Deep learning-guided adjuvant chemotherapy selection for elderly patients with breast cancer.

PURPOSE: The efficacy of adjuvant chemotherapy in elderly breast cancer patients is currently controversial. This study aims to provide personalized adjuvant chemotherapy recommendations using deep learning (DL). METHODS: Six models with various causal inference approaches were trained to make individualized chemotherapy recommendations. Patients who received actual treatment recommended by DL models were compared with those who did not. Inverse probability treatment weighting (IPTW) was used to reduce bias. Linear regression, IPTW-adjusted risk difference (RD), and SurvSHAP(t) were used to interpret the best model. RESULTS: A total of 5352 elderly breast cancer patients were included. The median (interquartile range) follow-up time was 52 (30-80) months. Among all models, the balanced individual treatment effect for survival data (BITES) performed best. Treatment according to following BITES recommendations was associated with survival benefit, with a multivariate hazard ratio (HR) of 0.78 (95% confidence interval (CI): 0.64-0.94), IPTW-adjusted HR of 0.74 (95% CI: 0.59-0.93), RD of 12.40% (95% CI: 8.01-16.90%), IPTW-adjusted RD of 11.50% (95% CI: 7.16-15.80%), difference in restricted mean survival time (dRMST) of 12.44 (95% CI: 8.28-16.60) months, IPTW-adjusted dRMST of 7.81 (95% CI: 2.93-11.93) months, and p value of the IPTW-adjusted Log-rank test of 0.033. By interpreting BITES, the debiased impact of patient characteristics on adjuvant chemotherapy was quantified, which mainly included breast cancer subtype, tumor size, number of positive lymph nodes, TNM stages, histological grades, and surgical type. CONCLUSION: Our results emphasize the potential of DL models in guiding adjuvant chemotherapy decisions for elderly breast cancer patients.

Development and validation of a clinical nomogram for differentiating hemorrhagic and ischemic stroke prehospital.

OBJECTIVES: The early detection and identification of stroke are essential to the prognosis of patients with suspected stroke symptoms out-of-hospital. We aimed to develop a risk prediction model based on the FAST score to identify the different types of strokes early for emergency medical services (EMS). METHODS: This retrospective observational study enrolled 394 stroke patients at a single center from January 2020 to December 2021. Demographic data, clinical characteristics, and stroke risk factors with patients were collected from the EMS record database. Univariate and multivariate logistic regression analysis was used to identify the independent risk predictors. The nomogram was developed based on the independent predictors, in which the discriminative value and calibration of the nomogram were verified by the receiver operator characteristic (ROC) curve and calibration plots. RESULTS: A total of 31.90% (88/276) of patients were diagnosed with hemorrhagic stroke in the training set, while 36.40% (43/118) in the validation set. The nomogram was developed based on the multivariate analysis, including age, systolic blood pressure, hypertension, vomiting, arm weakness, and slurred speech. The area under the curve (AUC) of the ROC with nomogram was 0.796 (95% CI: 0.740-0.852, P < 0.001) and 0.808 (95% CI:0.728-0.887, P < 0.001) in the training set and validation set, respectively. In addition, the AUC with the nomogram was superior to the FAST score in both two sets. The calibration curve showed a good agreement with the nomogram and the decision curves analysis also demonstrated that the nomogram had a wider range of threshold probabilities than the FAST score in the prediction risk of hemorrhagic stroke. CONCLUSIONS: This novel noninvasive clinical nomogram shows a good performance in differentiating hemorrhagic and ischemic stroke for EMS staff prehospital. Moreover, all of the variables of nomogram are acquired in clinical practice easily and inexpensively out-of-hospital.

Quantified treatment effect at the individual level is more indicative for personalized radical prostatectomy recommendation: implications for prostate cancer treatment using deep learning.

