Enhanced sleep staging with artificial intelligence: a validation study of new software for sleep scoring.

Manual sleep staging (MSS) using polysomnography is a time-consuming task, requires significant training, and can lead to significant variability among scorers. STAGER is a software program based on machine learning algorithms that has been developed by Medibio Limited (Savage, MN, USA) to perform automatic sleep staging using only EEG signals from polysomnography. This study aimed to extensively investigate its agreement with MSS performed during clinical practice and by three additional expert sleep technicians. Forty consecutive polysomnographic recordings of patients referred to three US sleep clinics for sleep evaluation were retrospectively collected and analyzed. Three experienced technicians independently staged the recording using the electroencephalography, electromyography, and electrooculography signals according to the American Academy of Sleep Medicine guidelines. The staging initially performed during clinical practice was also considered. Several agreement statistics between the automatic sleep staging (ASS) and MSS, among the different MSSs, and their differences were calculated. Bootstrap resampling was used to calculate 95% confidence intervals and the statistical significance of the differences. STAGER's ASS was most comparable with, or statistically significantly better than the MSS, except for a partial reduction in the positive percent agreement in the wake stage. These promising results indicate that STAGER software can perform ASS of inpatient polysomnographic recordings accurately in comparison with MSS.

Balance Rehabilitation with Peripheral Visual Stimulation in Patients with Panic Disorder and Agoraphobia: An Open-Pilot Intervention Study.

Given the involvement of balance system abnormalities in the pathophysiology of panic disorder and agoraphobia (PD-AG), we evaluated initial evidence for feasibility, acceptability, and potential clinical usefulness of 10 sessions of balance rehabilitation with peripheral visual stimulation (BR-PVS) in an open-pilot 5-week intervention study including six outpatients with PD-AG who presented residual agoraphobia after selective serotonin reuptake inhibitor (SSRI) treatment and cognitive-behavioral therapy, dizziness in daily life, and peripheral visual hypersensitivity measured by posturography. Before and after BR-PVS, patients underwent posturography, otovestibular examination (no patients presented peripheral vestibular abnormalities), and panic-agoraphobic symptom and dizziness evaluation with psychometric tools. After BR-PVS, four patients achieved postural control normalization measured by posturography, and one patient exhibited a favorable trend of improvement. Overall, panic-agoraphobic symptoms and dizziness decreased, even though to a lesser extent in one patient who had not completed the rehabilitation sessions. The study presented reasonable levels of feasibility and acceptability. These findings suggest that balance evaluation should be considered in patients with PD-AGO presenting residual agoraphobia and that BR-PVS might be an adjunctive therapeutic option worth being tested in larger randomized controlled studies.

Molecular Tailored Therapeutic Options for Advanced Gastrointestinal Stromal Tumors (GISTs): Current Practice and Future Perspectives.

Gastrointestinal stromal tumors (GISTs) are one of the most common mesenchymal tumors characterized by different molecular alterations that lead to specific clinical presentations and behaviors. In the last twenty years, thanks to the discovery of these mutations, several new treatment options have emerged. This review provides an extensive overview of GISTs' molecular pathways and their respective tailored therapeutic strategies. Furthermore, current treatment strategies under investigation and future perspectives are analyzed and discussed.

Cardiorespiratory Assessments in Panic Disorder Facilitated by Wearable Devices: A Systematic Review and Brief Comparison of the Wearable Zephyr BioPatch with the Quark-b2 Stationary Testing System.

