Can We Geographically Validate a Natural Language Processing Algorithm for Automated Detection of Incidental Durotomy Across Three Independent Cohorts From Two Continents?

BACKGROUND: Incidental durotomy is an intraoperative complication in spine surgery that can lead to postoperative complications, increased length of stay, and higher healthcare costs. Natural language processing (NLP) is an artificial intelligence method that assists in understanding free-text notes that may be useful in the automated surveillance of adverse events in orthopaedic surgery. A previously developed NLP algorithm is highly accurate in the detection of incidental durotomy on internal validation and external validation in an independent cohort from the same country. External validation in a cohort with linguistic differences is required to assess the transportability of the developed algorithm, referred to geographical validation. Ideally, the performance of a prediction model, the NLP algorithm, is constant across geographic regions to ensure reproducibility and model validity. QUESTION/PURPOSE: Can we geographically validate an NLP algorithm for the automated detection of incidental durotomy across three independent cohorts from two continents? METHODS: Patients 18 years or older undergoing a primary procedure of (thoraco)lumbar spine surgery were included. In Massachusetts, between January 2000 and June 2018, 1000 patients were included from two academic and three community medical centers. In Maryland, between July 2016 and November 2018, 1279 patients were included from one academic center, and in Australia, between January 2010 and December 2019, 944 patients were included from one academic center. The authors retrospectively studied the free-text operative notes of included patients for the primary outcome that was defined as intraoperative durotomy. Incidental durotomy occurred in 9% (93 of 1000), 8% (108 of 1279), and 6% (58 of 944) of the patients, respectively, in the Massachusetts, Maryland, and Australia cohorts. No missing reports were observed. Three datasets (Massachusetts, Australian, and combined Massachusetts and Australian) were divided into training and holdout test sets in an 80:20 ratio. An extreme gradient boosting (an efficient and flexible tree-based algorithm) NLP algorithm was individually trained on each training set, and the performance of the three NLP algorithms (respectively American, Australian, and combined) was assessed by discrimination via area under the receiver operating characteristic curves (AUC-ROC; this measures the model's ability to distinguish patients who obtained the outcomes from those who did not), calibration metrics (which plot the predicted and the observed probabilities) and Brier score (a composite of discrimination and calibration). In addition, the sensitivity (true positives, recall), specificity (true negatives), positive predictive value (also known as precision), negative predictive value, F1-score (composite of precision and recall), positive likelihood ratio, and negative likelihood ratio were calculated. RESULTS: The combined NLP algorithm (the combined Massachusetts and Australian data) achieved excellent performance on independent testing data from Australia (AUC-ROC 0.97 [95% confidence interval 0.87 to 0.99]), Massachusetts (AUC-ROC 0.99 [95% CI 0.80 to 0.99]) and Maryland (AUC-ROC 0.95 [95% CI 0.93 to 0.97]). The NLP developed based on the Massachusetts cohort had excellent performance in the Maryland cohort (AUC-ROC 0.97 [95% CI 0.95 to 0.99]) but worse performance in the Australian cohort (AUC-ROC 0.74 [95% CI 0.70 to 0.77]). CONCLUSION: We demonstrated the clinical utility and reproducibility of an NLP algorithm with combined datasets retaining excellent performance in individual countries relative to algorithms developed in the same country alone for detection of incidental durotomy. Further multi-institutional, international collaborations can facilitate the creation of universal NLP algorithms that improve the quality and safety of orthopaedic surgery globally. The combined NLP algorithm has been incorporated into a freely accessible web application that can be found at https://sorg-apps.shinyapps.io/nlp_incidental_durotomy/ . Clinicians and researchers can use the tool to help incorporate the model in evaluating spine registries or quality and safety departments to automate detection of incidental durotomy and optimize prevention efforts. LEVEL OF EVIDENCE: Level III, diagnostic study.

Advances in monitoring for acute spinal cord injury: a narrative review of current literature.

