Medical Information Extraction With NLP-Powered QABots: A Real-World Scenario.

The advent of computerized medical recording systems in healthcare facilities has made data retrieval tasks easier, compared to manual recording. Nevertheless, the potential of the information contained within medical records remains largely untapped, mostly due to the time and effort required to extract data from unstructured documents. Natural Language Processing (NLP) represents a promising solution to this challenge, as it enables the use of automated text-mining tools for clinical practitioners. In this work, we present the architecture of the Virtual Dementia Institute (IVD), a consortium of sixteen Italian hospitals, using the NLP Extraction and Management Tool (NEMT), a (semi-) automated end-to-end pipeline that extracts relevant information from clinical documents and stores it in a centralized REDCap database. After defining a common Case Report Form (CRF) across the IVD hospitals, we implemented NEMT, the core of which is a Question Answering Bot (QABot) based on a modern NLP model. This QABot is fine-tuned on thousands of examples from IVD centers. Detailed descriptions of the process to define a common minimum dataset, Inter-Annotator Agreement calculated on clinical documents, and NEMT results are provided. The best QABot performance show an Exact Match score (EM) of 78.1%, a F1-score of 84.7%, a Lenient Accuracy (LAcc) of 0.834, and a Mean Reciprocal Rank (MRR) of 0.810. EM and F1 scores outperform the same metrics obtained with ChatGPTv3.5 (68.9% and 52.5%, respectively). With NEMT the IVD has been able to populate a database that will contain data from thousands of Italian patients, all screened with the same procedure. NEMT represents an efficient tool that paves the way for medical information extraction and exploitation for new research studies.

Clinical validity of the Italian adaptation of the Uniform Data Set Neuropsychological Test Battery (I-UDSNB) in Mild Cognitive Impairment and Alzheimer's Disease.

BACKGROUND: The identification and staging of Alzheimer's Disease (AD) represent a challenge, especially in the prodromal stage of Mild Cognitive Impairment (MCI), when cognitive changes can be subtle. Worldwide efforts were dedicated to select and harmonize available neuropsychological instruments. In Italy, the Italian Network of Neuroscience and Neuro-Rehabilitation has promoted the adaptation of the Uniform Data Set Neuropsychological Test Battery (I-UDSNB), collecting normative data from 433 healthy controls (HC). Here, we aimed to explore the ability of I-UDSNB to differentiate between a) MCI and HC, b) AD and HC, c) MCI and AD. METHODS: One hundred thirty-seven patients (65 MCI, 72 AD) diagnosed after clinical-neuropsychological assessment, and 137 HC were included. We compared the I-UDSNB scores between a) MCI and HC, b) AD and HC, c) MCI and AD, with t-tests. To identify the test(s) most capable of differentiating between groups, significant scores were entered in binary logistic and in stepwise regressions, and then in Receiver Operating Characteristic curve analyses. RESULTS: Two episodic memory tests (Craft Story and Five Words test) differentiated MCI from HC subjects; Five Words test, Semantic Fluency (vegetables), and TMT-part B differentiated AD from, respectively, HC and MCI. CONCLUSIONS: Our findings indicate that the I-UDSNB is a suitable tool for the harmonized and concise assessment of patients with cognitive decline, showing high sensitivity and specificity for the diagnosis of MCI and AD.

The novel I213S mutation in PSEN1 gene is located in a hotspot codon associated with familial early-onset Alzheimer's disease.

Mutations in presenilin 1 gene (PSEN1) are the most common causes of autosomal dominant early-onset Alzheimer's disease (EOAD). We report a novel PSEN1 mutation (I213S) that was discovered in an Italian patient with a family history of early-onset dementia, who developed a slowly progressive cognitive decline since the age of 40 years. Clinical investigations, including neuropsychological assessment, brain MRI and 18-fluorodeoxyglucose PET, as well as cerebrospinal fluid biomarkers, supported the diagnosis of EOAD. Genetic studies identified a novel missense mutation at codon 213 (I213S). Three other mutations at the same codon have been described in association with EOAD. Previous in silico, in vitro and in vivo studies indicated that these mutations affect the functional properties of γ-secretase and are most likely pathogenic. In silico algorithms suggested that even the I213S mutation has similar deleterious effects on PSEN1 structure and function. Overall, these data strongly support a role of hotspot site for the codon 213 of PSEN1, and provide evidence that the genetic variants located on this site cause EOAD.