Abnormalities in cardiorespiratory measurements have repeatedly been found in patients with panic disorder (PD) during laboratory-based assessments. However, recordings performed outside laboratory settings are required to test the ecological validity of these findings. Wearable devices, such as sensor-imbedded garments, biopatches, and smartwatches, are promising tools for this purpose. We systematically reviewed the evidence for wearables-based cardiorespiratory assessments in PD by searching for publications on the PubMed, PsycINFO, and Embase databases, from inception to 30 July 2022. After the screening of two-hundred and twenty records, eight studies were included. The limited number of available studies and critical aspects related to the uncertain reliability of wearables-based assessments, especially concerning respiration, prevented us from drawing conclusions about the cardiorespiratory function of patients with PD in daily life. We also present preliminary data on a pilot study conducted on volunteers at the Villa San Benedetto Menni Hospital for evaluating the accuracy of heart rate (HR) and breathing rate (BR) measurements by the wearable Zephyr BioPatch compared with the Quark-b2 stationary testing system. Our exploratory results suggested possible BR and HR misestimation by the wearable Zephyr BioPatch compared with the Quark-b2 system. Challenges of wearables-based cardiorespiratory assessment and possible solutions to improve their reliability and optimize their significant potential for the study of PD pathophysiology are presented.

Reliability of patient-reported toxicities during adjuvant chemotherapy.

BACKGROUND: Patient-reported outcomes (PROs) are validated tools to assess the impact of efficacy and toxicities of cancer treatments on patients' health status. Because of the demonstrated little reliability of humans in reporting memories of painful experiences, this work explores the reliability of cancer patients in reporting chemotherapy-related toxicities. AIM: This study aims to evaluate the concordance between toxicities experienced by the patients during chemotherapy and toxicities reported to the doctor at the end of the cycles. METHODS: Questionnaires concerning chemotherapy-related toxicities were administered on days 2, 5, 8, 11, 14, and 17 of each chemo cycle and at the end of the same cycle to patients undergoing adjuvant chemotherapy. The co-primary end-points were Lins's concordance correlation coefficient (CCC) and mean difference between real-time and retrospective toxicity assessments. RESULTS: In total, 7182 toxicity assessments were collected from 1096 questionnaires. Concordance was observed between the retrospective evaluations and the toxicity assessments at early (day 2), peak (maximum toxicity), late (day 14 or 17), and mean real-time evaluations for each chemotherapy cycle (CCC for mean ranging from 0.52 to 0.77). No systematic discrepancy was found between real-time and retrospective evaluations, except for peak, which was systematically underestimated retrospectively. CONCLUSIONS: Toxicities reported by the patients to the doctor at the end of each chemotherapy cycle reflect what they actually experienced without any substantial distortion. This result is very relevant both for the clinical implications in daily patients' management and in the light of the current growing impact on digital monitoring of PROs.

High prevalence of major depression in US sleep clinics: the need for routine depression screening in sleep services.

Depression screening is not part of routine clinical practice in US sleep clinics. Our study aimed to report the prevalence of depression among individuals referred to US sleep clinics. According to our findings, approximately 21% of patients had depression, with about 4% reporting severe symptoms, 9% had frequent death and/or self-harming thoughts, and 61% were taking antidepressants. Our results highlighted a considerable risk of prevalent depression in sleep clinics and supported the limited existing data on this topic. Our study advocates for the need for routine depression screening in sleep services to reduce the detrimental consequences of a delayed depression diagnosis and the risk of a worse prognosis for both depression and sleep-wake disorders. CITATION: Daccò S, Caldirola D, Grassi M, Alciati A, Perna G, Defillo A. High prevalence of major depression in US sleep clinics: the need for routine depression screening in sleep services. J Clin Sleep Med. 2023;19(4):835-836.

Clinical Significance of Germline Pathogenic Variants among 51 Cancer Predisposition Genes in an Unselected Cohort of Italian Pancreatic Cancer Patients.

