RESUMEN
ObjectiveTo develop a bilateral rehabilitation robot motion assistance strategy based on admittance control, so that rehabilitation physicians can assist patients in rehabilitation training through remote teaching. MethodsA bilateral remote rehabilitation platform with upper limb terminal traction was constructed. Based on the velocity admittance control, the interactive movement between the master robot and the rehabilitation physician was realized, and the position information transmission of the master-slave robot was realized through the communication framework built. The slave robot received the position coordinates of the main robot, and drove the patient to carry out rehabilitation exercises under the attitude admittance controller. ResultsThe robot could drive the patient to accurately track the trajectory of the doctor's teaching in real time, and improve the safety and compliance of the training and human-computer interaction. ConclusionBy introducing two admittance controllers, the trajectory of the physician's end can be accurately tracked when driving the patient's movement from the robotic arm, which effectively avoids the discomfort of the patient's arm in process of rehabilitation.
RESUMEN
Research ObjectiveHealth systems use clinical predictive algorithms to allocate resources to high-risk patients. Such algorithms are trained using historical data and are later implemented in clinical settings. During this implementation period, predictive algorithms are prone to performance changes ("drift") due to exogenous shocks in utilization or shifts in patient characteristics. Our objective was to examine the impact of sudden utilization shifts during the SARS-CoV-2 pandemic on the performance of an electronic health record (EHR)-based prognostic algorithm. Study DesignWe studied changes in the performance of Conversation Connect, a validated machine learning algorithm that predicts 180-day mortality among outpatients with cancer receiving care at medical oncology practices within a large academic cancer center. Conversation Connect generates mortality risk predictions before each encounter using data from 159 EHR variables collected in the six months before the encounter. Since January 2019, Conversation Connect has been used as part of a behavioral intervention to prompt clinicians to consider early advance care planning conversations among patients with [≥]10% mortality risk. First, we descriptively compared encounter-level characteristics in the following periods: January 2019-February 2020 ("pre-pandemic"), March-May 2020 ("early-pandemic"), and June-December 2020 ("later-pandemic"). Second, we quantified changes in high-risk patient encounters using interrupted time series analyses that controlled for pre-pandemic trends and demographic, clinical, and practice covariates. Our primary metric of performance drift was false negative rate (FNR). Third, we assessed contributors to performance drift by comparing distributions of key EHR inputs across periods and predicting later pandemic utilization using pre-pandemic inputs. Population Studied237,336 in-person and telemedicine medical oncology encounters. Principal FindingsAge, race, average patient encounters per month, insurance type, comorbidity counts, laboratory values, and overall mortality were similar among encounters in the pre-, early-, and later-pandemic periods. Relative to the pre-pandemic period, the later-pandemic period was characterized by a 6.5-percentage-point decrease (28.2% vs. 34.7%) in high-risk encounters (p<0.001). FNR increased from 41.0% (95% CI 38.0-44.1%) in the pre-pandemic period to 57.5% (95% CI 51.9-63.0%) in the later pandemic period. Compared to the pre-pandemic period, the early and later pandemic periods had higher proportions of telemedicine encounters (0.01% pre-pandemic vs. 20.0% early-pandemic vs. 26.4% later-pandemic) and encounters with no preceding laboratory draws (17.7% pre-pandemic vs. 19.8% early-pandemic vs. 24.1% later-pandemic). In the later pandemic period, observed laboratory utilization was lower than predicted (76.0% vs 81.2%, p<0.001). In the later-pandemic period, mean 180-day mortality risk scores were lower for telemedicine encounters vs. in-person encounters (10.3% vs 11.2%, p<0.001) and encounters with no vs. any preceding laboratory draws (1.5% vs. 14.0%, p<0.001). ConclusionsDuring the SARS-CoV-2 pandemic period, the performance of a machine learning prognostic algorithm used to prompt advance care planning declined substantially. Increases in telemedicine and declines in laboratory utilization contributed to lower performance. Implications for Policy or PracticeThis is the first study to show algorithm performance drift due to SARS-CoV-2 pandemic-related shifts in telemedicine and laboratory utilization. These mechanisms of performance drift could apply to other EHR clinical predictive algorithms. Pandemic-related decreases in care utilization may negatively impact the performance of clinical predictive algorithms and warrant assessment and possible retraining of such algorithms.
