To the editor,


We read the paper by Reinert et al. with interest, which was published on January 17, 2020 [1]. The authors evaluated the potential value of CT texture analysis (CTTA) in differential diagnosis of pancreatic ductal adenocarcinomas (PDACs) from pancreatic neuroendocrine tumors (PNETs) in the portal-venous phase. Dr. Reinert and colleagues concluded that 8 of 92 texture features were statistically significant in the differentiation between PDACs and PNETs. This work is valuable as these two entities share similar imaging findings but have different prognosis and require different treatment [2]. Surgical excision remains the primary treatment for any localized PNETs and timely diagnosis confers a high 5-year survival at 75% [3]. However, the median survival of resected PDAC patients after adjuvant therapy ranges from 20.1 to 28.0 months even in optimal clinical trial conditions [4].

The objectives of this paper are valuable, but there a few methodological points worth investigating. The properties of the vascular network such as microvessel density are considered important in tumor development and work as one of the major prognostic factors in PNETs [5, 6]. In Dr. Reinert’s study, PDACs and PNETs were classified into hypo-, iso-, or hyperdense in 50%/50%/0% and 45%/45%/10% of cases, respectively. Subsequently, they compared PDACs and PNETs using CTTA in the portal-venous phase. In our opinion, PDACs are generally iso- or hypodense compared to the adjacent pancreatic tissues. Typical hypervascular PNETs are easily differentiated from PDACs using conventional contrast-enhanced CT imaging features due to their high amount of vascularization. Therefore, 10% of hyperdense PNETs should be excluded when comparing PDACs with PNETs. Moreover, the authors also evaluated the value of CTTA in differentiating G1 from G2/3 tumors. In our opinion, the treatment strategies are not the same for G2/3 tumors. The 5-year survival rates differ for G1, G2, and G3 tumors (75% vs 62% vs 7%) [7]. Therefore, it would be better to assess the role of CTTA in differentiating G1/2 from G3 tumors. In conclusion, Dr. Reinert and colleagues have done interesting and valuable work; however, it would be more valuable to differentiate PDACs from PNETs after excluding hyperdense PNETs and also of G1/2 from G3 PNETs using CTTA in the portal-venous phase.