Predictive modeling for pituitary adenomas: single center experience in 501 consecutive patients

Area of Science:

• Neurosurgery
• Endocrinology
• Oncology

Background:

• Personalized postoperative management of pituitary adenomas necessitates early risk stratification.
• Developing effective risk stratification systems is crucial for optimizing patient care.

Purpose of the Study:

• To develop and validate predictive models for early risk stratification in pituitary adenoma patients.
• To identify key parameters predicting long-term outcomes and recurrence.

Main Methods:

• Reviewed 501 pituitary adenoma cases operated by a single neurosurgeon.
• Utilized Cox proportional hazards models incorporating age, gender, tumor diameter, cavernous sinus invasion (CSI), diagnostic category, and proliferation markers (ki-67).
• Developed predictive models and employed Kaplan-Meier survival curves for time-dependent analyses.

Main Results:

• Tumor diameter, CSI, and ki-67 index were significant predictors of long-term events.
• Model 1 (CSI, diameter ≥2.9 cm, ki-67 >3%) showed high specificity (98.7%) in predicting persistent hypersecretory syndrome and residual tumor.
• Model 2 (ki-67 >3%, CSI) identified smaller at-risk adenomas with 93.1% specificity.
• Silent corticotroph adenomas exhibited lower event-free survival compared to ACTH-negative nonfunctional adenomas.

Conclusions:

• Predictive models incorporating CSI, ki-67, and tumor diameter facilitate tailored surveillance and management of pituitary adenoma patients.
• These models improve the ability to stratify risk and personalize postoperative care.