IKOSA Prisma in Action: Assessing Angiogenesis Hallmarks Within the CAM Assay Model  

25 May, 2023 | CAM, Prisma, Use case

We present to you a simple AI-driven method to analyze image data collected from the CAM assay model. Bear with us as we will explain to you how to effortlessly automate the entire image analysis process with the help of the IKOSA CAM Assay Application.

The CAM Assay Model

The chicken embryo chorioallantoic membrane (CAM) assay model was first developed in the early 1970s and has since then become one of the most popular in-vivo methods to study the process of angiogenesis.

The CAM assay method involves the engraftment of tumor tissue on the extraembryonic membrane of a developing chick embryo through a small opening in the eggshell. Using this technique, growth processes in the vascular network of the tumor can be examined in a living environment. The CAM assay model is widely used to study angiogenic and anti-angiogenic responses to different types of substances. 

The angiogenic response is typically measured based on changes observed in the total vessel length, total vessel area, mean vessel thickness, and vessel branching point count.    

Using artificial intelligence to examine the properties of a vascular network of the CAM has become automatic, reliable, precise, and faster. Our deep learning CAM Assay Application available on the IKOSA Platform is doing exactly that. We will now show you how this is done. 

Did you know?

Vessel branching points: These are the points where blood vessels divide into two or more branches. The value range is ≥ 0.

Total vessel length: The total vessel length refers to the overall length of all detected vessels in pixels. The value range is ≥ 0.

Vessel area: This refers to the total area covered by detected vessels in Pixel^2. The value range is ≥ 0.

Mean vessel thickness: This parameter refers to the mean thickness of the detected vessels in the image in pixels. The value range is ≥ 0.

Materials and Methods

Our sample consists of 32 CAM images kindly provided by Dr. Nassim Ghaffari Tabrizi-Wizsy and Lorenz Faihs M.D. who are doing pioneering work in the field of angiogenesis research. Dr. Ghaffari Tabrizi-Wizsy is a researcher at the Institute of Pathophysiology and Immunology at the Medical University of Graz. Dr. Lorenz Faihs specialized in angiogenesis research during his studies at the Medical University of Vienna.

For this showcase analysis, we selected 5 representative images based on criteria such as the presence of particular features and imaging artifacts. We aimed for a homogenous sample but also included images with reflection artifacts to show you how this might influence the analysis.

CAM image in the IKOSA Image Viewer.

To assess the various signs of angiogenic growth within the context of the CAM model, we used the CAM Assay Application available in the IKOSA Prisma portfolio. The CAM Assay App is an AI-powered software solution specifically designed to help you automatically quantify angiogenic processes in CAM images. All you need to do to start the analysis is upload your CAM image data on the IKOSA Platform and submit it for analysis with this specialized image analysis app.

Did you know?

Besides 2D images, this type of analysis can also be performed on time-lapse recordings (time series), z-stacks, or multichannel images, uploaded as multipage TIFF files.

To help you out setting up this process, we would like to guide you through our data analysis journey step by step. 

First, we had a close look at the characteristics of our data to see if the requirements to run the CAM Assay App regarding image file format, size, resolution, and color scheme have been met. Make sure to check the following aspects before you run the analysis:

  • Image type: 2D (standard and/or WSI), time series, multichannel, z-stack
  • Color channels: 3 (RGB) or 1 (Gray)
  • Color depth (per channel): 8 bit
  • Size (px): WSI formats (arbitrary), standard images (max. 25.000 x 25.000)
  • Resolution (μm/px): typically 1-7

Z-stack images cannot be uploaded directly to IKOSA. However, they can still be analyzed via IKOSA Prisma API.

Please check the CAM Assay App Documentation to find out about the different image data formats and modalities required so that the application runs optimally.

You need to take additional requirements into account as well:

  • We recommend the vessel thickness be between 3 and 220 pixels.
  • The vessels that you want to be detected have to be in focus.

Optionally, you can define Regions of Interest (ROIs):

  • Using ROIs allows you to define an area of interest within an image.
  • The analysis can then be performed on single or multiple ROIs.

