Endocrine Abstracts

Objective: To provide a latent class regression model predictive of surgical outcome in patients with non-functioning macroadenomas (NFMA) of the pituitary gland.

Patients and methods: Seventy-five patients with surgically treated NFMA of the pituitary gland were analyzed retrospectively. Thirty-two patients were male. Median age at the time of surgery was 58 years (range: 16–84 years). Median follow-up time was 90 months (range: 14–208 months). Gender, age, tumor size, parasellar tumor growth, surgical approach, tumor remnant, tumor recurrence, number of surgical procedures, adjuvant radiotherapy, postoperative course of vision as assessed by an ophthalmologist, and postoperative functioning of the adenohypophysis, each as documented in our patient files, were chosen as potentially predictive covariates. Affection of nasal airways, impairment of vision, neccessity of hormone substitution, dependency on nursing, return to work, quality of life and overall satisfaction with the surgical result, each as recorded in a standardized patient questionnaire at follow-up, were chosen as outcome variables. Multiple latent class models (LCM) were fitted. Covariates with a probability of <0.2 to exceed the absolute t value in the respective LCM were considered predictive of class membership. The lowest Bayesian information criterion (BIC) of all LCM was considered indicative of best fit.

Results: The most parsimonious LCM consisted of two classes ‘satisfactory outcome’ (estimated patient share (EPS): 60.5%; proportion of overall satisfied patients (POSP) in class: 100.0%), and ‘potentially unsatisfactory outcome’ (EPS: 39.5% and POSP: 73.8%), with age, tumor size, parasellar tumor growth, surgical approach, postoperative course of vision, and postoperative functioning of the adenohypophysis as predictive covariates.

Conclusion: The presented model allows to identify individuals at risk for an unsatisfactory outcome after surgery for NFMA of the pituitary gland.