In the present study, we observed that age, gender, T-stage, N-stage and M-stage were valuable clinical characters that could be used to distinguish pNENs from SPTs. However, in order to obtaine the information of T-stage, N-stage and M-stage, the tumor tissues should be isolated from the pancrease by invasonal technologies, such as ultrasound-guided fine-needle biopsy (EUS-FNB) or surgical resection. We, thererfore, excluded these variabes from the present study and developed a non-invasive nomogram by age and gender. As indicated in Fig. 3A, a female (68 points) and under the age of 33 years old (100 points) will have a score of 168 points, which predicts the probility of SPT is 70%. This nomogram might have some clinical implications. For example, it might help the clinicians to accurately distinct pNENs from SPTs and to determine an appropriate diagnostic or treatment strategy for patients.
Although pNENs and SPTs are rare tumors, some studies suggested that the incidence of these tumors significantly increased in the last decade. Thus, pNENs and SPTs received attention in the publications [8]. As same as pNENs, SPTs patients have non-specific clinical manifestations, for example, abdominal pain, abdominal discomfort, and weight loss [14]. Usually, the patients are hospitalized due to abdominal masses or accidentally found a tumor in the pancreas. As presented in Fig. 1B, the survival of SPTs patients is significantly superior to pNENs. These observations are supported by other studies [15, 16]. In addition, compared to pNENs, the SPTs have relatively low malignant biological behavior. Thus, aggressive surgical resection might give rise to survival benefits in SPTs patients, even when patients with distant metastasis. Indeed, Wang et al. reported that surgical resections of the primary and metastatic lesions, as completely as possible, could give improve the prognosis of SPTs patients. However, surgical resection is contraindicated in patients with metastatic pNENs [17]. Therefore, an accurate preoperative diagnosis of SPTs and pNENs will help the clinicians to make optimal decisions and chose the appropriate treatments for SPTs and pNENs patients, respectively.As mentioned above, SPTs and pNENs have the same clinical symptoms and signs. In addition, the previous study reported that SPTs exhibit neuroendocrine differentiation and in these tumor tissues the author also observed chromogranin A, CA19-9, and vimentin which are used to diagnose pNENs [18]. This suggests that these tumor markers could not distinguish SPTs from pNENs. Notably, Li et al. evaluated the clinical and immunohistochemical characteristics of 37 SPTs, and they observed intracytoplasmic dot-like immunoreactivity of CD99 in these tumors [19]. This is in contrast to pNENs tumors in which the CD99 was observed in the membrane [19]. In addition, the authors found a loss of E-cadherin and aberrant nuclear expression of β-catenin in SPTs. Thus, the expression of CD99 in combination with E-cadherin and β-catenin might be valuable combinational tumor markers for the diagnosis of SPTs and pNENs.
Because sometimes both SPTs and pNENs have cystic degeneration and calcification, it is difficult to distinguish between SPTs and pNENs by computed tomography (CT). It is reported that MRI is a valuable strategy for the diagnosis of pancreatic tumors. Compared to CT, MRI could appropriately exhibit the soft-tissue characteristics, theretofore MRI could be used to evaluate the functional and metabolic of tumors. Notably, both SPTs and cystic endocrine tumors have the same features, such as areas of cystic change, enhancing components, and well-defined contours [20]. Although positron emission tomography/computed tomography (PET/CT) is widely used in the diagnosis of malignant tumors and pancreatic disease, there are very few studies that evaluated the accuracy of PET/CT in the diagnosis of SPTs [21,22,23]. François et al. reported that 18F-FDOPA PET/CT was a promising approach for the diagnosis of pNENs and SPTs. The 18F-FDOPA PET-positive/SRS-negative lesions might be the SPTs. However, these should be verified in a large cohort [21]. EUS-FNB might be another promising diagnostic tool for distincitng SPTs from PNENs [24, 25]. Recently, some studies reported that the diagnostic accuracy of EUS-FNB in combination with the immunohistochemical staining of β-catenin or cadherin in solid pancreatic lesions is high than 90% [26,27,28]. Therefore, EUS-FNB should be recommmeded as the standard of care for differential diagnosis of pancreatic lesions.
Notably, there are some limitations to our study. This is a retrospective study and the confounding bias might distort the association between the variables and SPTs. In addition, SPTs and pNENs are rare tumors. To collect as many patients as possible, we used the data in the Surveillance, Epidemiology, and End Results database which encompasses approximately 28% of the USA population. However, after a strict selection process, only 30 SPTs patients were included in the present study. In addition, based on the pathology and Ki-67 index, the grade of pNENs is classified into four groups: pNET G1 (Ki-67<2%), pNET G2 (3%<Ki-67<20%), pNET G3 (Ki-67>20%) and pNEC [4]. While this WHO classification and the Ki-67 index are not recorded in the SEER database, and the tumore is graded according to morphological description, for example, well differentiated, moderately differentiated and poorly differentiated. Moreover, 40.7% cases lost the information of the tumor grade. These limitations of SEER database might also cause bias when interpreting the results.