Exploring telediagnostic procedures in child neuropsychiatry: addressing ADHD diagnosis and autism symptoms through supervised machine learning.

Recently, there has been an increase in telemedicine applied to child neuropsychiatry, such as the use of online platforms to collect remotely case histories and demographic and behavioral information. In the present proof-of-concept study, we aimed to understand to what extent information parents and teachers provide through online questionnaires overlaps with clinicians' diagnostic conclusions on attention-deficit/hyperactivity disorder (ADHD). Moreover, we intended to explore a possible role that autism spectrum disorders (ASD) symptoms played in this process. We examined parent- and teacher-rated questionnaires collected remotely and an on-site evaluation of intelligence quotients from 342 subjects (18% females), aged 3-16 years, and referred for suspected ADHD. An easily interpretable machine learning model-decision tree (DT)-was built to simulate the clinical process of classifying ADHD/non-ADHD based on collected data. Then, we tested the DT model's predictive accuracy through a cross-validation approach. The DT classifier's performance was compared with those that other machine learning models achieved, such as random forest and support vector machines. Differences in ASD symptoms in the DT-identified classes were tested to address their role in performing a diagnostic error using the DT model. The DT identified the decision rules clinicians adopt to classify an ADHD diagnosis with an 82% accuracy rate. Regarding the cross-validation experiment, our DT model reached a predictive accuracy of 74% that was similar to those of other classification algorithms. The caregiver-reported ADHD core symptom severity proved the most discriminative information for clinicians during the diagnostic decision process. However, ASD symptoms were a confounding factor when ADHD severity had to be established. Telehealth procedures proved effective in obtaining an automated output regarding a diagnostic risk, reducing the time delay between symptom detection and diagnosis. However, this should not be considered an alternative to on-site procedures but rather as automated support for clinical practice, enabling clinicians to allocate further resources to the most complex cases.

Bayesian optimization and deep learning for steering wheel angle prediction.

Automated driving systems (ADS) have undergone a significant improvement in the last years. ADS and more precisely self-driving cars technologies will change the way we perceive and know the world of transportation systems in terms of user experience, mode choices and business models. The emerging field of Deep Learning (DL) has been successfully applied for the development of innovative ADS solutions. However, the attempt to single out the best deep neural network architecture and tuning its hyperparameters are all expensive processes, both in terms of time and computational resources. In this work, Bayesian optimization (BO) is used to optimize the hyperparameters of a Spatiotemporal-Long Short Term Memory (ST-LSTM) network with the aim to obtain an accurate model for the prediction of the steering angle in a ADS. BO was able to identify, within a limited number of trials, a model-namely BO_ST-LSTM-which resulted, on a public dataset, the most accurate when compared to classical end-to-end driving models.

Efficient Kernel-Based Subsequence Search for Enabling Health Monitoring Services in IoT-Based Home Setting.

This paper presents an efficient approach for subsequence search in data streams. The problem consists of identifying coherent repetitions of a given reference time-series, also in the multivariate case, within a longer data stream. The most widely adopted metric to address this problem is Dynamic Time Warping (DTW), but its computational complexity is a well-known issue. In this paper, we present an approach aimed at learning a kernel approximating DTW for efficiently analyzing streaming data collected from wearable sensors, while reducing the burden of DTW computation. Contrary to kernel, DTW allows for comparing two time-series with different length. To enable the use of kernel for comparing two time-series with different length, a feature embedding is required in order to obtain a fixed length vector representation. Each vector component is the DTW between the given time-series and a set of "basis" series, randomly chosen. The approach has been validated on two benchmark datasets and on a real-life application for supporting self-rehabilitation in elderly subjects has been addressed. A comparison with traditional DTW implementations and other state-of-the-art algorithms is provided: results show a slight decrease in accuracy, which is counterbalanced by a significant reduction in computational costs.

Muscle shortening manoeuvre reduces pain and functional impairment in shoulder impingement syndrome: clinical and ultrasonographic evidence.

