Diffusion-weighted MR imaging for characterizing mediastinal lymph nodes in children

partially justify the different results of the study by Razek et al. compared with previous similar works, in addition to other causes listed by the authors (i.e., different methods of ADC measurement; histological findings; patient age) [1]. Second, selecting the region of interest (ROI) within tissue is essential for obtaining reliable quantitative measurements. Placement of a single ROI that does not encompass the entire lesion—as the authors did—is generally used for chemical-shift MRI of the anterior mediastinum to detect microscopic fat in tissues by selecting axial slices in which the soft tissue to be measured exhibits the highest signal intensity on opposed-phase imaging, so as not to miss solid lesions within normal or hyperplastic thymus and with clear fat content on opposed-phase imaging [3]. Conversely, for DW-MRI in parametric ADC maps, the freehand ROI is generally obtained on more sections by including the entire volume of the solid portion of the lesion, and the ADC is the mean or median of the ADC values obtained [4]. The latter method should better reflect the real ADC value of tissue rather than the former, which was used by the authors, as stated in their “Methods” section and shown on ADC maps in both figures [1]. Indeed, tumors are often heterogeneous in their spatial ADC distribution, an aspect that cannot be adequately detected from a single ROI placed in a limited portion of the tissue. Because the heterogeneity of tissue is well evident at visual analysis in the ADC maps shown by the authors (a case of non-Hodgkin’s lymphoma and mediastinal sarcoidosis), it seems from illustrations that positioning the ROI in a different portion of tissue on the ADC map would give a different ADC value. In addition, when considering a quantitative parameter such as a cancer biomarker in the clinical setting, intraand interreader variability of a manual measurement method is essential to determine the limits of error when obtaining quantitative data. The use of a consensus analysis, without reporting Dear Editor, In a recent article published in the Japanese Journal of Radiology by Razek and colleagues, diffusion-weighted magnetic resonance imaging (DW-MRI) was found to be a reliable and accurate imaging modality, through quantitative assessment, for distinguishing benign from malignant mediastinal lymph nodes in children [1]. Indeed, in a cohort of 29 children, the authors found a significantly lower apparent diffusion coefficient (ADC) in malignant lymphadenopathies compared with benign lymph nodes (mean ADC of 0.99 and 1.35 × 10−3 mm/s, respectively; optimal cutpoint for group discrimination of 1.22 × 10−3 mm/s). However, although such quantitative information could improve characterization of mediastinal lymph nodes in children, some considerations should be pointed out. First, although an echo time as short as possible is recommended for DW-MRI (usually 50–65 ms for the chest), the longer echo time of 108 ms used by the authors could have led to an underestimation of ADC values, which is higher for pathological tissue with greater diffusion restriction (such as malignant lymph nodes) compared with normal tissue. In fact, in an experimental model, Schmidt and colleagues recently found an underestimation of ADC values up to 30 % by using an echo time of 92 ms compared with 52 ms, and the same set of b values for compartments with T2 relaxation time of 400 ms [2]. This finding may