has some limitations in correctly identifying patients within groups with different prognoses, as false positive and false negative results can be found [1, 2]. So, semi-quantitative parameters, including lesion-tobackground SUV ratios, have been recently proposed to evaluate residual activity in interim FDG-PET/CT [1–4]. In their letter, Laffon and Marthan [5] remarked on some aspects of two different lesion-to-background SUV ratios: qPET by Hasenclever et al. [3] (defined as the ratio between target lesion SUVmean of the four hottest voxels and liver SUVmean in a well defined volume); and rPET by Annunziata et al. [4] (defined as the ratio of target lesion SUVmax to liver SUVmax). Both ratios could have technical advantages over visual analysis with DS, including the intraexam normalization of SUV sources of error and the conversion of a visual qualitative scale in a continuous semiquantitative scale. At the same time, ratios could have clinical advantages over DS related to the use of well-determined cutpoints, subsequent diagnostic standardization and added prognostic power [3–5]. Nevertheless, Laffon and Marthan [5] found an Boriginal sin^ in the use of both ratios in terms of measurement uncertainty (MU), which is calculated as an amplification of single SUV MU and is mathematically higher in ratios with respect to a single SUV measurement. According to their previous works [5], MU of rPET could be evaluated in a dynamic scan as the Brelative difference between a single estimate of a parameter and its average true value^. In our previous study [4], we retrospectively evaluated static PET/CT scans acquired at approximately sixty minutes after FDG injection by collecting a single SUVmax measurement for each target lesion and the liver right lobe. For this reason, MU cannot be evaluated in our study by the method by Laffon andMarthan.We are aware that every kind of measurement is affected by a sort of MU, and ratios have their own MU. On the other hand, the clinical advantages of ratios probably exceed theoretical MU associated to these measurements, from a practical point of view. Concerning this, Laffon and Marthan suggested that metabolic tumor volume (MTV) could improve the target-tobackground approach [5], since it is a Btrue quantitative parameter .̂ However, this is questionable. In fact, MTV is defined as the volume of a tumor with a higher SUV than an established threshold [1, 2]. As well as absolute SUV values, MTVis dependent upon a number of factors, such as the threshold used, the type of volume selection and, again, upon MU. In conclusion, lesion-to-background SUV ratios can improve visual evaluation of interim FDG-PET/CT in patients with HL, in terms of both diagnostic standardization and added prognostic power. We are in accord with Laffon and Marthan [5], that to overcome the possibleMU related to these measurements, further prospective studies with dynamic acquisitions are desirable to collect several SUV measurements at different time points following FDG injection. * Salvatore Annunziata [email protected]