What does cancer look like?

Disclaimer

To be clear, there is NO SINGLE WAY to describe all cancer from any given standpoint, otherwise, radiology would have been solved by now ¹

It can take many FORMS, and here are just a few examples. So, claiming I can give you a wordy way to teach you HOW to spot cancer from medical images would be a HUGE OVERSTATEMENT. Instead, my goal here is to help you understand WHY cancer LOOKS that way, and this builds the intuition of HOW to build a system that spots cancer.

What Is Cancer?

Cancer is a group of diseases involving abnormal cell growth with the potential to invade or spread to other parts of the body. ² This simple description from Wikipedia actually contains three key characteristics:

1. Growth: Cells grow rapidly
2. Invasion: They invade surrounding organs
3. Metastasis: They move and replicate elsewhere

Your body becomes “corrupted” in several ways:

4. Consolidation: Tissue becomes hard due to rapid growth and fighting with surrounding tissue (desmoplasia)
5. Vascularization: Promotes blood vessel growth for nutrient supply
6. Lymphatic spread: Surrounding lymph nodes become infected

The Telltale Signs

These characteristics are reflected in medical imaging, though they may appear differently across various modalities, body regions, and patients.

1. Growth: Look for a conspicuous “object”. This is especially evident in follow-up scans (longitudinal imaging studies) where you can measure actual growth.
2. Invasion: Boundaries become obscured or spiculated .
3. Metastasis: Multiple occurrences beyond a single location.
4. Consolidation: This is subtle and varies A LOT across different modalities. The general theme is inhomogeneity and patterned calcifications .
5. Vascularization: Irregular vessel patterns or blood flow (in ultrasound, you can visually see blood flow into the foreign mass).
6. Lymph nodes: Look for surrounding lymph nodes that are accumulated, enlarged, or distorted in shape and texture.

Building the System

All these observations and knowledge pave the way for us to build a system that detects cancer effectively.

Multimodal Approach

A robust cancer detection system should combine multiple imaging modalities:

• Structural imaging (CT, MRI): Reveals growth patterns and invasion
• Functional imaging (PET, ultrasound): Shows metabolic activity and blood flow
• Longitudinal analysis: Tracks changes over time to detect growth

Detection Pipeline

1. Growth detection: Identify suspicious masses and measure size changes
2. Boundary analysis: Assess margin characteristics (smooth vs. spiculated)
3. Pattern recognition: Look for consolidation and calcification patterns
4. Vascular assessment: Analyze blood vessel patterns and flow
5. Lymph node evaluation: Check for regional spread indicators

Integration Strategy

The system should leverage anatomical priors and clinical context to improve accuracy while reducing false positives through structured reasoning about cancer characteristics.