A standardised pathway for the surveillance of stable vestibular schwannoma

The Real-World Impact of Vestibular Schwannoma Fully Automated Volume Measures on the Evaluation of Size Change and Clinical Management Outcomes in a Multidisciplinary Meeting Setting

Background

Vestibular schwannoma (VS) management decisions are made within multidisciplinary meetings (MDMs). The improved accuracy of volumetric compared to linear tumor measurements is well-recognized, but current volumetric evaluation methods are too time-intensive. The aim was to determine if the availability of fully automated volumetric tumor measures during MDM preparation resulted in different radiological outcomes compared to a standard approach with linear dimensions, and whether this impacted the clinical management decisions.

Methods

A prospective cohort study evaluated 50 adult patients (mean age 64.6, SD 12.8; 24 male, 26 female) with unilateral sporadic VS. Two simulated MDMs were convened using different methods to measure tumor size during radiology preparation: MDM-mlm used linear tumor dimensions, while MDM-avm was provided with fully automated deep learning-based volume measurements. Interval changes in VS size from the index to final and penultimate to final magnetic resonance imaging (MRI) studies defined the radiological outcomes. The subsequent clinical MDM outcomes were classified. Wilcoxon signed rank tests compared the radiological classification of VS size change and the management outcomes between the MDM-mlm and the MDM-avm.

Results

The 57 interval MRI comparisons in 33 patients showed a significant difference in the classification of VS size change between the MDM-mlm and MDM-avm for all intervals (z=2.49, P=.01). However, there was no significant difference in the resulting management decisions between the 2 MDMs (z=0.30, P= .76).

Conclusion

Provision of fully automated VS volume measurements to "real-world" MDM preparation significantly impacted the radiological classification of VS size change but did not influence management decisions.

Mapping the Clinical Pathway for Patients Undergoing Vestibular Schwannoma Resection

BACKGROUND

The introduction of the electronic health record (EHR) has improved the collection and storage of patient information, enhancing clinical communication and academic research. However, EHRs are limited by data quality and the time-consuming task of manual data extraction. This study aimed to use process mapping to help identify critical data entry points within the clinical pathway for patients with vestibular schwannoma (VS) ideal for structured data entry and automated data collection to improve patient care and research.

METHODS

A 2-stage methodology was used at a neurosurgical unit. Process maps were developed using semi-structured interviews with stakeholders in the management of VS resection. Process maps were then retrospectively validated against EHRs for patients admitted between August 2019 and December 2021, establishing critical data entry points.

RESULTS

In the process map development, 20 stakeholders were interviewed. Process maps were validated against EHRs of 36 patients admitted for VS resection. Operative notes, surgical inpatient reviews (including ward rounds), and discharge summaries were available for all patients, representing critical data entry points. Areas for documentation improvement were in the preoperative clinics (30/36; 83.3%), preoperative skull base multidisciplinary team (32/36; 88.9%), postoperative follow-up clinics (32/36; 88.9%), and postoperative skull base multidisciplinary team meeting (29/36; 80.6%).

CONCLUSIONS

This is a first use to our knowledge of a 2-stage methodology for process mapping the clinical pathway for patients undergoing VS resection. We identified critical data entry points that can be targeted for structured data entry and for automated data collection tools, positively impacting patient care and research.

The impact of multiple deprivation on the management of vestibular schwannomas

PURPOSE

While some factors have been well-shown to affect the decision-making in treating patients with vestibular schwannomas (VS), little is known on the role of deprivation. Our objective was to assess the effect of socioeconomic background on the management of patients with VS.

METHODS

This retrospective cohort study included 460 patients with sporadic VS from West of Scotland. The postcode-based, multifactorial Scottish Index of Multiple Deprivation (SIMD) was used to assess the socioeconomic background of each patient. We performed a multivariate analysis including tumour size, growth and patient age with management modality (observation, stereotactic radiotherapy, microsurgery) being the main outcome measure and outcome (need for additional treatment) an additional measure.

RESULTS

We found no significant difference in the demographics, tumour characteristics and primary treatment choice between patients with different SIMD scores. In addition, there was no statistically significant difference in the growth occurrence rates following first-line treatment (p = 0.964) and in the second-line treatment choice (p = 0.460).

