Metastatic Colorectal Cancer and Contiguous Spread: How CT, MRI, PET/CT, Ultrasound, and AI Redefine the Radiologist’s Role
1. Clinical Hook
A 73-year-old man with colorectal malignancy presents an imaging problem that is more complicated than simply identifying a primary tumor.
The critical question is not merely:
“Where is the cancer?”
It is:
“How far has the cancer traveled, by which route, and does that pattern change management?”
That distinction is fundamental in advanced colorectal cancer.
The supplied case material describes metastatic disease and contiguous spread through several possible pathways, including direct extension, lymphatic dissemination, peritoneal spread, fascial-plane extension, and hematogenous dissemination.
The imaging set illustrates why no single modality can answer every clinical question.
Contrast-enhanced CT provides the broad anatomic map. MRI contributes superior soft-tissue characterization, particularly in the pelvis. FDG-PET/CT identifies metabolically active disease. Ultrasound can detect selected hepatic lesions and support bedside assessment. Colonoscopy directly visualizes the mucosal tumor and permits biopsy.
The radiologist therefore becomes more than an image interpreter.
The radiologist becomes the person who reconstructs the route of tumor spread.
That reconstruction can determine whether disease is potentially resectable, whether adjacent organs are threatened, whether systemic therapy should be prioritized, and whether apparently separate abnormalities represent true metastases or unrelated findings.
2. Learning Objectives
After reading this article, the reader should be able to:
Explain the major pathways of metastatic and contiguous colorectal cancer spread.
Recognize the key CT, MRI, PET/CT, ultrasound, and endoscopic findings illustrated in this case.
Distinguish direct pelvic invasion from distant metastatic disease.
Understand why CT and MRI provide complementary rather than competing information.
Identify the major imaging pitfalls in metastatic colorectal cancer staging.
Understand where radiomics, foundation models, vision-language models, and clinical AI may augment—not replace—the radiologist.
3. Anatomy Review
Colorectal tumor spread is strongly influenced by anatomy.
The colon and rectum are not isolated tubes. Their lymphatic drainage, vascular supply, mesenteric attachments, serosal surfaces, pelvic fascial planes, and relationships to adjacent organs determine the routes through which malignant cells can travel.
The major pathways include:
Direct transmural extension
Lymphatic dissemination
Portal venous hematogenous spread
Systemic hematogenous spread
Peritoneal dissemination
Fascial-plane or contiguous extension
The liver has particular importance because venous drainage from much of the colon enters the portal circulation. This creates a biologically plausible pathway for hepatic metastasis. The supplied source specifically identifies the liver as the most commonly affected distant organ in colorectal cancer, followed by lung and peritoneal disease.
Figure 1. Anatomy of Metastatic Colorectal Cancer Spread
4. Case Presentation
History
The supplied case identifies a 73-year-old man with colorectal cancer. Detailed presenting history is not provided in the source material.
Symptoms
The source describes common manifestations of colorectal cancer and metastatic disease, including abdominal discomfort, altered bowel habits, rectal bleeding, weight loss, fatigue, anemia, and symptoms related to hepatic, pulmonary, peritoneal, osseous, or cerebral metastases.
The actual symptom profile of this individual patient is not documented and therefore should not be reconstructed.
Physical Examination
No patient-specific physical examination findings are provided.
Clinical Question
The imaging question is best framed as:
What is the extent of the primary colorectal malignancy, and is there evidence of metastatic or contiguous tumor spread that changes staging or treatment strategy?
Laboratory Findings
Patient-specific laboratory values are not provided.
In clinical practice, CBC, liver function tests, renal function, carcinoembryonic antigen, and molecular biomarkers may contribute to management, but these cannot be attributed to this case without documentation.
CT Findings
The supplied CT demonstrates a measurable colorectal tumor. The original figure includes axial and coronal reformations and quantitative measurements of tumor dimensions and CT attenuation. The source describes measurement of the maximum tumor diameter perpendicular to the bowel wall, tumor length on coronal reconstruction, and manually delineated tumor area with attenuation statistics.
