Preoperative assessment of pituitary adenoma consistency using MRI and multiparametric models: a pilot study highlighting the TCTI ratio and perivascular fibrosis.

OBJECTIVES

Intraoperative consistency (IoC) of pituitary adenomas (PAs) impacts surgical outcomes. We evaluated individual parameters and MRI multiparametric models (MPMs) using apparent diffusion coefficient (ADC), tumor to cerebellar peduncle T2-weighted intensity (TCTI) ratio, and contrast-enhanced FIESTA (CE-FIESTA) to assess IoC of PAs. Additionally, collagen content (CC) and perivascular fibrosis (PF) in tumor specimens were evaluated and compared to IoC.

METHODS

We prospectively included 27 patients who underwent preoperative MRI with ADC, TCTI ratio, and CE-FIESTA. IoC was assessed during surgical resection. Each MRI sequence and combinatorial MPMs were evaluated to identify their association with IoC. Subsequently, the degree of CC and the presence of PF in tumor samples were compared to IoC.

RESULTS

Individually, TCTI ratio demonstrated a statistically significant association with IoC (p = 0.049), whereas ADC and CE-FIESTA did not. Among the combinatorial strategies, the comprehensive Full-MPM (TCTI ratio + ADC + CE-FIESTA) demonstrated statistically significant association (p = 0.044). MPM1 showed a trend towards significance, while MPM2 and MPM3 did not. Histologically, high CC (p = 0.045) and the presence of PF (p = 0.03) were significantly associated with firm consistency.

CONCLUSIONS

While multiparametric models show exploratory potential for assessing tumor consistency, their current statistical uncertainty limits definitive application. Consequently, this pilot study suggests that TCTI may represent the most practical tool for anticipating tumor firmness or confirming soft consistency. Furthermore, both high collagen content and perivascular fibrosis provide crucial histopathological anchors for the low TCTI ratios. Ultimately, these findings provide neurosurgeons with accessible imaging parameters to support preoperative planning.