CT perfusion in characterizing anterior mediastinal solid tumors

Dear Editor,

In the January–February 2017 issue of Diagnostic and Interventional Radiology, Bakan et al. (1) reported on the usefulness of quantitative parameters derived from computed tomography (CT) perfusion in characterizing anterior mediastinal solid tumors. Notably, on a cohort of 25 patients, they found significantly lower blood volume and blood flow in “nonthymoma malignancies” compared with thymomas (1). Their results add valuable information to optimize diagnosis and clinical decision-making. However, some considerations should be pointed out.

First, in the introduction, the authors reported that thymectomy is not therapeutic for thymic hyperplasia. This information is inaccurate because, although myasthenia gravis is an exclusion criterion, Wolfe et al. (2) recently proved the benefit of thymectomy in patients with nonthymomatous myasthenia gravis, including lymphoid thymic hyperplasia, for improving clinical outcomes and reducing the need for immunosuppressive therapy.

Second, the cohort included eight cases of thymic hyperplasia with diagnosis obtained through histology in four cases and chemical-shift magnetic resonance imaging (MRI) in the remaining cases. However, the authors should provide the type of hyperplasia (lymphoid or rebound) at histology and should declare clinical condition inducing rebound hyperplasia if histology demonstrated normal thymic tissue (3). Similarly, they should provide clinical information on the remaining four patients of this group to suppose lymphoid or rebound type. Moreover, it should be stressed that an alternative diagnosis of normal thymus cannot be excluded in the latter cases because normal thymus with soft tissue attenuation at CT is also seen in adulthood. This differentiation may be important since, based on the different histologic composition of lymphoid hyperplasia compared with normal thymus, we found different apparent diffusion coefficient values at diffusion-weighted imaging between these entities in patients with myasthenia gravis (4). Hence, similar differences cannot be excluded by using quantitative data from CT perfusion.

Third, the authors did not provide WHO and Masaoka-Koga classifications of thymomas with relative quantitative metrics at CT perfusion. Despite the small thymoma group, this information would help reader in the interpretation of results because the wide range of histologic types and different clinical behavior of thymomas could be the basis for different values at CT perfusion. Moreover, the term “benign” used by Bakan et al. for thymic epithelial tumors should be avoided because of unfavorable prognosis in advanced thymoma, with 10-year overall survival rates from 84% in stage I to 42% in stage IVA.

Lastly, among limitations of the study, the authors declared a relatively high standard deviation in thymoma group because of different stages of thymoma in their cohort. However, standard deviation of groups or diseases was not reported in Table or in Results and, anyhow, it should not be used if sample data are not normally distributed. Indeed, the non-normal distribution of data in the study can be supposed by the use of median values and nonparametric Mann-Whitney U test to compare medians between groups. Rather than standard deviation, the authors should provide, together with the absolute range in the Table, a percentiles range (e.g., the interquartile range) to accurately describe distribution and spread of data.