This is important because tumor measurement is not merely descriptive.
In advanced disease, reproducible measurements can become the foundation for:
Baseline staging
Treatment response assessment
Surgical planning
Longitudinal surveillance
Quantitative imaging analysis
Radiomics
The CT also provides the global survey necessary to evaluate liver, peritoneal, nodal, and other distant disease.
MRI Findings
The supplied CT-versus-MRI image demonstrates the complementary role of MRI in pelvic assessment.
MRI provides superior soft-tissue contrast and can more clearly define relationships between a pelvic tumor and adjacent structures. The source specifically emphasizes its value for evaluating soft tissues, liver metastases, peritoneal spread, and adjacent organs such as the bladder and small bowel.
For suspected rectal involvement, high-resolution pelvic MRI is particularly important because local staging depends on anatomical relationships that are difficult to characterize reliably with CT alone.
Current ACR criteria rate pelvic MRI as usually appropriate for locoregional staging of rectal cancer.
PET/CT Findings
The supplied PET/CT demonstrates focal metabolic activity in the pelvis/sacral region on the provided fused and component images.
However, the source does not provide SUV values, histologic confirmation, or a formal statement that the osseous focus represents metastasis.
Therefore, the correct radiologic interpretation is:
Hypermetabolic pelvic/sacral abnormality suspicious for metabolically active disease, requiring correlation with the CT morphology, MRI findings, clinical history, and—when clinically necessary—additional confirmation.
PET positivity alone should not automatically be equated with metastatic disease.
Pathology
No histopathologic report is provided in the source.
Final Diagnostic Interpretation
The available material supports:
Metastatic colorectal malignancy with a pelvic/rectal tumor component demonstrated on imaging, associated distant metastatic disease including hepatic involvement illustrated on ultrasound, with additional metabolically active pelvic/sacral abnormality on PET/CT.
The exact TNM stage cannot be assigned from the supplied material alone.
5. Pathophysiology
Colorectal carcinogenesis is a multistep process involving genomic and epigenomic alterations.
The source specifically identifies APC, KRAS, TP53, microsatellite instability, DNA methylation, and histone modification as mechanisms relevant to tumor progression and metastatic behavior.
The biological sequence can be conceptualized as:
Mucosal transformation → adenomatous or malignant progression → invasion through the bowel wall → lymphovascular access → regional or distant dissemination → metastatic colonization.
Once tumor cells breach the muscularis propria and reach deeper layers, access to lymphatic and vascular channels becomes increasingly important.
The radiologic appearance therefore reflects biology.
Why does the tumor become irregular?
Irregular tumor margins can reflect asymmetric infiltration of the bowel wall and surrounding tissues rather than simple expansile growth.
Why does contiguous spread occur?
Direct extension follows anatomical resistance planes. The tumor may cross the serosal surface and infiltrate neighboring structures when the anatomical interface is lost.
Why does the liver become involved?
Portal venous drainage provides a direct vascular route from colorectal tumors to hepatic tissue.
Why can MRI show more than CT in the pelvis?
MRI separates soft tissues based on intrinsic relaxation characteristics, allowing improved visualization of tumor boundaries, pelvic fascia, muscles, neurovascular structures, and adjacent organs.
Figure 2. Tumor Biology and Imaging Phenotype
6. Epidemiology
Colorectal cancer remains one of the world's major cancer burdens.
The source material cites approximately 1.9 million new cases globally in 2020 and describes colorectal cancer as the third most common malignancy.
