Shortcomings and areas for improvement in digital pathology image segmentation challenges

Foucart, Adrien; Debeir, Olivier; Decaestecker, Christine

doi:10.1016/j.compmedimag.2022.102155

Cited by 8 publications

(6 citation statements)

References 59 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The IoU metric does not seem a priori controversial for evaluating a segmentation task. It is widely used, including in many digital pathology challenges and benchmarks 13 . However, it also has known weaknesses, particularly when used on small objects 14 .…”

Section: Theoretical Analysismentioning

confidence: 99%

Panoptic quality should be avoided as a metric for assessing cell nuclei segmentation and classification in digital pathology

Foucart

Debeir

Decaestecker

2023

Sci Rep

Self Cite

View full text Add to dashboard Cite

Panoptic Quality (PQ), designed for the task of “Panoptic Segmentation” (PS), has been used in several digital pathology challenges and publications on cell nucleus instance segmentation and classification (ISC) since its introduction in 2019. Its purpose is to encompass the detection and the segmentation aspects of the task in a single measure, so that algorithms can be ranked according to their overall performance. A careful analysis of the properties of the metric, its application to ISC and the characteristics of nucleus ISC datasets, shows that is not suitable for this purpose and should be avoided. Through a theoretical analysis we demonstrate that PS and ISC, despite their similarities, have some fundamental differences that make PQ unsuitable. We also show that the use of the Intersection over Union as a matching rule and as a segmentation quality measure within PQ is not adapted for such small objects as nuclei. We illustrate these findings with examples taken from the NuCLS and MoNuSAC datasets. The code for replicating our results is available on GitHub (https://github.com/adfoucart/panoptic-quality-suppl).

show abstract

Section: Theoretical Analysismentioning

confidence: 99%

Panoptic quality should be avoided as a metric for assessing cell nuclei segmentation and classification in digital pathology

Foucart

Debeir

Decaestecker

2023

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…An accuracy of less than 90%, barely outscoring more basic thresholding algorithms, is not an uncommon occurrence in these segmentation models. Increases in model performance in the last 10 years has been driven by grand challenges in segmentation tasks offered by societies with a vested interest in medical cancer pathology, such as the International Conference on Pattern Recognition (ICPR) and Medical Image Computing and Computer Assisted Intervention (MICCAI) conferences (Foucart et al, 2022). U-Net, which was developed in 2015, still holds a top scoring position in the GLaS competition for glandular segmentation of colorectal adenocarcinoma tissue images.…”

Section: Segmentationmentioning

confidence: 99%

The state of applying artificial intelligence to tissue imaging for cancer research and early detection

Robben,

Hajighasemi,

Nasr

et al. 2023

F1000Res

View full text Add to dashboard Cite

Artificial intelligence (AI) represents a new frontier in human medicine that could save more lives and reduce the costs, thereby increasing accessibility. As a consequence, the rate of advancement of AI in cancer medical imaging and more particularly tissue pathology has exploded, opening it to ethical and technical questions that could impede its adoption into existing systems. In order to chart the path of AI in its application to cancer tissue imaging, we review current work and identify how it can improve cancer pathology diagnostics and research. In this review, we identify 5 core tasks that models are developed for, including regression, classification, segmentation, generation, and compression tasks. We address the benefits and challenges that such methods face, and how they can be adapted for use in cancer prevention and treatment. The studies looked at in this paper represent the beginning of this field and future experiments will build on the foundations that we highlight.

show abstract

“…Third, the number of mitotic nuclei is significantly lower than non-mitotic nuclei, making it challenging to extract useful features for deep learning models due to the imbalance of positive and negative samples. 21 Fourth, datasets for training deep learning models are limited, with most public datasets originating from research challenges like the MItosis DOmain Generalization Challenge (MIDOG′22) rather than directly from hospitals, resulting in differences in image quality, lab environments, tissue types, and availability of patches rather than WSIs. 10 , 12 , 13 Lastly, the interpretability of deep learning models is one of the key challenges in the medical domain, especially in clinical settings.…”

Section: Introductionmentioning

confidence: 99%

Toward interpretable and generalized mitosis detection in digital pathology using deep learning

Farooq,

Saleem,

Aleem

et al. 2024

DIGITAL HEALTH

View full text Add to dashboard Cite

Objective The mitotic activity index is an important prognostic factor in the diagnosis of cancer. The task of mitosis detection is difficult as the nuclei are microscopic in size and partially labeled, and there are many more non-mitotic nuclei compared to mitotic ones. In this paper, we highlight the challenges of current mitosis detection pipelines and propose a method to tackle these challenges. Methods Our proposed methodology is inspired from recent research on deep learning and an extensive analysis on the dataset and training pipeline. We first used the MiDoG′22 dataset for training, validation, and testing. We then tested the methodology without fine-tuning on the TUPAC′16 dataset and on a real-time case from Shaukat Khanum Memorial Cancer Hospital and Research Centre. Results Our methodology has shown promising results both quantitatively and qualitatively. Quantitatively, our methodology achieved an F1-score of 0.87 on the MiDoG'22 dataset and an F1-score of 0.83 on the TUPAC dataset. Qualitatively, our methodology is generalizable and interpretable across various datasets and clinical settings. Conclusion In this paper, we highlight the challenges of current mitosis detection pipelines and propose a method that can accurately predict mitotic nuclei. We illustrate the accuracy, generalizability, and interpretability of our approach across various datasets and clinical settings. Our methodology can speed up the adoption of computer-aided digital pathology in clinical settings.

show abstract

Shortcomings and areas for improvement in digital pathology image segmentation challenges

Cited by 8 publications

References 59 publications

Panoptic quality should be avoided as a metric for assessing cell nuclei segmentation and classification in digital pathology

Panoptic quality should be avoided as a metric for assessing cell nuclei segmentation and classification in digital pathology

The state of applying artificial intelligence to tissue imaging for cancer research and early detection

Toward interpretable and generalized mitosis detection in digital pathology using deep learning

Contact Info

Product

Resources

About