Development of machine learning and natural language processing algorithms for preoperative prediction and automated identification of intraoperative vascular injury in anterior lumbar spine surgery

Natural Language Processing-Driven Artificial Intelligence Models for the Diagnosis of Lumbar Disc Herniation with L5 and S1 Radiculopathy: A Preliminary Evaluation

OBJECTIVE

To develop and validate natural language processing-driven artificial intelligence (AI) models for the diagnosis of lumbar disc herniation (LDH) with L5 and S1 radiculopathy using electronic health records (EHRs).

METHODS

EHRs of patients undergoing single-level percutaneous endoscopic lumbar discectomy for the treatment of LDH at the L4/5 or L5/S1 level between June 1, 2013, and December 31, 2021, were collected. The primary outcome was LDH with L5 and S1 radiculopathy, which was defined as nerve root compression recorded in the operative notes. Datasets were created using the history of present illness text and positive symptom text with radiculopathy (L5 or S1), respectively. The datasets were randomly split into a training set and a testing set in a 7:3 ratio. Two machine learning models, the long short-term memory network and Extreme Gradient Boosting, were developed using the training set. Performance evaluation of the models on the testing set was done using measures such as the receiver operating characteristic curve, area under the curve, accuracy, recall, F1-score, and precision.

RESULTS

The study included a total of 1681 patients, with 590 patients having L5 radiculopathy and 1091 patients having S1 radiculopathy. Among the 4 models developed, the long short-term memory model based on positive symptom text showed the best discrimination in the testing set, with precision (0.9054), recall (0.9405), accuracy (0.8950), F1-score (0.9226), and area under the curve (0.9485).

CONCLUSIONS

This study provides preliminary validation of the concept that natural language processing-driven AI models can be used for the diagnosis of lumbar disease using EHRs. This study could pave the way for future research that may develop more comprehensive and clinically impactful AI-driven diagnostic systems.

Artificial Intelligence in Spinal Imaging and Patient Care: A Review of Recent Advances

Artificial intelligence (AI) is transforming spinal imaging and patient care through automated analysis and enhanced decision-making. This review presents a clinical task-based evaluation, highlighting the specific impact of AI techniques on different aspects of spinal imaging and patient care. We first discuss how AI can potentially improve image quality through techniques like denoising or artifact reduction. We then explore how AI enables efficient quantification of anatomical measurements, spinal curvature parameters, vertebral segmentation, and disc grading. This facilitates objective, accurate interpretation and diagnosis. AI models now reliably detect key spinal pathologies, achieving expert-level performance in tasks like identifying fractures, stenosis, infections, and tumors. Beyond diagnosis, AI also assists surgical planning via synthetic computed tomography generation, augmented reality systems, and robotic guidance. Furthermore, AI image analysis combined with clinical data enables personalized predictions to guide treatment decisions, such as forecasting spine surgery outcomes. However, challenges still need to be addressed in implementing AI clinically, including model interpretability, generalizability, and data limitations. Multicenter collaboration using large, diverse datasets is critical to advance the field further. While adoption barriers persist, AI presents a transformative opportunity to revolutionize spinal imaging workflows, empowering clinicians to translate data into actionable insights for improved patient care.

Applications of natural language processing tools in the surgical journey

Background

Natural language processing tools are becoming increasingly adopted in multiple industries worldwide. They have shown promising results however their use in the field of surgery is under-recognised. Many trials have assessed these benefits in small settings with promising results before large scale adoption can be considered in surgery. This study aims to review the current research and insights into the potential for implementation of natural language processing tools into surgery.

Methods

A narrative review was conducted following a computer-assisted literature search on Medline, EMBASE and Google Scholar databases. Papers related to natural language processing tools and consideration into their use for surgery were considered.

Results

Current applications of natural language processing tools within surgery are limited. From the literature, there is evidence of potential improvement in surgical capability and service delivery, such as through the use of these technologies to streamline processes including surgical triaging, data collection and auditing, surgical communication and documentation. Additionally, there is potential to extend these capabilities to surgical academia to improve processes in surgical research and allow innovation in the development of educational resources. Despite these outcomes, the evidence to support these findings are challenged by small sample sizes with limited applicability to broader settings.

Conclusion

With the increasing adoption of natural language processing technology, such as in popular forms like ChatGPT, there has been increasing research in the use of these tools within surgery to improve surgical workflow and efficiency. This review highlights multifaceted applications of natural language processing within surgery, albeit with clear limitations due to the infancy of the infrastructure available to leverage these technologies. There remains room for more rigorous research into broader capability of natural language processing technology within the field of surgery and the need for cross-sectoral collaboration to understand the ways in which these algorithms can best be integrated.

Exploring the mystery of colon cancer from the perspective of molecular subtypes and treatment

The molecular categorization of colon cancer patients remains elusive. Gene set enrichment analysis (GSEA), which investigates the dysregulated genes among tumor and normal samples, has revealed the pivotal role of epithelial-to-mesenchymal transition (EMT) in colon cancer pathogenesis. In this study, we employed multi-clustering method for grouping data, resulting in the identification of two clusters characterized by varying prognostic outcomes. These two subgroups not only displayed disparities in overall survival (OS) but also manifested variations in clinical variables, genetic mutation, and gene expression profiles. Using the nearest template prediction (NTP) method, we were able to replicate the molecular classification effectively within the original dataset and validated it across multiple independent datasets, underscoring its robust repeatability. Furthermore, we constructed two prognostic signatures tailored to each of these subgroups. Our molecular classification, centered on EMT, hold promise in offering fresh insights into the therapy strategies and prognosis assessment for colon cancer.