OBJECTIVES: To evaluate the short-term efficacy of muscle shortening manoeuvre (MSM), by inducing an increase in strength of the shoulder muscles, for the treatment of shoulder impingement syndrome (SIS). METHODS: Sixty subjects (mean age: 58.6 years) with SIS were assigned to one of 3 different treatment interventions: 1) MSM: a series of fast accelerations in the upward direction was applied to the upper limb that was also submitted to forces acting in the opposite direction (added mass); 2) traditional physiotherapeutic technique: scapulothoracic gliding; 3) simple traction: the added mass was applied to the limb without the series of fast accelerations. Pain intensity, Neer's impingement sign, range of motion and muscle strength were assessed. Ultrasound (US) examination was performed before, immediately after and 30 days after each treatment to study the width of the subacromial-subdeltoid bursa, long biceps tendon sheath and acromioclavicular joint. Impingement was evaluated by dynamic examination. RESULTS: After treatment with MSM, pain was significantly reduced (p<0.001), Neer's impingement sign was negative, range of motion and muscle strength were increased. US examination showed that the widths of the subacromial-subdeltoid bursa (p<0.001), long biceps tendon sheath (p<0.001) and acromioclavicular joint (p<0.001) were significantly reduced; impingement was no more detected. After 30 days, improvement in clinical and US findings was maintained. In the two control groups, no significant changes were observed after treatment. CONCLUSIONS: Clinical and US findings demonstrate that MSM, by inducing an increase in muscle strength, is effective in the short-term treatment of SIS.

Clinical and radiological evaluation of sacroiliac joints compared with ultrasound examination in early spondyloarthritis.

OBJECTIVE: To compare clinical and X-ray examinations with US findings of SI joints (SIJ) in early SpA patients. METHODS: Twenty-three early SpA patients, diagnosed according to Assessment of SpondyloArthritis international Society criteria, were investigated clinically [sacral sulcus tenderness, BASMI, BASFI, BASDAI, pain and fatigue visual analogue scale (VAS), morning stiffness and sleep disturbance], with SIJ X-rays (New York score) and with My Lab70 US 7-10 MHz US (Esaote, Genoa, Italy), evaluating the width of the SIJ capsule and posterior sacroiliac (PSL) and sacrotuberosus (STL) ligament thickness and comparing the results with 23 healthy controls. RESULTS: SIJ width [right 2.2 (0.6) and left 2.3 (0.7) in SpA vs 1.6 (0.1) and 1.7 (0.2) in healthy controls, respectively, expressed as mean (s.d.)] and STL thickness [right 3.9 (1.3) and left 3.4 (1.0) vs 1.8 (0.1) and 1.8 (0.1), respectively, expressed as mean (s.d.)] were higher in SpA patients than in controls (P < 0.001 and P < 0.05, respectively). PSL thickness was similar in patients and controls. Only STL thickness was higher when SIJ was tender at clinical examination (P < 0.01) and correlated with pain VAS (P < 0.001) and BASFI (P < 0.05). Furthermore, SIJ US results were unrelated to X-ray findings (similar when X-ray sacroiliitis was present and not). CONCLUSION: Our exploratory study suggested that in early SpA patients US might be a promising method, complementary to other imaging techniques, to study articular and soft tissue periarticular involvement of SIJ, independent of clinical and X-ray examination.

Ultrasound detects occult entheseal involvement in early psoriatic arthritis independently of clinical features and psoriasis severity.

OBJECTIVES: Entheseal involvement is a frequent and distinctive feature of psoriatic arthritis (PsA), and is often under-diagnosed. The aim of the present study is to investigate using ultrasound (US), lower limb entheseal abnormalities in patients with early psoriatic arthritis (ePsA) and to evaluate their correlation with ePsA clinical characteristics. METHODS: Ninety-two ePsA patients (with duration of symptoms less than 1 year), diagnosed according to CASPAR criteria, were consecutively scored with Glasgow Ultrasound Enthesitis Scoring System (GUESS) and Power Doppler (PD) US (My Lab 70 Esaote) of lower limbs entheses (quadriceps, patellar, achilles tendons and plantar fascia). Patients were clinically examined by palpation of lower limbs entheses, Maastricht Ankylosing Spondylitis Enthesitis Index (MASES) and total Psoriasis Area and Severity Index (PASI). Correlations were investigated between GUESS and PD with other ePsA clinical characteristics (duration of symptoms and morning stiffness, pain and fatigue visual analogue scale [VAS], Health Assessment Questionnaire SpA-modified [S-HAQ]). RESULTS: All patients had GUESS>1 and 40.2% showed positive PD signal on entheses, at a higher percentage than tenderness revealed by clinical examination (29.3%). GUESS and PD did not correlate with MASES, PASI and other clinical characteristics. No significant differences in GUESS and PD were detected between positive or negative findings of MASES and PASI. CONCLUSIONS: US detects subclinical entheseal involvement in ePsA, independently of ePsA clinical examination and symptoms.

