Multicenter external validation of prediction models for clinical outcomes after spinal fusion for lumbar degenerative disease.

BACKGROUND: Clinical prediction models (CPM), such as the SCOAP-CERTAIN tool, can be utilized to enhance decision-making for lumbar spinal fusion surgery by providing quantitative estimates of outcomes, aiding surgeons in assessing potential benefits and risks for each individual patient. External validation is crucial in CPM to assess generalizability beyond the initial dataset. This ensures performance in diverse populations, reliability and real-world applicability of the results. Therefore, we externally validated the tool for predictability of improvement in oswestry disability index (ODI), back and leg pain (BP, LP). METHODS: Prospective and retrospective data from multicenter registry was obtained. As outcome measure minimum clinically important change was chosen for ODI with ≥ 15-point and ≥ 2-point reduction for numeric rating scales (NRS) for BP and LP 12 months after lumbar fusion for degenerative disease. We externally validate this tool by calculating discrimination and calibration metrics such as intercept, slope, Brier Score, expected/observed ratio, Hosmer-Lemeshow (HL), AUC, sensitivity and specificity. RESULTS: We included 1115 patients, average age 60.8 ± 12.5 years. For 12-month ODI, area-under-the-curve (AUC) was 0.70, the calibration intercept and slope were 1.01 and 0.84, respectively. For NRS BP, AUC was 0.72, with calibration intercept of 0.97 and slope of 0.87. For NRS LP, AUC was 0.70, with calibration intercept of 0.04 and slope of 0.72. Sensitivity ranged from 0.63 to 0.96, while specificity ranged from 0.15 to 0.68. Lack of fit was found for all three models based on HL testing. CONCLUSIONS: Utilizing data from a multinational registry, we externally validate the SCOAP-CERTAIN prediction tool. The model demonstrated fair discrimination and calibration of predicted probabilities, necessitating caution in applying it in clinical practice. We suggest that future CPMs focus on predicting longer-term prognosis for this patient population, emphasizing the significance of robust calibration and thorough reporting.

A critical moment in machine learning in medicine: on reproducible and interpretable learning.

Over the past two decades, advances in computational power and data availability combined with increased accessibility to pre-trained models have led to an exponential rise in machine learning (ML) publications. While ML may have the potential to transform healthcare, this sharp increase in ML research output without focus on methodological rigor and standard reporting guidelines has fueled a reproducibility crisis. In addition, the rapidly growing complexity of these models compromises their interpretability, which currently impedes their successful and widespread clinical adoption. In medicine, where failure of such models may have severe implications for patients' health, the high requirements for accuracy, robustness, and interpretability confront ML researchers with a unique set of challenges. In this review, we discuss the semantics of reproducibility and interpretability, as well as related issues and challenges, and outline possible solutions to counteracting the "black box". To foster reproducibility, standard reporting guidelines need to be further developed and data or code sharing encouraged. Editors and reviewers may equally play a critical role by establishing high methodological standards and thus preventing the dissemination of low-quality ML publications. To foster interpretable learning, the use of simpler models more suitable for medical data can inform the clinician how results are generated based on input data. Model-agnostic explanation tools, sensitivity analysis, and hidden layer representations constitute further promising approaches to increase interpretability. Balancing model performance and interpretability are important to ensure clinical applicability. We have now reached a critical moment for ML in medicine, where addressing these issues and implementing appropriate solutions will be vital for the future evolution of the field.

How can we optimize the long-term outcome in children with intracranial cavernous malformations? A single-center experience of 61 cases.