We didn’t use any ROIs in this use case example, because we didn’t want to reduce the complexity of our image data.

If you want to learn more about how to draw ROIs and run an analysis on them, please have a look at our Knowledge Base article.

After you’ve made sure that all the image data requirements have been met, you can start uploading the images into the image library of your project.

Upload your CAM images into the Image Library of your project.

You can proceed with the following steps, once you’ve successfully uploaded your CAM image data:

  1. Select the images you want to include in your analysis.
  2. Next, click the analysis button in the Image Library of your project.
  3. Finally, select the CAM Assay App from the list provided.
  4. Start the analysis job.

Easy, isn’t it?

Submit your image data for analysis with the CAM Assay App.

Try the CAM Assay App and the full functionalities of IKOSA at no cost.


Once the analysis is complete, the CAM Assay App automatically provides a ready-made CSV or Excel file for you to download. This file contains the results for the analyzed input images or ROIs.

The CAM Assay App provides you with a CSV/Excel file containing analysis results such as vessel total area, vessel total length, vessel mean thickness, and the number of vessel branching points. The table shows the calculated parameters for the analyzed images. 
Downloading the results.

Depending on your experimental design, the app generates a visualization of the analysis results for a specific time point (of a time series), z-layer (of a z-stack), or channel (of a multichannel image). This visualization is either based on the whole image (if no ROIs have been selected for analysis) or represents each ROI.

Visualization of detected vessels.

The results visualization contains the following information:

  • Areas of detected vessels are shown in blue overlay.
  • Vessel paths are displayed as green lines.
  • Vessel branching points are indicated by red dots.

CAM image data often contains light reflections. That is why researchers are often faced with the choice as to whether they should include such images or areas in the analysis. We ran the app on a couple of images with light artifacts to see how it would affect the results. This test showed that the CAM Assay App works reliably on images with little reflection. This is a huge plus because we don’t want the information in our dataset to be wasted. However, we have to be careful with images that are heavily affected by light reflections, as in this case the performance drops in areas with “white spots”.

Analysis output of a CAM image slightly affected by light artifacts (see red arrows).
Analysis output of a CAM image heavily affected by light artifacts (see red arrow).

In addition, you also get a JSON file containing the detected vessel structures. The position is measured from the left upper corner (1,1) of the image. Another JSON file containing all information regarding the analysis job (application name, version, project, etc.) is provided as well, to ensure the reproducibility of results.

In case you are working with ROIs, the CAM Assay App generates another JSON file containing all information regarding the ROIs defined for the analysis job. 

Once your analysis job is finished, you will be notified on the IKOSA Platform and via Email. You can then easily download the results in a bundle or for each image individually.

Try the CAM Assay App and the full functionalities of IKOSA at no cost.


Our specialized CAM Assay App offers you valuable and easy-to-grasp information on several reporting parameters central to the CAM methodology. In addition to being user-friendly and requiring no coding skills or previous AI experience, it allows you to count the number of blood vessels, branching points, and vessel paths as well as to gather quantitative data on the total length and mean thickness of vessels on the CAM. 

These quantitative measurements can be easily incorporated into experimental designs, offering diverse possibilities for analysis and interpretation.

Examples of how to use those metrics in your study design:

  • Compare changes in the vascular network of the CAM based on a control group and a treatment group.
  • Compare changes in the vascular network of the CAM at different time points after treatment with an (anti-)angiogenic substance.
  • Compare differences in the vascular network of the CAM after treatment with different (anti-)angiogenic substances.

It is important to note that the CAM Assay App provides quantitative measurements based on predefined parameters, which may not capture the full complexity of angiogenic processes. Therefore, it is recommended to interpret the results with other experimental observations in mind and consider the limitations of the selected parameters. If your research design is more complex than that and you want to include additional parameters, it is worth training your own application with our specialized software solution IKOSA AI.

The CAM Assay App is not limited to one particular file format and supports a variety of imaging modalities. It can run on 2D-, multichannel-, and time-series images, as well as on z-stacks. However, be careful to meet the data requirements as described above.

