The Impact of Using Voxel-Level Segmentation Metrics on Evaluating Multifocal Prostate Cancer Localisation

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

3 Scopus Citations
View graph of relations

Author(s)

  • Qianye Yang
  • Tom Syer
  • Zhe Min
  • Shonit Punwani
  • Mark Emberton
  • Dean Barratt
  • Yipeng Hu

Related Research Unit(s)

Detail(s)

Original languageEnglish
Title of host publicationApplications of Medical Artificial Intelligence
Subtitle of host publicationFirst International Workshop, AMAI 2022, Proceedings
EditorsShandong Wu, Behrouz Shabestari, Lei Xing
PublisherSpringer, Cham
Pages128-138
Edition1
ISBN (Electronic)978-3-031-17721-7
ISBN (Print)978-3-031-17720-0
Publication statusPublished - 30 Sept 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13540 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Title1st International Workshop on Applications of Medical Artificial Intelligence, AMAI 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022
CityVirtual, Online
Period18 September 2022

Abstract

Dice similarity coefficient (DSC) and Hausdorff distance (HD) are widely used for evaluating medical image segmentation. They have also been criticised, when reported alone, for their unclear or even misleading clinical interpretation. DSCs may also differ substantially from HDs, due to boundary smoothness or multiple regions of interest (ROIs) within a subject. More importantly, either metric can also have a nonlinear, non-monotonic relationship with outcomes based on Type 1 and 2 errors, designed for specific clinical decisions that use the resulting segmentation. Whilst cases causing disagreement between these metrics are not difficult to postulate, one might argue that they may not necessarily be substantiated in real-world segmentation applications, as a majority of ROIs and their predictions often do not manifest themselves in extremely irregular shapes or locations that are prone to such inconsistency. This work first proposes a new asymmetric detection metric, adapting those used in object detection, for planning prostate cancer procedures. The lesion-level metrics is then compared with the voxel-level DSC and HD, whereas a 3D UNet is used for segmenting lesions from multiparametric MR (mpMR) images. Based on experimental results using 877 sets of mpMR images, we report pairwise agreement and correlation 1) between DSC and HD, and 2) between voxel-level DSC and recall-controlled precision at lesion-level, with Cohen’s κ∈ [ 0.49, 0.61 ] and Pearson’s r∈ [ 0.66, 0.76 ] (p-values<0.001) at varying cut-offs. However, the differences in false-positives and false-negatives, between the actual errors and the perceived counterparts if DSC is used, can be as high as 152 and 154, respectively, out of the 357 test set lesions. We therefore carefully conclude that, despite of the significant correlations, voxel-level metrics such as DSC can misrepresent lesion-level detection accuracy for evaluating localisation of multifocal prostate cancer and should be interpreted with caution.

Research Area(s)

  • Lesion-level localisation metrics, Multi-parametric MR, Prostate cancer, Voxel-level segmentation metrics

Citation Format(s)

The Impact of Using Voxel-Level Segmentation Metrics on Evaluating Multifocal Prostate Cancer Localisation. / Yan, Wen; Yang, Qianye; Syer, Tom et al.
Applications of Medical Artificial Intelligence: First International Workshop, AMAI 2022, Proceedings. ed. / Shandong Wu; Behrouz Shabestari; Lei Xing. 1. ed. Springer, Cham, 2022. p. 128-138 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13540 LNCS).

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review