Does AI unlock genetic keys to improve colorectal cancer survival?

Last Updated on March 21, 2025 by Joseph Gut – thasso

March 20, 2025 – Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths worldwide, with approximately 20% of patients presenting with metastatic disease at the time of diagnosis. Despite the cancer’s long progression time, often 5–10 years, is the majority found late, making early detection crucial for improving survival rates.

Early screening, combined with genetic testing, enables clinicians to make more informed decisions, ultimately improving patient survival. Dr. Lin Cong, a leading research scientist from BGI Genomics’ Institute of Intelligent Medical Research (IIMR), is at the forefront of a transformative CRC study. This collaborative research with Sweden’s Uppsala University was published in the Journal Nature in 2024. The studied patient cohort involved 1,063 CRC patients and genetic profiling, which could revolutionize the  approach of prevention, diagnosis, and treatment, and may provide insights into how research is shaping the future of personalized cancer care.

Novel driver genes for early intervention

Through the study, Dr. Lin’s team identified several key genetic alterations specific to CRC, including mutations in pathways like WNT, EGFR, TGFβ, and even mitochondrial genes. These mutations play a crucial role in disease progression and could be used for better early detection and treatment. The researchers discovered mutations in genes like APC, TP53, and KRAS that tend to appear early in cancer’s development, and also identified mutations that emerge later, which could be targets for more advanced therapies. These findings may offer critical guidance for early detection strategies and identify key events relevant to CRC invasion and metastasis.

New prognostic subtypes guide personalized treatment

In their research, Lin Cong’s team analyzed tumor gene expression profiles and identified five prognostic subtypes with distinct molecular characteristics, known as Colorectal Cancer Prognostic Subtypes (CRPS). This new system allows for more accurate prediction of patient outcomes compared to existing methods. The widely used Consensus Molecular Subtypes (CMS) classification system categorizes CRC into four molecular subtypes based on gene expression patterns. While CMS classification provides a useful framework, the CRPS system offers more precise prognostic predictions by integrating independent cohort transcriptome data. Thus, by analyzing CRC at a deeper genetic level, researchers  can personalize treatment more effectively.

The further developed system CRPS goes beyond the Consensus Molecular Subtypes (CMS) classification, providing a clearer picture of each patient’s prognosis. Some tumors classified under CMS4, which are typically considered aggressive, were reclassified as CRPS2, a subtype with a better prognosis.

Early screening for better survival outcomes

Early detection of CRC, which is often diagnosed too late due to its subtle onset is important.. She noted that, despite the disease’s long progression time, patients frequently present with metastasis at diagnosis, which significantly reduces treatment effectiveness. Thus, a large Swedish study, published in JAMA Network Open in February 2024, revealed that cancer-specific mortality decreased significantly by 14% in patients who underwent early screening for CRC, as compared with a control group that received late or no invitations to participate in screening. Excess mortality decreased by 16% in the exposure group. Another study exploiting the time lag in the implementation of organized screening in Sweden found that areas such as Stockholm and Gotland, which began offering screening in 2008, showed a 14% reduction in long-term colorectal cancer mortality compared to regions that started screening later. The current trend in public education about colorectal cancer is encouraging, but  to do more to raise awareness of early screening is still needed. In addition to conventional treatments, specific subtypes of early-stage CRC can be treated with immunotherapy, improving patient survival rates;

An AI-driven future for diagnosis and treatment

This is where artificial intelligence (AI) is coming into play. Dr. Lin’s research group aims to develop a universally applicable AI model in the next three to five years that  shall enable faster, more accurate diagnoses, offering clinicians the ability to make data-driven decisions that can transform patient care. The AI model will be trained using a vast amount of spatial-temporal -omics data, including information on patient cell morphologies, genetic alterations, and gene expression patterns. This technology could one day be used to determine whether a patient has cancer, identify the stage or grade of the disease, and even pinpoint specific target gene mutations from a single biopsy. Once such a AI model for CRC is established, it cold be adapted for use in other cancers and diseases. This technology will allow clinicians to predict disease occurrence, monitor progression, and select the best treatments for each individual, leading to more effective, personalized health care.

Needless to mentioned that thasso had already had several contributions on possible roles of AI (and Chatbots) in medicine and patient care here, here, here, here, and here, to name just a few.

See here a sequenco on colorectal cancar facing AI:

Disclaimer: Images and/or videos (if any) and some text passages in this blog may be copyrighted. All rights remain with the owner of such rights.