New AI-Based Technique to Classify Colorectal Cancer with IR Imaging

Automatic detection of intestinal tumor type can be performed by using infrared microscopy in just 30 minutes. Then, the results are utilized to make targeted therapy decisions.

At Ruhr-Universität Bochum (RUB), a research group from the Prodi Centre for Protein Diagnostics has made use of quantum cascade laser-based infrared (IR) microscopes to categorize colorectal cancer tissue samples from regular clinical operations in an automatic and marker-free manner.

Using artificial intelligence, the researchers were able to distinguish between several tumor types with high precision within about 30 minutes.

Depending on the category, doctors would be able to predict which path the disease will take, and thus, the suitable therapy can be selected. The research group published its report in the Scientific Reports journal on June 23^rd, 2020.

Microsatellite Status Facilitates Prognosis

Microsatellite instable (MSI) and microsatellite stable (MSS) tumors in the colon and other types of cancers are differentiated. In general, microsatellites frequently repeated short DNA sequences that do not have any function.

The survival rate of patients who have MSI tumors is higher. This is because the mutation rate of cancer cells is around 1,000 times higher, making their growth not so successful. Furthermore, novel immunotherapy turns out to be highly successful in patients with MSI tumors.

It is therefore important for the prognosis and the therapy decision to know what kind of tumor we are dealing with.

Anke Reinacher-Schick, Professor and Head of the Department of Haematology and Oncology, St. Josef Hospital, Ruhr-Universität Bochum

Until now, a differential diagnosis has been performed by immunohistochemical staining of tissue samples followed by complex genetic analysis.

Fast and Reliable Measurement

Previous studies by the research group under the guidance of Professor Klaus Gerwert from the Department of Biophysics, RUB, have already demonstrated the potential of using IR imaging as a diagnostic tool for the classification of tissue, also called label-free digital pathology.

The technique identifies cancer tissue without previous staining or other marking, and thus functions automatically with the help of artificial intelligence. In contrast to the traditional differential diagnosis of microsatellite status, which takes around a day, the latest technique takes only about 30 minutes.

The protein research team has considerably enhanced the technique by improving it for the detection of a molecular variation in the tissue. Earlier, the tissue could only be visualized morphologically.

This is a big step that shows that IR imaging can become a promising method in future diagnostics and therapy prediction.

Klaus Gerwert, Professor, Department of Biophysics, Ruhr-Universität Bochum

Encouraging Feasibility Study

In association with the Institute of Pathology at RUB under the guidance of Professor Andrea Tannapfel and the Department of Haematology and Oncology at the RUB St. Josef Hospital, the research group carried out a feasibility study involving 100 patients.

It exhibited a specificity of 93% and sensitivity of 100%—all MSI tumors were accurately classified with the latest technique, and very few samples were wrongly classified as MSI tumors.

At present, an extended clinical trial has been initiated, which will be performed on samples from the Colopredict Plus 2.0 registry study. The registry study was started by Andrea Tannapfel and Anke Reinacher-Schick and enables the validation of the outcomes from the published study.

According to Andrea Tannapfel, “The methodology is also of great interest to us, because very little sample material is used, which can be a decisive advantage in today’s diagnostics with an increasing number of applicable techniques.”

Another Step Toward Personalized Healthcare

In the forthcoming days, the technique will be used in the clinical workflow to evaluate its prospects for precision oncology.

Following an increasingly targeted therapy of oncological diseases, it is very important to provide rapid and precise diagnostics.

Anke Reinacher-Schick, Professor and Head of the Department of Haematology and Oncology, St. Josef Hospital, Ruhr-Universität Bochum

The study was performed at the Prodi Research Centre for Protein Diagnostics (funding code 111.08.03.05-133974) and formerly the Pure Consortium (funding code: 233-1.08.03.03-031-68079), funded by the Ministry of Culture and Science of the State of North Rhine-Westphalia.

Journal Reference:

Kallenbach-Thieltges, A., et al. (2020) Label-free, automated classification of microsatellite status in colorectal cancer by infrared imaging. Scientific Reports. doi.org/10.1038/s41598-020-67052-z.