Digital Health Innovation and Informatics

PV QA 2 - Poster Viewing Q&A 2

MO_17_2465 - Pre- and postoperative prediction of long-term meningioma outcomes

Monday, October 22
10:45 AM - 12:15 PM
Location: Innovation Hub, Exhibit Hall 3

Pre- and postoperative prediction of long-term meningioma outcomes
E. D. Gennatas1, A. Wu1, S. E. Braunstein2, O. Morin2, W. C. Chen1, S. Magill3, J. Villanueva-Meyer4, A. Perry5, M. W. McDermott3, T. D. Solberg1, G. Valdes2, and D. Raleigh6; 1University of California San Francisco, Department of Radiation Oncology, San Francisco, CA, 2University of California, San Francisco, San Francisco, CA, 3University of California San Francisco, Department of Neurological Surgery, San Francisco, CA, 4University of California San Francisco, Department of Radiology, San Francisco, CA, 5University of California San Francisco, Department of Pathology, San Francisco, CA, 6University of California, San Francisco, Department of Radiation Oncology, San Francisco, CA

Purpose/Objective(s):

Meningiomas are conventionally stratified according to tumor grade and extent of resection, commonly in isolation from other clinical variables. Here, we use machine learning to integrate demographic, clinical, radiographic and pathologic data to develop predictive models for meningioma outcomes.

Materials/Methods:

We developed a comprehensive database containing information from 235 patients who underwent surgery for 257 meningiomas at a single institution from 1990 to 2015. The median follow-up was 4.3 years, and resection specimens were re-evaluated according to current diagnostic criteria, revealing 128 WHO grade I, 104 grade II and 25 grade III meningiomas. Classification and regression trees, logistic regression, regularized logistic regression, support vector machines, MediBoost, random forest and gradient boosting were trained and tuned by nested resampling to build models based on preoperative features, conventional postoperative features, or both.

Results:

Models restricted to preoperative information, such as patient demographics and radiographic features, had similar accuracy for predicting local failure (AUC=0.74) or overall survival (AUC=0.68) as models based on meningioma grade and extent of resection (AUC=0.73 and AUC=0.72, respectively). Integrated models incorporating all available demographic, clinical, radiographic and pathologic data provided the most accurate estimates (AUC=0.78 and AUC=0.74, respectively). From these models, we developed decision trees, nomograms, and interactive online apps to estimate the risk of local failure and overall survival for meningioma patients.

Conclusion:

Multivariate predictive models can be used to predict meningioma outcomes and stratify meningioma patients by risk. Preoperative clinical features have been underutilized to date, but can predict meningioma outcomes with accuracy comparable to conventional models based on tumor grade and extent of resection.

Author Disclosure: E.D. Gennatas: None. A. Wu: None. S.E. Braunstein: Advisory Board; Radiation Oncology Questions, LLC. O. Morin: None. W.C. Chen: None. J. Villanueva-Meyer: None. M.W. McDermott: None. T.D. Solberg: Speaker's Bureau; Brainlab. Stock; Global Radiosurgery LLC. Deputy Editor-in-Chief; JACMP. D. Raleigh: None.

Efstathios Gennatas, PhD, MBBS

Presentation(s):

Send Email for Efstathios Gennatas


Assets

MO_17_2465 - Pre- and postoperative prediction of long-term meningioma outcomes



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Pre- and postoperative prediction of long-term meningioma outcomes