The primary difficulty particularly with these non-epithelial neuroendocrine tumors is that there are no established standards for determining their malignancy, and patients must undergo long-term monitoring and face uncertainty since metastases can develop even many years after the initial tumor [
10,
31,
32]. The primary objective during histological examination following surgery is to accurately diagnose a malignant tumor that has not yet spread to other parts of the body. Numerous studies have been conducted, and efforts are being made to develop an algorithmic model based on known factors that indicate malignancy. In many studies, the PASS score is a reliable prognostic factor, even though the evaluation of certain histological parameters such as hyperchromasia, nuclear pleomorphism, and cellularity are subjective. Tumors with a PASS score less than 4 did not develop metastases, indicating that only patients with a PASS score of 4 or higher would require follow-up, including annual radiological imaging and biochemical measurements of metanephrines [
18]. In recent assessments of the Pheochromocytoma of the Adrenal Gland Scaled Score (PASS), as well as other scoring systems that have been introduced since its development, the World Health Organization (WHO) Classification of Endocrine Tumors recognizes that while these systems have their advantages and disadvantages, their use should not be discouraged in further research, but instead should be optimized in conjunction with molecular analyses [
1]. Considering this, we employed the PASS score as a suitable standard for our diagnostic algorithm. In our study, we categorized our cases based on the value of the PASS score and analyzed additional prognostic factors. In numerous studies, angiogenesis has been recognized as a critical factor for the proliferation of tumors and the development of metastases [
19]. Given the highly vascular nature of paragangliomas as a solid tumor, there has been interest in assessing the significance of angiogenesis in its pathogenesis, and investigations were prompted. Gao X. et al investigated CD31 as an angiogenic marker, by counting the number of CD31 positive vessels within the highest expressed tumor area (MVD), and the number was significantly higher in pheochromocytomas with PASS<4 than in those with PASS≥4. Gao X. et al. also detected that intratumoral hemorrhage was significantly higher in pheochromocytomas with PASS≥4, than PASS<4, which could also result in relatively low CD31 status in histologically low-grade tumors [
16]. H. Ohji et al counted CD34 labeled blood vessels (MVD) and found no statistical association between the MVD and malignancy, while Q. Liu et al, which assessed MVD by staining endothelial cells with antibody to factor VIII/von Willebrand factor antigen, found a highly significant difference between the groups of benign and malignant [
20,
21]. In the former study the cases were divided into benign and malignant according to the presence of metastatic disease and in the latter group, besides the presence of metastasis, tumors with evidence of capsular or vascular invasion were included, which are histological criteria included in PASS scoring scale. Q. Lui also noticed that areas of highest vascularization were usually along the periphery of the tumor [
20]. M. Bialas et al studied angiogenesis status in pheochromocytomas, including MVD and vascular architecture, after immunostaining endothelial cells with antibodies CD31 and CD105 [
33]. Vascular architecture patterns were highly heterogeneous within pheochromocytomas, both benign and malignant. The MVD in the central areas of the tumor was higher than in the subcapsular areas. Secondary changes in tumors, like hemorrhage and cystic degeneration, influenced counting MVD and assessing vascular architecture [
33]. Favier et al used immunostaining for CD34 and α-actin smooth muscle cells to define vascular architecture, and they reported differences, noticing two different patterns, mainly benign pheochromocytomas having more regular patterns, with short and straight vascular segments, while malignant presented with an irregular pattern, longer vascular segments, of irregular length [
34]. The density of vascular segments was lower in irregular patterns, but blood vessel counting was not considered appropriate for these tumors [
34]. L. Oudiijk et al found that the mean sensitivity and specificity of vascular architecture as a predictor of potential malignant behavior of paragangliomas was 59,7% and 72,9%, with a significant agreement between the observers [
35]. To assess tumor angiogenesis and to minimize the potential influence of marker choice on our results we chose to evaluate intratumoral MVD using several IHC markers at the same time. By making TMA we tried to select areas of the tumor with a high density of blood vessels without hemorrhages and cystic degeneration. While most studies, as was mentioned, have utilized CD31 and CD34, which are pan-endothelial cell markers that react with both proliferating and pre-existing vessels in tumors, CD105 (endoglin) is an endothelial antibody that primarily binds to activated endothelial cells in angiogenesis. Thus, it may be a more specific marker for tumor angiogenesis, as indicated by prior research [