MicroRNA expression profile helps to distinguish benign nodules from papillary thyroid carcinomas starting from cells of fine-needle aspiration.

OBJECTIVE: MicroRNAs (miRNAs) are small endogenous noncoding RNAs that pair with target messengers regulating gene expression. Changes in miRNA levels occur in thyroid cancer. Fine-needle aspiration (FNA) with cytological evaluation is the most reliable tool for malignancy prediction in thyroid nodules, but cytological diagnosis remains undetermined for 20% of nodules. DESIGN: In this study, we evaluated the expression of seven miRNAs in benign nodules, papillary thyroid carcinomas (PTCs), and undetermined nodules at FNA. METHODS: The prospective study included 141 samples obtained by FNA of thyroid nodules from 138 patients. miRNA expression was evaluated by quantitative RT-PCR and statistical analysis of data was performed. Genetic analysis of codon 600 of BRAF gene was also performed. RESULTS: Using data mining techniques, we obtained a criterion to classify a nodule as benign or malignant on the basis of miRNA expression. The decision model based on the expression of miR-146b, miR-155, and miR-221 was valid for 86/88 nodules with determined cytology (97.73%), and adopting cross-validation techniques we obtained a reliability of 78.41%. The prediction was valid for 31/53 undetermined nodules with 16 false-positive and six false-negative predictions. The mutated form V600E of BRAF gene was demonstrated in 19/43 PTCs and in 1/53 undetermined nodules. CONCLUSIONS: The expression profiles of three miRNAs allowed us to distinguish benign from PTC starting from FNA. When the assay was applied to discriminate thyroid nodules with undetermined cytology, a low sensitivity and specificity despite the low number of false-negative predictions was obtained, limiting the practical interest of the method.

Ankylosing spondylitis: how diagnostic and therapeutic delay have changed over the last six decades.

OBJECTIVES: Ankylosing spondylitis (AS) is a chronic, progressive, and disabling disease, but the diagnosis is often missed and markedly delayed. An early diagnosis is important to establish a treatment to reduce disability and modify the natural course of disease. The aim of this study was to investigate the diagnostic (DD) and therapeutic (TD) delay according to the decade of diagnosis. The DD and TD correlation with radiological severity score and the new imaging techniques used in diagnosis (magnetic resonance [MRI], computerised tomography, scintigraphy for sacroiliac joints) were also investigated. METHODS: 135 AS patients (45 female and 90 male, 36.5±10.2 years old at diagnosis) with disease onset between 1950 and 2008, were investigated; the time from onset to diagnosis (DD) and treatment (TD), the New York and ASAS criteria fulfilment, the New York sacroiliac radiological score, bamboo spine presence at first visit and the new imaging technique used at diagnosis were recorded and their correlations were analysed. RESULTS: The New York and ASAS criteria were met at the first visit, by 87% and 96%, respectively. The delay from onset of symptoms to diagnosis and treatment was 9±8 and 12±11 years, respectively, but decreased significantly between different decades (p<0.001). The severity of sacroiliitis (mean 2±1; 17/135, 12.5% - IV grade sacroiliitis at diagnosis) and bamboo spine (3.7% at diagnosis) correlated with DD and TD (p<0.001). Sacroiliac MRI use at diagnosis significantly decreased both DD and TD (p>0.001 and p<0.05, respectively). CONCLUSIONS: DD and TD were correlated to radiological severity; they progressively decreased over 6 decades.

Longitudinal ultrasound and clinical follow-up of Baker's cysts injection with steroids in knee osteoarthritis.

This study was conducted to assess ultrasound (US) and clinical changes of Baker's cyst (BC) of patients with knee osteoarthritis (OA) after steroid injection. Patients with knee OA complicated with symptomatic BC (40) were treated with US-guided direct (posterior) aspiration. The injection of 40 mg triamcynolone acetonide was in 20 patients direct into the BC and in other 20 subjects intra-articular (anterior). BC diameters (longitudinal, transverse, and thickness) were measured and followed up with US at baseline, 2, 4, and 8 weeks after injection. Swelling, pain, and range motion were scored at clinical examination with Rauschning and Lindgren classification (RLC, since 0 normal to 3 maximal signs). All US measures of BC and RLC significantly decreased after treatment, in comparison to baseline (p < 0.001) and during the follow-up, did not change through the time (no significant difference between 2, 4, and 8 weeks). At 4 and 8 weeks, diameters measured at US are lower when BC is directly infiltrated in comparison to intra-articular injection (p < 0.01). US steroid direct injection reduces US measures and clinics of BC in knee OA, in particular, when steroid is directly infiltrated into BC.

