External Validation of Clinical Prediction Models in Unilateral Primary Aldosteronism

Abstract

Background Targeted treatment of primary aldosteronism (PA) is informed by adrenal vein sampling (AVS), which remains limited to specialized centers. Clinical prediction models have been developed to help select patients who would most likely benefit from AVS. Our aim was to assess the performance of these models for PA subtyping. Methods This external validation study evaluated consecutive patients referred for PA who underwent AVS at a tertiary care referral center in Alberta, Canada during 2006-2018. In alignment with the original study designs and intended uses of the clinical prediction models, the primary outcome was the presence of lateralization on AVS. Model discrimination was evaluated using the C-statistic. Model calibration was assessed by comparing the observed vs. predicted probability of lateralization in the external validation cohort. Results The validation cohort included 342 PA patients who underwent AVS (mean age, 52.1 years [SD, 11.5]; 201 [58.8%] male; 186 [54.4%] with lateralization). Six published models were assessed. All models demonstrated low-to-moderate discrimination in the validation set (C-statistics; range, 0.60-0.72), representing a marked decrease compared to the derivation sets (range, 0.80-0.87). Comparison of observed and predicted probabilities of unilateral PA revealed significant miscalibration. Calibration-in-the-large for every model was >0 (range, 0.35-1.67), signifying systematic underprediction of lateralizing disease. Calibration slopes were consistently <1 (range, 0.35-0.87), indicating poor performance at the extremes of risk. Conclusions Overall, clinical prediction models did not accurately predict AVS lateralization in this large cohort. These models cannot be reliably used to inform the decision to pursue AVS for most patients.

Time to Benefit of Surgery vs Targeted Medical Therapy for Patients With Primary Aldosteronism: A Meta-analysis

“…However, access to adrenal vein sampling can be challenging because it is limited to highly specialized centers. Indeed, neither clinical characteristics, biochemical criteria, nor cross-sectional imaging (even in combination) can be used to reliably bypass adrenal vein sampling for most people ( 69 ). While routine access to adrenal vein sampling may be a present limiting factor for patients desiring surgery, there is growing evidence to support nuclear imaging as a feasible alternative to adrenal vein sampling, thus potentially increasing the number of people likely to be referred for surgery in the near future ( 70 ).…”

A new machine learning-based prediction model for subtype diagnosis in primary aldosteronism

“…Meanwhile, innovative imaging techniques, such as 6β-131-iodomethyl-19-norcholesterol (NP-59) scanning and 11C-metomidate positron emission tomography (PET)-CT scanning, have been attempted, yet they currently lack adequate effectiveness ( 4 , 5 , 21 ). Over the past decade, conventional scoring systems were developed to discriminate UPA from BPA based on demographics, biochemical parameters or imaging, with a sensitivity ranging from 32-95% and specificity ranging from 46-100% ( 4 , 22 – 26 ); however, these scoring systems either lacked external validation or demonstrated low-to-modest performance when validated ( 17 ), indicating that they lack generalizability and are not reliable ( 27 ).…”