Multigene germline panel testing is recommended for Pancreatic Cancer (PC) patients; however, for non-BRCA1/2 genes, the clinical utility is unclear. A comprehensive multi-gene assessment in unselected Italian PC patients is missing. We evaluated the prevalence and impact of Pathogenic Variants (PV) in 51 PC susceptibility genes in a real-world series of 422 Italian PC patients unselected for Family History (FH), compared the clinical characteristics and conducted survival analyses. 17% of patients had PVs (70/422), mainly in BRCA1/2 (4.5%, all <70 y), CDKN2A (4.5%, all >50 y), ATM (2.1%). PV carriers were younger (64 vs. 67; p = 0.02) and had more frequent personal/FH of PC, melanoma and breast/ovarian cancer (all p < 0.05). The Overall Survival (OS) was longer in patients carrying PVs (HR 0.78; p = 0.090), comprising ATM carriers (HR 0.33; p = 0.054). In the oxaliplatin-treated subset, PV carriers showed better control of the disease, although this was not statistically significant (67% vs. 56%). CDKN2A, BRCA2 and ATM were the most frequently altered genes. ATM PVs were positively associated with OS in 41% of PV carriers, 60% of whom carried CDKN2A,BRCA2 or ATM PVs, had negative FH and would have been missed by traditional referral. Thus, CDKN2A and ATM should be added to BRCA1/2 testing regardless of FH.

A Circulating Risk Score, Based on Combined Expression of Exo-miR-130a-3p and Fibrinopeptide A, as Predictive Biomarker of Relapse in Resectable Non-Small Cell Lung Cancer Patients.

To date, the 5-year overall survival rate of 60% for early-stage non-small cell lung cancer (NSCLC) is still unsatisfactory. Therefore, reliable prognostic factors are needed. Growing evidence shows that cancer progression may depend on an interconnection between cancer cells and the surrounding tumor microenvironment; hence, circulating molecules may represent promising markers of cancer recurrence. In order to identify a prognostic score, we performed in-depth high-throughput analyses of plasma circulating markers, including exosomal microRNAs (Exo-miR) and peptides, in 67 radically resected NSCLCs. The miRnome profile selected the Exo-miR-130a-3p as the most overexpressed in relapsed patients. Peptidome analysis identified four progressively more degraded forms of fibrinopeptide A (FpA), which were depleted in progressing patients. Notably, stepwise Cox regression analysis selected Exo-miR-130a-3p and the greatest FpA (2-16) to build a score predictive of recurrence, where high-risk patients had 18 months of median disease-free survival. Moreover, in vitro transfections showed that higher levels of miR-130a-3p lead to a deregulation of pathways involved in metastasis and angiogenesis, including the coagulation process and metalloprotease increase which might be linked to FpA reduction. In conclusion, by integrating circulating markers, the identified risk score may help clinicians predict early-stage NSCLC patients who are more likely to relapse after primary surgery.

First-onset major depression during the COVID-19 pandemic: A predictive machine learning model.

BACKGROUND: This study longitudinally evaluated first-onset major depression rates during the pandemic in Italian adults without any current clinician-diagnosed psychiatric disorder and created a predictive machine learning model (MLM) to evaluate subsequent independent samples. METHODS: An online, self-reported survey was released during two pandemic periods (May to June and September to October 2020). Provisional diagnoses of major depressive disorder (PMDD) were determined using a diagnostic algorithm based on the DSM criteria of the Patient Health Questionnaire-9 to maximize specificity. Gradient-boosted decision trees and the SHapley Additive exPlanations technique created the MLM and estimated each variable's predictive contribution. RESULTS: There were 3532 participants in the study. The final sample included 633 participants in the first wave (FW) survey and 290 in the second (SW). First-onset PMDD was found in 7.4% of FW participants and 7.2% of the SW. The final MLM, trained on the FW, displayed a sensitivity of 76.5% and a specificity of 77.8% when tested on the SW. The main factors identified in the MLM were low resilience, being an undergraduate student, being stressed by pandemic-related conditions, and low satisfaction with usual sleep before the pandemic and support from relatives. Current smoking and taking medication for medical conditions also contributed, albeit to a lesser extent. LIMITATIONS: Small sample size; self-report assessment; data covering 2020 only. CONCLUSIONS: Rates of first-onset PMDD among Italians during the first phases of the pandemic were considerable. Our MLM displayed a good predictive performance, suggesting potential goals for depression-preventive interventions during public health crises.