Predicting osteoporotic fractures post-vertebroplasty: a machine learning approach with a web-based calculator

PURPOSE

The aim of this study was to develop and validate a machine learning (ML) model for predicting the risk of new osteoporotic vertebral compression fracture (OVCF) in patients who underwent percutaneous vertebroplasty (PVP) and to create a user-friendly web-based calculator for clinical use.

METHODS

A retrospective analysis of patients undergoing percutaneous vertebroplasty: A retrospective analysis of patients treated with PVP between June 2016 and June 2018 at Liuzhou People's Hospital was performed. The independent variables of the model were screened using Boruta and modelled using 9 algorithms. Model performance was assessed using the area under the receiver operating characteristic curve (ROC_AUC), and clinical utility was assessed by clinical decision curve analysis (DCA). The best models were analysed for interpretability using SHapley Additive exPlanations (SHAP) and the models were deployed visually using a web calculator.

RESULTS

Training and test groups were split using time. The SVM model performed best in both the training group tenfold cross-validation (CV) and validation group AUC, with an AUC of 0.77. DCA showed that the model was beneficial to patients in both the training and test sets. A network calculator developed based on the SHAP-based SVM model can be used for clinical risk assessment ( https://nicolazhang.shinyapps.io/refracture_shap/ ).

CONCLUSIONS

The SVM-based ML model was effective in predicting the risk of new-onset OVCF after PVP, and the network calculator provides a practical tool for clinical decision-making. This study contributes to personalised care in spinal surgery.

An Integrated Voice Recognition and Natural Language Processing Platform to Automatically Extract Thoracolumbar Injury Classification Score Features From Radiology Reports

BACKGROUND

Many predictive models for estimating clinical outcomes after spine surgery have been reported in the literature. However, implementation of predictive scores in practice is limited by the time-intensive nature of manually abstracting relevant predictors. In this study, we designed natural language processing (NLP) algorithms to automate data abstraction for the thoracolumbar injury classification score (TLICS).

METHODS

We retrieved the radiology reports of all Mayo Clinic patients with an International Classification of Diseases, 9th or 10th revision, code corresponding to a fracture of the thoracolumbar spine between January 2005 and October 2020. Annotated data were used to train an N-gram NLP model using machine learning methods, including random forest, stepwise linear discriminant analysis, k-nearest neighbors, and penalized logistic regression models.

RESULTS

A total of 1085 spine radiology reports were included in our analysis. Our dataset included 483 compression, 401 burst, 103 translational/rotational, and 98 distraction fractures. A total of 103 reports had documented an injury of the posterior ligamentous complex. The overall accuracy of the random forest model for fracture morphology feature detection was 76.96% versus 65.90% in the stepwise linear discriminant analysis, 50.69% in the k-nearest neighbors, and 62.67% in the penalized logistic regression. The overall accuracy to detect posterior ligamentous complex integrity was highest in the random forest model at 83.41%. Our random forest model was implemented in the backend of a web application in which users can dictate reports and have TLICS features automatically extracted.

CONCLUSIONS

We have developed a machine learning NLP model for extracting TLICS features from radiology reports, which we deployed in a web application that can be integrated into clinical practice.

A practical guide to the development and deployment of deep learning models for the orthopaedic surgeon: Part III, focus on registry creation, diagnosis, and data privacy

Deep learning is a subset of artificial intelligence (AI) with enormous potential to transform orthopaedic surgery. As has already become evident with the deployment of Large Language Models (LLMs) like ChatGPT (OpenAI Inc.), deep learning can rapidly enter clinical and surgical practices. As such, it is imperative that orthopaedic surgeons acquire a deeper understanding of the technical terminology, capabilities and limitations associated with deep learning models. The focus of this series thus far has been providing surgeons with an overview of the steps needed to implement a deep learning-based pipeline, emphasizing some of the important technical details for surgeons to understand as they encounter, evaluate or lead deep learning projects. However, this series would be remiss without providing practical examples of how deep learning models have begun to be deployed and highlighting the areas where the authors feel deep learning may have the most profound potential. While computer vision applications of deep learning were the focus of Parts I and II, due to the enormous impact that natural language processing (NLP) has had in recent months, NLP-based deep learning models are also discussed in this final part of the series. In this review, three applications that the authors believe can be impacted the most by deep learning but with which many surgeons may not be familiar are discussed: (1) registry construction, (2) diagnostic AI and (3) data privacy. Deep learning-based registry construction will be essential for the development of more impactful clinical applications, with diagnostic AI being one of those applications likely to augment clinical decision-making in the near future. As the applications of deep learning continue to grow, the protection of patient information will become increasingly essential; as such, applications of deep learning to enhance data privacy are likely to become more important than ever before. Level of Evidence: Level IV.

