Bydon, Mohamad

OBJECTIVE The aim of this study was to identify predictors of the best 24-month improvements in patients undergoing surgery for cervical spondylotic myelopathy (CSM). For this purpose, the authors leveraged a large prospective cohort of surgically treated patients with CSM to identify factors predicting the best outcomes for disability, quality of life, and functional status following surgery. METHODS This was a retrospective analysis of prospectively collected data. The Quality Outcomes Database (QOD) CSM dataset (1141 patients) at 14 top enrolling sites was used. Baseline and surgical characteristics were compared for those reporting the top and bottom 20th percentile 24-month Neck Disability Index (NDI), EuroQol-5D (EQ-5D), and modified Japanese Orthopaedic Association (mJOA) change scores. A multivariable logistic model was constructed and included candidate variables reaching p ≤ 0.20 on univariate analyses. Least important variables were removed in a stepwise manner to determine the significant predictors of the best outcomes (top 20th percentile) for 24-month NDI, EQ-5D, and mJOA change. RESULTS A total of 948 (83.1%) patients with 24-month follow-up were included in this study. For NDI, 204 (17.9%) had the best NDI outcome and 200 (17.5%) had the worst NDI outcome. Factors predicting the best NDI outcomes included symptom duration less than 12 months (OR 1.5, 95% CI 1.1–1.9; p = 0.01); procedure other than posterior fusion (OR 1.5, 95% CI 1.03–2.1; p = 0.03); higher preoperative visual analog scale neck pain score (OR 1.2, 95% CI 1.1–1.3; p < 0.001); and higher baseline NDI (OR 1.06, 95% CI 1.05–1.07; p < 0.001). For EQ-5D, 163 (14.3%) had the best EQ-5D outcome and 169 (14.8%) had the worst EQ-5D outcome. Factors predicting the best EQ-5D outcomes included arm pain–only complaints (compared to neck pain) (OR 1.9, 95% CI 1.3–2.9; p = 0.002) and lower baseline EQ-5D (OR 167.7 per unit lower, 95% CI 85.0–339.4; p < 0.001). For mJOA, 222 (19.5%) had the best mJOA outcome and 238 (20.9%) had the worst mJOA outcome. Factors predicting the best mJOA outcomes included lower BMI (OR 1.03 per unit lower, 95% CI 1.004–1.05; p = 0.02; cutoff value of ≤ 29.5 kg/m 2 ); arm pain–only complaints (compared to neck pain) (OR 1.7, 95% CI 1.1–2.5; p = 0.02); and lower baseline mJOA (OR 1.6 per unit lower, 95% CI 1.5–1.7; p < 0.001). CONCLUSIONS Compared to the worst outcomes for EQ-5D, the best outcomes were associated with patients with arm pain–only complaints. For mJOA, lower BMI and arm pain–only complaints portended the best outcomes. For NDI, those with the best outcomes had shorter symptom durations, higher preoperative neck pain scores, and less often underwent posterior spinal fusions. Given the positive impact of shorter symptom duration on outcomes, these data suggest that early surgery may be beneficial for patients with CSM.

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.

Study Design: Retrospective review of a prospectively maintained database. Objective: Assess differences in preoperative status and postoperative outcomes among patients of different educational backgrounds undergoing surgical management of cervical spondylotic myelopathy (CSM). Summary of Background Data: Patient education level (EL) has been suggested to correlate with health literacy, disease perception, socioeconomic status (SES), and access to health care. Methods: The CSM data set of the Quality Outcomes Database (QOD) was queried for patients undergoing surgical management of CSM. EL was grouped as high school or below, graduate-level, and postgraduate level. The association of EL with baseline disease severity (per patient-reported outcome measures), symptoms >3 or ≤3 months, and 24-month patient-reported outcome measures were evaluated. Results: Among 1141 patients with CSM, 509 (44.6%) had an EL of high school or below, 471 (41.3%) had a graduate degree, and 161 (14.1%) had obtained postgraduate education. Lower EL was statistically significantly associated with symptom duration of >3 months (odds ratio=1.68), higher arm pain numeric rating scale (NRS) (coefficient=0.5), and higher neck pain NRS (coefficient=0.79). Patients with postgraduate education had statistically significantly lower Neck Disability Index (NDI) scores (coefficient=-7.17), lower arm pain scores (coefficient=−1), and higher quality-adjusted life-years (QALY) scores (coefficient=0.06). Twenty-four months after surgery, patients of lower EL had higher NDI scores, higher pain NRS scores, and lower QALY scores (P<0.05 in all analyses). Conclusions: Among patients undergoing surgical management for CSM, those reporting a lower educational level tended to present with longer symptom duration, more disease-inflicted disability and pain, and lower QALY scores. As such, patients of a lower EL are a potentially vulnerable subpopulation, and their health literacy and access to care should be prioritized.

