Association of Patient Frailty With Vestibular Schwannoma Resection Outcomes and Machine Learning Development of a Vestibular Schwannoma Risk Stratification Score

Predictors of Prolonged Hospital Stay After Microsurgery for Vestibular Schwannoma: Analysis of a Decade of Data

IMPORTANCE

Microsurgical resection is one of the treatments for vestibular schwannomas (VS). While several factors have been linked to increased length of stay (LOS) for VS patients undergoing microsurgery, a better understanding of these factors is important to provide prognostic information for patients.

OBJECTIVE

Determine predictors of increased LOS for VS patients undergoing microsurgical resection.

DESIGN

Retrospective analysis using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2010 to 2020.

SETTING

Database review.

PARTICIPANTS

All patients who underwent microsurgery (CPT codes 61520, 61526/61596) for the management of vestibular schwannoma (ICD9 and ICD10 codes 225.1, D33.3) were included.

MAIN OUTCOMES AND MEASURES

Analyzing perioperative factors that can predict prolonged hospital stay.

RESULTS

A total of 2096 cases were identified and 1,188 (57%) of these patients were female. The mean age was 51.0 ± 14.0 years. Factors contributing to prolonged LOS included African American race (OR = 2.11, 95% CI: 1.32-3.36, p = 0.002), insulin-dependent diabetes mellitus (OR = 2.12, 95% CI: 1.09-4.4.11, p = 0.026), hypertension (OR = 1.26, 95% CI: 1-1.58, p = 0.046), functional dependency (OR = 5.22, 95% CI: 2.31-11.79, p = 0.001), prior steroid use (OR = 1.96, 95% CI: 1.18-3.15, p = 0.009), ASA class III (OR = 2.06, 95% CI: 1.18-3.6, p < 0.011), ASA class IV (OR = 6.34, 95% CI: 2.62-15.33, p < 0.001), and prolonged operative time (OR = 2.14, 95% CI: 1.76-2.61). Microsurgery by a translabyrinthine (TL) approach compared to a retrosigmoid (RSG) approach had lower odds of prolonged LOS (OR = 0.67, 95% CI: 0.54-0.82, p < 0.001). In a separate analysis regarding patients receiving reoperation, operative time was the only predictor of prolonged LOS (OR = 2.77, 95% CI: 1.39-5.53, p = 0.004.).

CONCLUSIONS AND RELEVANCE

Our analysis offers an analysis of the factors associated with a prolonged LOS for the surgical management of VS. By identifying healthcare disparities, targeting modifiable factors, and applying risk stratification based on demographics and comorbidities, we can work toward reducing disparities in LOS and enhancing patient outcomes.

Machine learning application in otology

This review presents a comprehensive history of Artificial Intelligence (AI) in the context of the revolutionary application of machine learning (ML) to medical research and clinical utilization, particularly for the benefit of researchers interested in the application of ML in otology. To this end, we discuss the key components of ML-input, output, and algorithms. In particular, some representation algorithms commonly used in medical research are discussed. Subsequently, we review ML applications in otology research, including diagnosis, influential identification, and surgical outcome prediction. In the context of surgical outcome prediction, specific surgical treatments, including cochlear implantation, active middle ear implantation, tympanoplasty, and vestibular schwannoma resection, are considered. Finally, we highlight the obstacles and challenges that need to be overcome in future research.

Clinical characteristics and outcomes after trigeminal schwannoma resection: a multi-institutional experience

Given their rarity, the clinical course of patients undergoing trigeminal schwannoma (TS) resection remains understudied. The objective of this study is to describe clinical characteristics and outcomes in patients undergoing surgical resection for TS in a multi-institutional cohort. This is a retrospective study of patients undergoing TS resection at two institutions between 2004 and 2022. Patient, radiographic, and clinical characteristics were reviewed and analyzed with standard statistical methods. Thirty patients were included. The median patient age was 43 (IQR: 35-52) years, and 14 (47%) patients were female. Median clinical and radiographic follow-ups were 43 (IQR: 20-81) and 47 (IQR: 27-97) months respectively. The most common presenting symptoms were trigeminal hypesthesia (57%) and headaches (30%), diplopia (30%), and ataxia/cerebellar signs (30%). The median maximum tumor diameter was 3.3 (IQR: 2.5-5.4) cm. Most tumors were Samii type C (50%) and mixed cystic-solid (63%). Surgical approaches included endoscopic endonasal (33%), supratentorial (30%), combined/staged (20%), infratentorial (10%), and anterior petrosal (7%) approaches. Gross-total resection was achieved in 16 (53%) patients. Radiographic tumor recurrence was noted in four patients at a median of 79 (range 5-152) months. Twenty-six (87%) patients reported improvements in at least one symptom by last follow-up. The most common perioperative complication was new cranial nerve deficit, with 17% of patients having a transient deficit and 10% having a permanent cranial nerve deficit. Surgical resection of TS showed good progression-free survival and symptom improvement, but was associated with cranial nerve deficits.

