2026 Poster Presentations
P207: HISTOPATHOLOGICAL PREDICTORS OF CANCER IN ACROMEGALY
Arjun R Adapa, MD; Alex Hernandez, BS; Misha Amini, MD; Jeffrey Bruce, MD; Pamela Freda, MD; Columbia University
Background: Acromegaly is caused by growth hormone (GH)-secreting pituitary adenomas. Despite advances in surgical technique, medical therapy, and radiation, the disease is associated with significant morbidity. Of particular significance is the risk of malignancy due to persistent GH and insulin-like growth factor-1 (IGF-1). Although multicenter studies have characterized overall remission rates and prognostic factors, fewer reports describe the relationship between histopathologic features and cancer outcomes within a single institutional cohort, where consistency in surgical approach, medical management, and follow-up protocols provide unique insights.
Methods: A retrospective analysis of prospectively collected data of 598 acromegaly patients who underwent transsphenoidal surgery between 1981 and 2019 was conducted. Acromegaly was diagnosed by elevated IGF-1 level or abnormal GH suppression after oral glucose. All patients had pathological confirmation of a somatotroph. From this cohort, 133 patients with complete histopathological data were analyzed. Each tumor specimen underwent immunohistochemistry (IHC), including assessments of granulation patterns by CAM5.2 (densely granulated [DG], sparsely granulated [SG], and mixed) and MIB-1 labeling indices. Detailed health and medication history, physical examination, and GH/IGF-1 levels were collected at the initial and each follow-up visit. Pairwise comparisons and logistic regression for predictors of any cancer diagnosis were performed using SPSS v30.0.0.0.
Results: The mean age at diagnosis, follow-up time, adenoma size, and rate of GTR of our population was 46, 112 months, 16.8±9 mm, and 73%, respectively. Mean %IGF-1 age-adjusted upper limit of normal (ULN) at diagnosis and last follow up were 243.6 +/-17.6 and 100.5 +/-12.2, respectively, reflecting an average decrease in %IGF-1 ULN of 142.9 between diagnosis and last follow-up. There were 77, 47, and 10, DG, SG, and mixed patients, respectively. Adenoma size was not dependent on granulation (p=0.453). Granulation was not correlated with %IGF-1 ULN or GH levels. SG adenomas were inversely correlated with MIB-1 (R2=0.04, p=0.03). Notably, patients with DG adenomas had a 7-fold higher odds of cancer relative to SG (95% CI 1.5-31.9, p=0.012) on univariate analysis. Notably, smoking status, alcohol use, number of follow-up visits, GH level at diagnosis, %IGF-1 ULN at diagnosis or last follow-up, PRL positivity, MIB-1 index, GTR, diabetes, hypertension, and obesity were not independent predictors of a cancer diagnosis. On multivariate analysis, DG adenomas (OR 20.0, 95% CI 1.7-234.8, p=0.017), duration of follow-up (OR 1.01, 95% CI 1.0-1.024, p=0.022), and age at diagnosis (OR 1.11, 95% CI 1.03-1.21, p=0.009) were predictive of cancer, with DG adenomas conferring the strongest association. MIB-1 index was a positive predictor of the need for immediate post-operative somatostatin analog therapy (OR 1.320, 95% CI 1.04-1.67, p=0.022).
Conclusion: Histopathologic features of GH-secreting pituitary adenomas, particularly granulation patterns, may be related to cancer risk independent of GH or IGF-1 levels. While literature supports an association between SG and lower serum GH/IGF-1 levels, granulation and %IGF-1 ULN were not correlated in this study. Larger studies are needed to validate these predictors and their relationships to known cancer risk factors. Nevertheless, integrating pathology with clinical data may improve risk stratification, guide surveillance, and advance understanding of tumor biology.