Proteomics-Enhanced AI-Digital Pathology in Metastatic Mucinous Colorectal Carcinoma: A Case Report

This case report integrates AI-assisted digital pathology with spatial proteomics to characterize the architectural evolution, immune evasion mechanisms, and signaling pathway reprogramming driving therapy resistance in metastatic mucinous colorectal carcinoma, ultimately proposing a personalized therapeutic framework targeting mucin-associated barriers.

The Gist

The Big Picture: A Case of the "Slippery, Invisible" Cancer

Imagine a patient (let's call him "Mr. K") who has a very tricky type of colon cancer. This isn't your average tumor; it's a Mucinous Colorectal Carcinoma.

Think of a normal tumor like a dense brick wall. It's solid, and while it's hard to break down, you can usually see where the bricks are.

Mr. K's tumor, however, is like a giant, sticky blob of jelly. It is made mostly of mucus (a slimy substance). This "jelly" does two dangerous things:

1. It hides the cancer: The cancer cells float inside this jelly, making them hard for the body's security guards (immune cells) to find.
2. It blocks the medicine: When doctors try to send chemotherapy drugs to kill the cancer, the drugs get stuck in the jelly and can't reach the cells.

This paper is a detective story about how a team of scientists used AI (Artificial Intelligence) and Proteomics (studying the tiny proteins that make up the body) to figure out exactly how this "jelly cancer" works, how it changes over time, and how we might finally beat it.

---

The Story of Mr. K's Journey

Mr. K was a 58-year-old man. His story started in 2019 with a polyp in his colon. Even after surgery and chemotherapy, the cancer kept coming back.

• 2021: It spread to his belly lining (peritoneum).
• 2023: It moved to the ligaments near his liver.
• 2024: It had spread all over his abdomen and even into the skin under his belly button.

Despite multiple surgeries and a special "hot bath" of chemotherapy (HIPEC) to cook the cancer cells, the tumor kept regrowing. It was a battle of "whack-a-mole," but the mole was made of jelly.

The Detective Work: AI and Microscopes

The scientists didn't just look at the tissue under a normal microscope. They used a high-tech approach:

1. The Digital Eye (AI): They took photos of the entire tissue samples and used AI software (called QuPath) to act like a super-powered mapmaker.

   • The Analogy: Imagine a city map where the AI automatically colors the roads, the parks, the houses, and the "jelly" zones in different colors.
   • What they found: The AI showed that the immune cells (the body's police) were standing outside the jelly zones, looking in but unable to get in. The cancer cells were hiding safely inside the slime.

2. The Protein Fingerprint (Proteomics): They analyzed over 6,000 different proteins in the tumor samples.

   • The Analogy: If the tumor is a car, the proteins are the engine parts. By looking at the engine, the scientists could tell what the car was doing: Was it speeding up? Was it trying to hide? Was it running out of gas?

The Three Stages of Evolution

The most exciting part of the paper is how the cancer changed as it moved from the colon to the belly and then to the liver area. It was like the cancer was learning new tricks to survive.

• Stage 1: The "Sleepy" Primary Tumor (In the Colon)

  • What it was doing: It was busy building its "jelly fortress" (Mucin). It was turning off its immune alarms so the body wouldn't attack it. It was like a spy hiding in a bunker, conserving energy.
  • The Problem: The jelly was so thick that drugs couldn't get in.

• Stage 2: The "Angry" Belly Tumor (Peritoneal Metastasis)

  • What it was doing: When it moved to the belly, it changed. It started shouting (inflammation) and building bridges to new areas (angiogenesis). It was still hiding in jelly, but now it was actively fighting off the immune system and remodeling its surroundings.
  • The Shift: It became more aggressive and started using different energy sources.

• Stage 3: The "Super-Adapted" Liver Tumor (Hepatoduodenal Metastasis)

  • What it was doing: This was the most dangerous version. It had completely re-engineered its engine. It was now very good at processing fats and energy, allowing it to survive in a tough new environment (near the liver). It had turned off the "cell cycle" brakes and was ready to explode with growth.
  • The Lesson: The cancer didn't stay the same; it evolved into a different monster for every new location.

The "Aha!" Moment: The Jelly Barrier

The scientists realized that the mucin (jelly) wasn't just a passive byproduct; it was an active weapon.

