The clinical problem the review addresses is stark. Immune checkpoint inhibitors, the drugs that have revolutionized treatment of many advanced cancers, work spectacularly well in only a small subset of colorectal cancer patients. Tumors classified as microsatellite instability-high, which carry abundant mutations and are therefore highly visible to the immune system, respond robustly to pembrolizumab and similar agents, with five-year follow-up from the randomized phase III KEYNOTE-177 study confirming durable benefit over chemotherapy. But the vast majority of metastatic colorectal cancers are microsatellite-stable, meaning their mutation burden is low and their immune landscape is cold: few T cells infiltrate the tumor, and suppressive cells and signaling molecules actively keep the immune system at bay. For these patients, checkpoint blockade alone offers little, and the search for ways to convert cold tumors into hot ones has become one of the field’s most urgent priorities.

The authors frame their solution around what they call the tumor-immune-soil nexus, a conceptual model in which the tumor, the immune system, and the surrounding microenvironment form an interdependent ecosystem. Within this framework, Traditional Chinese Medicine formulas are proposed to function as formula-defined ecological conditioners, multi-target systems that simultaneously adjust several parameters of the ecosystem rather than hitting a single molecular switch. The blueprint rests on three actionable pillars: remodeling the gut microbiome to restore immune-supportive bacterial communities and metabolic flux, reprogramming the myeloid compartment away from immunosuppressive states, and inducing immunogenic cell death in tumor cells to generate the danger signals needed to ignite an immune response. Each pillar is grounded in a growing body of preclinical and early clinical evidence.

The first pillar, microbiome remodeling, is perhaps the most provocative. The gut harbors trillions of bacteria that shape systemic immunity, and disruption of this community has been linked to colorectal cancer development and to poor responses to immunotherapy. Several classical herbal formulas have now been shown in experimental systems to reshape gut microbial composition. Gegen Qinlian decoction, for example, was reported to enhance the effect of PD-1 blockade in microsatellite-stable colorectal cancer models by remodeling both the gut microbiota and the tumor microenvironment. Huang-Lian-Jie-Du decoction enhanced the efficacy of capecitabine and oxaliplatin through the bacterium Akkermansia muciniphila and CD8-positive T cells, while Shenling Baizhu powder has been implicated in potentiating immunotherapy response through gut microbial remodeling and fatty acid metabolism modulation. Herbal compounds can also repair the intestinal barrier, whose breakdown allows bacterial products and even live bacteria to disseminate and promote metastasis, particularly to the liver.

Metabolism provides the mechanistic bridge between microbes and immune cells. Bacterial metabolites such as short-chain fatty acids and bile acids profoundly influence T cell differentiation, including the balance between inflammatory Th17 cells and suppressive regulatory T cells. Secondary bile acids produced by gut bacteria have been linked to Western diet-associated colorectal cancer and to liver metastasis through altered neutrophil recruitment. The review argues that by steering microbial metabolism, herbal formulas could shift these metabolite pools toward configurations that favor anti-tumor immunity. The authors also invoke the classical concept of the Gan Pi axis, a traditional framework linking liver and spleen physiology, and propose translating it into modern immunometabolism, suggesting that ancient syndrome categories may map onto measurable metabolic and immune states that can be targeted and monitored.

The second pillar targets the myeloid compartment, the army of innate immune cells that in colorectal cancer too often works for the tumor rather than against it. Tumor-associated macrophages polarized toward the M2 phenotype and myeloid-derived suppressor cells both suppress cytotoxic T cells and correlate with poor outcomes. TGF-beta signaling builds what researchers have described as a dual immune barrier by impairing T cell recruitment and instructing immunosuppressive macrophages, while molecules such as MNDA promote immunosuppression by facilitating infiltration of polymorphonuclear myeloid-derived suppressor cells. Here again, herbal pharmacology offers candidate tools: Astragalus polysaccharide has been shown to induce macrophage polarization toward the pro-inflammatory M1 state via Notch signaling, curcumin has been reported to inhibit and redifferentiate myeloid-derived suppressor cells, and artesunate modulates macrophage inflammatory signaling through the TLR4 pathway. Senescent fibroblasts, which drive T cell dysfunction through CD36-mediated lipid transfer, represent another stromal target within this pillar.