Predicting New-Onset Psychiatric Disorders Throughout the COVID-19 Pandemic: A Machine Learning Approach.

OBJECTIVE: The investigators estimated new-onset psychiatric disorders (PsyDs) throughout the COVID-19 pandemic in Italian adults without preexisting PsyDs and developed a machine learning (ML) model predictive of at least one new-onset PsyD in subsequent independent samples. METHODS: Data were from the first (May 18-June 20, 2020) and second (September 15-October 20, 2020) waves of an ongoing longitudinal study, based on a self-reported online survey. Provisional diagnoses of PsyDs (PPsyDs) were assessed via DSM-based screening tools to maximize assessment specificity. Gradient-boosted decision trees as an ML modeling technique and the SHapley Additive exPlanations technique were applied to identify each variable's contribution to the model. RESULTS: From the original sample of 3,532 participants, the final sample included 500 participants in the first wave and 236 in the second. Some 16.0% of first-wave participants and 18.6% of second-wave participants met criteria for at least one new-onset PPsyD. The final best ML predictive model, trained on the first wave, displayed a sensitivity of 70% and a specificity of 73% when tested on the second wave. The following variables made the largest contributions: low resilience, being an undergraduate student, and being stressed by pandemic-related conditions. Living alone and having ceased physical activity contributed to a lesser extent. CONCLUSIONS: Substantial rates of new-onset PPsyDs emerged among Italians throughout the pandemic, and the ML model exhibited moderate predictive performance. Results highlight modifiable vulnerability factors that are suitable for targeting by public campaigns or interventions to mitigate the pandemic's detrimental effects on mental health.

Prediction of illness remission in patients with Obsessive-Compulsive Disorder with supervised machine learning.

INTRODUCTION: The course of OCD differs widely among OCD patients, varying from chronic symptoms to full remission. No tools for individual prediction of OCD remission are currently available. This study aimed to develop a machine learning algorithm to predict OCD remission after two years, using solely predictors easily accessible in the daily clinical routine. METHODS: Subjects were recruited in a longitudinal multi-center study (NOCDA). Gradient boosted decision trees were used as supervised machine learning technique. The training of the algorithm was performed with 227 predictors and 213 observations collected in a single clinical center. Hyper-parameter optimization was performed with cross-validation and a Bayesian optimization strategy. The predictive performance of the algorithm was subsequently tested in an independent sample of 215 observations collected in five different centers. Between-center differences were investigated with a bootstrap resampling approach. RESULTS: The average predictive performance of the algorithm in the test centers resulted in an AUROC of 0.7820, a sensitivity of 73.42%, and a specificity of 71.45%. Results also showed a significant between-center variation in the predictive performance. The most important predictors resulted related to OCD severity, OCD chronic course, use of psychotropic medications, and better global functioning. LIMITATIONS: All recruiting centers followed the same assessment protocol and are in The Netherlands. Moreover, the sample of the data recruited in some of the test centers was limited in size. DISCUSSION: The algorithm demonstrated a moderate average predictive performance, and future studies will focus on increasing the stability of the predictive performance across clinical settings.

Elevated C-reactive protein levels across diagnoses: The first comparison among inpatients with major depressive disorder, bipolar disorder, or obsessive-compulsive disorder.