RESUMEN
Coronavirus disease 2019 (COVID-19) has infected more than 1.3 million individuals all over the world and caused more than 106,000 deaths. One major hurdle in controlling the spreading of this disease is the inefficiency and shortage of medical tests. There have been increasing efforts on developing deep learning methods to diagnose COVID-19 based on CT scans. However, these works are difficult to reproduce and adopt since the CT data used in their studies are not publicly available. Besides, these works require a large number of CTs to train accurate diagnosis models, which are difficult to obtain. In this paper, we aim to address these two problems. We build a publicly-available dataset containing hundreds of CT scans positive for COVID-19 and develop sample-efficient deep learning methods that can achieve high diagnosis accuracy of COVID-19 from CT scans even when the number of training CT images are limited. Specifically, we propose a Self-Trans approach, which synergistically integrates contrastive self-supervised learning with transfer learning to learn powerful and unbiased feature representations for reducing the risk of overfitting. Extensive experiments demonstrate the superior performance of our proposed Self-Trans approach compared with several state-of-the-art baselines. Our approach achieves an F1 of 0.85 and an AUC of 0.94 in diagnosing COVID-19 from CT scans, even though the number of training CTs is just a few hundred.
RESUMEN
Objective:To explore the effect of high-frequency and low-frequency transcranial magnetic stimulation (rTMS) on the unaffected pharyngeal motor cortex of dysphagic stroke survivors.Methods:Forty-two stroke survivors with dysphagia were enrolled and randomly divided into a high-frequency stimulation group ( n=14), a low-frequency stimulation group ( n=13), and a sham group ( n=15). All received conventional swallowing training. The high- and low-frequency stimulation groups additionally received 250 pulses of 5Hz or 1Hz rTMS over the cortical representation of the mylohyoid muscle on the unaffected side daily for 2 consecutive weeks. In the sham group, sham rTMS was applied with identical protocols. Before and after the intervention, all subjects were subjected to a videofluoroscopic swallowing study and surface electromyography (sEMG). They were also evaluated using the fuctional dysphagia scale (FDS) and the penetration aspiration scale (PAS). Results:After the intervention, a significant improvement was observed in the average PAS, FDS and sEMG results in both rTMS groups compared with the sham control group. The average FDS score of the high-frequency stimulation group had improved significantly more than that of the low-frequency group.Conclusions:rTMS of the contra-lesional cortical representation of the mylohyoid muscle at either 5Hz or 1Hz can effectively improve dysphagia post-stroke. The higher frequency gives superior results.
RESUMEN
Objective:To observe the effect of contralateral controlled functional electrical stimulation (CCFES) on the recovery of upper limb motor function after a stroke.Methods:Stroke survivors 1 to 6 months after onset were randomly divided into a CCFES group (14 cases, group A), a CCFES intensive group (14 cases, group B) and a neuromuscular electrical stimulation group (15 cases , group C). In addition to routine rehabilitation training, groups A and B received contralateral controlled functional electrical stimulation, while Group C received routine neuromuscular electrical stimulation. Group B was treated twice daily, while the other two groups were given 1 session each day, 5 times a week for 3 weeks. Each session lasted 20 minutes. The Fugl-Meyer assessment (FMA), the Modified Barthel Index (MBI), surface electromyography and the active range of motion for wrist dorsiflexion were used to evaluate the subjects′ upper limb function before and after the treatment.Results:There was no significant difference in any of the measurements among the three groups before the treatment. After the treatment, all of them had improved significantly, with the improvements in groups A and B significantly greater than in group C, on average.Conclusions:Both normal and intensive contralateral controlled functional electrical stimulation have significant advantages over neuromuscular electrical stimulation in promoting functional recovery of the upper limbs. Intensive CCFES training is superior to routine training in improving muscle strength and range of motion.