36,
37,
38,
39]. In addition to the established antibodies used to measure intratumoral MVD, we examined ERG, a transcription factor belonging to the ETS family that is expressed by endothelial cells involved in angiogenesis. This approach has not been previously explored, to the best of our knowledge, on paragangliomas [
40]. All four antibodies used for intratumoral MVD indicated that these were well-vascularized tumors. There was no statistically significant difference in intratumoral MVD with regard to PASS potentially malignant behavior. However, all four endothelial antibodies showed lower intratumoral MVD in the group of malignant tumors, suggesting that cells may be able to overcome hypoxia. Hypoxia has been mentioned as essential for angiogenesis in paraganliomas and due to the pseudohypoxia signaling pathway, they are well known for increased angiogenesis [
41,
42]. Although neoplastic angiogenesis allows tumor spread, it does not imply that these well-vascularized tumors will metastasize. Additional knowledge is needed to understand the crucial features that drive metastasis of paraganglioma and the role of angiogenesis in this process. In light of the high replicative potential observed in many malignant tumors, several studies have focused on the proliferative Ki67 index [
43,
44,
45,
46]. However, it has been found that this index has low sensitivity when it comes to intra-adrenal and extra-adrenal paraganglioma despite its high specificity. Paragangliomas are slow-growing tumors, and the most show very low Ki-67 labeling [
18]. Proliferative index Ki67 have high specificity which implies poor sensitivity, with only 50% of malignant tumors have a score greater than 2-3%, so the cut-off value for malignancy is low [
18,
47]. Our study confirmed the low sensitivity of Ki67 in this context. Nevertheless, Ki67 is an immunohistochemical marker that should be measured, as high Ki67 values may warrant attention for other prognostic factors, such as genetic predisposition and careful consideration of significant histologic predictors of malignancy, such as tumor necrosis [
18]. Paragangliomas consist of chief cells that are visible on H&E staining, as well as sustentacular cells, that often necessitate a specific immunohistochemical antibody. While their precise role in metastasis development is not yet clear, several studies, including our own, have demonstrated a decrease in the number of sustentacular cells in malignant tumors [
43,
48,
49,
50]. In some cases in our study these cells were entirely absent. These tumors have the strongest genetic contribution with nearly 40% of them associated with mutations in one of the known susceptibility genes, including NF1, VHL, RET, EPAS1, EGLN1, SDHA, SDHB, SDHC, SDHD, SDHAF2, FH, TMEM127, MAX, MDH2, GOT2, SLC25A11, DLST, H3F3A, DNMT3A, MET, MERTK, and KIF1B. There are different types of mutations including germline only, germline and somatic, somatic and somatic mosaicism [
3,
4,
5,
51]. Germline mutations involving genes coding for succinate dehydrogenases (SDHx) are the most common genetic cause of paraganglioma, occurring in up to 25% of cases [
52,
53]. They are followed by genes VHL (4-10%), RET (1-5%), and NF1 (1-5%) [
54]. Testing for the germline SDHB mutation in patients with paraganglioma is recommended by Clinical Practice Guidelines [
6,
55]. Immunohistochemically SDHB negative staining in tumors carries a high risk for the presence of SDHx mutations, whether in SDHB, SDHC, SDHD or SDHA [
14]. Loss of SDHB immunoreactivity in tumor cells with SDHx mutations is reported with 100% sensitivity and 84% specificity, with a positive predictive value of 92% and a negative predictive value of 100% [56]. In our study the majority of cases retained immunopositivity to SDHB, and only one had a confirmed mutation. Since genetic testing was not available in all cases, and unfortunately still is not routinely done, immunohistochemistry helped to potentially rule out an SDHx mutation. It was found that the malignant paraganliomas were larger than the benign ones with a median tumor size of 61mm and 44mm, respectively. However, the precise cutoff value that could precisely predict malignant behavior is not established. Since they are all potentially malignant, a staging system is introduced, with size 5cm as cut off value for T1 and T2 [
22]. Tumor size and weight are routinely reported, which makes them available for risk stratification, but they are not considered an independent parameter. Invasive features, like an invasion of periadrenal fatty tissue despite being more common in malignant tumors, were not significantly correlated regarding potential malignant behavior in some studies, but it is a criterion for T3 [
15,
18,
22]. The prognosis of malignant paragangliomas remains poor with overall 5-year survival less than 50% [57]. Management of malignant tumor includes surgery, radiation therapy or chemotherapy and several targeted therapies have been investigated, but currently, there are no curative therapy options [58].