Spinal cord injury (SCI) is a devastating condition that affects about 17,000 individuals every year in the United States, with approximately 294,000 people living with the ramifications of the initial injury. After the initial primary injury, SCI has a secondary phase during which the spinal cord sustains further injury due to ischemia, excitotoxicity, immune-mediated damage, mitochondrial dysfunction, apoptosis, and oxidative stress. The multifaceted injury progression process requires a sophisticated injury-monitoring technique for an accurate assessment of SCI patients. In this narrative review, we discuss SCI monitoring modalities, including pressure probes and catheters, micro dialysis, electrophysiologic measures, biomarkers, and imaging studies. The optimal next-generation injury monitoring setup should include multiple modalities and should integrate the data to produce a final simplified assessment of the injury and determine markers of intervention to improve patient outcomes.

Split cord malformation in adults: Literature review and classification.

The objective of this study was to summarize the available literature describing the presentation, diagnostic evaluation, and management for adults with Type 1 and Type 2 split spinal cord malformations. A review of the literature was performed using the CINAHL, PubMed, Embase, and Web of Science database, alongside all associated bibliographies, to include studies describing Type 1 and Type 2 split cord malformations diagnosed in patients above the age of 18. All relevant studies of split cord malformations were included, regardless of the year published and terminology used to describe the dysraphism. Clinical case series (≥ 2 patients), cohort studies, and review articles comprising adult patients with radiographically diagnosed diastematomyelia, diplomyelia, or dimyelia were included (Class of Evidence I-IV). A total of 17 unique articles, describing 146 unique adult spinal cord malformation subjects, were included. The most common associated condition was tethered cord syndrome (59.8 %). Operative management for symptomatic split cord malformation was performed in 72.3 % of cases. For those with preoperative neurologic deficits, operative management resulted in symptomatic improvement in 96.6 %, compared to 0 % conservative management (p < 0.05). For those with pain alone, operative management resulted in improvement of 91.1 %, compared to 12.5 % conservative management (p < 0.05). To date, this is the only literature review to include all split cord malformations (SCM Types I and II) presenting in adulthood, with clinical characteristics, associated conditions, and long-term treatment outcomes.

Treatment approaches for Scheuermann kyphosis: a systematic review of historic and current management.

OBJECTIVE: Scheuermann kyphosis (SK) is an idiopathic kyphosis characterized by anterior wedging of ≥ 5° at 3 contiguous vertebrae managed with either nonoperative or operative treatment. Nonoperative treatment typically employs bracing, while operative treatment is performed with either a combined anterior-posterior fusion or posterior-only approach. Current evidence for these approaches has largely been derived from retrospective case series or focused reviews. Consequently, no consensus exists regarding optimal management strategies for patients afflicted with this condition. In this study, the authors systematically review the literature on SK with respect to indications for treatment, complications of treatment, differences in correction and loss of correction, and changes in treatment over time. METHODS: Using PubMed, Embase, CINAHL, Web of Science, and the Cochrane Library, all full-text publications on the operative and nonoperative treatment for SK in the peer-reviewed English-language literature between 1950 and 2017 were screened. Inclusion criteria involved fully published, peer-reviewed, retrospective or prospective studies of the primary medical literature. Studies were excluded if they did not provide clinical outcomes and statistics specific to SK, described fewer than 2 patients, or discussed results in nonhuman models. Variables extracted included treatment indications and methodology, maximum pretreatment kyphosis, immediate posttreatment kyphosis, kyphosis at last follow-up, year of treatment, and complications of treatment. RESULTS: Of 659 unique studies, 45 met our inclusion criteria, covering 1829 unique patients. Indications for intervention were pain, deformity, failure of nonoperative treatment, and neural impairment. Among operatively treated patients, the most common complications were hardware failure and proximal or distal junctional kyphosis. Combined anterior-posterior procedures were additionally associated with neural, pulmonary, and cardiovascular complications. Posterior-only approaches offered superior correction compared to combined anterior-posterior fusion; both groups provided greater correction than bracing. Loss of correction was similar across operative approaches, and all were superior to bracing. Cross-sectional analysis suggested that surgeons have shifted from anterior-posterior to posterior-only approaches over the past two decades. CONCLUSIONS: The data indicate that for patients with SK, surgery affords superior correction and maintenance of correction relative to bracing. Posterior-only fusion may provide greater correction and similar loss of correction compared to anterior-posterior approaches along with a smaller complication profile. This posterior-only approach has concomitantly gained popularity over the combined anterior-posterior approach in recent years.

