Distinct Sarcoma Microenvironments Predict Benefit from Addition of Pembrolizumab to Preoperative Radiotherapy and Surgery in SU2C-SARC032

The SU2C-SARC032 trial demonstrates that adding pembrolizumab to preoperative radiotherapy improves disease-free survival for patients with high-risk UPS and LPS by modulating the tumor microenvironment, with distinct immune-cold and immune-hot sarcoma ecotypes both showing benefit through increased T cell activation and diversity.

The Gist

The Big Picture: A High-Stakes Game of Defense

Imagine a patient has a large, dangerous tumor (a sarcoma) in their arm or leg. The standard playbook for doctors is to shrink the tumor with radiation (like a heavy bombardment) and then cut it out with surgery.

However, for many patients, the cancer comes back later, often spreading to other parts of the body. To stop this, doctors tried adding a new weapon: Pembrolizumab. This is a type of immunotherapy that acts like a "coach" for the body's immune system, teaching it to recognize and attack the cancer.

In a major clinical trial called SU2C-SARC032, researchers found that adding this "coach" to the radiation and surgery plan helped patients live longer without the cancer returning. But there was a catch: The drug is expensive and has serious side effects. Not everyone benefited from it; for some, it was just extra stress with no reward.

The Question: How do we know exactly which patients need this extra "coach" and which ones can get by with just the standard radiation and surgery?

The Discovery: Two Different Types of "Fortresses"

The researchers looked closely at the "neighborhood" inside the tumors (called the Tumor Microenvironment). They realized that not all tumors are built the same way. They found two very different types of tumor "fortresses" that both benefited from the immunotherapy, but for completely opposite reasons.

1. The "Empty Desert" (Sarcoma Ecotype 1 / SE1)

• The Analogy: Imagine a fortress that is completely empty. There are no soldiers inside, no guards, and no spies. The cancer cells are just hanging out in a quiet, empty room. Because there are no immune cells there, the body's natural defenses can't find the cancer to fight it.
• What happened: When these patients got radiation plus the immunotherapy, the radiation acted like a loud alarm system. It blew up some cancer cells and released signals that woke up the immune system. The immunotherapy (the coach) then rushed in, recruited fresh soldiers (T-cells), and filled the empty desert with an army that started fighting the cancer.
• The Result: The "empty" tumor became a "hot" battlefield where the immune system finally showed up and won.

2. The "Exhausted Ghetto" (Sarcoma Immune Class E / SIC E)

• The Analogy: Imagine a fortress that is already packed with soldiers. There are tons of immune cells inside, trying to fight the cancer. But, the cancer has put up "Do Not Disturb" signs and has locked the soldiers in a state of exhaustion. The soldiers are there, but they are too tired or confused to do their job.
• What happened: These patients already had a strong immune presence. The radiation helped clear some obstacles, and the immunotherapy acted like caffeine for the exhausted soldiers. It removed the "Do Not Disturb" signs, woke the tired soldiers up, and gave them the energy to finish the job.
• The Result: The "exhausted" army was re-energized and successfully cleared out the cancer.

The "Goldilocks" Rule: Grade Matters

The researchers also noticed that the drug worked best on the most aggressive tumors (Grade 3). It's like saying the "coach" is most effective when the enemy is the strongest. If the tumor is less aggressive (Grade 2), the standard radiation might be enough, and adding the drug might just cause unnecessary side effects.

How They Figured This Out

The team didn't just guess; they used high-tech detective work:

• Digital Cytometry: They took a snapshot of the tumor's genetic code and used a computer program to count exactly what kind of cells were inside (like counting how many police, firefighters, and civilians are in a city).
• Single-Cell Atlas: They built a massive library of 65,000 individual cells from sarcoma patients to understand what a "healthy" immune response looks like versus a "broken" one.
• Blood Monitoring: They checked the patients' blood over time to see if the treatment was changing the immune system, though they found that looking at the tumor itself was much more revealing than looking at the blood.

The Bottom Line

This study is a game-changer because it moves us away from "one size fits all" medicine.

