Radio-Pathomic Maps of Cell Density Identify Brain Tumor Invasion beyond Traditional MRI-Defined Margins

Abstract

BACKGROUND AND PURPOSE: Currently, contrast-enhancing margins on T1WI are used to guide treatment of gliomas, yet tumor invasion beyond the contrast-enhancing region is a known confounding factor. Therefore, this study used postmortem tissue samples aligned with clinically acquired MRIs to quantify the relationship between intensity values and cellularity as well as to develop a radio-pathomic model to predict cellularity using MR imaging data. MATERIALS AND METHODS:This single-institution study used 93 samples collected at postmortem examination from 44 patients with brain cancer. Tissue samples were processed, stained with H&E, and digitized for nuclei segmentation and cell density calculation. Preand postgadolinium contrast T1WI, T2 FLAIR, and ADC images were collected from each patient's final acquisition before death. In-house software was used to align tissue samples to the FLAIR image via manually defined control points. Mixed-effects models were used to assess the relationship between single-image intensity and cellularity for each image. An ensemble learner was trained to predict cellularity using 5 Â 5 voxel tiles from each image, with a two-thirds to one-third train-test split for validation.RESULTS: Single-image analyses found subtle associations between image intensity and cellularity, with a less pronounced relationship in patients with glioblastoma. The radio-pathomic model accurately predicted cellularity in the test set (root mean squared error ¼ 1015 cells/mm 2 ) and identified regions of hypercellularity beyond the contrast-enhancing region. CONCLUSIONS:A radio-pathomic model for cellularity trained with tissue samples acquired at postmortem examination is able to identify regions of hypercellular tumor beyond traditional imaging signatures.

Non-invasive tumor probability maps developed using autopsy tissue identify novel areas of tumor beyond the imaging-defined margin

“…Our prior studies of autopsy tissue samples have highlighted how imaging signatures validated within contrast enhancement, such as an inverse relationship between ADC and cellularity, tend to show much weaker relationships when validated using tissue from beyond the enhancing region 14,17 . Our prior paper also developed a proof-of-concept model for predicting cellularity using autopsy tissue samples, which, while insightful, does not directly correlate with areas of tumor, as not all tumor regions (i.e., areas of pseudopalisading necrosis) are strictly hypercellular and not all areas of increased cellularity (i.e., areas of immune response) are tumor 25 . This study sought to expand upon this method by incorporating multiple tissue segmentations to predict tumor presence itself.…”

“…Bootstrap aggregating (bagging) random forest models were trained to predict voxel-wise densities using 5 by 5 voxel tiles from the T1, T1+C, FLAIR and ADC as input. This model framework was selected based on our previous publication, which developed a proof-of-concept model for predicting cellularity using MRI data in a smaller sample of patients 25 . These models were developed on 2/3rds of the full dataset (N=43) and tested on the held-out test set (n=22) to assess generalizability.…”

“…Despite the promise of these recent techniques, identifying areas of tumor beyond the contrast- and FLAIR-enhancing margins remain particularly difficult. Studies of autopsy tissue samples aligned to MRI acquired near death have found areas of active tumor as far as 10 cm beyond the treated margin, indicating the need for improvements in tumor tracking 17,25,26 . However, identifying regions of tumor beyond current signatures requires ground truth tissue samples from beyond the suspected tumor region, which is beyond the capabilities of studies using biopsy tissue and radiologist annotations as the gold standard.…”

Cellular Density in Adult Glioma, Estimated with MR Imaging Data and a Machine Learning Algorithm, Has Prognostic Power Approaching World Health Organization Histologic Grading in a Cohort of 1181 Patients

“…These models produce graphic mapping of CD that characterizes the full tumor heterogeneity and shows promising clinical applications such as identifying hypercellular regions outside contrast enhancement. [6][7][8][9] Our study differs from the current literature in that we directly evaluated the correlation between measures of cellularity and survival outcome. We focused specifically on simple, interpretable, first-order measures of cellularity rather than complex nonlinear feature combinations like texture analysis or deep filter features.…”

“…Several recent works have developed models capable of estimating heightened cellularity using MR imaging data. [6][7][8][9] However, the actual prognostic value of these model estimates has not been directly validated.…”