Synergistic prognostic value of transcription factor multiplicity and Ki-67 in pituitary neuroendocrine tumour recurrence after endoscopic transsphenoidal surgery: a novel tiered molecular risk model.

BACKGROUND

WHO 2022 classification underscores lineage-defining transcription factors (TFs) as key determinants of pituitary neuroendocrine tumour (PitNET) biology. We hypothesise that interaction between lineage "driver" (TFs) and proliferative "Движок" (Ki-67) provides superior prognostic power for predicting recurrence compared to either factor in isolation or to traditional clinicopathological parameters.

MATERIALS AND

METHODS

This retrospective cohort study analysed 302 patients who underwent primary Endoscopic Endonasal Transsphenoidal Surgery for PitNETs. Five prognostic models were constructed to simulate diverse pathology laboratory capabilities for global utility: specific TF/Ki-67 combinations (Model A), specific TFs only (B), unspecific TF multiplicity/Ki-67 (C), unspecific TF multiplicity only (D), and Ki-67 alone (E). Multivariate Cox regression, adjusted for key clinical confounders, assessed associations with recurrence.

RESULTS

Eighty-one patients (26.82%) experienced recurrence, median follow-up: 30 months (range: 2-69). Model A & C demonstrated superior prognostic performance, identifying multiple high-risk profiles (adjusted HR [adj.HR] range: 5.36-17.70, all P<0.05). The practical Model C performed nearly equivalently to the granular Model A. Their findings were synthesised into simplified 4-tiered molecular risk stratification. Using Tier I as reference, adj.HR for recurrence escalated significantly across tiers: Model A, Tier II (adj.HR=3.65, P = 0.081), Tier III (adj.HR=6.65, P = 0.009), Tier IV (adj.HR=15.44, P = 0.001); and Model C, Tier II (adj.HR=3.47, P = 0.097), Tier III (adj.HR=6.50, P = 0.010), Tier IV (adj.HR=17.78, P = 0.002). In models lacking Ki-67, only SF-1(+) & T-PIT(+) (Model B, adj.HR=2.75, P = 0.037) and triple-TF positivity (Models B & D, adj.HR~3.80, P = 0.012) retained significance. Model E, Ki-67 ≥3% (adj.HR=1.76, P = 0.015) remained significant.

CONCLUSION

This study validates critical prognostic synergy between TF multiplicity and elevated Ki-67 in PitNET recurrence. We propose a novel tiered molecular risk model anchored in this interaction. Whilst this synergy is best captured by the comprehensive Model A, the full spectrum of models (A-E) provides a flexible diagnostic framework. This ensures risk stratification is accessible to pathology services with varying technical capabilities, to guide personalised post-operative management and surveillance.