Tampere University of Technology

TUTCRIS Research Portal

Metastasis detection from whole slide images using local features and random forests

Research output: Contribution to journalArticleScientificpeer-review

Standard

Metastasis detection from whole slide images using local features and random forests. / Valkonen, Mira; Kartasalo, Kimmo; Liimatainen, Kaisa; Nykter, Matti; Latonen, Leena; Ruusuvuori, Pekka.

In: Cytometry Part A, Vol. 91, No. 6, 2017, p. 555-565.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

Valkonen, M, Kartasalo, K, Liimatainen, K, Nykter, M, Latonen, L & Ruusuvuori, P 2017, 'Metastasis detection from whole slide images using local features and random forests', Cytometry Part A, vol. 91, no. 6, pp. 555-565. https://doi.org/10.1002/cyto.a.23089

APA

Vancouver

Author

Valkonen, Mira ; Kartasalo, Kimmo ; Liimatainen, Kaisa ; Nykter, Matti ; Latonen, Leena ; Ruusuvuori, Pekka. / Metastasis detection from whole slide images using local features and random forests. In: Cytometry Part A. 2017 ; Vol. 91, No. 6. pp. 555-565.

Bibtex - Download

@article{29d84edcb53e46c58d7fdaca4d3e8c8b,
title = "Metastasis detection from whole slide images using local features and random forests",
abstract = "Digital pathology has led to a demand for automated detection of regions of interest, such as cancerous tissue, from scanned whole slide images. With accurate methods using image analysis and machine learning, significant speed-up, and savings in costs through increased throughput in histological assessment could be achieved. This article describes a machine learning approach for detection of cancerous tissue from scanned whole slide images. Our method is based on feature engineering and supervised learning with a random forest model. The features extracted from the whole slide images include several local descriptors related to image texture, spatial structure, and distribution of nuclei. The method was evaluated in breast cancer metastasis detection from lymph node samples. Our results show that the method detects metastatic areas with high accuracy (AUC=0.97-0.98 for tumor detection within whole image area, AUC=0.84-0.91 for tumor vs. normal tissue detection) and that the method generalizes well for images from more than one laboratory. Further, the method outputs an interpretable classification model, enabling the linking of individual features to differences between tissue types.",
keywords = "Breast cancer, Computer aided diagnosis, Digital pathology, Machine learning, Metastasis detection, Random forest, Sentinel lymph nodes, Whole slide images",
author = "Mira Valkonen and Kimmo Kartasalo and Kaisa Liimatainen and Matti Nykter and Leena Latonen and Pekka Ruusuvuori",
note = "INT=tut-bmt,{"}Valkonen, Mira{"}",
year = "2017",
doi = "10.1002/cyto.a.23089",
language = "English",
volume = "91",
pages = "555--565",
journal = "Cytometry Part A",
issn = "1552-4922",
publisher = "Wiley",
number = "6",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Metastasis detection from whole slide images using local features and random forests

AU - Valkonen, Mira

AU - Kartasalo, Kimmo

AU - Liimatainen, Kaisa

AU - Nykter, Matti

AU - Latonen, Leena

AU - Ruusuvuori, Pekka

N1 - INT=tut-bmt,"Valkonen, Mira"

PY - 2017

Y1 - 2017

N2 - Digital pathology has led to a demand for automated detection of regions of interest, such as cancerous tissue, from scanned whole slide images. With accurate methods using image analysis and machine learning, significant speed-up, and savings in costs through increased throughput in histological assessment could be achieved. This article describes a machine learning approach for detection of cancerous tissue from scanned whole slide images. Our method is based on feature engineering and supervised learning with a random forest model. The features extracted from the whole slide images include several local descriptors related to image texture, spatial structure, and distribution of nuclei. The method was evaluated in breast cancer metastasis detection from lymph node samples. Our results show that the method detects metastatic areas with high accuracy (AUC=0.97-0.98 for tumor detection within whole image area, AUC=0.84-0.91 for tumor vs. normal tissue detection) and that the method generalizes well for images from more than one laboratory. Further, the method outputs an interpretable classification model, enabling the linking of individual features to differences between tissue types.

AB - Digital pathology has led to a demand for automated detection of regions of interest, such as cancerous tissue, from scanned whole slide images. With accurate methods using image analysis and machine learning, significant speed-up, and savings in costs through increased throughput in histological assessment could be achieved. This article describes a machine learning approach for detection of cancerous tissue from scanned whole slide images. Our method is based on feature engineering and supervised learning with a random forest model. The features extracted from the whole slide images include several local descriptors related to image texture, spatial structure, and distribution of nuclei. The method was evaluated in breast cancer metastasis detection from lymph node samples. Our results show that the method detects metastatic areas with high accuracy (AUC=0.97-0.98 for tumor detection within whole image area, AUC=0.84-0.91 for tumor vs. normal tissue detection) and that the method generalizes well for images from more than one laboratory. Further, the method outputs an interpretable classification model, enabling the linking of individual features to differences between tissue types.

KW - Breast cancer

KW - Computer aided diagnosis

KW - Digital pathology

KW - Machine learning

KW - Metastasis detection

KW - Random forest

KW - Sentinel lymph nodes

KW - Whole slide images

U2 - 10.1002/cyto.a.23089

DO - 10.1002/cyto.a.23089

M3 - Article

VL - 91

SP - 555

EP - 565

JO - Cytometry Part A

JF - Cytometry Part A

SN - 1552-4922

IS - 6

ER -