Evaluating Computer Vision, Large Language, and Genome-Wide Association Models in a Limited Sized Patient Cohort for Pre-Operative Risk Stratification in Adult Spinal Deformity Surgery

Background: Adult spinal deformities (ASD) are varied spinal abnormalities, often necessitating surgical intervention when associated with pain, worsening deformity, or worsening function. Predicting post-operative complications and revision surgery is critical for surgical planning and patient counseling. Due to the relatively small number of cases of ASD surgery, machine learning applications have been limited to traditional models (e.g., logistic regression or standard neural networks) and coarse clinical variables. We present the novel application of advanced models (CNN, LLM, GWAS) using complex data types (radiographs, clinical notes, genomics) for ASD outcome prediction. Methods: We developed a CNN trained on 209 ASD patients (1549 radiographs) from the Stanford Research Repository, a CNN pre-trained on VinDr-SpineXR (10,468 spine radiographs), and an LLM using free-text clinical notes from the same 209 patients, trained via Gatortron. Additionally, we conducted a GWAS using the UK Biobank, contrasting 540 surgical ASD patients with 7355 non-surgical ASD patients. Results: The LLM notably outperformed the CNN in predicting pulmonary complications (F1: 0.545 vs. 0.2881), neurological complications (F1: 0.250 vs. 0.224), and sepsis (F1: 0.382 vs. 0.132). The pre-trained CNN showed improved sepsis prediction (AUC: 0.638 vs. 0.534) but reduced performance for neurological complication prediction (AUC: 0.545 vs. 0.619). The LLM demonstrated high specificity (0.946) and positive predictive value (0.467) for neurological complications. The GWAS identified 21 significant (p < 10-5) SNPs associated with ASD surgery risk (OR: mean: 3.17, SD: 1.92, median: 2.78), with the highest odds ratio (8.06) for the LDB2 gene, which is implicated in ectoderm differentiation. Conclusions: This study exemplifies the innovative application of cutting-edge models to forecast outcomes in ASD, underscoring the utility of complex data in outcome prediction for neurosurgical conditions. It demonstrates the promise of genetic models when identifying surgical risks and supports the integration of complex machine learning tools for informed surgical decision-making in ASD.

Limitations in Evaluating Machine Learning Models for Imbalanced Binary Outcome Classification in Spine Surgery: A Systematic Review

Clinical prediction models for spine surgery applications are on the rise, with an increasing reliance on machine learning (ML) and deep learning (DL). Many of the predicted outcomes are uncommon; therefore, to ensure the models' effectiveness in clinical practice it is crucial to properly evaluate them. This systematic review aims to identify and evaluate current research-based ML and DL models applied for spine surgery, specifically those predicting binary outcomes with a focus on their evaluation metrics. Overall, 60 papers were included, and the findings were reported according to the PRISMA guidelines. A total of 13 papers focused on lengths of stay (LOS), 12 on readmissions, 12 on non-home discharge, 6 on mortality, and 5 on reoperations. The target outcomes exhibited data imbalances ranging from 0.44% to 42.4%. A total of 59 papers reported the model's area under the receiver operating characteristic (AUROC), 28 mentioned accuracies, 33 provided sensitivity, 29 discussed specificity, 28 addressed positive predictive value (PPV), 24 included the negative predictive value (NPV), 25 indicated the Brier score with 10 providing a null model Brier, and 8 detailed the F1 score. Additionally, data visualization varied among the included papers. This review discusses the use of appropriate evaluation schemes in ML and identifies several common errors and potential bias sources in the literature. Embracing these recommendations as the field advances may facilitate the integration of reliable and effective ML models in clinical settings.

Towards a practical use of text mining approaches in electrodiagnostic data

Healthcare professionals produce abounding textual data in their daily clinical practice. Text mining can yield valuable insights from unstructured data. Extracting insights from multiple information sources is a major challenge in computational medicine. In this study, our objective was to illustrate how combining text mining techniques with statistical methodologies can yield new insights and contribute to the development of neurological and neuromuscular-related health information. We demonstrate how to utilize and derive knowledge from medical text, identify patient groups with similar diagnostic attributes, and examine differences between groups using demographical data and past medical history (PMH). We conducted a retrospective study for all patients who underwent electrodiagnostic (EDX) evaluation in Israel's Sheba Medical Center between May 2016 and February 2022. The data extracted for each patient included demographic data, test results, and unstructured summary reports. We conducted several analyses, including topic modeling that targeted clinical impressions and topic analysis to reveal age- and sex-related differences. The use of suspected clinical condition text enriched the data and generated additional attributes used to find associations between patients' PMH and the emerging diagnosis topics. We identified 6096 abnormal EMG results, of which 58% (n = 3512) were males. Based on the latent Dirichlet allocation algorithm we identified 25 topics that represent different diagnoses. Sex-related differences emerged in 7 topics, 3 male-associated and 4 female-associated. Brachial plexopathy, myasthenia gravis, and NMJ Disorders showed statistically significant age and sex differences. We extracted keywords related to past medical history (n = 37) and tested them for association with the different topics. Several topics revealed a close association with past medical history, for example, length-dependent symmetric axonal polyneuropathy with diabetes mellitus (DM), length-dependent sensory polyneuropathy with chemotherapy treatments and DM, brachial plexopathy with motor vehicle accidents, myasthenia gravis and NMJ disorders with botulin treatments, and amyotrophic lateral sclerosis with swallowing difficulty. Summarizing visualizations were created to easily grasp the results and facilitate focusing on the main insights. In this study, we demonstrate the efficacy of utilizing advanced computational methods in a corpus of textual data to accelerate clinical research. Additionally, using these methods allows for generating clinical insights, which may aid in the development of a decision-making process in real-life clinical practice.

Artificial Intelligence in Neurosurgery: A State-of-the-Art Review from Past to Future

In recent years, there has been a significant surge in discussions surrounding artificial intelligence (AI), along with a corresponding increase in its practical applications in various facets of everyday life, including the medical industry. Notably, even in the highly specialized realm of neurosurgery, AI has been utilized for differential diagnosis, pre-operative evaluation, and improving surgical precision. Many of these applications have begun to mitigate risks of intraoperative and postoperative complications and post-operative care. This article aims to present an overview of the principal published papers on the significant themes of tumor, spine, epilepsy, and vascular issues, wherein AI has been applied to assess its potential applications within neurosurgery. The method involved identifying high-cited seminal papers using PubMed and Google Scholar, conducting a comprehensive review of various study types, and summarizing machine learning applications to enhance understanding among clinicians for future utilization. Recent studies demonstrate that machine learning (ML) holds significant potential in neuro-oncological care, spine surgery, epilepsy management, and other neurosurgical applications. ML techniques have proven effective in tumor identification, surgical outcomes prediction, seizure outcome prediction, aneurysm prediction, and more, highlighting its broad impact and potential in improving patient management and outcomes in neurosurgery. This review will encompass the current state of research, as well as predictions for the future of AI within neurosurgery.