More recent GLOBOCAN 2024 estimates continue to place colorectal cancer among the most frequently diagnosed cancers worldwide, accounting for approximately 9.9% of all new cancers and approximately 9.4% of cancer deaths.
| Epidemiologic Feature | Clinical Interpretation |
|---|---|
| Global incidence | Among the highest of all malignancies |
| Age | Predominantly older adults, but early-onset disease is increasing |
| Sex | Affects both sexes |
| Major risk factors | Age, family history, hereditary syndromes, obesity, diet, alcohol, smoking, and inflammatory bowel disease |
| Common metastatic site | Liver |
| Other major sites | Lung, peritoneum, distant lymph nodes |
| Less common sites | Bone and brain |
The supplied material estimates that approximately 25–30% of patients may have metastatic disease at diagnosis and that 40–50% may develop metastases during the disease course.
These figures vary according to population, stage distribution, screening, tumor biology, and follow-up duration and should therefore not be treated as universal fixed probabilities.
7. Clinical Presentation
Symptoms of the primary tumor can include:
Abdominal pain
Altered bowel habits
Rectal bleeding
Weight loss
Fatigue
Iron-deficiency anemia
Metastatic disease can produce organ-specific manifestations.
Hepatic metastases may produce jaundice or right upper-quadrant discomfort. Pulmonary disease may cause cough or dyspnea. Peritoneal dissemination can produce abdominal distention and ascites. Bone metastases may cause pain or fracture, while brain metastases can produce headache, seizure, or neurological deficit.
Red Flags
Particularly concerning findings include:
Unexplained iron-deficiency anemia
Rectal bleeding
Progressive change in bowel habits
Unintentional weight loss
Persistent abdominal pain
Bowel obstruction
Palpable mass
New hepatic lesions
Multiple indeterminate pulmonary nodules
Peritoneal nodularity or ascites
New destructive osseous lesion
8. Imaging Features: The Radiologist’s Perspective
Figure 3. Contrast-Enhanced Abdominal CT
The original CT demonstrates quantitative tumor assessment using axial and coronal images.
Radiologist Interpretation
The critical task is not simply to measure the mass.
The radiologist should determine:
Exact tumor location.
Longitudinal tumor extent.
Circumferential involvement.
Depth of wall invasion.
Adjacent organ involvement.
Regional lymphadenopathy.
Liver metastases.
Peritoneal disease.
Distant lymph nodes.
Osseous or pulmonary abnormalities within the field of view.
Imaging Pearl
Never interpret the primary colorectal tumor in isolation.
A technically excellent CT that identifies the primary mass but fails to characterize the liver, peritoneum, mesentery, and adjacent organs is clinically incomplete.
9. CT Versus MRI
Figure 4. CT and MRI Comparison
The supplied comparison demonstrates why CT and MRI answer different questions.
| Modality | Strength | Limitation | Clinical Value |
|---|---|---|---|
| CT | Rapid whole-body anatomical survey | Lower soft-tissue contrast | Initial staging and systemic disease mapping |
| MRI | Superior soft-tissue contrast | Longer examination and motion sensitivity | Pelvic local staging and liver characterization |
| DWI MRI | Sensitive to cellularity changes | Susceptible to artifacts | Tumor detection and response assessment |
| ADC | Quantitative diffusion information | Requires appropriate acquisition | Supports interpretation of restricted diffusion |
| PET/CT | Metabolic assessment | Limited spatial resolution and false positives | Problem solving and selected systemic staging |
| Ultrasound | Accessible and real-time | Operator dependent | Liver lesion detection and targeted assessment |
| Endoscopy | Direct mucosal visualization | Does not provide complete extraluminal staging | Diagnosis and biopsy |
Current ACR guidance supports contrast-enhanced CT of the chest, abdomen, and pelvis for systemic staging of colon cancer, while MRI has a complementary role.
For rectal cancer, pelvic MRI is usually appropriate for locoregional staging.
10. PET/CT
Figure 5. FDG PET/CT
The supplied PET/CT demonstrates a focal hypermetabolic abnormality in the pelvic/sacral region.