BACKGROUND: There are potential uncertainties and overtreatment existing in radical prostatectomy (RP) for prostate cancer (PCa) patients, thus identifying optimal candidates is quite important. PURPOSE: This study aims to establish a novel causal inference deep learning (DL) model to discern whether a patient can benefit more from RP and to identify heterogeneity in treatment responses among PCa patients. METHODS: We introduce the Self-Normalizing Balanced individual treatment effect for survival data (SNB). Six models were trained to make individualized treatment recommendations for PCa patients. Inverse probability treatment weighting (IPTW) was used to avoid treatment selection bias. RESULTS: 35,236 patients were included. Patients whose actual treatment was consistent with SNB recommendations had better survival outcomes than those who were inconsistent (multivariate hazard ratio (HR): 0.76, 95% confidence interval (CI), 0.64-0.92; IPTW-adjusted HR: 0.77, 95% CI, 0.61-0.95; risk difference (RD): 3.80, 95% CI, 2.48-5.11; IPTW-adjusted RD: 2.17, 95% CI, 0.92-3.35; the difference in restricted mean survival time (dRMST): 3.81, 95% CI, 2.66-4.85; IPTW-adjusted dRMST: 3.23, 95% CI, 2.06-4.45). Keeping other covariates unchanged, patients with 1 ng/mL increase in PSA levels received RP caused 1.77 months increase in the time to 90% mortality, and the similar results could be found in age, Gleason score, tumor size, TNM stages, and metastasis status. CONCLUSIONS: Our highly interpretable and reliable DL model (SNB) may identify patients with PCa who could benefit from RP, outperforming other models and clinical guidelines. Additionally, the DL-based treatment guidelines obtained can provide priori evidence for subsequent studies.

Development and validation of a prediction nomogram for academic burnout among Chinese adolescents: a cross-sectional study.

OBJECTIVES: This study aimed to screen the potential risk factors for academic burnout among adolescents during the COVID-19 pandemic, develop and validate a predictive tool based on the risk factors for predicting academic burnout. DESIGN: This article presents a cross-sectional study. SETTING: This study surveyed two high schools in Anhui Province, China. PARTICIPANTS: A total of 1472 adolescents were enrolled in this study. OUTCOME MEASURES: The questionnaires included demographic characteristic variables, living and learning states and adolescents' academic burnout scale. Least absolute shrinkage and selection operator and multivariate logistic regression analyses were employed to screen the risk factors for academic burnout and develop a predictive model. The receiver operating characteristic (ROC) curve and decision curve analysis (DCA) were used to assess the accuracy and discrimination of the nomogram. RESULTS: In this study, 21.70% of adolescents reported academic burnout. Multivariable logistic regression analysis showed that single-child family (OR=1.742, 95% CI: 1.243 to 2.441, p=0.001), domestic violence (OR=1.694, 95% CI: 1.159 to 2.476, p=0.007), online entertainment (>8 hours/day, OR=3.058, 95% CI: 1.634 to 5.720, p<0.001), physical activity (<3 hours/week, OR=1.686, 95% CI: 1.032 to 2.754, p=0.037), sleep duration (<6 hours/night, OR=2.342, 95% CI: 1.315 to 4.170, p=0.004) and academic performance (<400 score, OR=2.180, 95% CI: 1.201 to 3.958, p=0.010) were independent significant risk factors associated with academic burnout. The area under the curve of ROC with the nomogram was 0.686 in the training set and 0.706 in the validation set. Furthermore, DCA demonstrated that the nomogram had good clinical utility for both sets. CONCLUSIONS: The developed nomogram was a useful predictive model for academic burnout among adolescents during the COVID-19 pandemic. It is essential to emphasise the importance of mental health and promote a healthy lifestyle among adolescents during the future pandemic.

Comparative efficacy of different eating patterns in the management of type 2 diabetes and prediabetes: An arm-based Bayesian network meta-analysis.