Extending the resection beyond the contrast-enhancement for glioblastoma: feasibility, efficacy, and outcomes.

OBJECT: It is becoming well-established that increasing extent of resection with decreasing residual volume is associated with delayed recurrence and prolonged survival for patients with glioblastoma (GBM). These prior studies are based on evaluating the contrast-enhancing (CE) tumour and not the surrounding fluid attenuated inversion recovery (FLAIR) volume. It therefore remains unclear if the resection beyond the CE portion of the tumour if it translates into improved outcomes for patients with GBM. METHODS: Adult patients who underwent resection of a primary glioblastoma at a tertiary care institution between January 1, 2007 and December 31, 2012 and underwent radiation and temozolomide chemotherapy were retrospectively reviewed. Pre and postoperative MRI images were measured for CE tumour and FLAIR volumes. Multivariate proportional hazards were used to assess associations with both time to recurrence and death. Values with p < 0.05 were considered statistically significant. RESULTS: 245 patients met the inclusion criteria. The median [IQR] preoperative CE and FLAIR tumour volumes were 31.9 [13.9-56.1] cm3 and 78.3 [44.7-115.6] cm3, respectively. Following surgery, the median [IQR] postoperative CE and FLAIR tumour volumes were 1.9 [0-7.1] cm3 and 59.7 [29.7-94.2] cm3, respectively. In multivariate analyses, the postoperative FLAIR volume was not associated with recurrence and/or survival (p > 0.05). However, the postoperative CE tumour volume was significantly associated with both recurrence [HR (95%CI); 1.026 (1.005-1.048), p = 0.01] and survival [HR (95%CI); 1.027 (1.007-1.032), p = 0.001]. The postoperative FLAIR volume was also not associated with recurrence and/or survival among patients who underwent gross total resection of the CE portion of the tumour as well as those who underwent supratotal resection. CONCLUSIONS: In this study, the volume of CE tumour remaining after resection is more important than FLAIR volume in regards to recurrence and survival for patients with GBM.

Volumetric tumor growth rates of meningiomas involving the intracranial venous sinuses.

OBJECT: There is currently no consensus as to whether meningiomas located inside the venous sinuses should be aggressively or conservatively treated. The goals of this study were to identify how sinus-invading meningiomas grow, report and compare growth rates of tumor components inside and outside the different venous sinuses, identify risk factors associated with increased tumor growth, and determine the effects of the extent of tumor resection on recurrence for meningiomas that invade the dural venous sinuses. METHODS: Adult patients who underwent primary, non-biopsy resection of a WHO grade 1 meningioma invading the dural venous sinuses at a tertiary care institution between 2007 and 2015 were retrospectively reviewed. Rates of tumor growth were fit to several growth models to evaluate the most accurate model. Cohen's d analysis was used to identify associations with increased growth of tumor in the venous sinuses. Logistic regression was used to compare extent of resection with recurrence. RESULTS: Of the 68 patients included in the study, 34 patients had postoperative residual tumors in the venous sinuses that were measured over time. The growth model that best fit the growth of intrasinus meningiomas was the Gompertzian growth model (r2 = 0.93). The annual growth rate of meningiomas inside the sinuses was 7.3%, compared to extrasinus tumors with 13.6% growth per year. The only factor significantly associated with increased tumor growth in sinuses was preoperative embolization (effect sizes (ES) [95% CI], 1.874 [7.633-46.735], p = 0.008). CONCLUSIONS: This study shows that meningiomas involving the venous sinuses have a Gompertzian-type growth with early exponential growth followed by a slower growth rate that plateaus when they reach a certain size. Overall, the growth rate of the intrasinus portion is low (7.3%), which is half of the reported growth rates for other studies involving primarily extrasinus tumors.

Minimally invasive versus conventional spine surgery for vertebral metastases: a systematic review of the evidence.

