AI Predictions for Effective Precision Oncology

According to the WHO, colorectal cancer (CRC) is the second biggest cause of cancer-related deaths worldwide. For the first time, researchers from Helmholtz Munich and the University of Technology Dresden (TU Dresden) demonstrate that artificial intelligence (AI)-based predictions could yield findings equal to clinical testing on CRC samples.

AI Predictions for Effective Precision Oncology
AI-Based Image Analysis for Biomarker Prediction. Image Credit: Helmholtz Munich

AI predictions can hasten tissue sample examination, leading to quicker treatment choices. A big step has been taken towards implementing precision therapeutic techniques in the field of cancer with the development of this new biomarker detection model. The procedure has just been released in the Cancer Cell.

Researchers led by Dr. Tingying Peng from Helmholtz Munich and Prof. Jakob N. Kather from TU Dresden demonstrate that AI can predict certain biomarkers in stained tissue samples from CRC patients. To find patterns and help diagnostic choices for cancer care, they employed so-called transformer networks, a contemporary deep learning (DL) method. The novel strategy for biomarker identification vastly improves on earlier techniques.

Large-Scale Evaluation Proves Better Generalization and Data-Efficiency

The study team created software that conducts analysis using transformer neural networks, a novel technique. By evaluating it on a large multicentric group of over 13,000 patients from 16 cohorts from seven countries (Australia, China, Germany, Israel, Netherlands, UK, and USA), some of which were contributed by researchers at the German Cancer Research Center (DKFZ) Heidelberg and the network of the National Centers for Tumor Diseases (NCT), they demonstrate that their approach significantly improves performance, generalizability, data efficiency, and interpretability.

The algorithm trained on the huge multicentric cohort yields high sensitivity on resection tissue samples collected after surgery. Surprisingly, even though their model was only trained on tissue samples from resections, the findings could yield good performance on biopsy tissue taken during colonoscopy.

The generalization to biopsy tissue increases the algorithm’s benefit for the patient when ultimately implemented in clinical routine.

Sophia J. Wagner, Study First Author and PhD Student, Helmholtz Munich

AI-Based Pre-screening for Biopsies Accelerate Diagnosis

Due to its great sensitivity to biopsy tissue, the algorithm could be used as a pre-screening technique before affirmative testing in circumstances where AI testing yielded a positive result. By utilizing AI-based biomarker prediction, it can be possible to shorten the testing period and accelerate the process of determining a patient’s genetic risk status after a biopsy, allowing for early immunotherapy treatment if necessary.

Journal Reference:

Wagner, S. J., et al. (2023) Transformer-based biomarker prediction from colorectal cancer histology: A large-scale multicentric study. Cancer Cell. doi:10.1016/j.ccell.2023.08.002

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.