PET/CT can be useful when conventional imaging leaves uncertainty regarding:
Distant metastatic disease
Recurrent tumor
Metabolically active lymph nodes
Indeterminate lesions
Treatment response
The supplied source identifies PET/CT as useful for evaluating both local and distant metastatic disease and for assessing metabolic activity.
But PET has an important limitation:
FDG uptake is not synonymous with cancer.
Inflammation, infection, postoperative change, fracture, and other benign processes can produce substantial FDG uptake.
11. Ultrasound
Figure 6. Abdominal Ultrasound
The supplied ultrasound demonstrates a hepatic lesion labeled as metastasis in the source image, alongside a pelvic/rectal abnormality.
Ultrasound remains valuable because it is:
Rapid
Radiation-free
Inexpensive
Dynamic
Suitable for targeted hepatic assessment
Useful for image-guided procedures
However, a negative ultrasound should not automatically exclude small or poorly conspicuous hepatic metastases.
MRI can provide superior lesion characterization in selected patients.
12. Endoscopy
Figure 7. Colonoscopy
The supplied endoscopic image demonstrates a prominent colorectal/rectal mucosal mass.
Colonoscopy has a unique role because it provides direct visualization of the mucosal lesion and enables biopsy. The source explicitly identifies endoscopy as a method for identifying the primary tumor and obtaining tissue.
Imaging and endoscopy therefore answer different questions:
Endoscopy: What does the mucosal tumor look like, and can we obtain tissue?
CT/MRI: How far has the tumor extended beyond the mucosa?
PET/CT: Where is metabolically active disease elsewhere?
13. Plain Abdominal Radiography
Figure 8. Supine Abdominal Radiograph
The supplied radiograph demonstrates abnormal bowel gas pattern and pelvic abnormalities highlighted by arrows.
Plain radiography has a limited role in modern colorectal cancer staging but may still be useful when evaluating:
Bowel obstruction
Perforation
Abnormal gas distribution
Selected osseous abnormalities
It should not replace cross-sectional staging when metastatic disease is suspected.
14. Imaging Differential Diagnosis
| Condition | CT | MRI | PET | Key Differentiator |
|---|---|---|---|---|
| Primary colorectal carcinoma | Irregular mural thickening/mass | Better local tissue definition | FDG uptake possible | Endoscopic confirmation |
| Recurrent tumor | Soft-tissue mass | MRI useful in pelvis | FDG uptake | Comparison with prior imaging |
| Liver metastasis | Hypoenhancing or variable lesion | High sensitivity for small lesions | FDG uptake variable | Multiphasic MRI and morphology |
| Abscess/inflammation | Rim enhancement, inflammatory change | Restricted diffusion may occur | Often FDG avid | Clinical and laboratory correlation |
| Lymphoma | Bulky nodes, wall thickening | Variable | Usually FDG avid | Distribution and morphology |
| GIST | Exophytic mass | Heterogeneous | Often FDG avid | Origin and growth pattern |
| Pelvic fibrosis | Infiltrative soft tissue | Low-signal fibrotic components | Usually low uptake | Stability and morphology |
The most important pitfall is assuming that every new FDG-avid lesion represents metastatic colorectal cancer.
15. Radiologist Reading Report
Findings
Irregular colorectal/pelvic mass compatible with known colorectal malignancy. The supplied CT demonstrates measurable tumor with axial and coronal assessment. Additional imaging demonstrates hepatic metastatic disease and a metabolically active pelvic/sacral abnormality on PET/CT. MRI provides improved delineation of pelvic soft-tissue relationships compared with CT.
The available material does not provide complete chest staging, histopathology, laboratory values, or formal TNM classification.
Impression
Known colorectal malignancy with locally extensive pelvic/rectal tumor component demonstrated on multimodality imaging.
Hepatic metastatic disease demonstrated on the supplied ultrasound.
Hypermetabolic pelvic/sacral abnormality on FDG PET/CT, suspicious for active malignant disease; correlation with cross-sectional morphology is recommended.
