BACKGROUND:Despite spinal cord stimulation's (SCS) proven efficacy, failure rates are high with no clear understanding of which patients benefit long term. Currently, patient selection for SCS is based on the subjective experience of the implanting physician.OBJECTIVE:To develop machine learning (ML)–based predictive models of long-term SCS response.METHODS:A combined unsupervised (clustering) and supervised (classification) ML technique was applied on a prospectively collected cohort of 151 patients, which included 31 features. Clusters identified using unsupervised K-means clustering were fitted with individualized predictive models of logistic regression, random forest, and XGBoost.RESULTS:Two distinct clusters were found, and patients in the cohorts significantly differed in age, duration of chronic pain, preoperative numeric rating scale, and preoperative pain catastrophizing scale scores. Using the 10 most influential features, logistic regression predictive models with a nested cross-validation demonstrated the highest overall performance with the area under the curve of 0.757 and 0.708 for each respective cluster.CONCLUSION:This combined unsupervised–supervised learning approach yielded high predictive performance, suggesting that advanced ML-derived approaches have potential to be used as a functional clinical tool to improve long-term SCS outcomes. Further studies are needed for optimization and external validation of these models.

Given that the Ebola virus (EBOV) infects a wide array of organs and cells yet displays a relative lack of neurotropism, we asked whether a chimeric vesicular stomatitis virus (VSV) expressing the EBOV glycoprotein (GP) might selectively target brain tumors. The mucin-like domain (MLD) of the EBOV GP may enhance virus immune system evasion. Here, we compared chimeric VSVs in which EBOV GP replaces the VSV glycoprotein, thereby reducing the neurotoxicity associated with wild-type VSV. A chimeric VSV expressing the full-length EBOV GP (VSV-EBOV) containing the MLD was substantially more effective and safer than a parallel construct with an EBOV GP lacking the MLD (VSV-EBOVΔMLD). One-step growth, reverse transcription-quantitative PCR, and Western blotting assessments showed that VSV-EBOVΔMLD produced substantially more progeny faster than VSV-EBOV. Using immunodeficient SCID mice, we focused on targeting human brain tumors with these VSV-EBOVs. Similar to the findings of our previous study in which we used an attenuated VSV-EBOV with no MLD that expressed green fluorescent protein (GFP) (VSV-EBOVΔMLD-GFP), VSV-EBOVΔMLD without GFP targeted glioma but yielded only a modest extension of survival. In contrast, VSV-EBOV containing the MLD showed substantially better targeting and elimination of brain tumors after intravenous delivery and increased the survival of brain tumor-bearing mice. Despite the apparent destruction of most tumor cells by VSV-EBOVΔMLD, the virus remained active within the SCID mouse brain and showed widespread infection of normal brain cells. In contrast, VSV-EBOV eliminated the tumors and showed relatively little infection of normal brain cells. Parallel experiments with direct intracranial virus infection generated similar results. Neither VSV-EBOV nor VSV-EBOVΔMLD showed substantive infection of the brains of normal immunocompetent mice. IMPORTANCE The Ebola virus glycoprotein contains a mucin-like domain which may play a role in immune evasion. Chimeric vesicular stomatitis viruses with the EBOV glycoprotein substituted for the VSV glycoprotein show greater safety and efficacy in targeting brain tumors in immunodeficient mice when the MLD was expressed within the EBOV glycoprotein than when EBOV lacked the mucin-like domain.

No abstract

No abstract

Over 50% of the 34 million people who suffer from diabetes mellitus (DM) are affected by diabetic neuropathy. Painful diabetic neuropathy (PDN) impacts 40–50% of that group (8.5 million patients) and is associated with a significant source of disability and economic burden. Though new neuromodulation options have been successful in recent clinical trials (NCT03228420), still there are many barriers that restrict patients from access to these therapies. We seek to examine our tertiary care center (Albany Medical Center, NY, USA) experience with PDN management by leveraging our clinical database to assess patient referral patterns and utilization of neuromodulation. We identified all patients with a diagnosis of diabetes type 1 (CODE: E10.xx) or diabetes type 2 (CODE: E11.xx) AND neuralgia/neuropathic pain (CODE: M79.2) or neuropathy (CODE: G90.09) or chronic pain (CODE: G89.4) or limb pain (CODE: M79.6) OR diabetic neuropathy (CODE: E11.4) who saw endocrinology, neurology, and/or neurosurgery from January 1, 2019, to December 31, 2019. We then determined which patients had received pain medications and/or neuromodulation to divide the cohort into three groups: no treatment, conservative treatment, and neuromodulation treatment. The cohorts were compared with chi-square or one-way ANOVA with multiple comparisons to analyze the differences. A total of 2,635 PDN patients were identified, of which 700 received no treatment for PDN, 1,906 received medication(s), and 29 received neuromodulation (intrathecal therapy, spinal cord stimulation, or dorsal root ganglion stimulation). The patients who received pain medications for PDN visited neurology more often than the pain specialists. Of the patients that received neuromodulation, 24 had seen neurology, 6 neurology pain, and 3 anesthesia pain. They averaged 2.78 pain medications prior to implant. Approximately 41% of the patients in the conservative management group were prescribed three or more medications. Of the 1,935 treated patients, only 1.5% of the patients received neuromodulation. The patients on three or more pain medications without symptomatic relief may be potential candidates for neuromodulation. An opportunity, therefore, exists to educate providers on the benefits of neuromodulation procedures.