Whole slide pathology images from regional lymph node metastasis in colon adenocarcinoma produced at Region Gävleborg Clinical Pathology and Cytology department and Region Östergötland Clinical Pathology department. Annotations for AI training produced as part of AIDA clinical fellowship project investigating AI decision support in metastasis detection.
This dataset has been extended with a second collection series in the LNCO2 dataset using different collection and annotation parameters.
Keywords: Pathology, Whole slide imaging, Annotated, Lymph nodes, Cancer, Colon, Adenocarcinoma.
Sample images with reduced image quality. Please click to preview.
|Cite as||Gordan Maras, Martin Lindvall, and Claes Lundstrom (2019) Regional lymph node metastasis in colon adenocarcinoma doi:10.23698/aida/lnco|
|Title||Regional lymph node metastasis in colon adenocarcinoma|
|Resolution||20x and 40x|
Aperio Scanscope (20x)
Hamamatsu NanoZoomer (40x)
|Stain||Hematoxylin and eosin.|
|Copyright||Copyright 2019 Linköping University, Claes Lundström|
Available under the following licenses, described in the License section below.
AIDA BY license
ROI boxes around lymph nodes (positive, negative or unasssed) and detailed tumor polygons
See details below.
The following labels exists:
roi_lgl_norm:: Region contains a lymph node section that do not contain tumor-positive areas, eg, is normal or negative. (but see also note below)
roi_lgl_unknown:: Region contains a lymph node for which ground truth has not been established.
roi_lgl_tumor:: Region contains a lymph node where at least some areas are tumor-positive areas. (Might not be annotated)
roi_lgl_exclude:: Contains lymph node that was determined to be of poor quality or contain severe artifacts from scanning. Would be excluded from clinial review (not suitable).
tumor:: region exclusively contains tumor pixels (but see also excl category below)
excl & excl_artifact:: these areas should be substracted from tumor-annotated regions to gain pure-tumor pixels.
bkg: region contains only background pixels
tumor annotations are not exhaustive, meaning, there might be more tumor areas in
lgl_roi_tumor lymph nodes than annotated with
roi_lgl_norm areas contain tumor cells that are from an accidental contamination of the glass during processing. This is normal during routine clinical work and a challenge that automated systems should deal with in an appropriate manner. This is a very easy task for human pathologists to discern.
tumor regions not within a
roi_lgl_* region should be assumed to be annotations of tumor in the the primary tumor. This can easily be verified visually, and the primary tumor is always on the initial blocks of the case (A..B…C etc)
Sampling procedure: Query for patients with confirmed adenocarcinomas that underwent colorectomy with subsequent lymphnode review as part of TNM-staging. Query selected N cases with a relatively high number of reported positive lymph nodes.
Source: Two medical clinics in Sweden.
LNCO1 patients: 39 cases: 39 images: 977 median dimensions(WxH px): 49799 x 38729 median resolution (micrometer per pixel): 0.4959 annotations: number: 7379 lymph node sections (roi) 1557 which are negative 712 which are positive 292 which are unassessed 525 images with at least one detailed tumor annotation: 165
Pixel position scaling
Coordinates given are relative to the image width. To get the correct pixel position, X coordinates (and Y coordinates!) should therefore be multiplied with the image width.
Free for use in legal and ethical medical diagnostics research. Please contact the dataset provider for terms of access.
You are invited to send an access request email from your institutional account.
Clicking the access request email link above should open a draft email message in a new window, to help you provide relevant information for efficient request evaluation. If the above link does not work for you, then please click to view the suggested email text.
Subject: Requesting access to dataset doi:10.23698/aida/lnco
I work at INSTITUTION in COUNTRY, emailing from my institutional account.
I would like to access the dataset doi:10.23698/aida/lnco, for use in ethical and legal medical diagnostics research. Could you please send me a data sharing agreement template?
Our planned use of the data can be summarized as:
Example agreement template: https://datahub.aida.scilifelab.se/sharing/templates/
Name of research PI ("Recipient Scientist"): PI_NAME (cc here)
Title of research PI: PI_TITLE (ph d or better, in relevant field)
Name of institution: INSTITUTION_NAME
Name of department: DEPARTMENT_NAME
Institution postal address: POSTAL_ADDRESS
Name of authorized signatory: SIGNATORY_NAME (cc here)
Title of authorized signatory: SIGNATORY_TITLE
Also, I would like to ALTERNATIVELY_WOULD_NOT_LIKE_TO be included in the public record of data sharing facilitated by AIDA:
which is used to facilitate scientific discussion, and to show what good AIDA has been to the global research community.
(If "not", then only institution/department information will be included.)
Click to create draft email (new window). Use your institutional account.
AIDA BY license
Copyright 2019 Linköping University, Claes Lundström
Permission to use, copy, modify, and/or distribute this data within Analytic Imaging Diagnostics Arena (AIDA) for the purpose of medical diagnostics research with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies, and that publications resulting from the use of this data cite the following works:
Gordan Maras, Martin Lindvall, and Claes Lundstrom (2019) Regional lymph node metastasis in colon adenocarcinoma doi:10.23698/aida/lnco.
THE DATA IS PROVIDED “AS IS” AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS DATA INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR CHARACTERISTICS OF THIS DATA.