Machine learning model provides superior performance over specialist interpretations and could reliably predict metastatic disease in most patients with pheochromocytomas and paragangliomas

 A recently completed machine learning modelling study at the University Hospital Carl Gustav Carus in Dresden, Germany indicates that incorporating "plasma methoxytyramine in machine learning models, along with other clinical features such as primary tumour location and size, provides a highly accurate, non-invasive approach to predict metastases in patients with pheochromocytomas and paragangliomas", which can guide customized management and follow-up strategies for patients. The modeling study indicated that plasma methoxytyramine could identify metastatic disease "at sensitivities of 52% and specificities of 85%", with the highest-rated machine learning model encompassing nine features: plasma methoxytyramine, metanephrine, normetanephrine, age, sex, previous history of pheochromocytoma or paraganglioma, location and size of primary tumours, and presence of multifocal disease.  

To read more about this study, click here. 

Study mentioned: Pamporaki C, Berends AMA, Filippatos A, et al. Prediction of metastatic pheochromocytoma and paraganglioma: a machine learning modelling study using data from a cross-sectional cohort. The Lancet Digital Health; Published online 18 July 2023. DOI: https://doi.org/10.1016/S2589-7500(23)00094-8