CONCLUSIONS

Multiple deprivation does not affect decision making in patients with VS in the examined cohort. This is probably linked to the centralisation and uniformity of the service and might not necessarily be applicable to other health services without centralisation.

Joint Vestibular Schwannoma Enlargement Prediction and Segmentation Using a Deep Multi-task Model

OBJECTIVE

To develop a deep-learning-based multi-task (DMT) model for joint tumor enlargement prediction (TEP) and automatic tumor segmentation (TS) for vestibular schwannoma (VS) patients using their initial diagnostic contrast-enhanced T1-weighted (ceT1) magnetic resonance images (MRIs).

METHODS

Initial ceT1 MRIs for VS patients meeting the inclusion/exclusion criteria of this study were retrospectively collected. VSs on the initial MRIs and their first follow-up scans were manually contoured. Tumor volume and enlargement ratio were measured based on expert contours. A DMT model was constructed for jointly TS and TEP. The manually segmented VS volume on the initial scan and the tumor enlargement label (≥20% volumetric growth) were used as the ground truth for training and evaluating the TS and TEP modules, respectively.

RESULTS

We performed 5-fold cross-validation with the eligible patients (n = 103). Median segmentation dice coefficient, prediction sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were measured and achieved the following values: 84.20%, 0.68, 0.78, 0.72, and 0.77, respectively. The segmentation result is significantly better than the separate TS network (dice coefficient of 83.13%, p = 0.03) and marginally lower than the state-of-the-art segmentation model nnU-Net (dice coefficient of 86.45%, p = 0.16). The TEP performance is significantly better than the single-task prediction model (AUC = 0.60, p = 0.01) and marginally better than a radiomics-based prediction model (AUC = 0.70, p = 0.17).

CONCLUSION

The proposed DMT model is of higher learning efficiency and achieves promising performance on TEP and TS. The proposed technology has the potential to improve VS patient management.

LEVEL OF EVIDENCE

NA Laryngoscope, 133:2754-2760, 2023.

Evidence-based surveillance protocol for vestibular schwannomas: a long-term analysis of tumor growth using conditional probability

OBJECTIVE

The growth characteristics of vestibular schwannomas (VSs) under surveillance can be studied using a Bayesian method of growth risk stratification by time after surveillance onset, allowing dynamic evaluations of growth risks. There is no consensus on the optimum surveillance strategy in terms of frequency and duration, particularly for long-term growth risks. In this study, the long-term conditional probability of new VS growth was reported for patients after 5 years of demonstrated nongrowth. This allowed modeling of long-term VS growth risks, the creation of an evidence-based surveillance protocol, and the proposal of a cost-benefit analysis decision aid.

METHODS

The authors performed an international multicenter retrospective analysis of prospectively collected databases from five tertiary care referral skull base units. Patients diagnosed with sporadic unilateral VS between 1990 and 2010 who had a minimum of 10 years of surveillance MRI showing VS nongrowth in the first 5 years of follow-up were included in the analysis. Conditional probabilities of growth were calculated according to Bayes' theorem, and nonlinear regression analyses allowed modeling of growth. A cost-benefit analysis was also performed.

RESULTS

A total of 354 patients were included in the study. Across the surveillance period from 6 to 10 years postdiagnosis, a total of 12 tumors were seen to grow (3.4%). There was no significant difference in long-term growth risk for intracanalicular versus extracanalicular VSs (p = 0.41). At 6 years, the residual conditional probability of growth from this point onward was seen to be 2.28% (95% CI 0.70%-5.44%); at 7 years, 1.35% (95% CI 0.25%-4.10%); at 8 years, 0.80% (95% CI 0.07%-3.25%); at 9 years, 0.47% (95% CI 0.01%-2.71%); and at 10 years, 0.28% (95% CI 0.00%-2.37%). Modeling determined that the remaining lifetime risk of growth would be less than 1% at 7 years 7 months, less than 0.5% at 8 years 11 months, and less than 0.25% at 10 years 4 months.