Successful implementation of a nurse-navigator-run program using natural language processing identifying patients with an abdominal aortic aneurysm

BACKGROUND

Abdominal aortic aneurysms (AAA) are often identified incidentally on imaging studies. Patients and/or providers are frequently unaware of these AAA and the need for long-term follow-up. We sought to evaluate the outcome of a nurse-navigator-run AAA program that uses a natural language processing (NLP) algorithm applied to the electronic medical record (EMR) to identify patients with imaging report-identified AAA not being followed actively.

METHODS

A commercially available AAA-specific NLP system was run on EMR data at a large, academic, tertiary hospital with an 11-year historical look back (January 1, 2010, to June 2, 2021), to identify and characterize AAA. Beginning June 3, 2021, a direct link between the NLP system and the EMR enabled for real-time review of imaging reports for new AAA cases. A nurse-navigator (1.0 full-time equivalent) used software filters to categorize AAA according to predefined metrics, including repair status and adherence to Society for Vascular Surgery imaging surveillance protocol. The nurse-navigator then interfaced with patients and providers to reestablish care for patients not being followed actively. The nurse-navigator characterized patients as case closed (eg, deceased, appropriate follow-up elsewhere, refuses follow-up), cases awaiting review, and cases reviewed and placed in ongoing surveillance using AAA-specific software. The primary outcome measures were yield of surveillance imaging performed or scheduled, new clinic visits, and AAA operations for patients not being followed actively.

RESULTS

During the prospective study period (January 1, 2021, to December 30, 2021), 6,340,505 imaging reports were processed by the NLP. After filtering for studies likely to include abdominal aorta, 243,889 imaging reports were evaluated, resulting in the identification of 6495 patients with AAA. Of these, 2937 cases were reviewed and closed, 1183 were reviewed and placed in ongoing surveillance, and 2375 are awaiting review. When stratifying those reviewed and placed in ongoing surveillance by maximum aortic diameter, 258 were 2.5 to 3.4 cm, 163 were 3.5 to 3.9 cm, 213 were 4 to 5 cm, and 49 were larger than 5 cm; 36 were saccular, 86 previously underwent open repair, 274 previously underwent endovascular repair, and 104 were other. This process yielded 29 new patient clinic visits, 40 finalized imaging studies, 29 scheduled imaging studies, and 4 AAA operations in 3 patients among patients not being followed actively.

CONCLUSIONS

The application of an AAA program leveraging NLP successfully identifies patients with AAA not receiving appropriate surveillance or counseling and repair. This program offers an opportunity to improve best practice-based care across a large health system.

Current Applications of Machine Learning for Spinal Cord Tumors

Spinal cord tumors constitute a diverse group of rare neoplasms associated with significant mortality and morbidity that pose unique clinical and surgical challenges. Diagnostic accuracy and outcome prediction are critical for informed decision making and can promote personalized medicine and facilitate optimal patient management. Machine learning has the ability to analyze and combine vast amounts of data, allowing the identification of patterns and the establishment of clinical associations, which can ultimately enhance patient care. Although artificial intelligence techniques have been explored in other areas of spine surgery, such as spinal deformity surgery, precise machine learning models for spinal tumors are lagging behind. Current applications of machine learning in spinal cord tumors include algorithms that improve diagnostic precision by predicting genetic, molecular, and histopathological profiles. Furthermore, artificial intelligence-based systems can assist surgeons with preoperative planning and surgical resection, potentially reducing the risk of recurrence and consequently improving clinical outcomes. Machine learning algorithms promote personalized medicine by enabling prognostication and risk stratification based on accurate predictions of treatment response, survival, and postoperative complications. Despite their promising potential, machine learning models require extensive validation processes and quality assessments to ensure safe and effective translation to clinical practice.

New Frontiers of Natural Language Processing in Surgery

The vast and ever-growing volume of electronic health records (EHR) have generated a wealth of information-rich data. Traditional, non-machine learning data extraction techniques are error-prone and laborious, hindering the analytical potential of these massive data sources. Equipped with natural language processing (NLP) tools, surgeons are better able to automate, and customize their review to investigate and implement surgical solutions. We identify current perioperative applications of NLP algorithms as well as research limitations and future avenues to outline the impact and potential of this technology for progressing surgical innovation.

Natural Language Processing in Spine Surgery: A Systematic Review of Applications, Bias, and Reporting Transparency

BACKGROUND

Natural language processing (NLP) is a discipline of machine learning concerned with the analysis of language and text. Although NLP has been applied to various forms of clinical text, the applications and utility of NLP in spine surgery remain poorly characterized. Here, we systematically reviewed studies that use NLP for spine surgery applications, and analyzed applications, bias, and reporting transparency of the studies.

METHODS

We performed a literature search using the PubMed, Scopus, and Embase databases. Data extraction was performed after appropriate screening. The risk of bias and reporting quality were assessed using the PROBAST and TRIPOD tools.

