SAT-105 Collagen-1 content as prognostic marker for recurrence in clinically non-functioning pituitary neuroendocrine tumours (NF-PitNETs)

Abstract Disclosure: A. Rai: None. A.L. Carreira: None. T. Rice: None. J. Williams: None. K. Magid: None. A. Mikalauskaite: None. H.J. Marcus: None. N. Dorward: None. J. Grieve: None. D. Khan: None. A. Kolias: None. C. Gaston-Massuett: None. B.D. Radotra: None. R. Chhabra: None. P. Dutta: None. M. Korbonits: Novo Nordisk. F. Begalli: Novo Nordisk. Background: Recurrence (10-50%) is common in clinically non-functioning pituitary neuroendocrine tumours (NF-PitNETs). However, we lack predictive biomarkers to help the management of these tumours. Collagen fibres are important structural elements of the extracellular matrix (ECM) which is a main component of the tumour microenvironment. While an increase in collagen fibres has been associated with recurrence and aggressiveness in other cancers, their role in NF-PitNETs and its relationship with recurrence is unknown. Method: We performed picrosirius red staining on a total of 73 first surgery NF-PitNETs divided in two cohorts, 49 as full slides and 24 tissue microarray, these 2 sets showed comparable results; 6 normal pituitaries (NP; full slides) were used as controls. Recurrence was defined as second intervention after 1 year from first surgery. Collagens I and III were assessed using polarisation microscope, at least 4 fields of view were analysed from each sample. Percentage of collagen-stained area was calculated for each high-power field. Collagen fibres characteristics were assessed using TWOMBLI pipeline: thickness and linearity (number of collagen fibre end points, number of fibre branch points, total length of fibres), the structural complexity (fibre curvature, alignment, proportion of high-density matrix, ECM fractal dimension), fragmentation (hyphal growth unit), and compactness (matrix gaps, and lacunarity). ROC analysis for diagnostic test evaluation was performed in Graphad Prism. Results and Discussion: We observed significantly higher percentages of collagen-stained area in recurrent NF-PitNETs compared to non-recurrent tumours and NP (p<0.0001 for both), showing that recurrent tumours present more fibrotic areas. Detailed analysis revealed that collagen I was significantly increased, while collagen III was significantly decreased in recurrent compared to non-recurrent tumours and NP (p<0.0001 for both). Fibres showed increased number, thickness, length, structural complexity, and compactness in recurrent tumours compared to non-recurrent tumours and NP (p<0.0001 for both), while there was a significant decrease in fragmentation in recurrent tumours (p<0.0001), highlighting densely packed fibres in recurrent NF-PitNETs. Altogether, these results suggest that recurrent NF-PitNETs are characterised by larger areas and higher density collagen matrix, both associated with aggressiveness. Finally, ROC analysis confirmed that the percentage of collagen area is a good prognostic biomarker for recurrence in NF-PitNETs (AUC=0.88, p=0.006). Conclusions: These data highlight the importance of ECM in modulating tumour behaviour; indeed, collagen provide stiffness rendering tumours more difficult to remove surgically. Importantly, collagen staining provides a good and easily applicable prognostic marker for recurrence in NF-PitNETs. Presentation: Saturday, July 12, 2025