Can imaging help young adults lower colorectal cancer rates?

A concern grows among young adults. Rising at a significant rate, colorectal cancer is now the leading cause of cancer-related deaths in men and women under 50-years-old. A contrast occurs as rates rise, though. Screening techniques have also advanced. Today’s non-invasive methods can, for example, more accurately predict outcomes from a single CT scan. Pattern recognition in new AI tools can identify at-risk individuals and improve upon demographic-based guidelines.

One of the major keys to success in cancer treatment is not a pharmaceutical discovery or some secret remedy. Early detection increases the chance of survival eightfold in colorectal cancer if treatment is delivered early, according to Statistics Canada. Physicians can perform an optical colonoscopy, for example, as a primary or follow-up examination for individuals at risk of colorectal cancer. Not all first-stage imaging techniques are so invasive, though. Advancements in testing can bring more young adults into clinics by making the process easier and more efficient.

Physicians and oncologists can use colonoscopies to extract a tissue sample (biopsy) that will reveal valuable, microscopic data with the right tools. The photo depicts a histopathologic image of a colorectal precursor known as a colonic polyp.

A smarter way to look at risk

Colorectal cancer is composed of two main groups: rectal and colon cancer. Virtual colonoscopies use CT Imaging and offer a non-invasive approach to detect polyps and colon cancer. CT Imaging does not require a camera or tools to be physically inserted into the body, meaning patients can avoid sedation. Visitation time and patient prep are also significantly shorter and less burdensome with virtual colonoscopies. Colorectal cancer also includes rectal tumors, which make up 30% of cases. These are visible with an MRI scan. Advancements in imaging techniques can lessen the demand for more expensive procedures.

Dr. Sean Raj, a Board Certified Radiologist, is the Chief Medical and Innovation Officer at SimonMed. It is a network of physician-owned outpatient imaging clinics across the United States. Dr. Raj completed residency at Baylor College of Medicine.

Uprooted Concepts asked Dr. Raj how modern tools, including AI, are influencing the landscape of colorectal cancer detection, screening uptake, and personal risk assessments.

Dr. Sean Raj: “AI is already starting to meaningfully change how we detect colorectal cancer. On the imaging side, it helps us pick up subtle or easily overlooked lesions and brings a level of consistency that reduces variability between readers.

For example, in CT colonography, machine learning models can improve detection of smaller or flatter polyps, exactly the ones that are traditionally the easiest to miss.

Where I think this really gets interesting is upstream. At SimonMed we are using AI to risk-stratify patients based on elements like demographics and clinical history to help identify individuals who may benefit from earlier or more frequent screening, rather than relying solely on age-based guidelines. This shift from reactive to more personalized, proactive care is where the biggest impact will be.”

Colon cancer

Testicular cancer was the most prevalent cancer in men ages 15 to 35 before 2017. Whereas, for women in the age group, breast carcinoma had the highest rate of mortality. It, too, was outpaced by colorectal cancer rates in 2021. Recent shifts in oncological data suggest that young adults are more commonly suffering from late-stage colon cancer.

U. C.: Is the increase in colorectal cancer deaths correlated with an overall increase in cases? Or, have you observed a relative increase in late-stage cases among young adults?

Dr. Raj: “What’s particularly concerning is that while overall colorectal cancer incidence has stabilized or declined in older populations due to screening, we are seeing a clear rise in incidence among adults under 50.

More troubling is how this younger population is presenting. Younger patients are more likely to show up with later-stage disease, like Stage III or IV disease, which is much further along and often already spread to other parts of the body. A big part of that is delayed recognition — symptoms get attributed to more benign causes, and screening isn’t happening early enough.

From an imaging standpoint, we’re seeing more cases where the first diagnosis comes during a workup for pain or obstruction, rather than through routine screening. That’s a very different, and much more challenging entry point into the system.”

U. C.: How are advances in detection methods and personalized diagnostics able to improve treatment protocols and patient outcomes, such as biomarker detection? (Eg. CD47 expression. 1-3)

     Dr. Raj: “Advances in molecular diagnostics are allowing us to better characterize tumors pre-treatment. High-resolution MRI and CT, combined with molecular markers like CD47 expression, can help identify more aggressive tumor biology and guide decisions around neoadjuvant therapy or immunotherapy.