Visual pursuit: within-day variability in the severe disorder of consciousness.

Visual pursuit marks substantial recuperation from a vegetative state and evolution into a minimally-conscious state, but its incidence in different studies suggests some unreliability in contrast with its established prognostic relevance. Subjects in vegetative (n=9) or minimally-conscious (n=13) states were tested for visual pursuit 6 times/day (9:30, 10:30, and 11:30 am, and 2:00, 3:00, and 4.00 pm, for a total of 132 determinations). Visual pursuit was observed at all testing times in 8 minimally-conscious patients, and never in 5 subjects in a vegetative state. Its incidence per subject ranged from 50-100% of testing times in the minimally-conscious state (83±23%), and 0-33% in a vegetative state (7%±12), with spontaneous fluctuations during the day and maximal levels at 10.30 am and 3.00 pm, and was never observed at the post-prandial time point (2:00 pm). The overall chance of observing visual tracking at least once during the day was ∼33% in the vegetative state, whereas that of not observing it in the minimally-conscious state was ∼38%. These percentages are congruent with the reported misdiagnosis rate between the two conditions, and document spontaneous variability possibly related to circadian rhythms.

Ultrasound discloses entheseal involvement in inactive and low active inflammatory bowel disease without clinical signs and symptoms of spondyloarthropathy.

OBJECTIVE: To investigate the presence of lower limb entheseal abnormalities in IBD patients without clinical signs and symptoms of SpA and their correlation with IBD clinical variables. METHODS: A total of 81 IBD patients [55 Crohn's disease (CD) and 26 ulcerative colitis (UC), 43 females and 38 males, mean age 41.3 (12.4) years, BMI 24 (2)] with low active (12) and inactive (67) disease were consecutively studied with US (LOGIQ5 General Electric 10-MHz linear array transducer) of lower limb entheses and compared with 40 healthy controls matched for sex, age and BMI. Quadriceps, patellar, Achilleon and plantar fascia entheses were scored according to the 0-36 Glasgow Ultrasound Enthesitis Scoring System (GUESS) and power Doppler (PD). Correlations of GUESS and PD with IBD features [duration, type (CD/UC) and activity (disease activity index for CD/Truelove score for UC)] were investigated. The intra- and inter-reader agreements for US were estimated in all images detected in patients and controls. RESULTS: Of the 81 patients, 71 (92.6%) presented almost one tendon alteration with mean GUESS 5.1 (3.5): 81.5% thickness (higher than controls P < 0.05), 67.9% enthesophytosis, 27.1% bursitis and 16.1% erosions. PD was positive in 13/81 (16%) patients. In controls, US showed only enthesophytes (5%) and no PD. GUESS and PD were independent of duration, activity or type (CD/UC) of IBD. The intra- and inter-reader agreements were high (>0.9 intra-class correlation variability). CONCLUSIONS: US entheseal abnormalities are present in IBD patients without clinical signs and symptoms of SpA. US enthesopathy is independent of activity, duration and type of gut disease.

Visual pursuit in the severe disorder of consciousness.

Visual pursuit is a key descriptor of the minimally conscious state (above 80% of cases). It is also observable in about 20% of subjects in vegetative state. Its reappearance after severe brain damage anticipates a favorable outcome, with recovery of consciousness in 73% of subjects (45% in the absence of it). We considered retrospectively 395 subjects in vegetative state because of traumatic (63%), massive acute vascular (30%), or diffuse anoxic-hypoxic (7%) brain damage consecutively admitted to one dedicated unit during the years 1998-2008. Visual tracking was observed in 290 subjects (73.4%) and was already detectable within 50 days from brain injury in about 60% of post-traumatic or vascular subjects and 21% of anoxic-hypoxic patients. After 230 days of follow-up or more, it was observed in 89% and 88% of post-traumatic and vascular subjects and in 67% of anoxic-hypoxic patients. Rating with the Glasgow Outcome Scale (GOS) was better in those subjects with recovered visual tracking and inversely correlated with the time of reappearance in post-traumatic and vascular subjects; also the subjects with late recovery of eye tracking (230 days or more) had better GOS outcome than those without it. The observation of visual tracking reappearing in subjects in vegetative state would reflect recuperation of the brainstem-cortical interaction and overall brain functional organization that are thought to sustain consciousness and are interfered with by the "functional disconnection," resulting in the vegetative state.