The third pillar addresses the fundamental requirement that the immune system must first perceive cancer as a threat. Immunogenic cell death is a specialized form of tumor cell demise that exposes calreticulin on the cell surface and releases danger signals, allowing dendritic cells to capture tumor antigens and prime T cell responses. Several compounds derived from Chinese herbs have been shown to induce this process: cantharidin-loaded nanomedicines triggered immunogenic cell death to enhance PD-1 blockade in preclinical models, and triptolide induced immunogenic cell death through endoplasmic reticulum stress and redox modulation. The review also highlights the STING pathway, a DNA-sensing circuit that is frequently epigenetically silenced in colorectal carcinoma, constraining DNA damage responses and enabling immune escape. Reactivating such danger-sensing machinery, the authors argue, is a prerequisite for converting the immune desert of microsatellite-stable tumors into productive anti-tumor immunity.

The authors are notably candid about the weaknesses of the existing evidence base, and this honesty distinguishes the review from more promotional treatments of the topic. A significant portion of the human data on Traditional Chinese Medicine in colorectal cancer remains observational, cross-sectional, or case-control in design, which limits causal inference and makes it impossible to reconstruct the long-term temporal sequence linking immune and microbiome dynamics to cancer progression or treatment response. Herb-drug interactions add a further layer of complexity: compounds such as baicalein and baicalin can alter the activity of drug-metabolizing enzymes like CYP3A4 and the transporter p-glycoprotein, changing the pharmacokinetics of co-administered drugs, a lesson underscored by the classic demonstration that St. John’s wort accelerates irinotecan metabolism. Without chemical standardization of formulas and careful pharmacokinetic monitoring, combination strategies risk being confounded or even dangerous.

To move the field forward, the review proposes a concrete methodological agenda. Future trials should evaluate standardized, chemically characterized formulas within adaptive platform designs that incorporate biomarker-driven endpoints, an approach aligned with the FDA’s botanical drug development guidance and with the emerging infrastructure of precision oncology trials. Validation tools should include spatial transcriptomics to map whether candidate formulas genuinely convert cold tumor regions into immune-inflamed ones, and microbiome-humanized mouse models that allow patient-derived microbial communities to be tested in controlled experiments. Network pharmacology, artificial intelligence-driven synergy prediction, and multi-omics integration are proposed as engines for identifying which combinations of herbal constituents produce the desired ecological shifts, while organoid platforms derived from patients could screen for resistance mechanisms before trials begin.

If the blueprint succeeds, the implications would extend well beyond colorectal cancer, offering a template for rationally integrating multi-component natural products with immunotherapy across tumor types. The vision is not herbal medicine as an alternative to mainstream oncology but as a priming agent, one that conditions the tumor ecosystem so that checkpoint inhibitors can work in the patients who currently derive no benefit from them. Whether classical formulas can survive the rigor of standardized chemistry, adaptive trials, and mechanistic validation remains an open question, but the review makes a compelling case that the answer is worth pursuing with the full arsenal of modern cancer science.

Subject of Research: Integrating Traditional Chinese Medicine with immune checkpoint inhibitors to convert immunotherapy-resistant microsatellite-stable metastatic colorectal cancer into an immune-responsive state

Article Title: Beyond the barrier: Engineering the tumor-immune-soil nexus–a mechanistic blueprint for integrating Traditional Chinese Medicine with immunotherapy in metastatic colorectal cancer

Article References: Beyond the barrier: Engineering the tumor-immune-soil nexus–a mechanistic blueprint for integrating Traditional Chinese Medicine with immunotherapy in metastatic colorectal cancer. (n.d.). https://doi.org/10.1007/s12032-026-03397-1