Advancements in Skull Base Surgery: Navigating Complex Challenges with Artificial Intelligence

Purpose

This narrative review examines the evolving landscape of artificial intelligence (AI) integration in skull base surgery, exploring its multifaceted applications and impact on various aspects of patient care.

Methods

Extensive literature review was conducted to gather insights into the role of AI in skull base surgery. Key aspects such as diagnosis, image analysis, surgical planning, navigation, predictive analytics, clinical decision-making, postoperative care, rehabilitation, and virtual simulations were explored. Studies were sourced from PubMed using keyword search strategy for relevant headings, sub-headings and cross-referencing.

Results

AI enhances early diagnosis through diagnostic algorithms that guide investigations based on clinical and radiological data. AI-driven image analysis enables accurate segmentation of intricate structures and extraction of radiomics data, optimizing preoperative planning and predicting treatment response. In surgical planning, AI aids in identifying critical structures, leading to precise interventions. Real-time AI-based navigation offers adaptive guidance, enhancing surgical accuracy and safety. Predictive analytics empower risk assessment, treatment planning, and outcome prediction. AI-driven clinical decision support systems optimize resource allocation and support shared decision-making. Postoperative care benefits from AI's monitoring capabilities and personalized rehabilitation protocols. Virtual simulations powered by AI expedite skill development and decision-making in complex procedures.

Conclusion

AI contributes to accurate diagnosis, surgical planning, navigation, predictive analysis, and postoperative care. Ethical considerations and data quality assurance are essential, ensuring responsible AI implementation. While AI serves as a valuable complement to clinical expertise, its potential to enhance decision-making, precision, and efficiency in skull base surgery is evident.

Anterior Skull Base Outcomes and Complications: A Propensity Score–Matched Evaluation of Age and Frailty as Measured by mFI-5 from the ACS-NSQIP Database

Background Frailty is increasingly recognized as a predictor of surgical outcomes; however, its utility in anterior cranial fossa (ACF) surgery remains unclear. We analyzed whether age and frailty are independent predictors of outcomes after ACF surgery using a retrospective cohort study.

Methods The American College of Surgeons National Surgical Quality Improvement Program database was queried, by Current Procedural Terminology codes, for ACF procedures in 2005 to 2020. Cases included open approaches, endoscopic approaches, and all tumor types except for pituitary adenoma. A propensity score–matched data set was analyzed via multiple logistic regression.

Results Unmatched multivariate analysis of ACF cases demonstrated that severe frailty (modified 5-item frailty index [mFI-5] ≥ 3) was independently associated with having any (odds ratio [OR] = 3.67) and minor (OR = 5.00) complications (both p < 0.001). Analysis of individual mFI-5 components demonstrated poor functional status was significantly associated with any (OR = 3.39), major (OR = 3.59), and minor (OR = 3.14) complications (all p < 0.001). After propensity score matching, only age was modestly impactful on minor complications (OR = 1.02) and extended length of stay (eLOS) (OR = 1.02) (p < 0.001). Frailty did not maintain its predictive ability after matching. Nonindependent functional status, as a subcomponent of mFI maintained significant predictive ability for any (OR = 4.94), major (OR = 4.68), and minor (OR = 4.80) complications and eLOS (OR = 2.92) (all p < 0.001).

Conclusion After propensity score matching, age demonstrated a greater ability to predict postoperative complications in ACF surgery than frailty. Rather than age or frailty, functional status served as a better outcome predictor and potential guide for patient counseling. Further validation of these findings in multicenter or disease-specific studies is warranted as well as aims to preoperatively improve functional status in ACF surgery.

Charlson Comorbidity Index and Frailty as Predictors of Resolution Following Middle Meningeal Artery Embolization for Chronic Subdural Hematoma

BACKGROUND

Research on variables associated with chronic subdural hematoma (cSDH) resolution following middle meningeal artery embolization (MMAE) is limited. This study investigated the clinical utility of age-adjusted Charlson Comorbidity Index (ACCI) and modified 5-item Frailty Index (mFI - 5) for predicting cSDH resolution following MMAE.

METHODS

We identified patients who underwent MMAE at our institution between January 2018 and December 2022, with at least 20 days of follow-up and one radiographic follow-up study. Patient demographics, characteristics, and outcomes were collected. Complete resolution was defined as absence of subdural collections on CT-scan at last follow-up. Nonage adjusted CCI (CCI), ACCI, and mFI - 5 scores were calculated. Univariate and multivariable logistic regression analyzed the relationship between cSDH resolution and variables. A receiver operating characteristic (ROC) curve established the utility of ACCI and mFI - 5 in predicting hematoma resolution.

