Predictive factors of malignancy in adrenal masses classified as indeterminate according to radiological criteria

Background: Adrenal masses are a very common finding in clinical practice, however some masses, before histopathological analysis, remain classified as indeterminate. These masses cannot be clearly classified as benign nor malignant. Objectives: The aim of our study was the assessment and management of indeterminate masses identified prior to histological confirmation, we focused on the identification of potential predictive factors for malignancy, that could significantly enhance diagnostic accuracy and therapeutic choices in future cases. Methods: Retrospective observational study performed on 191 adrenal lesions which were surgically removed at the “A.O.U. San Luigi Gonzaga” hospital in Turin, between the 2005 and 2023. Masses were classified, prior to histopathological analysis into lesions having a high probability of being benign, as well as those categorized as indeterminate, allowing for a comparative assessment between the two groups. Results: A comprehensive statistical analysis was performed to identify predictive factors for malignancy in adrenal masses. In the univariate analysis, DHEAS above the normal levels were found to be associated with malignancy (OR: 17.7, 95% CI: 4.23–93.4, p-value < 0.001), however this factor was not statistically significant when we performed the multivariable logistic regression model. Mass size and malignancy risk demonstrated a significant association both in the univariate analysis (OR: 1.03 per mm increase, 95% CI: 1.02-1.05, p-value < 0.001) and in the multivariable logistic regression model (OR: 1.27, 95% CI: 1.14–1.51, p = 0.001). A ROC curve analysis identified a mass size threshold of 36.5 mm, with a specificity of 100% and sensitivity of 68.4%, indicating that no benign lesions exceeded this cut-off, though 31.6% of malignant masses were below this threshold. The multivariable logistic regression model also confirmed the association of HU at non-enhanced CT (OR: 1.47, 95% CI: 1.28–1.84, p < 0.001), and HU >10 and/or lesion heterogeneity (OR: 363.44, 95% CI: 68.59–3757.09, p < 0.001) were independent predictors of malignancy. Age at surgery also emerged as a relevant factor (OR: 1.11 per year increase, 95% CI: 1.03–1.20, p = 0.008). The model demonstrated excellent discriminatory power (R² Tjur = 0.837), indicating strong predictive reliability. Conclusions: These findings emphasize the critical role of imaging characteristics and biochemical markers in assessing malignancy risk in adrenal masses, potentially guiding clinical decision-making and patient management strategies in patients that present with masses characterized as indeterminate prior to histological analysis. However, challenges remain in the management of truly indeterminate masses that neither clearly meet benign nor malignant criteria even following the results of the present study. In such cases, shared decision-making involving endocrinologists, radiologists, and surgeons remains crucial. Our study suggests that a multidisciplinary team approach optimizes patient outcomes while avoiding overtreatment.