Image Credits: AI Generated

DOI: 10.1007/s12032-026-03397-1

Keywords: metastatic colorectal cancer, Traditional Chinese Medicine, immune checkpoint inhibitors, gut microbiome, tumor microenvironment, immunotherapy resistance, microsatellite stability, immunogenic cell death, myeloid-derived suppressor cells, macrophage polarization, gut barrier, spatial transcriptomics

Adjuvant therapy, the chemotherapy given after surgical removal of a tumor to destroy any remaining cancer cells, has long been a blunt instrument in colon cancer. For decades, treatment intensity has been assigned largely on the basis of anatomical staging, the depth of tumor invasion through the bowel wall and the number of lymph nodes involved. Patients deemed high risk by these pathological criteria have typically received combination chemotherapy regimens, while those considered lower risk have received single-agent treatment or, in some cases, observation alone. This one-size-fits-many approach inevitably overtreats some patients, exposing them to toxicities such as peripheral neuropathy, diarrhea, fatigue and long-term nerve damage without any proven benefit, and undertreats others whose residual disease goes undetected until it reappears as metastatic recurrence.

Circulating tumor DNA offers a fundamentally different lens. Tumors continuously shed fragments of their DNA into the bloodstream, and highly sensitive assays can detect these fragments by looking for tumor-specific mutations or abnormal methylation patterns. In the context of colon cancer, the presence of detectable ctDNA after surgery is one of the most powerful predictors of eventual recurrence yet identified, outperforming conventional imaging and standard tumor markers. A positive postoperative ctDNA result indicates that microscopic disease almost certainly persists somewhere in the body, while a negative result suggests, with a high degree of confidence, that no measurable residual disease remains. The logical corollary is that patients with persistent ctDNA might benefit from more aggressive therapy, whereas those who clear their ctDNA might be spared unnecessary treatment.

Translating that logic into practice, however, has proven difficult. Most ctDNA studies to date have been observational, measuring the marker and correlating it with outcomes without letting it dictate treatment. The few interventional trials that have used ctDNA to assign therapy have generally done so in a static way, taking a single postoperative measurement and using it once. PEGASUS was designed to go further by making ctDNA a dynamic, longitudinal guide. In the trial, patients with microsatellite-stable, high-risk stage II or III colon cancer underwent serial blood testing, and the results were used to direct both the initial choice of adjuvant chemotherapy and subsequent decisions about whether to continue, de-escalate, or escalate treatment after the standard adjuvant period had ended.

The population studied was deliberately chosen to be clinically challenging. Microsatellite-stable tumors, which constitute the majority of colon cancers, do not respond to immune checkpoint inhibitors and carry a worse prognosis than their microsatellite-instability-high counterparts. High-risk stage II disease, defined by features such as bowel obstruction or perforation, poorly differentiated histology, lymphovascular invasion or inadequate lymph node sampling, sits in a particularly gray zone of practice, where the benefit of adding oxaliplatin to fluoropyrimidine chemotherapy is debated and treatment decisions vary widely between centers. Stage III disease, with its proven lymph node involvement, carries a substantial recurrence risk even after apparently complete resection. For all of these patients, the trade-off between the toxicity of intensive chemotherapy and the danger of undertreating occult disease is acute and personal.

The feasibility question that PEGASUS addressed is not trivial. A ctDNA-guided strategy demands that blood samples be drawn, processed and analyzed on a tight timeline so that results arrive in time to inform treatment decisions. It requires laboratories to maintain consistent assay sensitivity across repeated measurements for the same patient, because a test that flickers between positive and negative near its detection limit creates clinical ambiguity. It requires oncologists to accept a new kind of evidence, a molecular measurement, as the basis for withholding or adding cytotoxic drugs, and it requires patients to understand that a negative blood test does not guarantee cure but may justify a lighter treatment burden. Any one of these links could break the chain, and the trial was structured to determine whether the whole chain could hold together in practice.