Another aspect to keep in mind is the generalizability of the CAM Assay App’s performance across different experimental setups and conditions. While the app is designed to be versatile and applicable to various imaging modalities, it is advisable to validate its performance within specific experimental contexts. As with any study design, researchers should conduct thorough evaluations and assess accuracy and consistency across different experimental settings.

We know that the analysis and measurements can also be performed manually, however, this is a very time-consuming and tedious task that requires absolute precision. The human factor (inter- and intra-observer variability) which can affect the results and reproducibility of your study must also be taken into account. Using the CAM Assay App you can rule out human bias from your analysis.

On top of that, the IKOSA Platform provides a centralized hub for storing, analyzing, and sharing data, enabling seamless collaboration and access to your image data from anywhere in the world.

As you can see in this use case, the whole process does not require a lot of time or any specific AI-related skills. It took us about 25 min to prepare the data and run the app. Please consider that as a first-time user, it might take you a little longer than that. However, once you are familiar with our software, it saves valuable time, reduces human error, and can be applied to your entire data set.


Working in the life science industry, we understand the need for reproducible, reliable, and accurate results. By demonstrating to you how the CAM Assay App works, we want to share a simple and hassle-free AI-based method to gather data from CAM experiments. See what researchers who have already used our app have to say.

IKOSA users share their experience with the CAM Assay Application: 

“I’m very happy to automate the counting of new blood vessels process with IKOSA and get reliable results without wasting too many resources and being able to focus on our priorities.”

Dr. Nassim Ghaffari Tabrizi-Wizsy, the Institute of Pathophysiology and Immunology at the Medical University of Graz

“IKOSA provides a user-friendly premium software solution that meets high-quality standards at the same time. Thanks to it we have been able to bring our studies on the CAM Model a huge step ahead and we are eagerly looking forward to future joint projects.”

Prof. Dr. Silke Härteis, University of Regensburg, Germany

“We present a reliable method to quantify angiogenic and anti-angiogenic processes in the chick chorioallantoic membrane (CAM) that illuminate the mechanisms of vascular proliferation. Blood vessels on CAM images can be efficiently analyzed with the IKOSA CAM Assay App to measure their total area, length, mean thickness, and the number of branching points.”

Prof. Dr. Domenico Ribatti, University of Bari Aldo Moro, Italy

The IKOSA CAM Assay App has already been featured in several high-ranking biomedical publications:

The next breakthrough in angiogenesis research might be yours!

Try the CAM Assay App and the full functionalities of IKOSA at no cost.

Our authors:

KML Vision Team Benjamin Obexer Lead Content Writer

Benjamin Obexer

Lead content writer, life science professional, and simply a passionate person about technology in healthcare

KML Vision Team Elisa Opriessnig Content writer

Elisa Opriessnig

Content writer focused on the technological advancements in healthcare such as digital health literacy and telemedicine.

KML Vision Team Fanny Dobrenova Marketing Specialist

Fanny Dobrenova

Health communications and marketing expert dedicated to delivering the latest topics in life science technology to healthcare professionals.


Bichlmayer, E. M., Mahl, L., Hesse, L., Pion, E., Haller, V., Moehwald, A., … & Haerteis, S. (2022). A 3D In Vivo Model for Studying Human Renal Cystic Tissue and Mouse Kidney Slices. Cells, 11(15), 2269.

Faihs, L., Firouz, B., Slezak, P., Slezak, C., Weißensteiner, M., Ebner, T., … & Dungel, P. (2022). A Novel Artificial Intelligence-Based Approach for Quantitative Assessment of Angiogenesis in the Ex Ovo CAM Model. Cancers, 14(17), 4273.  

Kuri, P. M., Pion, E., Mahl, L., Kainz, P., Schwarz, S., Brochhausen, C., … & Haerteis, S. (2022). Deep Learning-Based Image Analysis for the Quantification of Tumor-Induced Angiogenesis in the 3D In Vivo Tumor Model—Establishment and Addition to Laser Speckle Contrast Imaging (LSCI). Cells, 11(15), 2321.