RESUMEN
Objective:To compare using a rat model of post-stroke depression (PSD) the effect of 28 days of high-intensity interval training (HIIT) with that of medium-intensity continuous training (MICT) on inflammation and neuron apoptosis induced by phosphatase and the deletion of a tensin homolog on chromosome ten (PTEN).Methods:Male Wistar rats were randomly divided into a SHAM, a PSD, an MICT and an HIIT group. Except in the SHAM group, blood flow in the middle cerebral artery was blocked for 90 minutes followed by reperfusion. Repeated but unpredictable mild stimulation was then applied to induce depression. The rats in the HIIT and MICT groups started 28 days of training 24 hours after the successful modeling. The running platform speed was adjusted according to the lactic acid threshold and the maximum speed was measured weekly. In the SHAM group the common carotid artery, internal carotid artery and external carotid artery were only separated without occlusion and there was no depressive stimulation. Any improvements in depression were detected using the sucrose preference test and the forced swimming test. The expression of PTEN, nuclear factor kappa B (NF-κB) and nod-like receptor protein 3 (NLRP3) were detected using western blotting. The expression of cysteine-containing, aspartate-specific proteases (caspase-3) was detected immunohistochemically.Results:Compared with the SHAM group, the PSD group′s average immobility time in the swimming test was significantly longer and its average sucrose consumption was significantly less. Compared with the PSD group, the rats in the MICT and HIIT groups showed, on average, significantly more sucrose consumption and shorter immobility time, indicating that their depression was ameliorated. The expression of PTEN, NF-κB and NLRP3 in the MICT and HIIT groups was, on average decreased significantly compared with the PSD group. The HIIT group′s averages were at the same time significantly lower than those of the MICT group. Average caspase-3 levels in the dentate gyrus of the MICT and HIIT group rats were significantly lower than in the PSD group, with the HIIT group′s average significantly lower than that of the MICT group.Conclusions:Both high- and medium-intensity interval training show neuroprotective effects. They inhibit activation of the PTEN/NF-κB/NLRP3 pathway, reducing the expression of apoptotic proteins in the hippocampus. Higher intensity training has a more obvious antidepressant effect.
RESUMEN
Objective To evaluate the effect of intensive electromyographic biofeedback ( EMGBF ) on swallowing and the velocity of hyolaryngeal excursion among stroke survivors with dysphagia. Methods Forty-two stroke survivors with dysphagia were randomly divided into a conventional rehabilitation group ( group A, n=15) , an EMGBF group (group B, n=14) and an intensive EMGBF group (group C, n=13).In addition to routine medica-tions and motor function rehabilitation training, all received 30 minutes of conventional swallowing training once a day, 5 days per week for 4 weeks. Group B additionally received 15 minutes of EMGBF once daily and group C twice daily. Videofluoroscopy was conducted to measure the superior and anterior excursion distances and the movement time of the hyoid bone when swallowing semi-liquid food, and the velocity was calculated. The swallowing dysfunction evaluation and a penetration-aspiration scale ( PAS) were also employed to evaluate the subjects'swallowing function before and after the treatment. Results There were no significant differences among the 3 groups in any of the measurements before the treatment. After the 4 weeks of treatment the average swallowing dysfunction evaluation and PAS scores of all three groups had improved significantly, as had the superior and anterior excursion velocity of the hyoid bone. Compared with group A, the average swallowing dysfunction evaluation and PAS scores of groups B and C had improved significantly more. The average swallowing dysfunction evaluation scores of groups B and C were not sig-nificantly different, but the average anterior and superior excursion velocity of the hyoid bone in group C was signifi-cantly higher than in group B. Conclusion EMGBF therapy has advantages over routine rehabilitation training in improving swallowing function after a stroke and speeding the velocity of the hyoid bone's excursions.
RESUMEN
Recently, with the research progress in molecular classification, the treatment of advanced non-small cell lung cancer (NSCLC) has been established as a model of anti-tumor treatment of precision medicine. The discovery of epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKI) has transformed the treatment of NSCLC from platinum based doublet chemotherapy into era of target therapy. EGFR-TKI, such as erlotinib and gefitinib, have been recommended as standard first-line treatment of patients with EGFR mutation. However, acquired resistance, defined as tumor progression after initial response, seems to be an inevitable consequence of this treatment approach. Clinical modes of EGFR-TKI failure are classified into three types: dramatic progression, gradual progression and local progression. A threonine-to-methionine substitution (T790M) in exon 20 of the EGFR gene is the most common mechanism of resistance. Other mechanisms of resistance include MET amplification, epithelial to mesenchymal transition, small cell transformation, and PIK3CA mutation. This brief comment will provide an overview of the complex and heterogeneous problem of acquired resistance to EGFR-TKI therapy in NSCLC, and the clinical treatment options and new targeted drugs overcoming EGFR-TKI acquired resistance.