RESULTS

The study included 85 MMAE procedures. In univariate analysis, patients without resolution were older, had higher CCI, higher ACCI, higher mFI - 5, and were more likely to have diabetes mellitus. In multivarible analysis, CCI (OR: 0.66, 95% CI: 0.48, 0.91) was independently associated with resolution controlling for age and antithrombotic resumption. The area under the ROC (AUROC) curve was 0.75 (95% CI: 0.65-0.85) for ACCI and 0.64 (95% CI: 0.52-0.76) for mFI - 5. The optimal cutoffs for predicting resolution were ACCI ≥5 (sensitivity = 0.63, specificity = 0.77), and mFI - 5 > 0 (sensitivity = 0.84, specificity = 0.43).

CONCLUSIONS

ACCI and mFI - 5 moderately predict MMAE resolution and may aid in medical decision-making.

Frailty and postoperative outcomes in brain tumor patients: a systematic review subdivided by tumor etiology

PURPOSE

Frailty has gained prominence in neurosurgical oncology, with more studies exploring its relationship to postoperative outcomes in brain tumor patients. As this body of literature continues to grow, concisely reviewing recent developments in the field is necessary. Here we provide a systematic review of frailty in brain tumor patients subdivided by tumor type, incorporating both modern frailty indices and traditional Karnofsky Performance Status (KPS) metrics.

METHODS

Systematic literature review was performed using PRISMA guidelines. PubMed and Google Scholar were queried for articles related to frailty, KPS, and brain tumor outcomes. Only articles describing novel associations between frailty or KPS and primary intracranial tumors were included.

RESULTS

After exclusion criteria, systematic review yielded 52 publications. Amongst malignant lesions, 16 studies focused on glioblastoma. Amongst benign tumors, 13 focused on meningiomas, and 6 focused on vestibular schwannomas. Seventeen studies grouped all brain tumor patients together. Seven studies incorporated both frailty indices and KPS into their analyses. Studies correlated frailty with various postoperative outcomes, including complications and mortality.

CONCLUSION

Our review identified several patterns of overall postsurgical outcomes reporting for patients with brain tumors and frailty. To date, reviews of frailty in patients with brain tumors have been largely limited to certain frailty indices, analyzing all patients together regardless of lesion etiology. Although this technique is beneficial in providing a general overview of frailty's use for brain tumor patients, given each tumor pathology has its own unique etiology, this combined approach potentially neglects key nuances governing frailty's use and prognostic value.

Validating the VS-5 Score for Predicting Outcomes After Vestibular Schwannoma Resection in an Institutional Cohort

BACKGROUND

The VS-5 index was recently proposed to predict complications, nonroutine discharge, length of stay (LOS), and cost after vestibular schwannoma (VS) resection. The VS-5 ranges from 0-17.86, and a score ≥2 was proposed as being predictive of postoperative adverse events. We sought to determine whether the VS-5 is predictive of nonroutine discharge and length of stay in an institutional cohort.

METHODS

This is a retrospective study of 100 patients undergoing VS resection. For each patient, a VS-5 score was calculated. Bivariate analyses were conducted to determine differences in postoperative outcomes between high- and low-risk subgroups. Area under the receiver operating characteristic curve sensitivity/specificity analysis using Youden's Index was conducted to evaluate the optimal cutoff.

RESULTS

Fifty-one (51%) patients were classified as high risk (VS-5 ≥ 2). Patients with VS-5 ≥ 2 had higher frequency of nonroutine discharge (22% vs. 4%, P = 0.0150) and no significant difference in postoperative LOS. The area under the receiver operating characteristic curve for predicting nonroutine discharge was 0.78 ± 0.15 (P < 0.0001). The optimal cutoff for nonroutine discharge was ≥6, higher than the published cutoff of ≥ 2. The new cutoff was predictive of nonroutine discharge (47% vs. 6%, P = 0 < 0.0001) and LOS (6 [3-11] days vs. 3 [1-28] days, P = 0.0001).

CONCLUSIONS

The VS-5 frailty index predicted nonroutine discharge but not LOS. Youden's index indicates that a cutoff of 6, not 2, is optimal for predicting nonroutine discharge and LOS.

Assessment of the impact of frailty on adverse surgical outcomes in patients undergoing surgery for intracranial tumors using modified frailty index: A systematic review and meta-analysis

BACKGROUND

Modified frailty index (MFI) is an emerging quantitative measure of frailty; however, the quantified risk of adverse outcomes in surgeries for intracranial tumors associated with increasing MFI scores has not been thoroughly reviewed in a comprehensive manner.