AIMS/INTRODUCTION: Diet therapy is a vital approach to manage type 2 diabetes and prediabetes. However, the comparative efficacy of different eating patterns is not clear enough. We aimed to compare the efficacy of various eating patterns for glycemic control, anthropometrics, and serum lipid profiles in the management of type 2 diabetes and prediabetes. MATERIALS AND METHODS: We conducted a network meta-analysis using arm-based Bayesian methods and random effect models, and drew the conclusions using the partially contextualized framework. We searched twelve databases and yielded 9,534 related references, where 107 studies were eligible, comprising 8,909 participants. RESULTS: Eleven diets were evaluated for 14 outcomes. Caloric restriction was ranked as the best pattern for weight loss (SUCRA 86.8%) and waist circumference (82.2%), low-carbohydrate diets for body mass index (81.6%), and high-density lipoprotein (84.0%), and low-glycemic-index diets for total cholesterol (87.5%) and low-density lipoprotein (86.6%). Other interventions showed some superiorities, but were imprecise due to insufficient participants and needed further investigation. The attrition rates of interventions were similar. Meta-regression suggested that macronutrients, energy intake, and weight may modify outcomes differently. The evidence was of moderate-to-low quality, and 38.2% of the evidence items met the minimal clinically important differences. CONCLUSIONS: The selection and development of dietary strategies for diabetic/prediabetic patients should depend on their holistic conditions, i.e., serum lipid profiles, glucometabolic patterns, weight, and blood pressure. It is recommended to identify the most critical and urgent metabolic indicator to control for one specific patient, and then choose the most appropriate eating pattern accordingly.

High serum amyloid A predicts risk of cognitive impairment after lacunar infarction: Development and validation of a nomogram.

Background: Post-stroke cognitive impairment (PSCI) after lacunar infarction was worth attention in recent years. An easy-to-use score model to predict the risk of PSCI was rare. This study aimed to explore the association between serum amyloid A (SAA) and cognitive impairment, and it also developed a nomogram for predicting the risk of PSCI in lacunar infarction patients. Methods: A total of 313 patients with lacunar infarction were enrolled in this retrospective study between January 2021 and December 2021. They were divided into a training set and a validation set at 70%:30% randomly. The Chinese version of the Mini-Mental State Examination (MMSE) was performed to identify cognitive impairment 3 months after discharge. Univariate and multivariate logistic regression analyses were used to determine the independent risk factors for PSCI in the training set. A nomogram was developed based on the five variables, and the calibration curve and the receiver operating characteristic (ROC) curve were drawn to assess the predictive ability of the nomogram between the training set and the validation set. The decision curve analysis (DCA) was also conducted in both sets. Results: In total, 52/313 (16.61%) participants were identified with PSCI. The SAA levels in patients with PSCI were significantly higher than non-PSCI patients in the training set (P < 0.001). After multivariate analysis, age, diabetes mellitus, white blood count, cystatin C, and SAA were independent risk predictors of PSCI. The nomogram demonstrated a good discrimination performance between the training set (AUC = 0.860) and the validation set (AUC = 0.811). The DCA showed that the nomogram had a well clinical utility in the two sets. Conclusion: The increased SAA is associated with PSCI in lacunar infarction patients, and the nomogram developed with SAA can increase prognostic information for the early detection of PSCI.

A novel algorithm for implementing time-frequency transform with low computation.

In this paper, a novel algorithm called two-dimensional sliding fast Fourier transform (2D SFFT) algorithm is proposed. This algorithm organizes one-dimensional data in two dimensions and calculates the spectrum of current data by using the existing spectrum and new collected data. The algorithm formula and accurate simulation results show the following: first, the computation required by the proposed 2D SFFT algorithm is lower than that required by the traditional sliding discrete Fourier transform algorithm when the sliding rate is larger than or equal to 4/M, where M is the sequence length. Moreover, the computation required by the proposed 2D SFFT algorithm is lower than that required by the fast Fourier transform (FFT) algorithm when the sliding rate is less than or equal to 6.25%. Finally, the error between the spectrum calculated by the 2D SFFT and FFT algorithms is less than 10-10. The 2D SFFT algorithm is used to increase the power of the ultra-short pulse, which is initially invisible in the frequency-domain window of the mixed-domain oscilloscope. Therefore, the 100% probability of intercept of the mixed-domain oscilloscope is lower.

Keratoacanthoma centrifugum marginatum: An unusual case.

Keratoacanthoma centrifugum marginatum (KCM) of the skin is a rare variant of cutaneous keratoacantoma. KCM was first reported in 1962 andpresents with progressive peripheral expansion , no spontaneous clearing and a bank-shaped outer wall with concurrent central healing. Treatment options include topical and systemic therapies.Surgical intervention is the preferred therapy for solitary KCM. We report on surgery and photodynamic therapy delivered sequentially to treat a giant facial Keratoacanthoma centrifugum marginatum patient. It was safe and effective .