CONCLUSIONS

This multicenter study evaluates the conditional probability of VS growth in patients with long-term VS surveillance (6-10 years). On the basis of these growth risks, the authors posited a surveillance protocol with imaging at 6 months (t = 0.5), annually for 3 years (t = 1.5, 2.5, 3.5), twice at 2-year intervals (t = 5.5, 7.5), and a final scan after 3 years (t = 10.5). This can be used to better inform patients of their risk of growth at particular points along their surveillance timeline, balancing the risk of missing late growth with the costs of repeated imaging. A cost-benefit analysis decision aid was also proposed to allow units to make their own decisions regarding the cessation of surveillance.

Wait and Scan Management of Intra-canalicular Vestibular Schwannomas: Analysis of Growth and Hearing Outcome

OBJECTIVE

To report on the results of intracanalicular vestibular schwannomas (ICVS) that were managed by wait and scan and to analyze the possible predictors of tumor growth and hearing deterioration throughout the observation period.

STUDY DESIGN

A retrospective case series.

SETTING

Quaternary referral center for skull base pathologies.

PATIENTS

Patients with sporadic ICVS managed by wait and scan.

INTERVENTION

Serial resonance imaging (MRI) with size measurement and serial audiological evaluation.

MAIN OUTCOME MEASURE

Tumor growth defined as 2 mm increase of maximal tumor diameter, further treatment, and hearing preservation either maintain initial modified Sanna hearing class, or maintain initial serviceable hearing (class A/B).

RESULTS

339 patients were enrolled. The mean follow-up was 36.5±31.7 months with a median of 24 months. Tumor growth occurred in 141 patients (40.6%) either as slow growth (SG) in 26.3% of cases or fast growth (FG) in 15.3% of cases. Intervention was performed in only 64 cases (18.8%). Out of 271 patients who underwent hearing analysis, 86 patients (33.5%) showed hearing deterioration to a lower hearing class of the modified Sanna classification. Tumor growth and older age were predictors of hearing deterioration. Of the 125 cases with initial serviceable hearing (Class A/B), 91 cases (72.8%) maintained serviceable hearing at last follow-up. Tumor growth and a worse initial pure tone average (PTA) were predictors of hearing deterioration.

CONCLUSIONS

Wait and scan management of ICVS is a viable option and only 18.8% of patients needed further treatment. Hearing tends to deteriorate over time.

Automated Koos Classification of Vestibular Schwannoma

Objective

The Koos grading scale is a frequently used classification system for vestibular schwannoma (VS) that accounts for extrameatal tumor dimension and compression of the brain stem. We propose an artificial intelligence (AI) pipeline to fully automate the segmentation and Koos classification of VS from MRI to improve clinical workflow and facilitate patient management.

Methods

We propose a method for Koos classification that does not only rely on available images but also on automatically generated segmentations. Artificial neural networks were trained and tested based on manual tumor segmentations and ground truth Koos grades of contrast-enhanced T1-weighted (ceT1) and high-resolution T2-weighted (hrT2) MR images from subjects with a single sporadic VS, acquired on a single scanner and with a standardized protocol. The first stage of the pipeline comprises a convolutional neural network (CNN) which can segment the VS and 7 adjacent structures. For the second stage, we propose two complementary approaches that are combined in an ensemble. The first approach applies a second CNN to the segmentation output to predict the Koos grade, the other approach extracts handcrafted features which are passed to a Random Forest classifier. The pipeline results were compared to those achieved by two neurosurgeons.

Results

Eligible patients (n = 308) were pseudo-randomly split into 5 groups to evaluate the model performance with 5-fold cross-validation. The weighted macro-averaged mean absolute error (MA-MAE), weighted macro-averaged F1 score (F₁), and accuracy score of the ensemble model were assessed on the testing sets as follows: MA-MAE = 0.11 ± 0.05, F₁ = 89.3 ± 3.0%, accuracy = 89.3 ± 2.9%, which was comparable to the average performance of two neurosurgeons: MA-MAE = 0.11 ± 0.08, F₁ = 89.1 ± 5.2, accuracy = 88.6 ± 5.8%. Inter-rater reliability was assessed by calculating Fleiss' generalized kappa (k = 0.68) based on all 308 cases, and intra-rater reliabilities of annotator 1 (k = 0.95) and annotator 2 (k = 0.82) were calculated according to the weighted kappa metric with quadratic (Fleiss-Cohen) weights based on 15 randomly selected cases.