METHODS

MEDLINE (PubMed), Scopus, Web of Science, and Embase were searched to identify observational studies on the association between 5 and 11 item-modified frailty index (MFI) and perioperative outcomes for neurosurgical procedures including complications, mortality, readmission, and reoperation rate. Primary analysis pooled all comparisons with MFI scores greater than or equal to 1 versus non-frail participants using mixed-effects multilevel model for each outcome.

RESULTS

In total, 24 studies were included in the review and 19 studies with 114,707 surgical operations were included in the meta-analysis. While increasing MFI scores were associated with worse prognosis for all included outcomes, reoperation rate was only significantly higher in patients with MFI ≥ 3. Among surgical pathologies, glioblastoma was influenced by a greater extent to the impact of frailty on complications and mortality that most other etiologies. In agreement with qualitative evaluation of the included studies, meta-regression did not reveal association between mean age of the comparisons and complications rate.

CONCLUSION

The results of this meta-analysis provides quantitative risk assessment of adverse outcomes in neuro-oncological surgeries with increased frailty. The majority of literature suggests that MFI is a superior and independent predictor of adverse outcomes compared to age.

Radiomics and machine learning for predicting the consistency of benign tumors of the central nervous system: A systematic review

PURPOSE

Predicting the consistency of benign central nervous system (CNS) tumors prior to surgery helps to improve surgical outcomes. This review summarizes and analyzes the literature on using radiomics and/or machine learning (ML) for consistency prediction.

METHOD

The Medical Literature Analysis and Retrieval System Online (MEDLINE) database was screened for studies published in English from January 1st 2000. Data was extracted according to the PRISMA guidelines and quality of the studies was assessed in compliance with the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2).

RESULTS

Eight publications were included focusing on pituitary macroadenomas (n = 5), pituitary adenomas (n = 1), and meningiomas (n = 2) using a retrospective (n = 6), prospective (n = 1), and unknown (n = 1) study design with a total of 763 patients for the consistency prediction. The studies reported an area under the curve (AUC) of 0.71-0.99 for their respective best performing model regarding the consistency prediction. Of all studies, four articles validated their models internally whereas none validated their models externally. Two articles stated making data available on request with the remaining publications lacking information with regard to data availability.

CONCLUSIONS

The research on consistency prediction of CNS tumors is still at an early stage regarding the use of radiomics and different ML techniques. Best-practice procedures regarding radiomics and ML need to be followed more rigorously to facilitate the comparison between publications and, accordingly, the possible implementation into clinical practice in the future.

Estimating the causal effect of frailty index on vestibular disorders: A two-sample Mendelian randomization

Background

Frailty index and vestibular disorders appear to be associated in observational studies, but causality of the association remains unclear.

Methods

A two-sample Mendelian randomization (MR) study was implemented to explore the causal relationship between the frailty index and vestibular disorders in individuals of European descent. A genome-wide association study (GWAS) of frailty index was used as the exposure (n = 175, 226), whereas the GWAS of vestibular disorders was the outcome (n = 462,933). MR Steiger filtering method was conducted to investigate the causal effect of the frailty index on vestibular disorders. An inverse variance weighted (IVW) approach was used as the essential approach to examine the causality. Additionally, the MR-Egger methods, the simple mode analysis, the weighted median analysis, and the weighted mode analysis were used as supplementary methods. The MR-PRESSO analysis, the MR-Egger intercept analysis, and Cochran's Q statistical analysis also were used to detect the possible heterogeneity as well as directional pleiotropy. To evaluate this association, the odds ratio (OR) with 95% confidence intervals (CIs) was used. All statistical analyses were performed in R. The STROBE-MR checklist for the reporting of MR studies was used in this study.

Results

In total, 14 single nucleotide polymorphisms (SNPs) were identified as effective instrumental variables (IVs) in the two sample MR analyses. The significant causal effect of the frailty index on vestibular disorders was demonstrated by IVW method [OR 1.008 (95% CI 1.003, 1.013), p = 0.001]. Results from the various sensitivity analysis were consistent. The “leave-one-out” analysis indicated that our results were robust even without a single SNP. According to the MR-Egger intercept test [intercept = −0.000151, SE = 0.011, p = 0.544], genetic pleiotropy did not affect the results. No heterogeneity was detected by Cochran's Q test. Results of MR Steiger directionality test indicated the accuracy of our estimate of the potential causal direction (Steiger p < 0.001).

Conclusion

The MR study suggested that genetically predicted frailty index may be associated with an increased risk of vestibular disorders. Notably, considering the limitations of this study, the causal effects between frailty index and vestibular disorders need further investigation. These results support the importance of effectively managing frailty which may minimize vestibular disorders and improve the quality of life for those with vestibular disorders.