Digital ulcers in scleroderma: staging, characteristics and sub-setting through observation of 1614 digital lesions.

OBJECTIVE: To evaluate in SSc, the frequency of digital lesions and the morphology, characteristics, natural course and time to healing of 1614 digital ulcers (DUs). METHODS: One hundred SSc patients were followed up for 4 years. In the first step, the digital lesions were observed and classified at the time of presentation [digital pitting scar (DPS); DU; calcinosis; gangrene]. In the second step, DUs were divided into subsets according to their origin and main features. In the third step, the time to healing was recorded for each DU and the influence of DU main characteristics on time to healing was also evaluated. RESULTS: In the first step, 1614 digital lesions were observed: DPS, 712 (44.1%) lesions; DU, 785 (48.6%); calcinosis, 110 (6.8%); and gangrene, 7 (0.8%). In the second step, DUs were subsetted as follows: DU developed on DPS (8.8%), pure DU; DU developed on calcinosis (60%); DU derived from gangrene. In the third step, the mean time to healing was 25.6 (15.6) days in DPS, 76.2 (64) days in pure DU, 93.6 (59.2) days in calcinosis ulcers and 281.1 (263.3) in gangrene. CONCLUSIONS: In SSc, digital lesions are represented by DPS, DU, calcinosis and gangrene, and provide an evidence-based DU subsetting according to their origin and main characteristics. Subsetting may be helpful for a precise DU evaluation and staging, and in randomized controlled trials for a precise identification of those DUs that are to be included in therapeutic studies.

A Hyper-Solution Framework for SVM Classification: Application for Predicting Destabilizations in Chronic Heart Failure Patients.

Support Vector Machines (SVMs) represent a powerful learning paradigm able to provide accurate and reliable decision functions in several application fields. In particular, they are really attractive for application in medical domain, where often a lack of knowledge exists. Kernel trick, on which SVMs are based, allows to map non-linearly separable data into potentially linearly separable one, according to the kernel function and its internal parameters value. During recent years non-parametric approaches have also been proposed for learning the most appropriate kernel, such as linear combination of basic kernels. Thus, SVMs classifiers may have several parameters to be tuned and their optimal values are usually difficult to be identified a-priori. Furthermore, combining different classifiers may reduce risk to perform errors on new unseen data. For such reasons, we present an hyper-solution framework for SVM classification, based on meta-heuristics, that searches for the most reliable hyper-classifier (SVM with a basic kernel, SVM with a combination of kernel, and ensemble of SVMs), and for its optimal configuration. We have applied the proposed framework on a critical and quite complex issue for the management of Chronic Heart Failure patient: the early detection of decompensation conditions. In fact, predicting new destabilizations in advance may reduce the burden of heart failure on the healthcare systems while improving quality of life of affected patients. Promising reliability has been obtained on 10-fold cross validation, proving our approach to be efficient and effective for an high-level analysis of clinical data.

High frequency ultrasound measurement of digital dermal thickness in systemic sclerosis.

BACKGROUND: Currently, assessment of dermal thickness in systemic sclerosis (SSc) is performed by palpation and assessment using the modified Rodnan skin score (mRSS). OBJECTIVE: To verify whether high frequency ultrasound (US) may be a reliable and a reproducible method to measure digital dermal thickness. METHODS: In 70 patients with SSc, skin thickness was evaluated with US by 2 observers at 2 different sites on the second digit of the dominant limb to determine the interobserver variability. Patients and controls were examined twice by the first observer for intraobserver variability. Patients were divided into three subgroups according to the phase of the disease (oedematous, fibrotic or atrophic). RESULTS: At both examined areas, US showed a significant dermal thickening (p<0.001) in the whole group of patients with SSc. A low intraobserver and interobserver variability was found. A highly significant correlation between the global mRSS and the local dermal thickness at the two examined sites (p=0.032, p=0.021) was detected. Skin thickness was significantly higher in the oedematous than in the fibrotic group (p<0.001) and significantly higher in the fibrotic and the oedematous group (p<0.001) than in the atrophic group (p<0.002). CONCLUSIONS: US is a reliable tool giving reproducible results, and is able to detect digital dermal thickening in SSc.