An innovative artificial intelligence-based method to compress complex models into explainable, model-agnostic and reduced decision support systems with application to healthcare (NEAR).

BACKGROUND AND OBJECTIVE: In everyday clinical practice, medical decision is currently based on clinical guidelines which are often static and rigid, and do not account for population variability, while individualized, patient-oriented decision and/or treatment are the paradigm change necessary to enter into the era of precision medicine. Most of the limitations of a guideline-based system could be overcome through the adoption of Clinical Decision Support Systems (CDSSs) based on Artificial Intelligence (AI) algorithms. However, the black-box nature of AI algorithms has hampered a large adoption of AI-based CDSSs in clinical practice. In this study, an innovative AI-based method to compress AI-based prediction models into explainable, model-agnostic, and reduced decision support systems (NEAR) with application to healthcare is presented and validated. METHODS: NEAR is based on the Shapley Additive Explanations framework and can be applied to complex input models to obtain the contributions of each input feature to the output. Technically, the simplified NEAR models approximate contributions from input features using a custom library and merge them to determine the final output. Finally, NEAR estimates the confidence error associated with the single input feature contributing to the final score, making the result more interpretable. Here, NEAR is evaluated on a clinical real-world use case, the mortality prediction in patients who experienced Acute Coronary Syndrome (ACS), applying three different Machine Learning/Deep Learning models as implementation examples. RESULTS: NEAR, when applied to the ACS use case, exhibits performances like the ones of the AI-based model from which it is derived, as in the case of the Adaptive Boosting classifier, whose Area Under the Curve is not statistically different from the NEAR one, even the model's simplification. Moreover, NEAR comes with intrinsic explainability and modularity, as it can be tested on the developed web application platform (https://neardashboard.pythonanywhere.com/). CONCLUSIONS: An explainable and reliable CDSS tailored to single-patient analysis has been developed. The proposed AI-based system has the potential to be used alongside the clinical guidelines currently employed in the medical setting making them more personalized and dynamic and assisting doctors in taking their everyday clinical decisions.

Thyroid gland: a rare site of metastasis.

PURPOSE: Despite its rich vascularity, metastasis (MTS) to the thyroid tissue is unusual, ranging from 1 to 3%. This entity is not usually considered as differential diagnosis and is not included in the upfront approach in patients with thyroid nodules. Once diagnosed, treatment is controversial. The aim of this study was to evaluate diagnosis, treatment, and outcome at the end of follow-up in patients with a diagnosis of MTS to the thyroid. METHODS: A retrospective multicenter study was designed from 1985 to 2022; 29 patients with MTS to the thyroid gland were included in the analysis. RESULTS: Clinical presentation included the presence of a nodular goiter (65.5%), compression symptoms (17.2%), diffuse goiter (10.3%), and suspicious lymph nodes in the neck (7%). Primary tumor sites were: kidney (44.8%), breast (24.1%), lung (13.8%), neuroendocrine system (6.9%), colon (3.4%), cervix (3.4%), and ovary (3.4%). In 18/23 patients, suspicious ultrasound criteria for malignancy were described. Preoperative diagnosis was made in 23/27 patients by FNA and confirmed in 18 cases by immunohistochemistry. Seventeen patients underwent surgery. At the end of the follow-up, 19 patients had died of oncological disease, and six were alive (2/6 disease-free with isolated intrathyroidal MTS). CONCLUSION: Renal carcinoma was the tumor that most frequently metastasized to the thyroid gland. Immunodiagnosis was a very useful tool for etiological confirmation. Patients with MTS to the thyroid gland as a unique site had a more favorable course compared to patients with multiple metastatic sites. Finally, outcomes and prognosis essentially depended on the biology of the primary tumor.

Action prediction in psychosis.

Aberrant motor-sensory predictive functions have been linked to symptoms of psychosis, particularly reduced attenuation of self-generated sensations and misattribution of self-generated actions. Building on the parallels between prediction of self- and other-generated actions, this study aims to investigate whether individuals with psychosis also demonstrate abnormal perceptions and predictions of others' actions. Patients with psychosis and matched controls completed a two-alternative object size discrimination task. In each trial, they observed reaching actions towards a small and a large object, with varying levels of temporal occlusion ranging from 10% to 80% of movement duration. Their task was to predict the size of the object that would be grasped. We employed a novel analytic approach to examine how object size information was encoded and read out across progressive levels of occlusion with single-trial resolution. Patients with psychosis exhibited an overall pattern of reduced and discontinuous evidence integration relative to controls, characterized by a period of null integration up to 20% of movement duration, during which they did not read any size information. Surprisingly, this drop in accuracy in the initial integration period was not accompanied by a reduction in confidence. Difficulties in action prediction were correlated with the severity of negative symptoms and impaired functioning in social relationships.

Cáncer de tiroides: En búsqueda del tratamiento individualizado / Thyroid cancer: In search of individualized treatment

La incidencia de cáncer de tiroides, principalmente carcinoma papilar, aumentó exponencialmente en todo el mundo. Este incremento podría reflejar los efectos combinados de las prácticas de detección , sumados a cambios en los factores de riesgo. A pesar de este aumento, la mortalidad específica se mantuvo estable en las últimas tres décadas. Dado que los pacientes con carcinoma papilar de tiroides a menudo tienen un muy buen pronóstico, con elevada supervivencia en el seguimiento a largo plazo en comparación con otro tipo de carcinomas, no ha habido una necesidad de cambiar el tratamiento convencional por mucho tiempo. Los pilares del tratamiento, cirugía (tiroidectomía total o casi total) y la terapia con radioyodo (131I), se encuentran en este momento en el centro de discusión en los foros mundiales. La tendencia actual es garantizar el tratamiento más eficaz y menos deletéreo para el paciente. En este momento lo más importante es individualizar el enfoque terapéutico considerando cuál es el estadio tumoral y el riesgo de recurrencia, para determinar qué pacientes se beneficiarán de una terapia más agresiva y cuáles podrán ser tratados con un enfoque más conservador.

The incidence of thyroid cancer has increased exponentially around the world (mostly papillary thyroid carcinoma). This growth may reflect the combined effects of increased screening practices, together with changes in risk factors for thyroid cancer. In spite of this, disease specific mortality remained stable in the last three decades. Due to the fact that patients with papillary thyroid carcinoma often have a very good prognosis, with high survival in the long term follow-up compared with other types of carcinomas, there has been no need to change the standard treatment. The mainstays of thyroid cancer treatment are surgery (total or near-total thyroidectomy) with or without the additional administration of radioiodine (131I). These approaches are now in the center of discussion in all global forums. The current trend is to ensure the most effective and less harmful treatment and the most important issue at this point is to individualize patients according to tumor stage and risk of recurrence, to define which patients will benefit of more aggressive therapy and who could be handled with a more conservative approach.