Spatial localization patterns in MRI-negative Cushing disease: a comparative pituitary heat map analysis.

OBJECTIVE

MRI-negative Cushing disease is challenging to treat due to difficulty with tumor localization and lower biochemical remission rates. Surgeons primarily rely on visual and tactile feedback to identify abnormal pituitary tissue. In this study, the authors analyzed intraoperative sketches of histology-confirmed corticotroph adenomas to compare the spatial distributions of MRI-positive and MRI-negative tumors. By characterizing these distribution patterns, the authors aimed to improve the localization and surgical targeting of MRI-negative pathology.

METHODS

The authors retrospectively reviewed patients with Cushing disease who underwent microscopic transsphenoidal resection by the late Dr. Edward H. Oldfield at the authors' institution from December 2007 to March 2017. Patients were stratified into MRI-positive and MRI-negative cohorts. A standardized pituitary sketch visualizing both coronal and sagittal planes was used for postoperative documentation. For every patient, Dr. Oldfield annotated the anatomical location of the histologically confirmed corticotroph adenoma. Image acquisition, processing, and heat map-based spatial analysis were performed to assess cumulative tumor density. Spatial differences were evaluated with cluster-based correction for multiple comparisons. Surgical outcomes were then compared between the 2 cohorts.

RESULTS

A total of 153 patients with Cushing disease and histologically confirmed pituitary corticotroph adenoma (133 MRI-positive and 20 MRI-negative adenomas) were analyzed. Baseline demographic and biochemical characteristics were similar across MRI status. Both cohorts demonstrated similar anatomical presentation, with over 50% of patients presenting with corticotroph adenomas in the anterior aspect of the pituitary gland. MRI-positive tumors were more likely to localize to the lateral and dorsal aspects of the anterior pituitary (p < 0.05). In contrast, MRI-negative adenomas showed no distinct area of increased prevalence. Postoperatively, the MRI-negative cohort had higher nadir cortisol levels (2.9 vs 1.5 μg/dl, p = 0.001) and increased reoperation rates (20% vs 3.8%, p = 0.027). Remission rates were lower in MRI-negative patients but not statistically significant (75.0% vs 91.0%, p = 0.050).

CONCLUSIONS

MRI-negative corticotroph adenomas were less frequently localized in the lateral and dorsal aspects of the anterior pituitary compared to MRI-positive tumors, with no distinct region of increased prevalence. These findings support a unique surgical strategy that leverages the anatomical distribution of histologically confirmed adenomas to guide gland exploration and resection of hidden tumors. While these observations may provide anatomical context for surgical exploration in radiographically occult disease, operative decision-making should remain guided by intraoperative findings and individual patient anatomy.