RESULTS

A total of 12 full-text articles were included. The most common diseases represented include spondylolisthesis (25%), scoliosis (17%), and lumbar disk herniation (17%). The most common procedures included spinal fusion (42%), imaging (e.g. magnetic resonance, X-ray) (25%), and scoliosis correction (17%). Reported outcomes were diverse and included incidental durotomy, venous thromboembolism, and the tone of social media posts regarding scoliosis surgery. Common sources of bias identified included the use of older methods that do not capture the nuance of a text, and not using a prespecified or standard outcome measure when evaluating NLP methods.

CONCLUSIONS

Although the application of NLP to spine surgery is expanding, current studies face limitations and none are indicated as ready for clinical use. Thus, for future studies we recommend an emphasis on transparent reporting and collaboration with NLP experts to incorporate the latest developments to improve models and contribute to further innovation.

Constructing a molecular subtype model of colon cancer using machine learning

Background: Colon cancer (CRC) is one of the malignant tumors with a high incidence in the world. Many previous studies on CRC have focused on clinical research. With the in-depth study of CRC, the role of molecular mechanisms in CRC has become increasingly important. Currently, machine learning is widely used in medicine. By combining machine learning with molecular mechanisms, we can better understand CRC’s pathogenesis and develop new treatments for it.

Methods and materials: We used the R language to construct molecular subtypes of colon cancer and subsequently explored prognostic genes with GEPIA2. Enrichment analysis is used by WebGestalt to obtain differential genes. Protein–protein interaction networks of differential genes were constructed using the STRING database and the Cytoscape tool. TIMER2.0 and TISIDB databases were used to investigate the correlation of these genes with immune-infiltrating cells and immune targets. The cBioportal database was used to explore genomic alterations.

Results: In our study, the molecular prognostic model of CRC was constructed to study the prognostic factors of CRC, and finally, it was found that Charcot–Leyden crystal galectin (CLC), zymogen granule protein 16 (ZG16), leucine-rich repeat-containing protein 26 (LRRC26), intelectin 1 (ITLN1), UDP-GlcNAc: betaGal beta-1,3-N-acetylglucosaminyltransferase 6 (B3GNT6), chloride channel accessory 1 (CLCA1), growth factor independent 1 transcriptional repressor (GFI1), aquaporin 8 (AQP8), HEPACAM family member 2 (HEPACAM2), and UDP glucuronosyltransferase family 2 member B15 (UGT2B15) were correlated with the subtype model of CRC prognosis. Enrichment analysis shows that differential genes were mainly associated with immune-inflammatory pathways. GFI1 and CLC were associated with immune cells, immunoinhibitors, and immunostimulator. Genomic analysis shows that there were no significant changes in differential genes.

Conclusion: By constructing molecular subtypes of colon cancer, we discovered new colon cancer prognostic markers, which can provide direction for new treatments in the future.

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.

Natural language processing in low back pain and spine diseases: A systematic review

Natural Language Processing (NLP) is a discipline at the intersection between Computer Science (CS), Artificial Intelligence (AI), and Linguistics that leverages unstructured human-interpretable (natural) language text. In recent years, it gained momentum also in health-related applications and research. Although preliminary, studies concerning Low Back Pain (LBP) and other related spine disorders with relevant applications of NLP methodologies have been reported in the literature over the last few years. It motivated us to systematically review the literature comprised of two major public databases, PubMed and Scopus. To do so, we first formulated our research question following the PICO guidelines. Then, we followed a PRISMA-like protocol by performing a search query including terminologies of both technical (e.g., natural language and computational linguistics) and clinical (e.g., lumbar and spine surgery) domains. We collected 221 non-duplicated studies, 16 of which were eligible for our analysis. In this work, we present these studies divided into sub-categories, from both tasks and exploited models’ points of view. Furthermore, we report a detailed description of techniques used to extract and process textual features and the several evaluation metrics used to assess the performance of the NLP models. However, what is clear from our analysis is that additional studies on larger datasets are needed to better define the role of NLP in the care of patients with spinal disorders.

Improving Surgical Triage in Spine Clinic: Predicting Likelihood of Surgery Using Machine Learning

BACKGROUND

Correctly triaging patients to a surgeon or a nonoperative provider is an important part of the referral process. Clinics typically triage new patients based on simple intake questions. This is time-consuming and does not incorporate objective data. Our goal was to use machine learning to more accurately screen surgical candidates seen in a spine clinic.

METHODS

Using questionnaire data and magnetic resonance imaging reports, a set of artificial neural networks was trained to predict whether a patient would be recommended for spine surgery. Questionnaire responses included demographics, chief complaint, and pain characteristics. The primary end point was the surgeon's determination of whether a patient was an operative candidate. Model accuracy in predicting this end point was assessed using a separate subset of patients.

RESULTS

The retrospective dataset included 1663 patients in cervical and lumbar cohorts. Questionnaire data were available for all participants, and magnetic resonance imaging reports were available for 242 patients. Within 6 months of initial evaluation, 717 (43.1%) patients were deemed surgical candidates by the surgeon. Our models predicted surgeons' recommendations with area under the curve scores of 0.686 for lumbar (positive predictive value 66%, negative predictive value 80%) and 0.821 for cervical (positive predictive value 83%, negative predictive value 85%) patients.

CONCLUSIONS

Our models used patient data to accurately predict whether patients will receive a surgical recommendation. The high negative predictive value demonstrates that this approach can reduce the burden of nonsurgical patients in surgery clinic without losing many surgical candidates. This could reduce unnecessary visits for patients, increase the proportion of operative candidates seen by surgeons, and improve quality of patient care.

