In Journal of medical imaging (Bellingham, Wash.)

PURPOSE : To bypass manual data preprocessing and optimize deep learning performance, we developed and evaluated CROPro, a tool to standardize automated cropping of prostate magnetic resonance (MR) images.

APPROACH : CROPro enables automatic cropping of MR images regardless of patient health status, image size, prostate volume, or pixel spacing. CROPro can crop foreground pixels from a region of interest (e.g., prostate) with different image sizes, pixel spacing, and sampling strategies. Performance was evaluated in the context of clinically significant prostate cancer (csPCa) classification. Transfer learning was used to train five convolutional neural network (CNN) and five vision transformer (ViT) models using different combinations of cropped image sizes ( 64 × 64 , 128 × 128 , and 256 × 256  pixels²), pixel spacing ( 0.2 × 0.2 , 0.3 × 0.3 , 0.4 × 0.4 , and 0.5 × 0.5    mm 2 ), and sampling strategies (center, random, and stride cropping) over the prostate. T2-weighted MR images ( N = 1475 ) from the online available PI-CAI challenge were used to train ( N = 1033 ), validate ( N = 221 ), and test ( N = 221 ) all models.

RESULTS : Among CNNs, SqueezeNet with stride cropping (image size: 128 × 128 , pixel spacing: 0.2 × 0.2    mm 2 ) achieved the best classification performance ( 0.678 ± 0.006 ). Among ViTs, ViT-H/14 with random cropping (image size: 64 × 64 and pixel spacing: 0.5 × 0.5    mm 2 ) achieved the best performance ( 0.756 ± 0.009 ). Model performance depended on the cropped area, with optimal size generally larger with center cropping ( ∼ 40    cm 2 ) than random/stride cropping ( ∼ 10    cm 2 ).

CONCLUSION : We found that csPCa classification performance of CNNs and ViTs depends on the cropping settings. We demonstrated that CROPro is well suited to optimize these settings in a standardized manner, which could improve the overall performance of deep learning models.

Patsanis Alexandros, Sunoqrot Mohammed R S, Bathen Tone F, Elschot Mattijs

2023-Mar

deep learning, image cropping, image processing, magnetic resonance imaging, prostate cancer

In Frontiers in psychology ; h5-index 92.0

OBJECTIVE : Given growing interest in companion robots to mitigate loneliness, large-scale studies are needed to understand peoples' perspectives on the use of robots to combat loneliness and attendant ethical issues. This study examines opinions about artificial companion (AC) robots regarding deception with dementia and impact on loneliness.

METHODS : Data are from a survey of 825 members of the OHSU Research via Internet Technology and Experience cohort (response rate = 45%). Sixty percent (n = 496) of the age diverse sample (range = 25-88; M = 64; SD = 13.17) is over 64, allowing us to compare across age and consider current and future older adults. Ordinal logistic regressions examined relationships between age, health, and other socio-demographic characteristics and perceptions of impact on loneliness and comfort with deception.

RESULTS : Most participants (68.7%) did not think an AC robot would make them feel less lonely and felt somewhat-to-very uncomfortable (69.3%) with the idea of being allowed to believe that an artificial companion is human. In adjusted models, one additional year of age was associated with lower likelihood of perceived benefit of reducing loneliness [Odds Ratio (OR) = 0.98; (0.97-0.99), p = 0.003] and lower comfort with deception [OR = 0.99; (0.97-1.00), p = 0.044]. Being female was associated with lower likelihood of comfort with deception [OR = 0.68; (0.50-0.93), p = 0.014] and high confidence using computers with greater comfort [OR = 2.18; (1.42-3.38), p < 0.001].

DISCUSSION : There was not strong support for AC robots to mitigate loneliness. Most participants were uncomfortable with this form of deception, indicating need for design solutions for those who want to avoid this possibility, as well as greater attentiveness to desirability and comfort across age and gender.

Berridge Clara, Zhou Yuanjin, Robillard Julie M, Kaye Jeffrey

2023

artificial intelligence, dementia, ethics, natural language processesing, robotics

In Frontiers in physiology

Lecce E, Nuccio S, Del Vecchio A, Conti A, Nicolò A, Sacchetti M, Felici F, Bazzucchi I

2023

HDsEMG, MU recruitment, WBV, muscle strength, neuromuscular control

In Frontiers in immunology ; h5-index 100.0

BACKGROUND : We developed a novel system for quantifying DNA damage response (DDR) to help diagnose and predict the risk of Alzheimer's disease (AD).

METHODS : We thoroughly estimated the DDR patterns in AD patients Using 179 DDR regulators. Single-cell techniques were conducted to validate the DDR levels and intercellular communications in cognitively impaired patients. The consensus clustering algorithm was utilized to group 167 AD patients into diverse subgroups after a WGCNA approach was employed to discover DDR-related lncRNAs. The distinctions between the categories in terms of clinical characteristics, DDR levels, biological behaviors, and immunological characteristics were evaluated. For the purpose of choosing distinctive lncRNAs associated with DDR, four machine learning algorithms, including LASSO, SVM-RFE, RF, and XGBoost, were utilized. A risk model was established based on the characteristic lncRNAs.