OBJECTIVE: We addressed elevated C-reactive protein level (eCRP) specificity comparing, for the first time, eCRP (i.e., serum CRP > 3 and ≤10 mg/L) in patients with major depressive disorder (MDD), bipolar disorder (BD), or obsessive-compulsive disorder (OCD). We also assessed to what extent multiple variables that can potentially increase inflammation may have influenced eCRP in our sample. METHODS: We performed a retrospective, observational, cross-sectional study using information documented in the electronic medical records (EMRs) of patients hospitalized for a 4-week psychiatric rehabilitation program. We collected all information according to the standardized procedures of the hospital's clinical practice and applied a logistic regression model (α = 0.05). RESULTS: We included 388 inpatients, that is, 156 (40.2%) with MDD, 135 (34.8%) with BD, and 97 (25.0%) with OCD, and found considerable eCRP rates among them (36.5%, 47.4%, and 29.9% in MDD, BD, and OCD, respectively) but without significant differences across groups. In the whole sample, eCRP variations were only partially attributable (approximately for one-third) to potential confounders. All groups presented considerable rates of cardiovascular risk factors, and we classified most patients as having medium or high CRP-based cardiovascular risk. CONCLUSION: This first study comparing eCRP in MDD, BD, and OCD suggests that eCRP may be a transdiagnostic feature of different psychiatric disorders, and other mechanisms beyond the effects of multiple confounders may explain the presence of eCRP in a substantial portion of psychiatric patients. Therefore, we encourage the routine measurement of CRP in psychiatric clinical practice.

Outstanding Response in a Patient With ROS1-Rearranged Inflammatory Myofibroblastic Tumor of Soft Tissues Treated With Crizotinib: Case Report.

Inflammatory myofibroblastic tumor (IMT) is a very rare subtype of sarcoma, which frequently harbor chromosomal rearrangements, including anaplastic lymphoma kinase (ALK) rearrangements (almost 50% of the IMTs) and other kinase fusions such as ROS1. ROS1 fusions are present in about 10% of IMT, almost half of the ALK-negative IMT patients. Apart from radical surgery for resectable tumors, there is no standard-of-care therapy for advanced IMTs. Nonetheless, the use of tyrosine kinase inhibitors has shown promising efficacy in IMT patients with targetable genomic alterations. We report the case of a 24-year-old patient with chemotherapy-refractory metastatic IMT harboring ROS1 kinase fusion, who experienced a significant clinical and pathological response to crizotinib. This clinical case highlights the need to assess all patients with unresectable IMTs for chromosomal abnormalities and gene mutations and address them to targeted agents as well as clinical trials.

Novel Emerging Molecular Targets in Non-Small Cell Lung Cancer.

In the scenario of systemic treatment for advanced non-small cell lung cancer (NSCLC) patients, one of the most relevant breakthroughs is represented by targeted therapies. Throughout the last years, inhibitors of the epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK), c-Ros oncogene 1 (ROS1), and V-raf murine sarcoma viral oncogene homolog B (BRAF) have been approved and are currently used in clinical practice. However, other promising molecular drivers are rapidly emerging as therapeutic targets. This review aims to cover the molecular alterations with a potential clinical impact in NSCLC, including amplifications or mutations of the mesenchymal-epithelial transition factor (MET), fusions of rearranged during transfection (RET), rearrangements of the neurotrophic tyrosine kinase (NTRK) genes, mutations of the Kirsten rat sarcoma viral oncogene (KRAS) and phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha (PIK3CA), as well as amplifications or mutations of human epidermal growth factor receptor 2 (HER2). Additionally, we summarized the current status of targeted agents under investigation for such alterations. This revision of the current literature on emerging molecular targets is needed as the evolving knowledge on novel actionable oncogenic drivers and targeted agents is expected to increase the proportion of patients who will benefit from tailored therapeutic approaches.

Systemic therapies in advanced epithelioid haemangioendothelioma: A retrospective international case series from the World Sarcoma Network and a review of literature.