De-escalation is the arm of the strategy with the greatest potential to change daily practice. If serial ctDNA testing shows that a patient has no detectable residual disease after surgery, or that ctDNA clears rapidly once chemotherapy begins, the rationale for continuing full-intensity combination chemotherapy weakens considerably. Oxaliplatin, the component of standard regimens most associated with permanent peripheral neuropathy, is the natural target of such de-escalation. Avoiding even a fraction of unnecessary oxaliplatin exposure across the large population of high-risk colon cancer survivors would represent a meaningful quality-of-life gain, sparing patients years of numbness, cold sensitivity and functional impairment, while reducing healthcare costs associated with managing chemotherapy toxicities.

Escalation, conversely, addresses the patients whom standard staging fails. A patient whose ctDNA remains detectable despite completing a full course of adjuvant chemotherapy is, by the logic of the biomarker, harboring active microscopic disease that current treatment has not eradicated. In conventional practice such a patient would simply be monitored, with treatment resuming only once relapse became visible on scans, often at a point when cure is no longer possible. A ctDNA-guided postadjuvant strategy opens a window in which additional or alternative therapy can be considered while the disease burden is still molecular rather than radiographic. Whether such escalation improves survival is a question that feasibility trials like PEGASUS set the stage for but cannot fully answer; establishing that the strategy can be delivered consistently is the necessary first step.

The significance of the trial extends beyond colon cancer. The concept of minimal residual disease guidance, in which molecular testing determines treatment intensity, is being pursued in breast cancer, pancreatic cancer, gastric cancer and a range of other tumor types, and the operational lessons from PEGASUS are likely to inform all of these efforts. The trial also contributes to a broader rethinking of how adjuvant therapy decisions are made, moving the field away from population averages and toward individualized risk. As ctDNA assays become more standardized and more affordable, the barrier to adopting such strategies shifts from technology to clinical governance: which laboratories, which thresholds, which timelines, and who bears responsibility for acting on a positive result.

What PEGASUS ultimately demonstrates, within the limits of a phase 2 design, is that the vision of blood-test-guided cancer treatment can be operationalized for one of the most common malignancies worldwide. Colon cancer affects well over a million people each year globally, and a substantial fraction undergo surgery with curative intent followed by adjuvant chemotherapy whose value in any individual case is uncertain. If serial ctDNA monitoring can reliably sort these patients into those who need maximal therapy and those who do not, the result would be a rare win on both fronts of oncology: better outcomes for patients with hidden residual disease and fewer toxic treatments for those already cured by surgery. The trial’s findings now form part of the growing evidence base that will determine whether molecular residual disease testing becomes a routine companion to the surgeon’s knife and the oncologist’s infusion chair, transforming follow-up after colon cancer surgery from a period of anxious waiting into a period of active, data-driven decision-making.

Subject of Research: Circulating tumor DNA-guided adjuvant and postadjuvant treatment in resected high-risk stage II and III microsatellite-stable colon cancer

Article Title: Circulating tumor DNA-guided de-escalation or escalation of adjuvant therapy in high-risk stage II and stage III colon cancer: the phase 2 PEGASUS trial

Article References: Marsoni, S., Montagut, C., Pietrantonio, F., Sartore-Bianchi, A., Lazzari, L., Bergamo, F., Zampino, M. G., Tarazona, N., Mandalà, M., Tamberi, S., Elez, E., Santos Vivas, C., Luraghi, P., Prisciandaro, M., Tosi, F., Ciardiello, D., Vidal, J., Seguì, V., Palazzo, M., … Lonardi, S. (2026). Circulating tumor DNA-guided de-escalation or escalation of adjuvant therapy in high-risk stage II and stage III colon cancer: the phase 2 PEGASUS trial. Nature Cancer. https://doi.org/10.1038/s43018-026-01237-9

Image Credits: AI Generated

DOI: 10.1038/s43018-026-01237-9

Keywords: circulating tumor DNA, colon cancer, adjuvant chemotherapy, PEGASUS trial, de-escalation, escalation, microsatellite stability, minimal residual disease, stage III colon cancer, precision oncology, oxaliplatin, biomarker-guided therapy