One of the major determinants of surgical candidacy in patients with symptomatic spinal metastases is the ability of the patient to tolerate the procedure-associated morbidity. In other pathologies, minimally invasive (MIS) procedures have been suggested to have lower intra-operative morbidity while providing similar outcomes. We conducted a systematic review of the PubMed library searching for articles that directly compared the operative and post-operative outcomes of patients treated for symptomatic spinal metastases. Inclusion criteria were articles reporting two or more cases of patients >18 years old treated with MIS or open approaches for spinal metastases. Studies reporting results in spinal metastases patients that could not be disentangled from other pathologies were excluded. Our search returned 1,568 articles, of which 9 articles met the criteria for inclusion. All articles were level III evidence. Patients treated with MIS approaches tended to have lower intraoperative blood loss, shorter operative times, shorter inpatient stays, and fewer complications relative to patients undergoing surgeries with conventional approaches. Patients in the MIS and open groups had similar pain improvement, neurological improvement, and functional outcomes. Recent advances in MIS techniques may reduce surgical morbidity while providing similar symptomatic improvement in patients treated for spinal metastases. As a result, MIS techniques may expand the pool of patients with spinal metastases who are candidates for operative management.

Epstein-Barr virus-associated smooth muscle tumor involving the spine of an HIV-infected patient: Case report and review of the literature.

Within the last two decades, there have been multiple reports of Epstein-Barr virus (EBV)-associated smooth muscle tumors in immunocompromised patients. This includes HIV-infected patients, post-transplant patients, and patients with congenital defects of their immune systems. Here we report the case of a 24-year-old African American female with congenital HIV presenting with progressive lower extremity weakness, constipation, aching pain in her shoulders, and subcostal anesthesia. Magnetic resonance imaging (MRI) revealed a large circumferential lesion extending from T1-T3 and a smaller left paraspinal lesion at C6-C7. The T1-T3 mass was excised via a right-sided costotransversectomy with laminectomy and fusion from T1-T3. Highly active antiretroviral therapy (HAART) was started postoperatively, and adjuvant radiotherapy was initiated but patient was lost to follow-up. Surgical pathology demonstrated a smooth muscle tumor diffuse nuclear positivity for EBV-encoded small RNA 1 by in situ hybridization. Although eight studies have reported HIV patients with EBV-associated smooth muscle tumors of the spine, to the author's knowledge, this is the first review comprised solely of patients with spinal involvement with the addition of our patient case.

The Radiographic Effects of Surgical Approach and Use of Retractors on the Brain After Anterior Cranial Fossa Meningioma Resection.

OBJECTIVE: There is an increasing trend toward skull base (SB) approaches and retractorless surgery to minimize brain manipulation during surgery. We evaluated the radiographic changes over time after surgical resection of anterior cranial fossa meningiomas with and without both SB approaches and/or fixed retractor systems. METHODS: All adults undergoing primary resection of an anterior cranial fossa World Health Organization grade I meningioma through a craniotomy at a single academic tertiary-care institution from 2010 to 2015 were retrospectively reviewed. Magnetic resonance imaging scans were reviewed and contrast-enhanced tumor and fluid-attenuated inversion recovery (FLAIR) volumes were measured. Matched-pair analyses between patients who underwent SB and non-SB approaches, as well as retractorless and retractor-assisted (RA) surgery, were made. RESULTS: Of the 136 total patients, 20 (15%), 12 (9%), 46 (34%), and 58 (43%) underwent SB/retractorless, SB/RA, non-SB/retractorless, and non-SB/RA surgery, respectively. Patients who underwent non-SB and RA surgery each independently had longer times to FLAIR resolution than those who underwent SB (20.9 vs. 5 months; P = 0.04) and retractorless (12 vs. 5.2 months; P = 0.02) surgery, respectively. Patients who underwent both non-SB and RA surgery had the longest median time to FLAIR resolution (30 months vs. 4 months in SB/retractorless, 3.6 months in SB/RA, and 3 months in non-SB/retractorless; P < 0.05). CONCLUSIONS: The use of SB approaches in combination with retractorless surgery may decrease the duration needed for FLAIR resolution after surgery. The results from this study therefore advocate SB approaches and retractorless surgery along the anterior SB when possible.