• It physically blocks drugs.
• It creates a wall that immune cells cannot cross.
• It traps growth factors that help the cancer grow.

The Proposed Solution: A Multi-Pronged Attack

Since the cancer is so tricky, the authors suggest we can't just use one hammer. We need a whole toolbox:

1. Dissolve the Jelly (Mucolytics):

   • The Idea: Before attacking the cancer, we need to melt the jelly. The paper suggests using a mix of Bromelain (an enzyme found in pineapples) and N-acetylcysteine (a common supplement).
   • The Analogy: Think of it like pouring hot water on a block of ice. Once the ice melts, the hidden treasure (the cancer cells) is exposed, and the police (immune system) and the soldiers (drugs) can finally get in.

2. Cut the Engine (Targeted Drugs):

   • Once the jelly is gone, we use drugs that specifically target the "engine parts" the cancer is using to grow (like the PI3K/AKT/mTOR pathway).

3. Wake Up the Police (Immunotherapy):

   • Once the barrier is down, we can use immunotherapy to wake up the immune cells that were stuck outside the wall, helping them recognize and eat the cancer.

The Bottom Line

This paper tells us that one size does not fit all in cancer treatment.

• The cancer in the colon is different from the cancer in the belly, which is different from the cancer near the liver.
• The "jelly" (mucin) is the main reason this cancer is so hard to treat.
• By combining AI mapping, protein analysis, and a new strategy of melting the jelly first, we might finally be able to treat these aggressive, recurring cancers effectively.

It's a hopeful message: by understanding the enemy's secrets (its shape, its proteins, and its defenses), we can build a better plan to defeat it.

Technical Summary

1. Problem Statement

Mucinous colorectal carcinoma (MAC) is a distinct and aggressive histomorphological subtype of colorectal cancer (CRC), characterized by abundant extracellular mucin (>50% of tumor volume). MAC is clinically challenging due to:

• Therapeutic Resistance: It often exhibits poor response to standard chemotherapy and targeted therapies.
• Immune Evasion: The dense mucin barrier physically excludes immune cells (lymphocytes) from the tumor core, creating an "immune-cold" microenvironment.
• Heterogeneity: There is a lack of comprehensive understanding regarding the molecular and spatial evolution of MAC from primary tumors to metastatic sites, particularly how phenotypic plasticity drives progression and resistance.

2. Methodology

The study employed a spatial multi-omics approach integrating advanced digital pathology with deep proteomic profiling on a single patient case.

• Case Cohort: A 58-year-old male patient with metastatic MAC. Samples were collected longitudinally over four years (2019–2024) from:
  1. Primary sigmoid colon tumor (post-resection).
  2. Peritoneal metastasis (post-HIPEC and recurrence).
  3. Hepatoduodenal ligament metastasis.
  4. Abdominal wall/subcutaneous metastasis.
• Digital Pathology & AI:
  • Formalin-Fixed Paraffin-Embedded (FFPE) tissues were digitized at 40× resolution.
  • QuPath (v0.4.x) was used for whole-slide imaging (WSI) analysis.
  • Custom Groovy Scripts: The authors developed a suite of scripts (c2e_v2.groovy, subclassifier_v4.groovy, etc.) to automate annotation, extract quantitative morphometric data (areas of tumor, mucin, stroma), and ensure reproducibility across annotators.
  • Annotation Strategy: Regions were segmented into specific compartments: tumor epithelia, mucin pools, stroma, vasculature, and immune infiltrates.
• Proteomics (Mass Spectrometry):
  • Sample Prep: Whole tissue sections (WTS) underwent deparaffinization, SDS/TEAB extraction, reduction/alkylation, and S-Trap digestion.
  • Acquisition: Data-Independent Acquisition (DIA) using nanoLC–MS/MS (Q-Exactive HF-X).
  • Analysis: Data processed via DIA-NN against the UniProt human proteome (1% FDR).
  • Functional Enrichment: Gene Set Enrichment Analysis (GSEA) was performed on Hallmark and Reactome pathways to identify signaling shifts (e.g., EMT, immune response, metabolism).