Multimodality imaging is required for complete staging and treatment planning.
16. Clinical Correlation
Imaging should be integrated with:
Histopathology
CEA
Liver function
Molecular biomarkers
MMR/MSI status
RAS status
BRAF status
HER2 status when clinically indicated
Tumor sidedness
Resectability assessment
Patient performance status
Modern colorectal cancer management is no longer based solely on anatomical stage.
It is increasingly based on the intersection of:
Anatomy + Biology + Molecular Profile + Treatment Response.
The 2026 ESMO metastatic colorectal cancer guideline reflects this evolution toward biologically informed treatment selection.
17. Treatment Strategy
Treatment depends on whether metastatic disease is:
Resectable
Potentially convertible to resectable
Unresectable
Limited to one organ
Multiorgan
Symptomatic
Biologically aggressive
The supplied source describes surgery, chemotherapy, targeted therapy, immunotherapy, radiation therapy, and palliative care as major treatment components.
For selected patients with resectable liver or lung metastases, local treatment can potentially provide long-term disease control.
For unresectable disease, systemic therapy is central.
Modern treatment selection requires molecular profiling. Current guidance recommends assessment of RAS and BRAF alterations and universal MMR/MSI testing in newly diagnosed colon cancer; NCCN materials also emphasize molecular testing as a foundation for treatment selection.
ASCO recommendations support biomarker-driven therapy, including pembrolizumab for MSI-H/dMMR metastatic disease and molecularly guided anti-EGFR or anti-VEGF strategies in appropriate RAS-defined and sidedness-defined tumors.
The ESMO metastatic colorectal cancer guideline has also continued to evolve, with updated recommendations published through 2025 and a new clinical practice guideline in 2026.
18. Prognosis
Prognosis is heterogeneous.
The source emphasizes three major determinants:
Extent of metastatic disease
Resectability
Response to treatment
A patient with isolated, technically resectable liver metastases can have a very different prognosis from a patient with diffuse hepatic, pulmonary, peritoneal, and osseous disease.
Therefore, the radiologist should avoid vague statements such as:
“Metastatic disease present.”
A more clinically useful report describes:
Number of metastases
Size
Distribution
Relationship to vessels
Organ involvement
New versus stable lesions
Resectability-relevant anatomy
Treatment response
The value of imaging lies in converting “metastatic” into a precise anatomical treatment map.
19. Artificial Intelligence Perspective
AI is particularly relevant to metastatic colorectal cancer because the disease generates large amounts of multimodal information.
A modern AI system could potentially analyze:
Radiomics
Radiomics converts images into quantitative features describing:
Intensity
Texture
Shape
Spatial heterogeneity
Tumor margins
Enhancement patterns
Recent research suggests potential for radiomics and machine learning to predict colorectal liver metastasis and treatment-related outcomes, but methodological heterogeneity and external-validation limitations remain important barriers.
Foundation Models
Radiology foundation models are designed to learn broad representations from large datasets and can potentially support multiple downstream tasks rather than a single narrow classification problem. However, hallucination, automation bias, domain shift, and validation remain major concerns.
Vision-Language Models
Vision-language models can combine medical images with reports and clinical text.
Potential applications include:
Image-grounded report drafting
Lesion-to-text explanation
Structured staging assistance
Longitudinal comparison
Clinical question answering
Multimodal decision support
Recent reviews emphasize both the rapid expansion of these models and their unresolved validation and dataset-bias problems.
AI Workflow
A clinically realistic architecture might be:
The key point is that AI should operate inside the clinical workflow, not as a disconnected algorithm.
Modern radiology integration increasingly depends on DICOM, PACS, IHE profiles, and interoperability architecture.
AI Limitations
AI may fail because of:
Poor image quality
Unusual tumor morphology
Rare metastatic patterns
Post-treatment changes
Inflammatory mimics
Domain shift
Incomplete clinical information
Incorrect segmentation
Dataset bias
Automation bias
Hallucinated explanations
Therefore:
AI can prioritize, quantify, compare, and assist. It should not independently determine the final oncologic diagnosis.
