Computational pathology

Analyze solid tumor tissue samples fast and enhance the quality and accuracy of macrodissection, nucleic acid extraction, and molecular profiling using TissueMark1.


7. Viewer Region Heatmap On

Accurate and reliable tumor estimation powered by
Deep Learning

TissueMark is a key offering in our computational pathology portfolio that assists the user to examine the region of interest for macro dissection by:

  • Visualizing the region of interest (ROI) and
  • Indicating the estimated cellular profile in the region of interest

TissueMark enables region of interest detection and cellular profile estimation in digital whole slide images of Breast, Prostate, Colon, Lung (Histology/ Cytology) and Ovarian formalin-fixed paraffin embedded, H&E stained tissue samples. The algorithms are trained to work on Philips iSyntax image format. The application provides three levels of visualization – a macrodissection boundary, a visual heat map of tumor density and, at higher magnification, cellular visualizations. Color-coded, this enables differentiation of the region of interest from stroma, inflammation, lymphocytes and necrosis thereby providing an accurate macrodissection boundary for further molecular testing.

Improve the quality of molecular tests with accurate ROI and cellularity guidance

Given the inter-pathologist variation that is widely acknowledged2 in the industry, TissueMark macrodissection boundary suggestions show a higher acceptance by pathologists. TissueMark also measures the percent3 of tumor nuclei to ensure sample quality and sufficient tumor nucleic acid for molecular profiling. TissueMark deep learning algorithms are trained to identify cellular structures from other morphology and classify identified cells into tumor vs non-tumor cells. This provides a reliable, accurate cellularity estimate that can help pathologists determine if the sample is sufficient for further molecular testing. Research studies have shown very high correlation (Pearson Correlation Coefficient > 0.95 across all supported tissue types) between TissueMark nuclei detection and the gold standard, hand counted estimations of pathologists.

7. viewer_region_heatmap_on


High throughput, intuitive workflow to save valuable time of lab personnel

TissueMark is an easy-to-use, intuitive tool that is aligned to the needs of the modern day molecular lab. TissueMark helps organize and dispatch whole slide images rather than slide trays with glass slides which require manual sorting, preparation and logistical transport. The application provides automatic organization of the worklist of a pathologist. In comparison to the current glass based workflow, the pathologist4 can view the entire worklist assigned to get an overview of the pending, completed, and urgent work.


TissueMark algorithms are designed for fast execution with a runtime of 60 seconds5 on every whole slide image across the tissue types supported. Fast algorithm execution combined with the workflow design ensures that the pathologist always has the results when they begin to review the slide, thereby saving valuable pathologist minutes in the lab.

1. labmanager_worklist_homepage


Interoperable with Philips IntelliSite Pathology Solution thereby providing a unified digital workflow 

TissueMark is inter-operable with Philips IntelliSite Pathology Solution (PIPS) and via PIPS with your Laboratory Information System (LIS), thereby enabling automatic execution of algorithms on whole slide images (WSI) that are chosen for molecular testing.


Minimum hardware requirements



Screen Resolution


Dual-core @1.65GHz

3GB of physical RAM memory


IE (11), Chrome, Firefox, Safari

Operating system

Other software


Most recent web browser

A PDF reader (e.g. Adobe Acrobat Reader)

100Mbit or 1Gbit Ethernet connection to internet/intranet

Related products

  1. TissueMark is not intended for diagnostic, monitoring or therapeutic purposes or in any other manner for regular medical practice.
  2. A Prospective, Multi-Institutional Diagnostic Trial to Determine Pathologist Accuracy in Estimation of Percentage of Malignant Cells, Viray et al. 
  3. Independent pathologists’ evaluation on 446 WSIs from multiple labs. Average boundary acceptance (including minor edits) of TissueMark generated macro-dissection boundaries, across Lung, Breast and Colon, at 80%
  4. Relies on the level of PIPS-LIS interoperability with TissueMark.
  5. As measured on whole slide images with tissue area of 15mmx15mm and run on server con_ gured with 128GB RAM, processor: Intel ® Xeonr CPU, E5-2640 v4@2.4 GHz, GPU Nvidia Tesla P4; measured without any inter-operability with the IMS.

Connect with Philips


© Koninklijke Philips N.V., 2004 – 2017. All rights reserved.

Select Country
United Kingdom – English >
United States – English >
Netherlands – Dutch >
Spain – Spanish >
France – French >