Differentiation of lumbar disc herniation and lumbar spinal stenosis using natural language processing–based machine learning based on positive symptoms

OBJECTIVE

The purpose of this study was to develop natural language processing (NLP)–based machine learning algorithms to automatically differentiate lumbar disc herniation (LDH) and lumbar spinal stenosis (LSS) based on positive symptoms in free-text admission notes. The secondary purpose was to compare the performance of the deep learning algorithm with the ensemble model on the current task.

METHODS

In total, 1921 patients whose principal diagnosis was LDH or LSS between June 2013 and June 2020 at Zhongda Hospital, affiliated with Southeast University, were retrospectively analyzed. The data set was randomly divided into a training set and testing set at a 7:3 ratio. Long Short-Term Memory (LSTM) and extreme gradient boosting (XGBoost) models were developed in this study. NLP algorithms were assessed on the testing set by the following metrics: receiver operating characteristic (ROC) curve, area under the curve (AUC), accuracy score, recall score, F1 score, and precision score.

RESULTS

In the testing set, the LSTM model achieved an AUC of 0.8487, accuracy score of 0.7818, recall score of 0.9045, F1 score of 0.8108, and precision score of 0.7347. In comparison, the XGBoost model achieved an AUC of 0.7565, accuracy score of 0.6961, recall score of 0.7387, F1 score of 0.7153, and precision score of 0.6934.

CONCLUSIONS

NLP-based machine learning algorithms were a promising auxiliary to the electronic health record in spine disease diagnosis. LSTM, the deep learning model, showed better capacity compared with the widely used ensemble model, XGBoost, in differentiation of LDH and LSS using positive symptoms. This study presents a proof of concept for the application of NLP in prediagnosis of spine disease.

An Evolution Gaining Momentum—The Growing Role of Artificial Intelligence in the Diagnosis and Treatment of Spinal Diseases

In recent years, applications using artificial intelligence have been gaining importance in the diagnosis and treatment of spinal diseases. In our review, we describe the basic features of artificial intelligence which are currently applied in the field of spine diagnosis and treatment, and we provide an orientation of the recent technical developments and their applications. Furthermore, we point out the possible limitations and challenges in dealing with such technological advances. Despite the momentary limitations in practical application, artificial intelligence is gaining ground in the field of spine treatment. As an applying physician, it is therefore necessary to engage with it in order to benefit from those advances in the interest of the patient and to prevent these applications being misused by non-medical partners.

Artificial Intelligence-Driven Prediction Modeling and Decision Making in Spine Surgery Using Hybrid Machine Learning Models

Healthcare systems worldwide generate vast amounts of data from many different sources. Although of high complexity for a human being, it is essential to determine the patterns and minor variations in the genomic, radiological, laboratory, or clinical data that reliably differentiate phenotypes or allow high predictive accuracy in health-related tasks. Convolutional neural networks (CNN) are increasingly applied to image data for various tasks. Its use for non-imaging data becomes feasible through different modern machine learning techniques, converting non-imaging data into images before inputting them into the CNN model. Considering also that healthcare providers do not solely use one data modality for their decisions, this approach opens the door for multi-input/mixed data models which use a combination of patient information, such as genomic, radiological, and clinical data, to train a hybrid deep learning model. Thus, this reflects the main characteristic of artificial intelligence: simulating natural human behavior. The present review focuses on key advances in machine and deep learning, allowing for multi-perspective pattern recognition across the entire information set of patients in spine surgery. This is the first review of artificial intelligence focusing on hybrid models for deep learning applications in spine surgery, to the best of our knowledge. This is especially interesting as future tools are unlikely to use solely one data modality. The techniques discussed could become important in establishing a new approach to decision-making in spine surgery based on three fundamental pillars: (1) patient-specific, (2) artificial intelligence-driven, (3) integrating multimodal data. The findings reveal promising research that already took place to develop multi-input mixed-data hybrid decision-supporting models. Their implementation in spine surgery may hence be only a matter of time.

Natural language processing for prediction of readmission in posterior lumbar fusion patients: which free-text notes have the most utility?

BACKGROUND CONTEXT

The increasing volume of free-text notes available in electronic health records has created an opportunity for natural language processing (NLP) algorithms to mine this unstructured data in order to detect and predict adverse outcomes. Given the volume and diversity of documentation available in spine surgery, it remains unclear which types of documentation offer the greatest value for prediction of adverse outcomes.

STUDY DESIGN/SETTING

Retrospective review of medical records at two academic and three community hospitals.

PURPOSE

The purpose of this study was to conduct an exploratory analysis in order to examine the utility of free-text notes generated during the index hospitalization for lumbar spine fusion for prediction of 90-day unplanned readmission.

PATIENT SAMPLE

Adult patients 18 years or older undergoing lumbar spine fusion for lumbar spondylolisthesis or lumbar spinal stenosis between January 1, 2016 and December 31, 2020.

OUTCOME MEASURES

The primary outcome was inpatient admission within 90-days of discharge from the index hospitalization.

METHODS

The predictive performance of NLP algorithms developed by using discharge summary notes, operative notes, nursing notes, physical therapy notes, case management notes, medical doctor (MD) (resident or attending), and allied practice professional (APP) (nurse practitioner or physician assistant) notes were assessed by discrimination, calibration, overall performance.

RESULTS

Overall, 708 patients were included in the study and 83 (11.7%) had 90-day inpatient readmission. In the independent testing set of patients (n=141) not used for model development, the area under the receiver operating curve of NLP algorithms for prediction of 90-day readmission using discharge summary notes, operative notes, nursing notes, physical therapy notes, case management notes, MD/APP notes was 0.70, 0.57, 0.57, 0.60, 0.60, and 0.49 respectively.

