COMMENTARY

Picking Out Patterns in Brain Cancer

Hossein Jadvar, MD, PhD, MPH, MBA

Disclosures

November 29, 2016

Magnetic Resonance Imaging-Based Radiomic Profiles Predict Patient Prognosis in Newly Diagnosed Glioblastoma Before Therapy

McGarry SD, Hurrell SL, Kaczmarowski AL, et al
Tomography. 2016;2:223-228

Summary

This retrospective investigation used radiomic profiling techniques to relate MRI-derived features of pretreatment glioblastoma multiforme (GBM) to outcome in 81 patients.

The authors found that five radiomic profiles predicted patient survival before therapy. The researchers were also able to pathologically confirm the presence of hypercellular tumor deposits in one radiomic profile associated with poor prognosis. Such radiomic profiling of the GBM may help with treatment decisions through more robust risk stratification.

Viewpoint

Radiomics is defined as "the extraction and analysis of advanced quantitative imaging features obtained with computed tomography (CT), positron emission tomography (PET) or magnetic resonance imaging (MRI) with data that are in a mineable form used to build models relating image features to phenotypes or gene-protein signatures in order to provide improved diagnostic, prognostic or predictive information."[1]

This new study by investigators at the Medical College of Wisconsin in Milwaukee retrospectively analyzed multiple-sequence MRI data obtained before treatment in patients with GBM in order to predict prognosis in these patients. The methods of image feature extraction were based on image segmentation and coding that were performed using computerized pattern recognition algorithms to arrive at radiomic profiles.

Using radiomics, the investigators were able to risk-stratify patients with respect to overall survival. As the investigators contend—and prior studies have shown—radiomics can be a powerful tool combining multimodality imaging features, which then can be pooled with other relevant data, such as underlying gene and protein signatures, to provide powerful multidimensional information on disease status and its behavior in time and with various interventions.

No doubt, this area of research using big data will advance as new strides are made in the biological, imaging, and computer sciences to further personalized and precision medicine.[2,3]

Abstract

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....