3. Key Contributions

• Integration of AI and Proteomics: Demonstrated a workflow where AI-driven histological annotation directly correlates with spatial proteomic data, linking morphological phenotypes to molecular mechanisms.
• Phenotypic Classification: Defined five distinct epithelial phenotypes within the mucinous tumor continuum:
  1. Dysplastic Epithelial (Invasive elements from adenoma).
  2. Primitive Gland-Forming.
  3. Epithelial Clumps/Layers.
  4. Floating Individual Tumor Cells.
  5. Complex Glandular/Cribriform (predominant in late metastasis).
• Mucin Barrier Mechanism: Provided molecular evidence that the mucin barrier is not just a physical block but an active modulator of the tumor microenvironment (TME), suppressing antigen presentation and immune infiltration.
• Longitudinal Evolution: Mapped the molecular trajectory of the tumor from a proliferative, cell-cycle-driven primary state to an inflammation- and metabolism-adapted metastatic state.

4. Key Results

A. Histopathological Findings

• Primary Tumor: Characterized by heterogeneous glandular patterns and extensive mucin pools. Immune cells were strictly excluded from mucin-rich regions, confined to stromal rims.
• Metastatic Evolution:
  • The Dysplastic phenotype disappeared in metastatic recurrences.
  • The Complex Glandular/Cribriform phenotype emerged and became dominant in late-stage peritoneal and abdominal wall metastases.
  • Immune Exclusion: Consistent across all samples; lymphocytes were unable to penetrate the dense mucin matrix, regardless of the metastatic site.

B. Proteomic Signatures

• Primary Tumor:
  • Overexpression: MUC2 (dominant mucin).
  • Suppression: Antigen processing/presentation, interferon-stimulated genes, T-cell chemokines, and metabolic pathways (oxidative phosphorylation, fatty acid metabolism).
  • State: "Immune-cold," stress-adapted, and cell-cycle driven.
• Peritoneal Metastasis:
  • Shift: Transitioned to an inflamed stromal program.
  • Upregulation: Inflammatory pathways (NF-κB, IL-6/JAK/STAT3), EMT, angiogenesis, and complement/coagulation cascades.
  • Metabolism: Reduced metabolic tone compared to primary, but increased stress adaptation (mTORC1, UPR).
• Hepatoduodenal Metastasis:
  • Shift: High metabolic plasticity and anabolic flux.
  • Upregulation: Oxidative phosphorylation, fatty acid metabolism, bile acid metabolism, and cholesterol homeostasis (adaptation to liver niche).
  • Signaling: Strong activation of PI3K/AKT/mTOR and MYC pathways, alongside EMT and coagulation.
  • Late-Stage Shift: A transition away from MYC dominance toward an inflammation-centered, lipid-oxidative state with elevated KRAS and Interferon responses.

C. The Mucin Barrier

The study confirmed that MUC2, MUC5AC, and MUC4 create a dense extracellular matrix. This barrier:

1. Physically prevents lymphocyte infiltration.
2. Hinders drug penetration.
3. Sequesters growth factors and cytokines, sustaining proliferative signaling.

5. Significance and Clinical Implications

• Precision Oncology Framework: The study proposes a personalized therapeutic strategy based on the specific molecular and spatial profile of the metastasis.
• Proposed Therapeutic Strategy:
  1. Mucolytics: Use of Bromelain and N-acetylcysteine (BromAc) to degrade the mucin barrier, improving drug delivery and immune access.
  2. Targeted Inhibitors:
     • CDK4/6 inhibitors (for cell cycle/E2F signatures).
     • mTOR inhibitors or MYC-targeted therapies (for hepatoduodenal metastases).
     • JAK/STAT3 inhibitors (for peritoneal inflammation).
  3. Immunotherapy: Potential for checkpoint blockade once the mucin barrier is disrupted.
  4. Metabolic Modulators: Metformin/phenformin to counteract metabolic adaptation.
• Future Directions: The authors advocate for the routine integration of spatial multi-omics in CRC management to guide dynamic treatment adjustments, particularly for aggressive mucinous subtypes that resist standard care.

Conclusion: This case report establishes that the progression of mucinous CRC is driven by a dynamic interplay between phenotypic plasticity, mucin-mediated immune evasion, and metabolic reprogramming. Integrating AI-enhanced pathology with proteomics provides a roadmap for overcoming chemoresistance in these difficult-to-treat tumors.