OBJECTIVE Diabetes mellitus (DM) is a known risk factor for postsurgical and systemic complications after lumbar spinal surgery. Smaller studies have also demonstrated diminished improvements in patient-reported outcomes (PROs), with increased reoperation and readmission rates after lumbar surgery in patients with DM. The authors aimed to examine longer-term PROs in patients with DM undergoing lumbar decompression and/or arthrodesis for degenerative pathology. METHODS The Quality Outcomes Database was queried for patients undergoing elective lumbar decompression and/or arthrodesis for degenerative pathology. Patients were grouped into DM and non-DM groups and optimally matched in a 1:1 ratio on 31 baseline variables, including the number of operated levels. Outcomes of interest were readmissions and reoperations at 30 and 90 days after surgery in addition to improvements in Oswestry Disability Index, back pain, and leg pain scores and quality-adjusted life-years at 90 days after surgery. RESULTS The matched decompression cohort comprised 7836 patients (3236 [41.3] females) with a mean age of 63.5 ± 12.6 years, and the matched arthrodesis cohort comprised 7336 patients (3907 [53.3%] females) with a mean age of 64.8 ± 10.3 years. In patients undergoing lumbar decompression, no significant differences in nonroutine discharge, length of stay (LOS), readmissions, reoperations, and PROs were observed. In patients undergoing lumbar arthrodesis, nonroutine discharge (15.7% vs 13.4%, p < 0.01), LOS (3.2 ± 2.0 vs 3.0 ± 3.5 days, p < 0.01), 30-day (6.5% vs 4.4%, p < 0.01) and 90-day (9.1% vs 7.0%, p < 0.01) readmission rates, and the 90-day reoperation rate (4.3% vs 3.2%, p = 0.01) were all significantly higher in the DM group. For DM patients undergoing lumbar arthrodesis, subgroup analyses demonstrated a significantly higher risk of poor surgical outcomes with the open approach. CONCLUSIONS Patients with and without DM undergoing lumbar spinal decompression alone have comparable readmission and reoperation rates, while those undergoing arthrodesis procedures have a higher risk of poor surgical outcomes up to 90 days after surgery. Surgeons should target optimal DM control preoperatively, particularly for patients undergoing elective lumbar arthrodesis.

In a period when the budding field of neurosurgery was believed to have little promise, Dr Alfred Washington Adson founded and led the first neurosurgical department at Mayo Clinic. He was not without reservations—surgical intervention for neurological conditions was rarely pursued because of poor outcomes and high complication rates, and Dr Adson acknowledged his early concerns about the future of neurosurgery in his memoirs. However, his education, mentorship, his training, and his first neurosurgical cases helped to shape the impact he ultimately had on the field and his legacy as a neurosurgeon. Dr Adson trained with several renowned Mayo general surgeons, notably his mentor Dr Emil Beckman, whose desire for operative precision shaped Dr Adson's drive to develop his own skills as a surgeon. Two years into his residency, he became the youngest staff surgeon and was tasked with managing the neurosurgical cases at Mayo. The five neurosurgical cases overseen by Dr Adson in the next year illuminated the opportunity for neurosurgery to drastically improve the lives of patients. Dr Adson, given the option of continuing as either a general surgeon or a neurosurgeon, ultimately chose to pursue neurosurgery. This article seeks to provide a historical perspective on the neurosurgeon Dr Alfred Washington Adson using primary and secondary accounts from the Mayo archives, highlighting his contributions to the early understanding of intracranial pathology and how his early experiences as a trainee developed into a personal passion for self-improvement, education, and advocacy for health care in America.

