Johannes Textor

Radboud UMC

 

In short, Johannes Textor's research entails the following: 

Transfer Learning for Immune Cell Landscape Analysis in Salivary Gland Cancer

'Salivary gland cancer is a rare type of head-and-neck cancer with 150-200 diagnoses per year in the Netherlands, and the most aggressive subtypes have poor prognosis. To develop new treatment options, we are imaging the interactions between immune system cells and tumor cells within patient biopsies using high-resolution digital microscopy. Machine learning approaches are the state of the art for analyzing such data, but they can require very large datasets to train on, which are usually not available for rare cancer types. In our project, we will address this problem using "transfer learning" methodology that allows machine learning algorithms to benefit from experience gained on larger datasets from more common cancer types and train more effectively on smaller datasets. Leveraging existing data and knowledge in this manner, we hope that our project will help to build a rationale for future immunotherapy treatments for salivary gland caner patients.'

 

Publications

• A Uniformly Bounded Correlation Function for Spatial Point Patterns
  • ACM Digital Library, August 24, 2024
  • Click here to view the publication.
• Detection of dendritic cell subsets in the tumor microenvironment by multiplex immunohistochemistry
  • PubMed, October 24, 2023
  • Click here to view the publication. 
• Multiplex Immunohistochemical Analysis of the Spatial Immune Cell Landscape of the Tumor Microenvironment
  • PubMed, August 18, 2023
  • Click here to view the publication.

 

Short summary of final results August 2025

Salivary gland cancer is a rare cancer in the head-and-neck region. In the Netherlands, between 150 to 200 people per year are diagnosed with salivary gland cancer . Of those, patients with salivary duct carcinoma (SDC) have an especially poor prognosis. For example, a typical patient with metastatic salivary duct carcinoma will live only 5 months. New therapy options for this rare and aggressive type of cancer are needed to improve this. There are now many new types of cancer treatments that use different parts of the patient’s own immune system to attack and destroy the cancer cells. Which of these new treatments might be the most promising for SDC? To answer this question, we have analyzed tissue samples of more than 200 SDC patients. Using artificial intelligence algorithms, we determined the frequency and type of different immune system cells in these tissue samples. We found that patients who had more “helper T cells” in their tissues tended to remain disease-free for longer than those with fewer helper T cells. If T cells play a role in determining the progression of disease, then therapy approaches that strengthen the T cells might therefore be beneficial for SDC patients.

The artificial intelligence algorithms we develop to perform our research are not only useful for analyzing SDC cancer tissue but can be applied to any kind of cancer tissue. The same is true for new methods that we developed to help interpret and explain the results of these AI algorithms to medical specialists. We have made these new methods available as free software that anyone can download and use. At the Radboudumc, we are already applying these new tools to other types of cancer – for example, in another project funded by the Hanarth foundation, we are applying these tools to study melanoma in the eye.

Short summary of progress / results May 2024

We have now completed the development of our machine learning algorithms and data analysis methods, and published three papers describing these methods in detail so that they can be used by other researchers. By using advanced statistical methods, we have now gathered information from 209 patients for detailed analysis. During the final year, we will combine the information found in the tissue samples of these patients with other clinical information to determine whether the spatial information can provide relevant information about prognosis for salivary duct cancer patients.

Short summary of progress / results May 2023

We have further developed our machine learning methods to analyze immune cells that play important roles in the immune response against cancer: dendritic cells and myeloid cells. To help other researchers and doctors make sense of the complicated data that is generated by analyzing large microsopy images, we have developed new methods to produce intuitive and interpretable visualizations and quantitative information.

Short summary of progress / results May 2022

We have gathered tissue biopsies from salivary gland cancer patients in the Netherlands and Japan. We have stained these tissue biopsies with multiple fluorescent markers to detect many types of immune cells and are now analyzing digital images of these biopsies using our custom-made artificial neural network.