BACKGROUND: This observational, retrospective effort across Europe, US, Australia, and Asia aimed to assess the activity of systemic therapies in EHE, an ultra-rare sarcoma, marked by WWTR1-CAMTA1 or YAP1-TFE3 fusions. METHODS: Twenty sarcoma reference centres contributed data. Patients with advanced EHE diagnosed from 2000 onwards and treated with systemic therapies, were selected. Local pathologic review and molecular confirmation were required. Radiological response was retrospectively assessed by local investigators according to RECIST. Progression free survival (PFS) and overall survival (OS) were estimated by Kaplan-Meier method. RESULTS: Overall, 73 patients were included; 21 had more than one treatment. Thirty-three patients received anthracyclines regimens, achieving 1 (3%) partial response (PR), 25 (76%) stable disease (SD), 7 (21%) progressive disease (PD). The median (m-) PFS and m-OS were 5.5 and 14.3 months respectively. Eleven patients received paclitaxel, achieving 1 (9%) PR, 6 (55%) SD, 4 (36%) PD. The m-PFS and m-OS were 2.9 and 18.6 months, respectively. Twelve patients received pazopanib, achieving 3 (25%) SD, 9 (75%) PD. The m-PFS and m-OS were.2.9 and 8.5 months, respectively. Fifteen patients received INF-α 2b, achieving 1 (7%) PR, 11 (73%) SD, 3 (20%) PD. The m-PFS and m-OS were 8.9 months and 64.3, respectively. Among 27 patients treated with other regimens, 1 PR (ifosfamide) and 9 SD (5 gemcitabine +docetaxel, 2 oral cyclophosphamide, 2 others) were reported. CONCLUSION: Systemic therapies available for advanced sarcomas have limited activity in EHE. The identification of new active compounds, especially for rapidly progressive cases, is acutely needed.

'Precision' or 'personalized' psychiatry: different terms - same content?

Due to the increased lifetime prevalence and personal, social, and economic burden of mental disorders, psychiatry is in need of a significant change in several aspects of its clinical and research approaches. Over the last few decades, the development of personalized / precision medicine in psychiatry focusing on tailored therapies that fit each patient's unique individual, physiological, and genetic profile has not achieved the same results as those obtained in other branches, such as oncology. The long-awaited revolution has not yet surfaced. There are various explanations for this including imprecise diagnostic criteria, incomplete understanding of the molecular pathology involved, absence of available clinical tools and, finally, the characteristics of the patient. Since then, the co-existence of the two terms has sparked a great deal of discussion around the definition and differentiation between the two types of psychiatry, as they often seem similar or even superimposable. Generally, the two terminologies are used indiscriminately, alternatively, and / or separately, within the same scientific works. In this paper, an overview is provided on the overlap between the application and meaning of the terms 'precision psychiatry' and 'personalized psychiatry'.

Precision Medicine for NSCLC in the Era of Immunotherapy: New Biomarkers to Select the Most Suitable Treatment or the Most Suitable Patient.

In recent years, the evolution of treatments has made it possible to significantly improve the outcomes of patients with non-small cell lung cancer (NSCLC). In particular, while molecular targeted therapies are effective in specific patient sub-groups, immune checkpoint inhibitors (ICIs) have greatly influenced the outcomes of a large proportion of NSCLC patients. While nivolumab activity was initially assessed irrespective of predictive biomarkers, subsequent pivotal studies involving other PD-1/PD-L1 inhibitors in pre-treated advanced NSCLC (atezolizumab within the OAK study and pembrolizumab in the Keynote 010 study) reported the first correlations between clinical outcomes and PD-L1 expression. However, PD-L1 could not be sufficient on its own to select patients who may benefit from immunotherapy. Many studies have tried to discover more precise markers that are derived from tumor tissue or from peripheral blood. This review aims to analyze any characteristics of the immunogram that could be used as a predictive biomarker for response to ICIs. Furthermore, we describe the most important genetic alteration that might predict the activity of immunotherapy.

Personalized Psychiatry and Depression: The Role of Sociodemographic and Clinical Variables.