Figure 9. Enterprise AI Architecture for Metastatic Colorectal Cancer
20. Future of Precision Medicine
The next generation of colorectal cancer imaging will increasingly connect phenotype with genotype.
Radiogenomics
Radiogenomics seeks relationships between imaging phenotype and molecular characteristics.
Digital Twins
A future digital twin could integrate:
Imaging
Pathology
Molecular data
Laboratory data
Treatment history
Longitudinal outcomes
to simulate disease trajectories.
Federated Learning
Hospitals may train models collaboratively without transferring raw patient data between institutions.
Synthetic Data
Synthetic imaging and multimodal datasets may help address limited access to rare metastatic cases.
Multimodal AI
The most powerful systems may not be purely imaging models.
They may simultaneously reason over:
The challenge is not simply building a larger model.
It is proving that the model is safe, reproducible, interpretable, and clinically useful.
21. Clinical Pearls
Never stop after identifying the primary colorectal tumor.
Always inspect the liver carefully.
Assess the peritoneal cavity systematically.
Examine regional and distant lymph nodes.
Look for direct invasion of adjacent organs.
Use pelvic MRI when local rectal staging is clinically relevant.
Do not equate FDG uptake automatically with metastasis.
Compare every suspicious lesion with prior imaging.
Describe metastatic disease in a treatment-oriented manner.
Identify whether lesions appear potentially resectable.
Quantitative tumor measurement should be reproducible.
Molecular information increasingly influences treatment decisions.
Radiomics is promising but not yet a substitute for validated clinical staging.
AI should augment radiologist workflow rather than eliminate expert review.
The most valuable radiology report explains not only what is present, but what it means for the next clinical decision.
Quiz
Question 1
Which imaging modality is generally most appropriate for locoregional staging of rectal cancer?
① Plain abdominal radiography
② Noncontrast CT
③ Pelvic MRI
④ Chest radiography
⑤ Whole-body ultrasound
Answer: ③ Pelvic MRI
Explanation: High-resolution pelvic MRI provides superior soft-tissue contrast and is generally appropriate for locoregional rectal cancer staging.
Question 2
Why is the liver a common site of colorectal cancer metastasis?
① Direct lymphatic drainage to the liver
② Portal venous drainage from the colorectal circulation
③ Direct communication with the biliary tree
④ Arterial embolization from the colon
⑤ Transdiaphragmatic spread
Answer: ② Portal venous drainage from the colorectal circulation
Explanation: Portal venous drainage provides an important hematogenous pathway from colorectal tumors to the liver.
Question 3
Which statement about FDG-PET/CT is most accurate?
① Every FDG-avid lesion is malignant.
② PET/CT completely replaces contrast-enhanced CT.
③ FDG uptake can occur in both malignant and inflammatory processes.
④ PET/CT cannot detect metastatic disease.
⑤ PET/CT is primarily an anatomical imaging modality.
Answer: ③
Explanation: FDG uptake reflects increased glucose metabolism and is not specific for malignancy.
Question 4
Which biomarker is particularly important when considering immune checkpoint therapy in metastatic colorectal cancer?
① Hemoglobin
② Serum sodium
③ MSI/MMR status
④ Amylase
⑤ Albumin alone
Answer: ③ MSI/MMR status
Explanation: MSI-H/dMMR metastatic colorectal cancer is a key setting in which immune checkpoint therapy can have major clinical relevance.
Question 5
What is the most appropriate role of AI in current metastatic colorectal cancer imaging?
① Autonomous final diagnosis without human review
② Replacement of pathology
③ Replacement of multidisciplinary decision-making
④ Human-supervised detection, quantification, comparison, and decision support
⑤ Elimination of radiology reporting
Answer: ④
Explanation: Current AI systems can support radiologists through detection, segmentation, quantification, workflow prioritization, and multimodal analysis, but robust validation and human oversight remain essential.