• Before: Doctors might give the expensive immunotherapy to everyone, hoping it works for some, while others suffer side effects for no reason.
• Now (and in the future): Doctors can look at a patient's tumor, check if it's an "Empty Desert" or an "Exhausted Ghetto," and check if it's aggressive (Grade 3).
  • If it fits one of these two profiles, they add the immunotherapy.
  • If it doesn't, they stick to the standard radiation and surgery.

In short: By understanding the unique "personality" of a tumor's neighborhood, doctors can now predict who will win the battle against sarcoma when given the right combination of weapons. This saves patients from unnecessary side effects and ensures the right people get the powerful help they need.

Technical Summary

1. Problem Statement

Soft tissue sarcomas (STS), specifically high-risk, localized undifferentiated pleomorphic sarcoma (UPS) and dedifferentiated/pleomorphic liposarcoma (LPS), have a high rate of metastatic recurrence despite standard care (preoperative radiotherapy [RT] followed by surgery). While immune checkpoint inhibitors (ICIs) like pembrolizumab have shown promise in metastatic settings, their efficacy in localized disease is variable. The SU2C-SARC032 phase 2 trial demonstrated that adding perioperative pembrolizumab to preoperative RT improved 2-year disease-free survival (DFS) from 52% to 67%, but this benefit was accompanied by increased toxicity (Grade 3+ adverse events rose from 31% to 56%) and was not uniform across all patients.

The Core Challenge: There is a critical lack of biomarkers to identify which specific patients with localized STS will derive a survival benefit from the combination of ICI and RT, versus those who will only experience toxicity without benefit. Current predictors like PD-L1 expression and tumor mutational burden (TMB) have failed to predict ICI response in STS.

2. Methodology

The authors performed a comprehensive multi-omics correlative analysis on samples from the SU2C-SARC032 trial (143 patients, 127 evaluable).

• Sample Collection:
  • Bulk RNA-seq: Performed on pre-treatment biopsies and post-treatment surgical resections (FFPE) for 119 patients.
  • Flow Cytometry: Performed on fresh post-treatment tumor resections (56 patients) to profile immune cell infiltration.
  • Cytometry by Time of Flight (CyTOF): Longitudinal profiling of peripheral blood mononuclear cells (PBMCs) at multiple time points (pre-treatment, during RT, post-surgery, 3/12 months).
• Computational & Analytical Frameworks:
  • Single-Cell Atlas Construction: Built a reference atlas of 65,786 single cells from UPS and LPS samples (using public data) to define 9 malignant, 7 mesenchymal, 17 lymphoid, and 12 myeloid cell states.
  • Digital Cytometry (Deconvolution): Used CIBERSORTx with the custom single-cell signature matrix to estimate cell state abundances in bulk RNA-seq samples.
  • TME Classification:
    • Sarcoma Immune Classes (SICs): Assigned patients to classes (A–E) based on bulk RNA-seq and MCP-Counter scores, focusing on SIC E (immune-hot, inflamed).
    • Sarcoma Ecotypes (SEs): Used EcoTyper to identify three transcriptional communities: SE1 (immune-cold, stromal-rich), SE2 (immune-hot, activated), and SE3 (immunosuppressive).
  • Repertoire Analysis: Reconstructed T-cell receptor (TCR) sequences from bulk RNA-seq using TRUST4 to assess clonality and diversity.
  • Statistical Analysis: Survival analysis (Kaplan-Meier, Cox proportional hazards) and differential gene expression (DESeq2) were used to correlate TME signatures with DFS and treatment response.