The objective is to provide a treatment algorithm for pediatric patients with intracranial cavernous malformations (CMs) based on our experience. Patients < 18 years of age who were treated either surgically or conservatively at the authors' institution between 1982 and 2019 were retrospectively evaluated. A total of 61 pediatric patients were treated at the authors' institution: 39 with lobar CMs; 18 with deep CMs, including 12 in the brainstem and 6 in the basal ganglia; and 4 with CMs in the cerebellar hemispheres. Forty-two patients underwent surgery, and 19 were treated conservatively. The median follow-up time was 65 months (1-356 months). In surgically treated patients, lesions were larger (2.4 cm vs 0.9 cm, p < 0.001). In patients with lobar CMs, seizures were more common (72% vs 21%, p = 0.003) in the surgery group than in conservatively managed patients. In deep CMs, modified Rankin scale (mRS) was higher (4 vs 1, p = 0.003) in the surgery group than in conservatively treated patients. At the time of last follow-up, no differences in Wieser outcome class I were seen (86% vs 67%) in lobar CMs, and mRS scores had aligned between the treatment groups in deep CMs (1 vs 0). We encountered no new permanent neurological deficit at time of last follow-up. We propose a treatment algorithm according to lesion location and size, burden of symptoms, epilepsy workup, and further clinical course during observation. A conservative management is safe in pediatric patients with asymptomatic CMs. Gross total resection should be the aim in patients with symptomatic lobar CMs. A less aggressive approach with subtotal resection, when required to prevent neurological compromise, sustainably improves neurological outcome in patients with deep CMs.

FUSE-ML: development and external validation of a clinical prediction model for mid-term outcomes after lumbar spinal fusion for degenerative disease.

BACKGROUND: Indications and outcomes in lumbar spinal fusion for degenerative disease are notoriously heterogenous. Selected subsets of patients show remarkable benefit. However, their objective identification is often difficult. Decision-making may be improved with reliable prediction of long-term outcomes for each individual patient, improving patient selection and avoiding ineffective procedures. METHODS: Clinical prediction models for long-term functional impairment [Oswestry Disability Index (ODI) or Core Outcome Measures Index (COMI)], back pain, and leg pain after lumbar fusion for degenerative disease were developed. Achievement of the minimum clinically important difference at 12 months postoperatively was defined as a reduction from baseline of at least 15 points for ODI, 2.2 points for COMI, or 2 points for pain severity. RESULTS: Models were developed and integrated into a web-app ( https://neurosurgery.shinyapps.io/fuseml/ ) based on a multinational cohort [N = 817; 42.7% male; mean (SD) age: 61.19 (12.36) years]. At external validation [N = 298; 35.6% male; mean (SD) age: 59.73 (12.64) years], areas under the curves for functional impairment [0.67, 95% confidence interval (CI): 0.59-0.74], back pain (0.72, 95%CI: 0.64-0.79), and leg pain (0.64, 95%CI: 0.54-0.73) demonstrated moderate ability to identify patients who are likely to benefit from surgery. Models demonstrated fair calibration of the predicted probabilities. CONCLUSIONS: Outcomes after lumbar spinal fusion for degenerative disease remain difficult to predict. Although assistive clinical prediction models can help in quantifying potential benefits of surgery and the externally validated FUSE-ML tool may aid in individualized risk-benefit estimation, truly impacting clinical practice in the era of "personalized medicine" necessitates more robust tools in this patient population.

Automated volumetric assessment of pituitary adenoma.

PURPOSE: Assessment of pituitary adenoma (PA) volume and extent of resection (EOR) through manual segmentation is time-consuming and likely suffers from poor interrater agreement, especially postoperatively. Automated tumor segmentation and volumetry by use of deep learning techniques may provide more objective and quick volumetry. METHODS: We developed an automated volumetry pipeline for pituitary adenoma. Preoperative and three-month postoperative T1-weighted, contrast-enhanced magnetic resonance imaging (MRI) with manual segmentations were used for model training. After adequate preprocessing, an ensemble of convolutional neural networks (CNNs) was trained and validated for preoperative and postoperative automated segmentation of tumor tissue. Generalization was evaluated on a separate holdout set. RESULTS: In total, 193 image sets were used for training and 20 were held out for validation. At validation using the holdout set, our models (preoperative / postoperative) demonstrated a median Dice score of 0.71 (0.27) / 0 (0), a mean Jaccard score of 0.53 ± 0.21/0.030 ± 0.085 and a mean 95th percentile Hausdorff distance of 3.89 ± 1.96./12.199 ± 6.684. Pearson's correlation coefficient for volume correlation was 0.85 / 0.22 and -0.14 for extent of resection. Gross total resection was detected with a sensitivity of 66.67% and specificity of 36.36%. CONCLUSIONS: Our volumetry pipeline demonstrated its ability to accurately segment pituitary adenomas. This is highly valuable for lesion detection and evaluation of progression of pituitary incidentalomas. Postoperatively, however, objective and precise detection of residual tumor remains less successful. Larger datasets, more diverse data, and more elaborate modeling could potentially improve performance.