RESULTS : The progression of AD was highly correlated with DDR levels. Single-cell studies confirmed that DDR activity was lower in cognitively impaired patients and was mainly enriched in T cells and B cells. DDR-related lncRNAs were discovered based on gene expression, and two different heterogeneous subtypes (C1 and C2) were identified. DDR C1 belonged to the non-immune phenotype, while DDR C2 was regarded as the immune phenotype. Based on various machine learning techniques, four distinctive lncRNAs associated with DDR, including FBXO30-DT, TBX2-AS1, ADAMTS9-AS2, and MEG3 were discovered. The 4-lncRNA based riskScore demonstrated acceptable efficacy in the diagnosis of AD and offered significant clinical advantages to AD patients. The riskScore ultimately divided AD patients into low- and high-risk categories. In comparison to the low-risk group, high-risk patients showed lower DDR activity, accompanied by higher levels of immune infiltration and immunological score. The prospective medications for the treatment of AD patients with low and high risk also included arachidonyltrifluoromethane and TTNPB, respectively.

CONCLUSIONS : In conclusion, immunological microenvironment and disease progression in AD patients were significantly predicted by DDR-associated genes and lncRNAs. A theoretical underpinning for the individualized treatment of AD patients was provided by the suggested genetic subtypes and risk model based on DDR.

Lai Yongxing, Lin Han, Chen Manli, Lin Xin, Wu Lijuan, Zhao Yinan, Lin Fan, Lin Chunjin

2023

Alzheimer’s disease, DNA damage response, immunity, machine learning, molecular subtypes, single-cell

In Journal of inflammation research

BACKGROUND : Since studies on systemic inflammatory response syndrome (SIRS) in patients with acute intracerebral hemorrhage (ICH) are insufficient. This study investigated the associations between SIRS on admission and clinical outcomes after acute ICH.

PATIENTS AND METHODS : The study included 1159 patients with acute spontaneous ICH from January 2014 to September 2016. In accordance with standard criteria, SIRS was defined as two or more of the following: (1) body temperature >38°C or <36°C, (2) respiratory rate >20 per minute, (3) heart rate >90 per minute, and (4) white blood cell count >12,000/μL or <4000/μL. The clinical outcomes of interest were death and major disability (defined as a modified Rankin Scale of 6 and 3-5), combined and separate at 1 month, 3 months and 1 year follow-up.

RESULTS : SIRS was observed in 13.5% (157/1159) of patients and independently increased the risk of death at 1 month, 3 months, or 1 year: hazard ratio (HR) 2.532 (95% confidence interval [CI] 1.487-4.311), HR 2.436 (95% CI 1.499-3.958), HR 2.030 (95% CI 1.343-3.068), respectively (P<0.05 for all). The relationship between SIRS and ICH mortality was more pronounced in older patients or patients with larger hematoma volumes. Patients with in-hospital infections were at greater risk of major disability. The risk was enhanced when SIRS was incorporated.

CONCLUSION : The presence of SIRS at the time of admission was associated with mortality in patients with acute ICH, particularly in older patients and those with large hematomas. SIRS may exacerbate the disability caused by in-hospital infections in patients with ICH.

Liu Lijun, Wang Anxin, Wang Dandan, Guo Jiahuan, Zhang Xiaoli, Zhao Xingquan, Wang Wenjuan

2023

death, disability, in-hospital infections, intracerebral hemorrhage, systemic inflammatory response syndrome

In Biomedical engineering online

BACKGROUND : Exercise electrocardiography (ECG) is a noninvasive test aiming at producing ischemic changes. However, resting ECG cannot be adopted in diagnosing myocardial ischemia till ST-segment depressions. Therefore, this study aimed to detect myocardial energy defects in resting ECG using the Hilbert-Huang transformation (HHT) in patients with angina pectoris.

METHODS : Electrocardiographic recordings of positive exercise ECG by performing coronary imaging test (n = 26) and negative exercise ECG (n = 47) were collected. Based on the coronary stenoses severity, patients were divided into three categories: normal, < 50%, and ≥ 50%. During the resting phase of the exercise ECG, all 10-s ECG signals are decomposed by HHT. The RT intensity index, composed of the power spectral density of the P, QRS, and T components, is used to estimate the myocardial energy defect.

RESULTS : After analyzing the resting ECG using HHT, the RT intensity index was significantly higher in patients with positive exercise ECG (27.96%) than in those with negative exercise ECG (22.30%) (p < 0.001). In patients with positive exercise ECG, the RT intensity index was gradually increasing with the severity of coronary stenoses: 25.25% (normal, n = 4), 27.14% (stenoses < 50%, n = 14), and 30.75% (stenoses ≥ 50%, n = 8). The RT intensity index of different coronary stenoses was significantly higher in patients with negative exercise ECG, except for the normal coronary imaging test.

CONCLUSIONS : Patients with coronary stenoses had a higher RT index at the resting stage of exercise ECG. Resting ECG analyzed using HHT could be a method for the early detection of myocardial ischemia.

Wang Chun-Lin, Wei Chiu-Chi, Tsai Cheng-Ting, Lee Ying-Hsiang, Liu Lawrence Yu-Min, Chen Kang-Ying, Lin Yu-Jen, Lin Po-Lin

2023-Mar-10

Exercise electrocardiography, Hilbert–Huang transform, Myocardial energy, Power spectral density