3. Key Contributions

• Identification of Dual Predictive Signatures: The study identifies two distinct, opposing tumor microenvironment (TME) states that predict benefit from pembrolizumab + RT:
  1. SIC E (Immune-Hot): Characterized by pre-existing inflammation, B-cells, tertiary lymphoid structures (TLS), and exhausted T-cells.
  2. SE1 (Immune-Cold): Characterized by a "desert" phenotype with low immune infiltration, high stromal/matrix remodeling, and pro-angiogenic features.
• Mechanistic Differentiation: The study elucidates that the mechanism of benefit differs based on the baseline TME:
  • In SIC E, pembrolizumab likely "reinvigorates" a pre-existing but dysfunctional (exhausted) anti-tumor response.
  • In SE1, the combination of RT and pembrolizumab induces a de novo immune response, converting an immune-cold tumor into an immune-hot state.
• Integration with Tumor Grade: The authors demonstrate that combining TME signatures with Grade 3 histology provides the most precise stratification for patient benefit.
• Longitudinal Systemic vs. Local Dynamics: The study highlights that while peripheral blood (PBMC) changes were modest, significant remodeling occurred within the tumor microenvironment, suggesting local tissue analysis is superior for monitoring response in localized STS.

4. Key Results

• Survival Outcomes by TME:
  • SIC E: Patients with SIC E tumors had better overall DFS. Within this group, those receiving pembrolizumab showed a trend toward improved DFS compared to RT alone (HR 0.37).
  • SE1: Surprisingly, patients with SE1 (immune-cold) tumors derived significant benefit from pembrolizumab (HR 0.39), whereas non-SE1 patients did not. This contradicts the general paradigm that immune-cold tumors do not respond to ICI alone, suggesting RT is essential to prime these tumors.
  • Combined Stratification: The most significant benefit was observed in patients with Grade 3 tumors that were either SIC E or SE1 (HR 0.26; 95% CI 0.09–0.8). Patients with Grade 2 or non-SIC E/SE1 tumors did not benefit.
• TME Remodeling (Post-Treatment):
  • SE1 Response: Pembrolizumab + RT led to a marked decrease in SE1 abundance and an increase in immune-hot SE2. There was a significant increase in CD8+ activated T cells, CD56low NK cells, and TCR diversity (Shannon entropy).
  • SIC E Response: Pembrolizumab + RT resulted in a reduction of exhausted PD-1+ T cells and an increase in effector memory T cells, suggesting reinvigoration.
  • Stromal Changes: The combination therapy reduced matrix-remodeling cancer-associated fibroblasts (CAFs) and ECM-remodeling sarcoma cells, which were abundant in baseline SE1.
• Peripheral Blood (CyTOF):
  • Changes in PBMC subsets were modest. There was a transient increase in memory T cells and a reduction in B cells during treatment, but no sustained, dramatic shifts in circulating immune profiles correlated with the deep intratumoral changes observed.
• TCR Repertoire: Only the combination arm (RT + Pembrolizumab) showed a significant post-treatment increase in TCR clonotype richness and diversity, indicating the recruitment and expansion of new T-cell clones.

5. Significance and Clinical Implications

• Precision Medicine in Sarcoma: This study provides a biologically grounded framework for patient selection. It suggests that immune-cold (SE1) sarcomas, previously considered refractory to immunotherapy, can be sensitized to ICI when combined with radiotherapy.
• Optimizing Risk/Benefit: By identifying that the benefit is concentrated in Grade 3 SIC E or SE1 patients, clinicians can potentially spare lower-risk (Grade 2) or non-responsive TME patients from the increased toxicity of adding pembrolizumab.
• Mechanistic Insight: The findings support a synergistic model where RT induces immunogenic cell death and inflammation (priming), which is then amplified by PD-1 blockade. This synergy is effective in both "re-igniting" hot tumors (SIC E) and "igniting" cold tumors (SE1).
• Future Directions: The authors note that these findings are hypothesis-generating and require validation in prospective trials. They mention the upcoming HARMONY phase 3 trial (CCTG), which aims to validate these biomarkers in a larger cohort of Grade 3 UPS/LPS patients.

In summary, this paper moves beyond simple "hot vs. cold" dichotomies, demonstrating that distinct biological states (SIC E and SE1) respond to the same combination therapy via different mechanisms, and that integrating these signatures with tumor grade is essential for optimizing the use of immunotherapy in localized sarcoma.