Skip to main content

01.12.2018 | Research | Ausgabe 1/2018 Open Access

Diagnostic Pathology 1/2018

Automatic evaluation of tumor budding in immunohistochemically stained colorectal carcinomas and correlation to clinical outcome

Diagnostic Pathology > Ausgabe 1/2018
Cleo-Aron Weis, Jakob Nikolas Kather, Susanne Melchers, Hanaa Al-ahmdi, Marion J. Pollheimer, Cord Langner, Timo Gaiser
Wichtige Hinweise

Electronic supplementary material

The online version of this article (https://​doi.​org/​10.​1186/​s13000-018-0739-3) contains supplementary material, which is available to authorized users.



Tumor budding, meaning a detachment of tumor cells at the invasion front of colorectal carcinoma (CRC) into single cells or clusters (<=5 tumor cells), has been shown to correlate to an inferior clinical outcome by several independent studies. Therefore, it has been discussed as a complementary prognostic factor to the TNM staging system, and it is already included in national guidelines as an additional prognostic parameter. However, its application by manual evaluation in routine pathology is hampered due to the use of several slightly different assessment systems, a time-consuming manual counting process and a high inter-observer variability. Hence, we established and validated an automatic image processing approach to reliably quantify tumor budding in immunohistochemically (IHC) stained sections of CRC samples.


This approach combines classical segmentation methods (like morphological operations) and machine learning techniques (k-means and hierarchical clustering, convolutional neural networks) to reliably detect tumor buds in colorectal carcinoma samples immunohistochemically stained for pan-cytokeratin. As a possible application, we tested it on whole-slide images as well as on tissue microarrays (TMA) from a clinically well-annotated CRC cohort.


Our automatic tumor budding evaluation tool detected the absolute number of tumor buds per image with a very good correlation to the manually segmented ground truth (R2 value of 0.86).
Furthermore the automatic evaluation of whole-slide images from 20 CRC-patients, we found that neither the detected number of tumor buds at the invasion front nor the number in hotspots was associated with the nodal status. However, the number of spatial clusters of tumor buds (budding hotspots) significantly correlated to the nodal status (p-value = 0.003 for N0 vs. N1/N2). TMAs were not feasible for tumor budding evaluation, as the spatial relationship of tumor buds (especially hotspots) was not preserved.


Automatic image processing is a feasible and valid assessment tool for tumor budding in CRC on whole-slide images. Interestingly, only the spatial clustering of the tumor buds in hotspots (and especially the number of hotspots) and not the absolute number of tumor buds showed a clinically relevant correlation with patient outcome in our data.
Additional file 1: Table S1. Architecture of the applied CNN. The 8-layer CNN has been designed to classify (100x100x3 pixel) images to the classes “tumor bud” and “no tumor” bud. It consists of two block of a combination of convolutional, rectifier and pooling layers and a fully connected layer. (DOCX 14 kb)
Additional file 2: Figure S3. Finding the corresponding core on two separate TMA-slides. Thumbnail of an HE-stained (A) and a pan-cytokeratin-stained (B) TMA-slide. The green circle highlights the same core on both slides, which has due morphological variations different numbers by the image processing based automatic counting. (PDF 6510 kb)
Additional file 3: Pseudocode 1 create TMA-map. Pseudocode 2 combine TMA-maps of different staining. Pseudocode 3 image analysis part I. Pseudocode 4 image analysis part II. (DOCX 16 kb)
Additional file 4: Figure S2. Dealing with spatial heterogeneity by different means. Histogram: On basis of the Z-score the vTMAs with values outside the underlying normal distribution could be identified. By doing so the histogram for the number of tumor buds per vTMA could be binarized into vTMA within and outside. Overlay WIS and heatmap for the ROI border: Furthermore by plotting the Z-score values against the coordinates on the WSI, a heatmap with the hotspot-probability could be obtained. In this map values > 1.67 are regarded as significant. (PDF 4012 kb)
Über diesen Artikel

Weitere Artikel der Ausgabe 1/2018

Diagnostic Pathology 1/2018 Zur Ausgabe

Neu im Fachgebiet Pathologie

21.08.2019 | Übersichten | Ausgabe 5/2019

Molekulare Biologie, Diagnostik und Therapie von Meningeomen

21.08.2019 | Arteriosklerose | CME Zertifizierte Fortbildung | Ausgabe 5/2019

Entstehung der Arteriosklerose

01.08.2019 | Originalien | Ausgabe 4/2019

Cocainkonsumintensität als Ursache erworbener Herzerkrankungen?

Korrelationen bei cocainassoziierten Drogentodesfällen