Automatic Segmentation of Stroke Lesions in Non-Contrast Computed Tomography Datasets With Convolutional Neural Networks

Anup Tuladhar,Serena Schimert,Helge C Kniep,Jens Fiehler,Deepthi Rajashekar,Nils D Forkert

doi:10.1109/access.2020.2995632

Abstract

Non-contrast computed tomography (NCCT) is commonly used for volumetric follow-up assessment of ischemic strokes. However, manual lesion segmentation is time-consuming and subject to high inter-observer variability. The aim of this study was to develop and establish a baseline convolutional neural network (CNN) model for automatic NCCT lesion segmentation. A total of 252 multi-center clinical NCCT datasets, acquired from 22 centers, and corresponding manual segmentations were used to train (204 datasets) and validate (48 datasets) a 3D multi-scale CNN model for lesion segmentation. Post-processing methods were implemented to improve the CNN-based lesion segmentations. The final CNN model and post-processing method was evaluated using 39 out-of-distribution holdout test datasets, acquired at seven centers that did not contribute to the training or validation datasets. Each test image was segmented by two or three neuroradiologists. The Dice similarity coefficient (DSC) and predicted lesion volumes were used to evaluate the segmentations. The CNN model achieved a mean DSC score of 0.47 on the validation NCCT datasets. Post-processing significantly improved the DSC to 0.50 (P <; 0.01). On the holdout test set, the CNN model achieved a mean DSC score of 0.42, which was also significantly improved to 0.45 (P <; 0.05) by post-processing. Importantly, the automatically segmented lesion volumes were not significantly different from the lesion volumes determined by the expert observers (P >0.05) and showed excellent agreement with manual lesion segmentation volumes (intraclass correlation coefficient, ICC = 0.88). The proposed CNN model can automatically and reliably segment ischemic stroke lesions in clinical NCCT datasets. Post-processing techniques can further improve accuracy. As the model was trained and evaluated on datasets from multiple centers, it is broadly applicable and is publicly available.

Highlights

Non-contrast computed tomography (NCCT) is the most common imaging modality for volumetric assessment of stroke lesions [1], [2]
As the out-of-distribution holdout test set was segmented by multiple observers, the manual segmentation lesion volume for each example was defined as the average volume calculated across observers
The median lesion volume for the test set was 20.9 [9.7–63.7] cm3, which is considerably lower compered to the training and validation sets

Summary

Introduction

Non-contrast computed tomography (NCCT) is the most common imaging modality for volumetric assessment of stroke lesions [1], [2]. Manual lesion segmentation in NCCT images is time consuming and associated with high inter-observer variability. Semi-automatic lesion segmentation tools have been developed [3], [4] but still. The associate editor coordinating the review of this manuscript and approving it for publication was Wei Wei. require human interaction, which could introduce a bias, while previous work on automatic NCCT lesion segmentation is limited [5], [6]. Deep convolutional neural networks (CNNs) have shown superior performance for various segmentation tasks in medical imaging because of their ability to learn complex patterns and relationships in the data [7]. Convolutional kernels in CNNs enable the learning of non-localized spatial relationships. The use of multi-scale features and three-dimensional (3D) kernels [8] allows an automated seg-

Objectives

Methods

Results

Discussion

Conclusion