Clinical implementation of artificial intelligence in neuroradiology with development of a novel workflow-efficient picture archiving and communication system-based automated brain tumor segmentation and radiomic feature extraction

Abstract

PurposePersonalized interpretation of medical images is critical for optimum patient care, but current tools available to physicians to perform quantitative analysis of patient’s medical images in real time are significantly limited. In this work, we describe a novel platform within PACS for volumetric analysis of images and thus development of large expert annotated datasets in parallel with radiologist performing the reading that are critically needed for development of clinically meaningful AI algorithms. Specifically, we implemented a deep learning-based algorithm for automated brain tumor segmentation and radiomics extraction, and embedded it into PACS to accelerate a supervised, end-to- end workflow for image annotation and radiomic feature extraction.Materials and methodsAn algorithm was trained to segment whole primary brain tumors on FLAIR images from multi-institutional glioma BraTS 2021 dataset. Algorithm was validated using internal dataset from Yale New Haven Health (YHHH) and compared (by Dice similarity coefficient [DSC]) to radiologist manual segmentation. A UNETR deep-learning was embedded into Visage 7 (Visage Imaging, Inc., San Diego, CA, United States) diagnostic workstation. The automatically segmented brain tumor was pliable for manual modification. PyRadiomics (Harvard Medical School, Boston, MA) was natively embedded into Visage 7 for feature extraction from the brain tumor segmentations.ResultsUNETR brain tumor segmentation took on average 4 s and the median DSC was 86%, which is similar to published literature but lower than the RSNA ASNR MICCAI BRATS challenge 2021. Finally, extraction of 106 radiomic features within PACS took on average 5.8 ± 0.01 s. The extracted radiomic features did not vary over time of extraction or whether they were extracted within PACS or outside of PACS. The ability to perform segmentation and feature extraction before radiologist opens the study was made available in the workflow. Opening the study in PACS, allows the radiologists to verify the segmentation and thus annotate the study.ConclusionIntegration of image processing algorithms for tumor auto-segmentation and feature extraction into PACS allows curation of large datasets of annotated medical images and can accelerate translation of research into development of personalized medicine applications in the clinic. The ability to use familiar clinical tools to revise the AI segmentations and natively embedding the segmentation and radiomic feature extraction tools on the diagnostic workstation accelerates the process to generate ground-truth data.

Early prognostication of overall survival for pediatric diffuse midline gliomas using MRI radiomics and machine learning: a two-center study

“…The mean Dice scores of 0.91 for TC and 0.88 for WT of cross-validation on the internal cohort were comparable to those reported for adult GBM segmentation using state-of-the-art deep learning models. 42,43 Although worse than the internal cohort, TC segmentation for the external cohort (mean Dice=0.74) is still comparable to the 0.62-0.74 Dice scores reported in a recent study of automatic segmentation of subregions of pediatric brain tumors 26 .…”

“…The mean Dice scores of 0.91 for TC and 0.88 for WT obtained on the internal cohort were comparable to those reported for adult GBM segmentation using state-of-the-art deep learning models. 43,44 The results on the external cohort were similar for WT, an indication of model generalizability and robustness when applied to independent data with different imaging and patient characteristics. Although results were inferior for TC segmentation for the external cohort (mean Dice=0.74), they are comparable to the 0.62-0.74 Dice scores reported in a recent study of automatic segmentation of subregions of pediatric brain tumors 26 .…”

Evolution and implementation of radiographic response criteria in neuro-oncology

“… 38–40 For example, incorporation of segmentation algorithms and volumetric tools directly into the picture archiving and communication system (PACS) can facilitate clinical practice workflows. 41 Furthermore, PACS-based tools to automatically track lesion growth can make response assessment more reliable and efficient. 38 One study showed that interrater agreement for time to progression was significantly greater with an AI-based tool for tracking longitudinal tumor volume changes compared to manual 2D measurements using the RANO criteria.…”

Artificial Intelligence in Brain Tumor Imaging: A Step toward Personalized Medicine

“…Finding the clinical uses of ML algorithms in clinical practice and identifying the areas of clinical care that can be enhanced by artificially generated algorithms are thus the next steps in neuro-oncology imaging [ 96 ].…”