Estimating Local Cellular Density in Glioma Using MR Imaging Data

BACKGROUND AND PURPOSE:Increased cellular density is a hallmark of gliomas, both in the bulk of the tumor and in areas of tumor infiltration into surrounding brain. Altered cellular density causes altered imaging findings, but the degree to which cellular density can be quantitatively estimated from imaging is unknown. The purpose of this study was to discover the best MR imaging and processing techniques to make quantitative and spatially specific estimates of cellular density.MATERIALS AND METHODS:We collected stereotactic biopsies in a prospective imaging clinical trial targeting untreated patients with gliomas at our institution undergoing their first resection. The data included preoperative MR imaging with conventional anatomic, diffusion, perfusion, and permeability sequences and quantitative histopathology on biopsy samples. We then used multiple machine learning methodologies to estimate cellular density using local intensity information from the MR images and quantitative cellular density measurements at the biopsy coordinates as the criterion standard.RESULTS:The random forest methodology estimated cellular density with R² = 0.59 between predicted and observed values using 4 input imaging sequences chosen from our full set of imaging data (T2, fractional anisotropy, CBF, and area under the curve from permeability imaging). Limiting input to conventional MR images (T1 pre- and postcontrast, T2, and FLAIR) yielded slightly degraded performance (R² = 0.52). Outputs were also reported as graphic maps.CONCLUSIONS:Cellular density can be estimated with moderate-to-strong correlations using MR imaging inputs. The random forest machine learning model provided the best estimates. These spatially specific estimates of cellular density will likely be useful in guiding both diagnosis and treatment.

Increased cellular density (CD) is a hallmark of cancer and a key feature in histologic glioma analysis.¹ Mapping cellular density throughout a tumor would be a valuable tool to probe how tumors infiltrate and analyze the transition between diseased and healthy brain. However, measuring CD requires tissue, which entails additional risks and is expensive to obtain. There is no currently accepted clinical algorithm to translate imaging data into quantitative assessments of CD.There is great need for a method to estimate CD noninvasively in human patients with gliomas. In this article, we describe the development of such a method using MR imaging data inputs by correlating with multiple biopsy specimens acquired during a prospective human clinical trial. We obtained comprehensive MR imaging, including conventional, diffusion, perfusion, and permeability imaging sequences. We used machine learning approaches to correlate imaging findings with CD measurements from pathology, devised an algorithm to estimate CD from MR imaging inputs, and generated CD maps for the visual display of the predictions. We identified the most informative imaging data subset. This work has multiple applications in the diagnosis and treatment of patients with gliomas: For example, the method can be used to guide biopsy, resection, and surgery and delineate tumor borderzones both pre- and postoperatively.²