A Case of Surgical Removal of a Giant and Multiple Acquired Plate-Like Osteoma Cutis After Implantation of a Dilator.

Osteoma cutis (OC) is a group of rare skin ossification diseases, most of which are secondary to inflammation, scarring, trauma, or tumors, but a small portion are primary. Plate-like osteoma cutis is rare, especially after puberty. This report documents a case of a 30-year-old female, who presented with multiple stone-hard plates on the forehead and bilateral temples, with no relevant family history, or abnormalities in metabolism. These lesions showed slow progression over the last 11 years. The pathological diagnosis confirmed osteoma cutis. The forehead lesions were treated surgically due to aesthetic problems. In addition, long-term follow-up and observations are still needed to determine progression to deeper levels of tissue.

Human Schlafen 5 Inhibits Proliferation and Promotes Apoptosis in Lung Adenocarcinoma via the PTEN/PI3K/AKT/mTOR Pathway.

BACKGROUND: Human Schlafen 5 (SLFN5) is reported to inhibit or promote the proliferation of several specific types of cancer cells by our lab and other researchers. We are curious about its implications in lung adenocarcinoma (LUAC), a malignant tumor with a high incidence rate and high mortality. METHOD: Lentiviral stable transfections of SLFN5-specific shRNA for knockdown and SLFN5 full-length coding sequence for overexpression were performed in LUAC cell for proliferation analysis in vitro and in vivo in nude mice. Clinical LUAC samples were collected for immunohistochemical analysis of SLFN5 protein levels. RESULTS: We found that knockdown of endogenous SLFN5 upregulates cancer cell proliferation while inhibiting apoptosis. Besides, SLFN5 inhibition on proliferation was also observed in a nude mouse xenograft model. In contrast, overexpression of exogenous SLFN5 inhibited cell proliferation in vitro and in vivo and promoted apoptosis. As to the signaling pathway, we found phosphatase and tensin homolog on chromosome 10 (PTEN) was positively regulated by SLFN5, while its downstream signaling pathway AKT/mammalian target of rapamycin (mTOR) was inhibited. Moreover, compared with adjacent normal tissues, SLFN5 protein levels were markedly decreased in lung adenocarcinoma tissues. In conclusion, these suggest that human SLFN5 plays inhibitory roles in LUAC progression through the PTEN/PI3K/AKT/mTOR pathway, providing a potential target for developing drugs for lung cancer therapy in the future.

Real-time self-adaptive calibration method for high speed acquisition system.

The main factors that enable capture of complex and transient signals in real-time are improved sampling rates and processing speeds. The time-interleaved architecture is an effective method that allows systems to break through the speed bottleneck of single analog-to-digital converters (ADCs) and go beyond the state-of-the-art process technology limit. However, the performance of the acquisition system may be reduced because of the offset, gain, and time mismatch errors that occur in time-interleaved ADC systems. To correct these errors, this paper first proposes a self-adaptive correction algorithm and then introduces real-time solutions for this algorithm. Finally, the proposed calibration method is implemented in a digital phosphor oscilloscope. Simulations and experimental testing indicate that this system shows good real-time performance and provides a high dynamic performance with an effective number of bits of 7.3 bits and a signal-to-noise ratio of 45.5574 dB.

Channelized broadband signal spectrum analysis based on weighted overlap-add structure.

The digital channelization technology has been applied in many electronic areas, and the real-time broadband spectrum analysis has been the research hotspot in the area of signal processing. This paper introduces the channelized broadband signal spectrum analysis method. Based on the weighted overlap-add (WOLA) structure, this method divides the input broadband signal into several sub-bands or channels, and then downconverts and decimates the sub-band signals to obtain the baseband signals with a low sampling rate. The spectrum analysis results of the input broadband signal are achieved by conducting further decimation, fast Fourier transform and spectrum splicing to the baseband signals. The Matlab simulation results verify the correctness of the WOLA structure, and finally, an experimental platform is designed in detail to verify the practicability of this broadband spectrum analysis method.