Hematologic and Inflammatory Predictors of Outcome in Patients with Brain Arteriovenous Malformations.

OBJECTIVE: This study investigated the prognostic value of admission blood counts for arteriovenous malformation (AVM) outcomes and compared admission blood counts for patients with ruptured and unruptured AVMs. METHODS: A retrospective analysis of patients who underwent surgical treatment for a ruptured cerebral AVM between February 1, 2014, and March 31, 2020, was conducted. The primary outcome was poor neurologic outcome, defined as a modified Rankin Scale score ≥2 in patients with unruptured AVMs or >2 in patients with ruptured AVMs. RESULTS: Of 235 included patients, 80 (34%) had ruptured AVMs. At admission, patients with ruptured AVMs had a significantly lower mean (SD) hemoglobin level (12.78 [2.07] g/dL vs. 13.71 [1.60] g/dL, P < 0.001), hematocrit (38.1% [5.9%] vs. 40.7% [4.6%], P < 0.001), lymphocyte count (16% [11%] vs. 26% [10%], P < 0.001), and absolute lymphocyte count (1.41 [0.72] × 103/µL vs. 1.79 [0.68] × 103/µL, P < 0.001), and they had a significantly higher mean (SD) white blood cell count (10.4 [3.8] × 103/µL vs. 7.6 [2.3] × 103/µL, P < 0.001), absolute neutrophil count (7.8 [3.8] × 103/µL vs. 5.0 [2.5] × 103/µL, P < 0.001), and neutrophil count (74% [14%] vs. 64% [13%], P < 0.001). Among patients with unruptured AVMs, white blood cell count ≥6.4 × 103/µL and absolute neutrophil count ≥3.4 × 103/µL were associated with a favorable neurologic outcome, whereas hemoglobin level ≥13.4 g/dL was associated with an unfavorable outcome. Among patients with ruptured AVMs, hypertension was associated with a 3-fold increase in odds of a poor neurologic outcome. CONCLUSIONS: Patients with ruptured and unruptured AVMs present with characteristic profiles of hematologic and inflammatory parameters evident in their admission blood work.

Optimal PHASES Scoring for Risk Stratification of Surgically Treated Unruptured Aneurysms.

OBJECTIVE: The PHASES (Population, Hypertension, Age, Size, Earlier subarachnoid hemorrhage, Site) score was developed to facilitate risk stratification for management of unruptured intracranial aneurysms (UIAs). This study aimed to identify the optimal PHASES score cutoff for predicting neurologic outcomes in patients with surgically treated aneurysms. METHODS: All patients who underwent microneurosurgical treatment for UIA at a large quaternary center from January 1, 2014, to December 31, 2020, were retrospectively reviewed. Inclusion criteria included a modified Rankin Scale (mRS) score of ≤2 at admission. The primary outcome was 1-year mRS score, with a "poor" neurologic outcome defined as an mRS score >2. RESULTS: In total, 375 patients were included in the analysis. The mean (SD) PHASES score for the entire study population was 4.47 (2.67). Of 375 patients, 116 (31%) had a PHASES score ≥6, which was found to maximize prediction of poor neurologic outcome. Patients with PHASES scores ≥6 had significantly higher rates of poor neurologic outcome than patients with PHASES scores <6 at discharge (58 [50%] vs. 90 [35%], P = 0.005) and follow-up (20 [17%] vs. 18 [6.9%], P = 0.002). After adjusting for age, Charlson Comorbidity Index score, nonsaccular aneurysm, and aneurysm size, PHASES score ≥6 remained a significant predictor of poor neurologic outcome at follow-up (odds ratio, 2.75; 95% confidence interval, 1.42-5.36, P = 0.003). CONCLUSIONS: In this retrospective analysis, a PHASES score ≥6 was associated with significantly greater proportions of poor outcome, suggesting that awareness of this threshold in PHASES scoring could be useful in risk stratification and UIA management.

