Clinicopathological and molecular determinants of pancreatic neuroendocrine tumors: toward precision risk stratification

Pancreatic neuroendocrine tumors (PanNETs) are clinically heterogeneous, and conventional risk stratification based on tumor size, grade, and stage often fails to capture the full spectrum of biological behavior, especially among small (≤ 2 cm), non-functioning tumors where management remains controversial. Molecular profiling has revealed that the most frequently altered gene, MEN1, carries prognostic information beyond its mutation status. MEN1‑mutant tumors with recurrent loss of heterozygosity across multiple chromosomes follow an aggressive course, whereas those with isolated chromosome 11 loss remain indolent. Besides, the increasing multi-omic profiling has shown that DNA methylation and transcriptomic classification powerfully refine prognosis beyond traditional staging. Different subtypes of MEN1 and DAXX or ATRX co-mutated tumors exhibiting ALT-driven chromosomal instability and a high risk of recurrence even in sub-2 cm low-grade lesions. Metastatic disease the same genotype is associated with remarkably prolonged survival and superior response to PRRT. These studies provide a systems-level view of tumor biology in PanNETs and could aid in classifying distinct subtypes. Integrating these molecular layers with conventional clinicopathological parameters offers a refined framework for risk-adapted management. Hence, this review discusses the latest research findings that build upon the traditional classification protocols and examines the prognosis value of traditional and novel molecular targets. Furthermore, it mainly emphasizes the urgent need to incorporate not just only tumor size, stage and grade but also molecular characteristics as they offer valuable insights into tumor heterogeneity and is recognized as a critical aspect in risk stratification and guiding therapy. Perspectives and actionable suggestions have been addressed for the development of more accurate and reliable classification systems and predictive models in the future. In this context, this review has proposed an estimated risk stratification by integrating clinicopathological-molecular findings and has suggested risk-adapted management, which requires further validation. Overall, integrating molecular features, such as MEN1 and DAXX or ATRX mutation patterns, ALT status, and multi-omic classifier with traditional clinicopathological parameters is essential to resolve the biological heterogeneity of PanNETs and to enable precise, risk-adapted clinical decision-making.