CONCLUSION

In this exploratory analysis, discharge summary, physical therapy, and case management notes had the most utility and daily MD/APP progress notes had the least utility for prediction of 90-day inpatient readmission in lumbar fusion patients among the free-text documentation generated during the index hospitalization.

Natural language processing for automated surveillance of intraoperative neuromonitoring in spine surgery

We sought to develop natural language processing (NLP) methods for automated detection and characterization of neuromonitoring documentation from free-text operative reports in patients undergoing spine surgery. We included 13,718 patients who received spine surgery at two tertiary academic medical centers between December 2000 - December 2020. We first validated a rule-based NLP method for identifying operative reports containing neuromonitoring documentation, comparing performance to standard administrative codes. We then trained a deep learning model in a subset of 993 patients to characterize neuromonitoring documentation and identify events indicating change in status or difficulty establishing baseline signals. Performance of the deep learning model was compared to gold-standard manual chart review. In our patient population, 3,606 (26.3%) patients had neuromonitoring documentation identified using NLP. Our NLP method identified notes containing neuromonitoring documentation with an F1-score of 1.0, surpassing performance of standard administrative codes which had an F1-score of 0.64. In the subset of 993 patients used for training, validation, and testing a deep learning model, the prevalence of change in status was 6.5% and difficulty establishing neuromonitoring baseline signals was 6.6%. The deep learning model had an F1-score = 0.80 and AUC-ROC = 1.0 for identifying change in status, and an F1-score = 0.80 and AUC-ROC = 0.97 for identifying difficulty establishing baseline signals. Compared to gold standard manual chart review, our methodology has greater efficiency for identifying infrequent yet important types of neuromonitoring documentation. This method may facilitate large-scale quality improvement initiatives that require timely analysis of a large volume of EHRs.

Machine learning in vascular surgery: a systematic review and critical appraisal

Machine learning (ML) is a rapidly advancing field with increasing utility in health care. We conducted a systematic review and critical appraisal of ML applications in vascular surgery. MEDLINE, Embase, and Cochrane CENTRAL were searched from inception to March 1, 2021. Study screening, data extraction, and quality assessment were performed by two independent reviewers, with a third author resolving discrepancies. All original studies reporting ML applications in vascular surgery were included. Publication trends, disease conditions, methodologies, and outcomes were summarized. Critical appraisal was conducted using the PROBAST risk-of-bias and TRIPOD reporting adherence tools. We included 212 studies from a pool of 2235 unique articles. ML techniques were used for diagnosis, prognosis, and image segmentation in carotid stenosis, aortic aneurysm/dissection, peripheral artery disease, diabetic foot ulcer, venous disease, and renal artery stenosis. The number of publications on ML in vascular surgery increased from 1 (1991-1996) to 118 (2016-2021). Most studies were retrospective and single center, with no randomized controlled trials. The median area under the receiver operating characteristic curve (AUROC) was 0.88 (range 0.61-1.00), with 79.5% [62/78] studies reporting AUROC ≥ 0.80. Out of 22 studies comparing ML techniques to existing prediction tools, clinicians, or traditional regression models, 20 performed better and 2 performed similarly. Overall, 94.8% (201/212) studies had high risk-of-bias and adherence to reporting standards was poor with a rate of 41.4%. Despite improvements over time, study quality and reporting remain inadequate. Future studies should consider standardized tools such as PROBAST and TRIPOD to improve study quality and clinical applicability.

Natural Language Processing Applications in the Clinical Neurosciences: A Machine Learning Augmented Systematic Review

Natural language processing (NLP), a domain of artificial intelligence (AI) that models human language, has been used in medicine to automate diagnostics, detect adverse events, support decision making and predict clinical outcomes. However, applications to the clinical neurosciences appear to be limited. NLP has matured with the implementation of deep transformer models (e.g., XLNet, BERT, T5, and RoBERTa) and transfer learning. The objectives of this study were to (1) systematically review NLP applications in the clinical neurosciences, and (2) explore NLP analysis to facilitate literature synthesis, providing clear examples to demonstrate the potential capabilities of these technologies for a clinical audience. Our NLP analysis consisted of keyword identification, text summarization and document classification. A total of 48 articles met inclusion criteria. NLP has been applied in the clinical neurosciences to facilitate literature synthesis, data extraction, patient identification, automated clinical reporting and outcome prediction. The number of publications applying NLP has increased rapidly over the past five years. Document classifiers trained to differentiate included and excluded articles demonstrated moderate performance (XLNet AUC = 0.66, BERT AUC = 0.59, RoBERTa AUC = 0.62). The T5 transformer model generated acceptable abstract summaries. The application of NLP has the potential to enhance research and practice in the clinical neurosciences.

Vascular Injury in Elective Anterior Surgery of the Lumbar Spine: A Narrative Review

The incidence of anterior lumbar surgery is increasing as the population is aging. Although adverse events regarding vasculature injury are uncommon, several have been described in the current literature. Complications can be categorized based on the time of occurrence, more specifically intraoperative or postoperative, but also regarding the nature of vascular damage such as thrombosis, occlusion, or rupture. The rate of complications is higher in the setting of revision anterior surgery than with primary anterior lumbar surgery. Moreover, the incidence of revision anterior surgery is also increasing in contrast to the past. Through this narrative review, an effort is made for a thorough understanding of the complications associated with anterior lumbar surgery, which will aid in the prevention, recognition, and management of this rare complication.