FAQ
FAQ 1. What is metastatic colorectal cancer?
It is colorectal cancer that has spread beyond the primary site to distant organs or other anatomical compartments.
FAQ 2. What is the most common distant metastatic site?
The liver is the most common distant metastatic site because of the portal venous drainage of the colorectal circulation.
FAQ 3. Is CT enough to stage colorectal cancer?
CT is central to systemic staging, but MRI provides important additional information, particularly for pelvic and liver assessment.
FAQ 4. When is pelvic MRI important?
It is particularly important when the primary tumor involves or is suspected to involve the rectum and detailed local staging is required.
FAQ 5. Does FDG uptake prove metastasis?
No. Inflammation and other benign processes can also be FDG-avid.
FAQ 6. Can liver metastases be surgically treated?
Selected patients with limited and technically resectable hepatic metastases may be candidates for local therapy.
FAQ 7. Why is molecular testing important?
RAS, BRAF, and MMR/MSI status can substantially influence systemic treatment selection.
FAQ 8. Can AI detect colorectal metastases?
AI can assist with lesion detection and quantitative analysis, but clinical deployment requires robust external validation and human oversight.
FAQ 9. What is radiomics?
Radiomics extracts quantitative imaging features that may provide information beyond conventional visual interpretation.
FAQ 10. Will AI replace radiologists?
The more realistic near-term model is radiologist-plus-AI rather than autonomous AI. The difficult cases remain precisely those in which clinical context, multimodality correlation, and expert judgment are most important.
Conclusion
Metastatic colorectal cancer is not a single imaging finding.
It is a spatial disease process.
A primary tumor can extend through the bowel wall, enter lymphatic channels, reach the liver through the portal circulation, disseminate through the peritoneum, invade adjacent pelvic organs, or develop distant systemic metastases.
The supplied case demonstrates the importance of multimodality imaging. CT provides the broad anatomical framework and quantitative tumor assessment. MRI clarifies soft-tissue relationships. PET/CT adds metabolic information. Ultrasound provides targeted assessment, particularly in the liver. Endoscopy establishes direct mucosal visualization and permits biopsy.
The modern radiologist must integrate these observations into a treatment-oriented interpretation.
The most useful question is no longer:
“What does the scan show?”
It is:
“What does the scan show, how certain are we, how does it relate to tumor biology, and what clinical decision could change because of it?”
That is also where AI becomes meaningful.
The future is unlikely to belong to systems that merely identify a tumor.
It will belong to systems capable of connecting image phenotype, molecular phenotype, treatment history, longitudinal change, and clinical context—while keeping the radiologist and multidisciplinary team responsible for the final clinical judgment.
25. Continue Learning
Recommended follow-up topics:
MRI of Rectal Cancer: A Practical Radiologist’s Guide to T and N Staging
CT Detection of Colorectal Liver Metastases
PET/CT in Recurrent and Metastatic Colorectal Cancer
Radiomics and AI Prediction of Colorectal Cancer Metastasis
Molecular Biomarkers and Precision Treatment in Metastatic Colorectal Cancer
References
C. Cremolini, M. Chalabi, E. Elez, et al., “Metastatic colorectal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up,” Annals of Oncology, vol. 37, no. 6, pp. 759–776, 2026, doi: 10.1016/j.annonc.2026.03.005.
A. Cervantes, R. Adam, S. Roselló, et al., “Metastatic colorectal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up,” Annals of Oncology, vol. 34, no. 1, pp. 10–32, 2023, doi: 10.1016/j.annonc.2022.10.003.
V. K. Morris, E. B. Kennedy, N. N. Baxter, et al., “Treatment of Metastatic Colorectal Cancer: ASCO Guideline,” Journal of Clinical Oncology, vol. 41, no. 3, pp. 678–700, 2023, doi: 10.1200/JCO.22.01690.