Conclusions

We developed the first AI framework to automatically classify VS according to the Koos scale. The excellent results show that the accuracy of the framework is comparable to that of neurosurgeons and may therefore facilitate management of patients with VS. The models, code, and ground truth Koos grades for a subset of publicly available images (n = 188) will be released upon publication.

Segmentation of vestibular schwannoma from MRI, an open annotated dataset and baseline algorithm

Automatic segmentation of vestibular schwannomas (VS) from magnetic resonance imaging (MRI) could significantly improve clinical workflow and assist patient management. We have previously developed a novel artificial intelligence framework based on a 2.5D convolutional neural network achieving excellent results equivalent to those achieved by an independent human annotator. Here, we provide the first publicly-available annotated imaging dataset of VS by releasing the data and annotations used in our prior work. This collection contains a labelled dataset of 484 MR images collected on 242 consecutive patients with a VS undergoing Gamma Knife Stereotactic Radiosurgery at a single institution. Data includes all segmentations and contours used in treatment planning and details of the administered dose. Implementation of our automated segmentation algorithm uses MONAI, a freely-available open-source framework for deep learning in healthcare imaging. These data will facilitate the development and validation of automated segmentation frameworks for VS and may also be used to develop other multi-modal algorithmic models.

Determinants of Vestibular Schwannoma Growth

OBJECTIVE

Management of vestibular schwannomas (VS) involves surgery, radiotherapy, or surveillance, based on patient and tumor factors. We recently described conditional probability as a more accurate method for stratifying VS growth risk. Building on this, we now describe determinants of VS growth, allowing clinicians to move toward a more personalized approach to growth-risk profiling.

METHODS

Retrospective analysis of a prospectively collected database in a tertiary referral skull base unit between 2005 and 2014. Inclusion of patients with unilateral VS managed on surveillance protocol for a minimum of 5 years. Analysis of patient age, sex, tumor location, tumor size, and symptomology using conditional probability.

RESULTS

A total of 340 patients met inclusion criteria. The conditional probability of growth of extracanalicular VS was significantly higher versus intracanalicular (IC) VS (30% versus 13%, p < 0.001) as was small-sized VS versus IC VS (28 versus 13%, p = 0.002), but only in the first year after diagnosis. Sex, age, and presenting symptoms did not significantly affect VS growth.

CONCLUSION

In our series, extracanalicular VS were more likely to grow than IC VS and small-sized VS more likely to grow than IC VS, but only in the first year after diagnosis. Conversely, sex, age, and presenting symptoms did not affect the conditional probability of VS growth.

An artificial intelligence framework for automatic segmentation and volumetry of vestibular schwannomas from contrast-enhanced T1-weighted and high-resolution T2-weighted MRI

OBJECTIVE

Automatic segmentation of vestibular schwannomas (VSs) from MRI could significantly improve clinical workflow and assist in patient management. Accurate tumor segmentation and volumetric measurements provide the best indicators to detect subtle VS growth, but current techniques are labor intensive and dedicated software is not readily available within the clinical setting. The authors aim to develop a novel artificial intelligence (AI) framework to be embedded in the clinical routine for automatic delineation and volumetry of VS.

METHODS

Imaging data (contrast-enhanced T1-weighted [ceT1] and high-resolution T2-weighted [hrT2] MR images) from all patients meeting the study's inclusion/exclusion criteria who had a single sporadic VS treated with Gamma Knife stereotactic radiosurgery were used to create a model. The authors developed a novel AI framework based on a 2.5D convolutional neural network (CNN) to exploit the different in-plane and through-plane resolutions encountered in standard clinical imaging protocols. They used a computational attention module to enable the CNN to focus on the small VS target and propose a supervision on the attention map for more accurate segmentation. The manually segmented target tumor volume (also tested for interobserver variability) was used as the ground truth for training and evaluation of the CNN. We quantitatively measured the Dice score, average symmetric surface distance (ASSD), and relative volume error (RVE) of the automatic segmentation results in comparison to manual segmentations to assess the model's accuracy.