Despite several pharmacological options, the clinical outcomes of major depressive disorder (MDD) are often unsatisfactory. Personalized psychiatry attempts to tailor therapeutic interventions according to each patient's unique profile and characteristics. This approach can be a crucial strategy in improving pharmacological outcomes in MDD and overcoming trial-and-error treatment choices. In this narrative review, we evaluate whether sociodemographic (i.e., gender, age, race/ethnicity, and socioeconomic status) and clinical [i.e., body mass index (BMI), severity of depressive symptoms, and symptom profiles] variables that are easily assessable in clinical practice may help clinicians to optimize the selection of antidepressant treatment for each patient with MDD at the early stages of the disorder. We found that several variables were associated with poorer outcomes for all antidepressants. However, only preliminary associations were found between some clinical variables (i.e., BMI, anhedonia, and MDD with melancholic/atypical features) and possible benefits with some specific antidepressants. Finally, in clinical practice, the assessment of sociodemographic and clinical variables considered in our review can be valuable for early identification of depressed individuals at high risk for poor responses to antidepressants, but there are not enough data on which to ground any reliable selection of specific antidepressant class or compounds. Recent advances in computational resources, such as machine learning techniques, which are able to integrate multiple potential predictors, such as individual/ clinical variables, biomarkers, and genetic factors, may offer future reliable tools to guide personalized antidepressant choice for each patient with MDD.

Multiple systemic treatment options in a patient with malignant tenosynovial giant cell tumour.

Tenosynovial giant cell tumour (TGCT) is a group of rare soft tissues neoplasia affecting synovial joints, bursae and tendon sheaths and is classified as localized type or diffuse type. The diffuse type (TGCT-D), also known as 'pigmented villonodular (teno)synovitis' is characterized by local aggressivity, with invasion and destruction of adjacent soft-tissue structures, and high local recurrence rate. Radical surgery remains the standard therapy while adjuvant radiotherapy may help to control local spread. Malignant TGCT is characterized by high rate of local recurrences and distant metastasis. Few cases of malignant TGCT and very few evidences on systemic therapies are described in the literature, so, to date, no systemic treatment is approved for this rare disease. We report the case of a malignant TGCT patient treated with many different systemic therapies, including chemotherapy and tyrosine-kinase inhibitors, and performed a review of the literature on the systemic treatment options of this rare tumour.

A Novel Ensemble-Based Machine Learning Algorithm to Predict the Conversion From Mild Cognitive Impairment to Alzheimer's Disease Using Socio-Demographic Characteristics, Clinical Information, and Neuropsychological Measures.

Background: Despite the increasing availability in brain health related data, clinically translatable methods to predict the conversion from Mild Cognitive Impairment (MCI) to Alzheimer's disease (AD) are still lacking. Although MCI typically precedes AD, only a fraction of 20-40% of MCI individuals will progress to dementia within 3 years following the initial diagnosis. As currently available and emerging therapies likely have the greatest impact when provided at the earliest disease stage, the prompt identification of subjects at high risk for conversion to AD is of great importance in the fight against this disease. In this work, we propose a highly predictive machine learning algorithm, based only on non-invasively and easily in-the-clinic collectable predictors, to identify MCI subjects at risk for conversion to AD. Methods: The algorithm was developed using the open dataset from the Alzheimer's Disease Neuroimaging Initiative (ADNI), employing a sample of 550 MCI subjects whose diagnostic follow-up is available for at least 3 years after the baseline assessment. A restricted set of information regarding sociodemographic and clinical characteristics, neuropsychological test scores was used as predictors and several different supervised machine learning algorithms were developed and ensembled in final algorithm. A site-independent stratified train/test split protocol was used to provide an estimate of the generalized performance of the algorithm. Results: The final algorithm demonstrated an AUROC of 0.88, sensitivity of 77.7%, and a specificity of 79.9% on excluded test data. The specificity of the algorithm was 40.2% for 100% sensitivity. Conclusions: The algorithm we developed achieved sound and high prognostic performance to predict AD conversion using easily clinically derived information that makes the algorithm easy to be translated into practice. This indicates beneficial application to improve recruitment in clinical trials and to more selectively prescribe new and newly emerging early interventions to high AD risk patients.