The Site of Origin of Medulloblastoma: Surgical Observations Correlated to Molecular Groups.

Developmental gene expression data from medulloblastoma (MB) suggest that WNT-MB originates from the region of the embryonic lower rhombic lip (LRL), whereas SHH-MB and non-WNT/non-SHH MB arise from cerebellar precursor matrix regions. This study aimed to analyze detailed intraoperative data with regard to the site of origin (STO) and compare these findings with the hypothesized regions of origin associated with the molecular group. A review of the institutional database identified 58 out of 72 pediatric patients who were operated for an MB at our department between 1996 and 2020 that had a detailed operative report and a surgical video as well as clinical and genetic classification data available for analysis. The STO was assessed based on intraoperative findings. Using the intraoperatively defined STO, "correct" prediction of molecular groups was feasible in 20% of WNT-MB, 60% of SHH-MB and 71% of non-WNT/non-SHH MB. The positive predictive values of the neurosurgical inspection to detect the molecular group were 0.21 (95% CI 0.08-0.48) for WNT-MB, 0.86 (95% CI 0.49-0.97) for SHH-MB and 0.73 (95% CI 0.57-0.85) for non-WNT/non-SHH MB. The present study demonstrated a limited predictive value of the intraoperatively observed STO for the prediction of the molecular group of MB.

Mortality After Microsurgical Treatment of Unruptured Intracranial Aneurysms in the Modern Era.

BACKGROUND: The incidence of mortality after treatment of unruptured intracranial aneurysms (UIAs) has been described historically. However, many advances in microsurgical treatment have since emerged, and most available data are outdated. We analyzed the incidence of mortality after microsurgical treatment of patients with UIAs treated in the past decade. METHODS: The medical records of all patients with UIAs who underwent elective treatment at our large quaternary center from January 1, 2014, to December 31, 2020, were reviewed retrospectively. We analyzed mortality at discharge and 1-year follow-up as the primary outcome using univariate to multivariable progression with P < 0.20 inclusion. RESULTS: During the 7-year study period, 488 patients (mean [SD] age = 58 [12] years) had UIAs treated microsurgically. Of these patients, 61 (12.5%) had a prior subarachnoid hemorrhage. One patient (0.2%) with a dolichoectatic vertebrobasilar aneurysm died while hospitalized, and 7 other patients (8 total; 1.6%) were determined to have died at 1-year follow-up (1 trauma, 2 myocardial infarction, 2 cerebrovascular accident, 1 pulmonary embolism, and 1 subdural hematoma complicated by abscess). On univariate analysis, significant risk factors for mortality at follow-up included diabetes mellitus, preoperative anticoagulant or antiplatelet use, aneurysm calcification, nonsaccular aneurysm, and higher American Society of Anesthesiologists grades (all P < 0.03). On multivariable logistic regression analysis, only nonsaccular aneurysms and higher American Society of Anesthesiologists grades were predictors of mortality. CONCLUSIONS: A low mortality rate is associated with recent microsurgical treatment of UIAs. However, nonsaccular aneurysms and higher American Society of Anesthesiologists grades appear to be predictors of mortality.