RESULTS

Imaging data from all eligible patients (n = 243) were randomly split into 3 nonoverlapping groups for training (n = 177), hyperparameter tuning (n = 20), and testing (n = 46). Dice, ASSD, and RVE scores were measured on the testing set for the respective input data types as follows: ceT1 93.43%, 0.203 mm, 6.96%; hrT2 88.25%, 0.416 mm, 9.77%; combined ceT1/hrT2 93.68%, 0.199 mm, 7.03%. Given a margin of 5% for the Dice score, the automated method was shown to achieve statistically equivalent performance in comparison to an annotator using ceT1 images alone (p = 4e-13) and combined ceT1/hrT2 images (p = 7e-18) as inputs.

CONCLUSIONS

The authors developed a robust AI framework for automatically delineating and calculating VS tumor volume and have achieved excellent results, equivalent to those achieved by an independent human annotator. This promising AI technology has the potential to improve the management of patients with VS and potentially other brain tumors.

Manual segmentation versus semi-automated segmentation for quantifying vestibular schwannoma volume on MRI

Purpose

Management of vestibular schwannoma (VS) is based on tumour size as observed on T1 MRI scans with contrast agent injection. The current clinical practice is to measure the diameter of the tumour in its largest dimension. It has been shown that volumetric measurement is more accurate and more reliable as a measure of VS size. The reference approach to achieve such volumetry is to manually segment the tumour, which is a time intensive task. We suggest that semi-automated segmentation may be a clinically applicable solution to this problem and that it could replace linear measurements as the clinical standard.

Methods

Using high-quality software available for academic purposes, we ran a comparative study of manual versus semi-automated segmentation of VS on MRI with 5 clinicians and scientists. We gathered both quantitative and qualitative data to compare the two approaches; including segmentation time, segmentation effort and segmentation accuracy.

Results

We found that the selected semi-automated segmentation approach is significantly faster (167 s vs 479 s, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p<0.001$$\end{document}p<0.001), less temporally and physically demanding and has approximately equal performance when compared with manual segmentation, with some improvements in accuracy. There were some limitations, including algorithmic unpredictability and error, which produced more frustration and increased mental effort in comparison with manual segmentation.

Conclusion

We suggest that semi-automated segmentation could be applied clinically for volumetric measurement of VS on MRI. In future, the generic software could be refined for use specifically for VS segmentation, thereby improving accuracy.

Electronic supplementary material

The online version of this article (10.1007/s11548-020-02222-y) contains supplementary material, which is available to authorized users.

Automated Detection of Vestibular Schwannoma Growth Using a Two-Dimensional U-Net Convolutional Neural Network

OBJECTIVES/HYPOTHESIS

To determine if an automated vestibular schwannoma (VS) segmentation model has comparable performance to using the greatest linear dimension to detect growth.

STUDY DESIGN

Case-control Study.

METHODS

Patients were selected from an internal database who had an initial gadolinium-enhanced T1-weighted magnetic resonance imaging scan and a follow-up scan captured at least 5 months later. Two observers manually segmented the VS to compute volumes, and one observer's segmentations were used to train a convolutional neural network model to automatically segment the VS and determine the volume. The results of automatic segmentation were compared to the observer whose measurements were not used in model development to measure agreement. We then examined the sensitivity, specificity, and area under the receiver-operating characteristic curve (AUC) to compare automated volumetric growth detection versus using the greatest linear dimension. Growth detection determined by the external observer's measurements served as the gold standard.

RESULTS

A total of 65 patients and 130 scans were studied. The automated method of segmentation demonstrated excellent agreement with the observer whose measurements were not used for model development for the initial scan (interclass correlational coefficient [ICC] = 0.995; 95% confidence interval [CI]: 0.991-0.997) and follow-up scan (ICC = 0.960; 95% CI: 0.935-0.975). The automated method of segmentation demonstrated increased sensitivity (72.2% vs. 63.9%), specificity (79.3% vs. 69.0%), and AUC (0.822 vs. 0.701) compared to using the greatest linear dimension for growth detection.

CONCLUSIONS

In detecting VS growth, a convolutional neural network model outperformed using the greatest linear dimension, demonstrating a potential application of artificial intelligence methods to VS surveillance.

LEVEL OF EVIDENCE

4 Laryngoscope, 131:E619-E624, 2021.