LRU – ISTITUTO IMAGING DELLA SVIZZERA ITALIANA (IIMSI)

The main research topics of the Clinics of Radiology refer to:

• the musculo-skeletal system (MSK) imaging, mainly focusing on the development and application of ultrafast MR sequences to study different joints;
• the oncologic imaging, with ongoing studies focusing on the association between the body composition evluated by imaging modalities and oncologic outcomes;
• the application of artificial intelligence (AI) apllied to imaging, with main current focus on the automatic evaluation of the bone age, on the body composition measurements, on radiomics.

The main research topics of the Clinics of Nuclear Medicine refer to:

• integrated diagnostics of different artificial intelligence systems and thyroid pathologies;
• radiomics exctracted from PET-TC in lymphomas;
• theranostics evauated in prostate cancers by PET-PSMA and in neuroendocrine tumours.

• For the MSK research, the MR vendors provide fast sequences to be apllied and the local researchers choose where and when to use them.
• For oncologic imaging, through the use of dedicated software, the different fat and muscle compartments of the body are assessed and quantified, and then the associations with clinical outcomes are evaluated with appropriate statistica analysis.
• For the AI studies, thanking the cooperation with the “Eulero Institute” of the Università della Svizzera italiana we are elaborating a convolutional neural network instructed tool, to automaticly calculate the bone age on elbow radiography. For body composition assessment, we are working on the possibility to automatically extract the needed values.
• For integrated diagnostics studies we serve as coordinator centre of a network of clinical laboratories, clinical units, imaging and pathology services and use advanced AI and deep learning technologies to develop rationale and cost-effective diagnostic and theragnostics working-flows for thyroid diseases and, more recently, prostate cancer.
• For multicentric internationl radiomics studies in nuclear medicine we serve as central revision center and evaluate all studies with a standard protocols and adedicated imaging software (PET‐Encore, MIM software Inc. version 6.7.10) workstation. Lesions are segmented and the mining of radiomics features (RF) from each PET/CT or SPET/CT study performed by the PyRadiomics open‐source Python package (version 2.2.0).We use the least absolute shrinkage and selection operator (LASSO) to select the most informative prognostic parameters in the testing cohorts.Such parameters are compared with other clinical and biological variables in survival endpoints (Kaplan–Meier analysis, log‐rank testing, concordance probability estimate (CPE)-Harrell’s C statistic.
• For theragnostic studies we participate in several international multicentric studies by including our patients affected by prostate cancer and neuroendocrine tumors.