SMART on FHIR in spine: integrating clinical prediction models into electronic health records for precision medicine at the point of care

Recent applications of artificial intelligence have shown great promise for improving the quality and efficiency of clinical care. Numerous clinical decision support tools exist in today's electronic health records (EHRs) such as medication dosing support, order facilitators (eg, procedure specific order sets), and point of care alerts. However, less has been done to integrate artificial intelligence (AI)-enabled risk predictors into EHRs despite wide availability of validated risk prediction tools. An interoperability standard known as SMART on FHIR (substitutable medical applications and reusable technologies on fast health interoperability resources) offers a promising path forward, enabling digital innovations to be seamlessly integrated with the EHR with regard to the user interface and patient data. For the next step in progress towards the goal of learning healthcare and informatics-enabled spine surgery, we propose the application of SMART on FHIR to integrate existing and new risk predictions tools in spine surgery through an EHR add-on-application.

Expedient Endovascular Hemorrhage Control During Anterior Lumbar Spinal Exposure Allows Procedural Completion in Rescued Patients

OBJECTIVES

To propose a contemporary management strategy for venous injury during anterior lumbar spinal exposure that incorporates endovascular treatment.

METHODS

Vein injuries suffered by patients treated in a single practice were reviewed. A treatment algorithm based on these experiences was formulated.

RESULTS

Between 2015 and 2018, 914 patients received anterior access procedures for indicated lumbar interbody fusions. Of these patients, 15 (1.6%) suffered minor vascular injuries treated with manual pressure or suture repair. Four (0.4%) patients undergoing anterior lumbar spine surgery suffered major venous injuries, all of whom received the indicated spinal hardware following endovascular rescue. Primary repair was attempted in three patients before endovascular control and not at all in one. Vascular access was obtained via the bilateral femoral veins in 2 patients, unilateral femoral in one, and bilateral femoral plus right internal jugular vein in one. Stent choice included both uncovered (5, 63%) and covered stents (3, 38%). Deep venous thrombosis occurred in 2 patient's post-treatment. 1 DVT was encountered in the setting of a covered stent and 1 uncovered stent thrombosis was treated with catheter-directed lysis 4 weeks post-operatively. Ultimately, 3 patients were therapeutically anticoagulated. Mean follow-up is 13 months (range 1-36) with duplex ultrasounds available at 6 months or later in 3 of 4 patients. There is no evidence of post-thrombotic syndrome in the 2 patients that developed DVT's or in-stent stenosis in the 3 patients with available follow-up imaging.

CONCLUSIONS

Endovascular techniques are important adjuncts when controlling large-volume hemorrhage associated with venous tears during anterior spinal exposure. Adequate direct compression allowing occlusion balloon inflation are key steps to reduce blood loss. Covered and uncovered stents are both appropriate choices to treat injuries. Patients must be anticoagulated post-operatively and surveilled for the sequelae of venous insufficiency. With expedient hemostasis, the indicated spinal surgery may be safely completed.

Postoperative adverse events secondary to iatrogenic vascular injury during anterior lumbar spinal surgery

BACKGROUND

Anterior lumbar spine surgery (ALSS) requires mobilization of the great vessels, resulting in a high risk of iatrogenic vascular injury (VI). It remains unclear whether VI is associated with increased risk of postoperative complications and other related adverse outcomes.

PURPOSE

The purpose of this study was to (1) assess the incidence of postoperative complications attributable to VI during ALSS, and (2) outcomes secondary to VI such as procedural blood loss, transfusion of blood products, length of stay (LOS), and in hospital mortality.

STUDY DESIGN

Retrospective propensity-score matched, case-control study at 2 academic and 3 community medical centers, PATIENT SAMPLE: Patients 18 years of age or older, undergoing ALSS between January 1st, 2000 and July 31st, 2019 were included in this analysis.

OUTCOME MEASURES

The primary outcome was the incidence of postoperative complications attributable to VI, such as venous thromboembolism, compartment syndrome, transfusion reaction, limb ischemia, and reoperations. The secondary outcomes included estimated operative blood loss (milliliter), transfused blood products, LOS (days), and in-hospital mortality.

METHODS

In total, 1,035 patients were identified, of which 75 (7.2%) had a VI. For comparative analyses, the 75 VI patients were paired with 75 comparable non-VI patients by propensity-score matching. The adequacy of the matching was assessed by testing the standardized mean differences (SMD) between VI and non-VI group (>0.25 SMD).

RESULTS

Two patients (2.7%) had VI-related postoperative complications in the studied period, which consisted of two deep venous thromboembolisms (DVTs) occurring on day 3 and 7 postoperatively. Both DVTs were located in the distal left common iliac vein (CIV). The VI these patients suffered were to the distal inferior vena cava and the left CIV, respectively. Both patients did not develop additional complications in consequence of their DVTs, however, did require systemic anticoagulation and placement of an inferior vena cava filter. There was no statistical difference with the non-VI group where no instances (0%) of postoperative complications were reported (p=.157). No differences were found in LOS or in hospital mortality between the two groups (p=.157 and p=.999, respectively). Intraoperative blood loss and blood transfusion were both found to be higher in the VI group in comparison to the non-VI group (650 mL, interquartile range [IQR] 300-1400 vs. 150 mL, IQR 50-425, p≤.001; 0 units, IQR 0-3 vs. 0 units, IQR 0-1, p=.012, respectively).

CONCLUSION

This study found a low number of serious postoperative complications related to VI in ALSS. In addition, these complications were not significantly different between the VI and matched non-VI ALSS cohort. Although not significant, the found DVT incidence of 2.7% after VI in ALSS warrants vigilance and preventive measures during the postoperative course of these patients.