Endocrine Abstracts

Background

Advances in imaging techniques combined with the increase in the number of patients undergoing radiological investigations have contributed to the increased detection of adrenal nodules. Prior to the publication of the European Society of Endocrinology (ESE) guidelines on the management of adrenal incidentalomas in 2016, there was marked variability in the management of adrenal nodules. The objective of this study is to validate the diagnostic performance of the radiological thresholds (size of nodule, pre-contrast attenuation and washout studies), as per the ESE guidelines 2016.

Methods

The histological diagnosis (gold standard) for 68 patients who had adrenolectomy between January 2011 – February 2020 was obtained from hospital records. CT reports of these patients were reviewed to collate data on size (n = 68), pre-contrast attenuation (n = 47) and absolute washout (AW) (n = 31) and relative washout (RW) (n = 31). Univariate models were used to analyse the diagnostic accuracy of these radiological thresholds. ROC analyses were conducted and standard thresholds were used to derive sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV). We propose new thresholds providing better estimation of these parameters.

Results

The malignancy prevalence in the dataset was 10.4%. Univariate analysis using the chi-squared test, demonstrated that size (P < 0.0001) and pre-contrast attenuation (P = 0.01) were significant predictors of benign histopathology. Receiver operating characteristic (ROC) area under the curve (AUC) was 0.89 (P = 0.005) for size and 0.88 for pre-contrast attenuation (P = 0.03). The AUC for RW (P = 0.08) and AW (P = 0.24) were not significant.

With the size cut off ≥4 cm (current standard) for diagnosing a malignant lesion, the NPV and PPV values were 0.96 and 0.17 respectively (sensitivity 0.80, specifity 0.53). A new threshold of ≥8cm for diagnosing a malignant lesion gave a NPV 0.98 and PPV 0.67 (sensitivity 0.80, specifity 0.95). Similarly, a pre-contrast cut off of >10HU (current standard) for a malignant lesion diagnosis had NPV 1.00 and PPV 0.18 (sensitivity 1.0, specifity 0.59). A new cut off 29.5HU had NPV 0.97 and PPV 0.40 (sensitivity 0.67, specifity 0.91).

Conclusions

These data show that size and pre-contrast attenuation are good predictors of benign histopathology. The low number of patients with cancer in the cohort has an impact on any test sensitivity, but these early data show that different thresholds might be required than currently in use. Larger cohort studies are required to investigate this further and to understand the role of washout studies in determining adrenal histopathology.