A. Cervantes and E. Martinelli, “Updated treatment recommendation for third-line treatment in advanced colorectal cancer from the ESMO Metastatic Colorectal Cancer Living Guideline,” Annals of Oncology, vol. 35, no. 2, pp. 241–243, 2024, doi: 10.1016/j.annonc.2023.10.129.
L. Candia, A. Cervantes, and E. Martinelli, “Updated treatment recommendations for third and further lines of treatment in advanced colorectal cancer,” Annals of Oncology, vol. 36, no. 3, pp. 342–344, 2025, doi: 10.1016/j.annonc.2024.11.015.
American College of Radiology, “ACR Appropriateness Criteria®: Staging and Disease Monitoring of Rectal Cancer,” 2025.
American College of Radiology, “ACR Appropriateness Criteria®: Staging and Disease Monitoring of Colon Cancer and Appendiceal Cancer,” 2025.
H. S. Ryu et al., “Colon cancer: the 2023 Korean clinical practice guidelines for diagnosis and treatment,” Annals of Coloproctology, vol. 40, no. 2, pp. 89–113, 2024, doi: 10.3393/ac.2024.00059.0008.
H. S. Ryu et al., “2023 Korean Multidisciplinary Guidelines for Colon Cancer Management: Summary of Radiological Points,” Korean Journal of Radiology, 2024, doi: 10.3348/kjr.2024.0575.
IARC/WHO, “Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries,” 2026.
A. S. Tejani et al., “Integrating and Adopting AI in the Radiology Workflow: A Primer for Standards and Integrating the Healthcare Enterprise (IHE) Profiles,” Radiology, vol. 311, no. 3, 2024, doi: 10.1148/radiol.232653.
“Foundation Models in Radiology: What, How, Why, and Why Not,” Radiology, 2025, doi: 10.1148/radiol.240597.
J. S. Ryu, H. Kang, Y. Chu, and S. Yang, “Vision-language foundation models for medical imaging: a review of current practices and innovations,” Biomedical Engineering Letters, vol. 15, pp. 809–830, 2025, doi: 10.1007/s13534-025-00484-6.
E. Abbaspour et al., “Application of radiomics for preoperative prediction of lymph node metastasis in colorectal cancer: a systematic review and meta-analysis,” International Journal of Surgery, 2024, doi: 10.1097/JS9.0000000000001239.
I. Seow-En et al., “Predictive modeling algorithms for liver metastasis in colorectal cancer: A systematic review of the current literature,” Annals of Hepato-Biliary-Pancreatic Surgery, vol. 28, no. 1, pp. 14–24, 2024, doi: 10.14701/ahbps.23-078.
“Prognostication of colorectal cancer liver metastasis by CE-based radiomics and machine learning,” Translational Oncology, vol. 47, 2024, art. no. 101997, doi: 10.1016/j.tranon.2024.101997.
I. Hartsock and G. Rasool, “Vision-language models for medical report generation and visual question answering: a review,” Frontiers in Artificial Intelligence, vol. 7, 2024, doi: 10.3389/frai.2024.1430984.
“Chatbots and Large Language Models in Radiology: A Practical Primer for Clinical and Research Applications,” Radiology, 2024, doi: 10.1148/radiol.232756.
Editorial Quality Note
This article deliberately separates documented case information from clinical interpretation. The supplied case documents a 73-year-old man, metastatic/contiguous colorectal disease, quantitative CT assessment, CT/MRI comparison, PET/CT, ultrasound, endoscopy, and abdominal radiography.
Patient-specific history, laboratory data, histopathology, complete TNM stage, and treatment response were not supplied and therefore have not been fabricated.
The epidemiologic figures in the original material were also updated against current IARC/WHO data, while imaging and treatment recommendations were cross-checked against current ACR, ASCO, Korean, and 2026 ESMO guidance.
Comments
Post a Comment