Natural language processing augments comorbidity documentation in neurosurgical inpatient admissions.

OBJECTIVE: To establish whether or not a natural language processing technique could identify two common inpatient neurosurgical comorbidities using only text reports of inpatient head imaging. MATERIALS AND METHODS: A training and testing dataset of reports of 979 CT or MRI scans of the brain for patients admitted to the neurosurgery service of a single hospital in June 2021 or to the Emergency Department between July 1-8, 2021, was identified. A variety of machine learning and deep learning algorithms utilizing natural language processing were trained on the training set (84% of the total cohort) and tested on the remaining images. A subset comparison cohort (n = 76) was then assessed to compare output of the best algorithm against real-life inpatient documentation. RESULTS: For "brain compression", a random forest classifier outperformed other candidate algorithms with an accuracy of 0.81 and area under the curve of 0.90 in the testing dataset. For "brain edema", a random forest classifier again outperformed other candidate algorithms with an accuracy of 0.92 and AUC of 0.94 in the testing dataset. In the provider comparison dataset, for "brain compression," the random forest algorithm demonstrated better accuracy (0.76 vs 0.70) and sensitivity (0.73 vs 0.43) than provider documentation. For "brain edema," the algorithm again demonstrated better accuracy (0.92 vs 0.84) and AUC (0.45 vs 0.09) than provider documentation. DISCUSSION: A natural language processing-based machine learning algorithm can reliably and reproducibly identify selected common neurosurgical comorbidities from radiology reports. CONCLUSION: This result may justify the use of machine learning-based decision support to augment provider documentation.

Optimizing surgical management of facet cysts of the lumbar spine: systematic review, meta-analysis, and local case series of 1251 patients.

OBJECTIVE: Lumbar facet cysts (LFCs) can cause neurological dysfunction and intractable pain. Surgery is the current standard of care for patients in whom conservative therapy fails, those with neurological deficits, and those with evidence of spinal instability. No study to date has comprehensively examined surgical outcomes comparing the multiple surgical treatment options for LFCs. Therefore, the authors aimed to perform a combined analysis of cases both in the literature and of patients at a single institution to compare the outcomes of various surgical treatment options for LFC. METHODS: The authors performed a literature review in accordance with PRISMA guidelines and meta-analysis of the PubMed, Embase, and Cochrane Library databases and reviewed all studies from database inception published until February 3, 2023. Studies that did not contain 3 or more cases, clearly specify follow-up durations longer than 6 months, or present new cases were excluded. Bias was evaluated using Cochrane Collaboration's Risk of Bias in Nonrandomised Studies-of Interventions (ROBINS-I). The authors also reviewed their own local institutional case series from 2015 to 2020. Primary outcomes were same-level cyst recurrence, same-level revision surgery, and perioperative complications. ANOVA, common and random-effects modeling, and Wald testing were used to compare treatment groups. RESULTS: A total of 1251 patients were identified from both the published literature (29 articles, n = 1143) and the authors' institution (n = 108). Patients were sorted into 5 treatment groups: open cyst resection (OCR; n = 720), tubular cyst resection (TCR; n = 166), cyst resection with arthrodesis (CRA; n = 165), endoscopic cyst resection (ECR; n = 113), and percutaneous cyst rupture (PCR; n = 87), with OCR being the analysis reference group. The PCR group had significantly lower complication rates (p = 0.004), higher recurrence rates (p < 0.001), and higher revision surgery rates (p = 0.001) compared with the OCR group. Patients receiving TCR (3.01%, p = 0.021) and CRA (0.0%, p < 0.001) had significantly lower recurrence rates compared with those undergoing OCR (6.36%). The CRA group (6.67%) also had significantly lower rates of revision surgery compared with the OCR group (11.3%, p = 0.037). CONCLUSIONS: While PCR is less invasive, it may have high rates of same-level recurrence and revision surgery. Recurrence and revision rates for modalities such as ECR were not significantly different from those of OCR. While concomitant arthrodesis is more invasive, it might lead to lower recurrence rates and lower rates of subsequent revision surgery. Given the limitations of our case series and literature review, prospective, randomized studies are needed.

Repetitive mild head trauma induces activity mediated lifelong brain deficits in a novel Drosophila model.

Mild head trauma, including concussion, can lead to chronic brain dysfunction and degeneration but the underlying mechanisms remain poorly understood. Here, we developed a novel head impact system to investigate the long-term effects of mild head trauma on brain structure and function, as well as the underlying mechanisms in Drosophila melanogaster. We find that Drosophila subjected to repetitive head impacts develop long-term deficits, including impaired startle-induced climbing, progressive brain degeneration, and shortened lifespan, all of which are substantially exacerbated in female flies. Interestingly, head impacts elicit an elevation in neuronal activity and its acute suppression abrogates the detrimental effects in female flies. Together, our findings validate Drosophila as a suitable model system for investigating the long-term effects of mild head trauma, suggest an increased vulnerability to brain injury in female flies, and indicate that early altered neuronal excitability may be a key mechanism linking mild brain trauma to chronic degeneration.