Author: Jeremy Moreau
Copyright © 2018, McGill University
meningioma.app is distributed under the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with your software, You can obtain one at https://mozilla.org/MPL/2.0/.
meningioma.app is also distributed under the terms of the following Healthcare Disclaimer.
In Canada, the United States, or any other jurisdictions where they may apply, the following additional disclaimer of warranty and limitation of liability are hereby incorporated into the terms and conditions of MPL 2.0:
No warranties of any kind whatsoever are made as to the results that You will obtain from relying upon the covered code (or any information or content obtained by way of the covered code), including but not limited to compliance with privacy laws or regulations or clinical care industry standards and protocols. Use of the covered code is not a substitute for a health care provider’s standard practice or professional judgment. Any decision with regard to the appropriateness of treatment, or the validity or reliability of information or content made available by the covered code, is the sole responsibility of the health care provider. Consequently, it is incumbent upon each health care provider to verify all medical history and treatment plans with each patient.
Under no circumstances and under no legal theory, whether tort (including negligence), contract, or otherwise, shall any Contributor, or anyone who distributes Covered Software as permitted by the license, be liable to You for any indirect, special, incidental, consequential damages of any character including, without limitation, damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other damages or losses, of any nature whatsoever (direct or otherwise) on account of or associated with the use or inability to use the covered content (including, without limitation, the use of information or content made available by the covered code, all documentation associated therewith, and the failure of the covered code to comply with privacy laws and regulations or clinical care industry standards and protocols), even if such party shall have been informed of the possibility of such damages.
License Information by OpenMRS / CC BY 4.0
The meningioma.app icon was designed by Jeremy Moreau and is made available under CC BY SA 4.0
All other icons were designed by Google, and are licensed under the Apache License Version 2.0 (https://google.github.io/material-design-icons/)
flask - BSD 3-Clause
joblib - BSD 3-Clause
matplotlib - Matplotlib license
numpy - BSD 3-Clause
pandas - BSD 3-Clause
scikit-learn - New BSD License
seaborn - BSD 3-Clause
jQuery - MIT License
OnsenUI - Apache License 2.0
Meningioma.app runs in the browser, but it can also be installed on IOS (iPhone/iPad) or Android devices. Note: an internet connection is always required.
Meningioma.app is the companion app to our paper, "Individual-patient prediction of meningioma malignancy and survival using the Surveillance, Epidemiology, and End Results database". We strongly recommend users read the (open access) paper to get a better understanding of the interpretation and limitations of the models included in this app: [Link to paper]. The app provides predictions of tumour behaviour and patient-specific survival estimates on the basis of a set of basic clinical variables. The models presented here fall under the "Development Stage" of Woo et al.'s framework of levels of evidence for predictive biomarkers and therefore will require further prospective validation to demonstrate true clinical utility. Our intention with the current app release is to give readers the ability to evaluate and probe these models for insights, whether with past historical cases, current, or hypothetical patients. We do however emphasise that at the current time the models should not be used to guide treatment decisions. If you are interested in undertaking prospective validation studies or have access to large datasets of imaging or molecular markers on which you would like to apply similar methods, please do not hesitate to contact us for potential collaborations.
The clinical variables used as input are based on the data collected by the Surveillance, Epidemiology, and End Results Program (SEER). The letters in brackets indicate whether the variable is used for predicting Malignancy (M), survival (S) or both (M, S).
Two graphs are produced as output. The first shows the predicted tumour behaviour class (benign or borderline malignancy/malignant) as well as the predicted probabilities of the model. At the selected thresholds (>5% chance of a non-benign meningioma), we obtained a sensitivity of 0.79 with specificity of 0.75 and a PPV of 0.14 with a negative predictive value (NPV) of 0.99 for distinguishing between benign and borderline malignant / malignant meningiomas. Refer to Fig. 3 of the paper for more details.
This graph shows individualised predicted survival curves for the inputted patient characteristics. The survival curves show the predicted probability of survival from time of diagnosis (e.g. 70% probability of survival at 60 months months / 5 years post-diagnosis). Refer to figures 4-5 of the paper for model scoring and additional details.