Abstract INTRODUCTION Glioblastoma (GBM) is the most common primary malignant brain tumor and has a poor clinical outcome. Extracellular vesicles (EVs) are promising biomarkers to assist the differential diagnosis between pseudoprogression and true tumor progression. We aimed to identify non-tumor plasma EV markers that predicted GBM occurrence using trained multiple tree-based machine learning algorithms. METHODS We included plasma samples of 40 GBM and 40 non-GBM (20 brain metastasis, 20 normal donors). Spectral flow cytometry was performed to determine plasma EV subpopulation frequency based on CD9,CD81,CD63,CD11b,CD45,CD31, and CD41a expression. Multiparametric t-SNE and FlowSOM analysis clustered plasma EVs based on these markers. Decision trees and random forest classifiers were trained and tested to predict the disease status (GBM vs non-GBM) based on plasma EV subpopulations. RESULTS Compared to non-GBM, GBM had increased frequency of CD9+ EVs (64.52% vs 40.69%, p&lt; 0.0001), CD9+CD63-CD81- EVs (54.45% vs 37.26%, p&lt; 0.0001) and myeloid-derived EVs (CD9+CD11b+ EVs, 20.77% vs 12.48%, p = 0.0241). In the decision tree analysis, low frequency of CD9+ EVs was found in 44% of all cases and represented a chance of 75% to be classified as non-GBM cases. An algorithm based on random forest analysis (n = 500 trees) using these markers had a sensitivity of 85.71% and specificity of 77.8% with an overall accuracy of 81.25% for identifying GBM. The frequencies of CD9+, CD9+CD63-CD81-, and CD9+CD63+CD81+ were the most critical factors in predicting GBM diagnosis. FlowSOM analysis identified 20 plasma EV subpopulations but had a lower accuracy (62.5%) when combined with our machine learning algorithm. CONCLUSIONS Non-tumor plasma EV expression profile combined with tree-based machine learning demonstrated high accuracy in predicting patient disease status. CD9+ EV frequency was one of the most promising biomarkers to assist in this diagnosis. We will continue to test our established algorithm and further validate our findings in a larger cohort.

OBJECTIVE The aim of this study was to explore the preoperative patient characteristics that affect surgical decision-making when selecting an anterior or posterior operative approach in patients diagnosed with cervical spondylotic myelopathy (CSM). METHODS This was a multi-institutional, retrospective study of the prospective Quality Outcomes Database (QOD) Cervical Spondylotic Myelopathy module. Patients aged 18 years or older diagnosed with primary CSM who underwent multilevel (≥ 2-level) elective surgery were included. Demographics and baseline clinical characteristics were collected. RESULTS Of the 841 patients with CSM in the database, 492 (58.5%) underwent multilevel anterior surgery and 349 (41.5%) underwent multilevel posterior surgery. Surgeons more often performed a posterior surgical approach in older patients (mean 64.8 ± 10.6 vs 58.5 ± 11.1 years, p < 0.001) and those with a higher American Society of Anesthesiologists class (class III or IV: 52.4% vs 46.3%, p = 0.003), a higher rate of motor deficit (67.0% vs 58.7%, p = 0.014), worse myelopathy (mean modified Japanese Orthopaedic Association score 11.4 ± 3.1 vs 12.4 ± 2.6, p < 0.001), and more levels treated (4.3 ± 1.3 vs 2.4 ± 0.6, p < 0.001). On the other hand, surgeons more frequently performed an anterior surgical approach when patients were employed (47.2% vs 23.2%, p < 0.001) and had intervertebral disc herniation as an underlying pathology (30.7% vs 9.2%, p < 0.001). CONCLUSIONS The selection of approach for patients with CSM depends on patient demographics and symptomology. Posterior surgery was performed in patients who were older and had worse systemic disease, increased myelopathy, and greater levels of stenosis. Anterior surgery was more often performed in patients who were employed and had intervertebral disc herniation.