On the radiology side, we’re also increasingly using imaging biomarkers, things like tumor heterogeneity or perfusion, to assess treatment response earlier than we could before. That matters because it allows oncologists to adjust therapy sooner, rather than waiting for traditional size-based changes. Ultimately, it’s about making treatment more dynamic and tailored to the individual patient.”

Lifestyle

Activities and habits that are considered high risk certainly increase the chance of developing polyps and tumors that can become cancerous. Inflammation is caused by numerous biological reactions that typically involve immune responses and promote cellular oxidation. These conditions create optimal pro-tumorigenic microenvironments.

Chronic stress, poor diet and sleep, or a lack of physical activity, for example, can induce inflammation. Sustained lifestyle habits that promote colorectal cancer can place people in a higher risk category. The importance of regular choices that reduce inflammatory foods and stress cannot be overstated. With that said, Dr. Raj suggested that disruptions to the health care system, such as the one caused by the Covid pandemic, caused more issues for patients than the virus itself.

When it’s time to go

Before opting for a virtual colonoscopy, consult with your healthcare provider to determine the most suitable screening method based on your medical history and specific circumstances. Older adults visit clinics for check-ups more frequently than individuals under 50. Modest success within the demographic, whether it is by guideline pressure or a greater health-conscious mindset, suggests that young adults can use available imaging resources to lower the growing number of colon cancer cases.

A note for industry experts

Radiomic analysis reveals real-time characteristics of colon and rectal tumors. Imaging can capture information on physical properties, molecular information, and genetic data. Research in the field, for example, continues to improve upon virtual detection of BRAF and KRAS mutations. (4, 5)

Studies have evaluated the accuracy of virtual CT and MRI in predicting outcomes of chemotherapy and immunotherapy. For example, Zhang et al. developed a prediction model for immunotherapy in higher-stage Colorectal Cancer with Microsatellite Instability. Researchers combined CT with RNA sequencing and characterized two immunophenotypes. They identified twenty-two radiomic features associated with neoadjuvant therapy. A model built from three of the most distinct features predicted PD-1 inhibitor outcomes with high accuracy and interpretability. Success does not come without limitations, though. Immunotherapy treatments can lag behind radiomic assessments, which can limit treatment viability in rapidly progressing tumors.

Wirth et al. conducted a meta-analysis that assessed twelve papers on AI-enhanced radiomic predictions in chemotherapy. The authors found excellent predictions from three algorithms and good-quality results from four studies in their analysis. Authors did, however, find limitations with algorithm standardization.

Sources

1. Li Q, Vignali P, Tang D, et al. Expression and Clinical Significance of CD47 in Colorectal Cancer: A Review. Cancers (Basel). 2025;18(1):54. Published 2025 Dec 24. doi:10.3390/cancers18010054

2. Zhang Y, Zhang X, Zhong X, et al. Immunophenotype-guided interpretable radiomics model for predicting neoadjuvant anti-PD-1 response in stage III-IV d-MMR/MSI-H colorectal cancer. J Immunother Cancer. 2025;13(8):e011569. Published 2025 Aug 4. doi:10.1136/jitc-2025-011569

3. Wirth L, Cooper E, Chan XQ, et al. Radiomic Artificial Intelligence Models Predicting the Response of Colorectal Cancer Liver Metastases to Chemotherapy-A Systematic Review. ANZ J Surg. 2026;96(3):432-439. doi:10.1111/ans.70412

4. Negreros-Osuna AA, Parakh A, Corcoran RB, et al. Radiomics Texture Features in Advanced Colorectal Cancer: Correlation with BRAF Mutation and 5-year Overall Survival. Radiol Imaging Cancer. 2020;2(5):e190084. Published 2020 Sep 18. doi:10.1148/rycan.2020190084

5. Badic B, Tixier F, Cheze Le Rest C, Hatt M, Visvikis D. Radiogenomics in Colorectal Cancer. Cancers (Basel). 2021;13(5):973. Published 2021 Feb 26. doi:10.3390/cancers13050973