Automatic contour segmentation of cervical cancer using artificial intelligence

Abstract

In cervical cancer treatment, radiation therapy is selected based on the degree of tumor progression, and radiation oncologists are required to delineate tumor contours. To reduce the burden on radiation oncologists, an automatic segmentation of the tumor contours would prove useful. To the best of our knowledge, automatic tumor contour segmentation has rarely been applied to cervical cancer treatment. In this study, diffusion-weighted images (DWI) of 98 patients with cervical cancer were acquired. We trained an automatic tumor contour segmentation model using 2D U-Net and 3D U-Net to investigate the possibility of applying such a model to clinical practice. A total of 98 cases were employed for the training, and they were then predicted by swapping the training and test images. To predict tumor contours, six prediction images were obtained after six training sessions for one case. The six images were then summed and binarized to output a final image through automatic contour segmentation. For the evaluation, the Dice similarity coefficient (DSC) and Hausdorff distance (HD) was applied to analyze the difference between tumor contour delineation by radiation oncologists and the output image. The DSC ranged from 0.13 to 0.93 (median 0.83, mean 0.77). The cases with DSC <0.65 included tumors with a maximum diameter < 40 mm and heterogeneous intracavitary concentration due to necrosis. The HD ranged from 2.7 to 9.6 mm (median 4.7 mm). Thus, the study confirmed that the tumor contours of cervical cancer can be automatically segmented with high accuracy.

Fully Automatic Whole-Volume Tumor Segmentation in Cervical Cancer

“…In the present study, the crude performance estimates are lower than that reported in some previous studies of CC tumor segmentation [10][11][12][13]. Bnouni et al reported a DSC of 0.93 (using T2-weighed MRI) [13] (n = 15), Kano et al reported a DSC score of 0.83 (using diffusion-weighted MRI) [11] (n = 98), and Lin et al reported a DSC score of 0.82 (using multiparametric MRI) [12] (n = 169). However, these studies all used k-fold cross-validation applied to a train/validation data set for performance estimation and hyperparameter tuning.…”

“…Thus, a seamless clinical integration of whole-volume radiomic tumor profiling requires the development of robust platforms for accurate automated tumor segmentation. Previous CC studies applying deep learning (DL) networks for automated primary tumor segmentation on MRI data report highly variable Dice scores (Dice scores: 0.44-0.93) between DL segmentation and tumor segmentation derived by radiologists [10][11][12][13]. Furthermore, poor reproducibility of certain radiomic parameters derived from automatic tumor segmentations have been reported [12].…”

Review of Cervix Cancer Classification Using Radiomics on Diffusion-Weighted Imaging

“…However, the automatic segmentation of the gross tumor volume (GTV) represents a difficult task and has no clinical applications so far. To the best of our knowledge, there is only one recent study that relies on deep learning for automatic segmentation of the GTV in CC from DWI [37].…”

Editorial for “A Multicenter Study on Preoperative Assessment of Lymphovascular Space Invasion in Early‐Stage Cervical Cancer Based on Multimodal MR Radiomics”

“…Attempts to automatically segment cervical cancer have already been reported with moderate success. 7,8 If computer vision continues to evolve on par with the pace of other artificial intelligence subfields, such as natural language processing, which recently amazed the world with the introduction of ChatGPT, that might happen sooner rather than later.…”

“…Once robust and validated neural networks obviate the manual segmentation step, the path will be cleared for integrating these radiomics markers and their interpretation (eg as the predicted risk by the model) directly into the scan report. Attempts to automatically segment cervical cancer have already been reported with moderate success 7,8 . If computer vision continues to evolve on par with the pace of other artificial intelligence subfields, such as natural language processing, which recently amazed the world with the introduction of ChatGPT, that might happen sooner rather than later.…”