OBJECTIVE The literature on non–small cell lung cancer (NSCLC) brain metastases (BMs) managed using stereotactic radiosurgery (SRS) relies mainly on single-institution studies or randomized controlled trials (RCTs). There is a literature gap on clinical and radiological outcomes of SRS for NSCLC metastases in real-world practice. The objective of this study was to benchmark mortality and progression outcomes in patients undergoing SRS for NSCLC BMs and identify risk factors for these outcomes using a national quality registry. METHODS The SRS Registry of the NeuroPoint Alliance was used for this study. This registry included patients from 16 enrolling sites who underwent SRS from 2017 to 2022. Data are prospectively collected without a prespecified research purpose. The main outcomes of this analysis were overall survival (OS), out-of-field recurrence, local progression, and intracranial progression. All time-to-event investigations included Kaplan-Meier analyses and multivariable Cox regressions. RESULTS Two hundred sixty-four patients were identified, with a mean age of 66.7 years and a female proportion of 48.5%. Most patients (84.5%) had a Karnofsky Performance Status (KPS) score of 80–100, and the mean baseline EQ-5D score was 0.539 quality-adjusted life years. A single lesion was present in 53.4% of the patients, and 29.1% of patients had 3 or more lesions. The median OS was 28.1 months, and independent predictors of mortality included no control of primary tumor (hazard ratio [HR] 2.1), KPS of 80 (HR 2.4) or lower (HR 2.4), coronary artery disease (HR 2.8), and 5 or more lesions present at the time of SRS treatment (HR 2.3). The median out-of-field progression-free survival (PFS) was 24.8 months, and the median local PFS was unreached. Intralesional hemorrhage was an independent risk factor of local progression, with an HR of 6.0. The median intracranial PFS was 14.0 months and was predicted by the number of lesions at the time of SRS (3–4 lesions, HR 2.2; 5–14 lesions, HR 2.5). CONCLUSIONS In this real-world prospective study, the authors used a national quality registry and found favorable OS in patients with NSCLC BMs undergoing SRS compared with results from previously published RCTs. The intracranial PFS was mainly driven by the emergence of new lesions rather than local progression. A greater number of lesions at baseline was associated with out-of-field progression, while intralesional hemorrhage at baseline was associated with local progression.

OBJECTIVE Chordomas are rare tumors that often recur regardless of surgery with negative margins and postoperative radiotherapy. The predictive accuracy of widely used immunohistochemical (IHC) markers in addressing the recurrence of skull base chordomas (SBCs) is yet to be determined. This study aimed to investigate IHC markers in the prediction of recurrence after SBC resection with adjuvant radiation therapy. METHODS The authors reviewed the records of patients who had treatment for SBC between January 2017 and June 2021 across the Mayo Clinic in Minnesota, Florida, and Arizona. Exclusion criteria included patients who had no histopathology or recurrence as an outcome. Histopathological markers included cytokeratin A1/A3 only, epithelial membrane antigen (EMA), S100 protein, pan-cytokeratin, IN1, GATA3, CAM5.2, OSCAR, and chondroid. Information from patient records was abstracted, including treatment, clinical and radiological follow-up duration, demographics, and histopathological factors. Decision tree and random forest classifiers were trained and tested to predict the recurrence based on unseen data using an 80/20 split. RESULTS A total of 38 patients with a diagnosis of SBC who underwent resection (gross-total resection: 42.1%; and subtotal resection: 57.9%) and radiation therapy were extracted from the medical records. The mean patient age was 48.2 (SD 19.6) years; most patients were male (n = 23; 60.5%) and White (n = 36; 94.7%). Pan-cytokeratin was associated with an increased risk of postoperative recurrence (OR 14.67, 95% CI 2.44–88.13; p = 0.00517) after resection and adjuvant radiotherapy. The decision tree analysis found pan-cytokeratin–positive tumors to have a 78% chance of being classified as a recurrence, with an accuracy of 75%. The distribution of minimal depth in the prediction of postoperative recurrence indicates that the most important variables were pan-cytokeratin, followed by cytokeratin A1/A3 and EMA. CONCLUSIONS The authors’ machine learning algorithm identified pan-cytokeratin as the largest contributor to recurrence among other IHC markers after SBC resection. Machine learning may facilitate the prediction of outcomes in rare tumors, such as chordomas.