Gut Bacteria Predict Melanoma Recurrence at Up to 94% Accuracy — NYU in Cell 2026

NYU Langone researchers analyzed gut bacteria from 674 patients enrolled in the CheckMate 915 melanoma trial and showed that matching patients by overall microbial fingerprint first lets gut bacteria predict post-surgery and post-immunotherapy recurrence with 83–94% accuracy depending on geographic region. The work, published in Cell on April 17, 2026, makes the case that single-marker microbiome biomarkers don’t generalize — but individualized fingerprints do.

Lead investigator Jiyoung Ahn, PhD, professor of population health at NYU Grossman School of Medicine, said the team grouped patients by overall microbial similarity regardless of location, then extracted recurrence-risk signatures within each group. The four taxa most associated with recurrence shifts were Eubacterium, Ruminococcus, Firmicutes, and Clostridium.

The matching method is the breakthrough

Why earlier work stalled:

• Markers derived in one region (e.g., North America) failed to transfer to European or Asian patients
• Diet, environment, and genetics fragment the microbiome geographically
• Single markers couldn’t carry a global predictor

The new approach:

• Sequence all bacterial taxa, compute overall pairwise similarity between patients
• Cluster patients with similar fingerprints first
• Extract risk signals within each cluster
• A North American–trained signature works on a European patient with a similar fingerprint

Accuracy:

• 83–94% across regions when fingerprint-matched
• Drops to 50–60% when applied naïvely

CheckMate 915 — the patient population

CheckMate 915 was a global phase 3 trial comparing nivolumab alone vs. nivolumab plus ipilimumab as adjuvant immunotherapy in stage 3–4 melanoma. Baseline stool samples were collected from 674 patients, with 1–2 years of post-surgery and post-immunotherapy recurrence follow-up. This study is a retrospective analysis of how well baseline gut bacteria predicted what happened next.

What the four key taxa mean

Eubacterium:

• Major butyrate producer (a short-chain fatty acid)
• Anti-inflammatory; protects gut epithelium
• Abundance generally a favorable signal

Ruminococcus:

• Ferments dietary fiber to butyrate
• Some species degrade mucin (a protective layer) — depends on the species
• Cuts both ways

Firmicutes and Clostridium:

• Broad phylum and genus
• Central to immune modulation and T-cell differentiation
• Direct link to checkpoint inhibitor response

The interpretation — ecosystem, not single bug:

• Resists “this microbe is good” reductionism
• Relative ratios within the patient’s whole fingerprint carry the signal
• The microbiome is the right unit for personalized medicine

Why the gut matters for melanoma specifically

Immunotherapy response:

• Melanoma is the cancer most responsive to checkpoint blockade (nivolumab, pembrolizumab)
• But only about 50% of patients sustain response
• A large portion of that variability traces to the gut — T cells are educated in the gut before they fight tumors

Earlier melanoma-microbiome data:

• 2018 Science: Akkermansia muciniphila abundance predicted better response
• 2021 Science: Fecal microbiota transplant converted non-responders to responders in some cases
• This Cell paper extends from “response” prediction to “recurrence” prediction

Limits and clinical timeline

Not a clinical tool yet:

• Retrospective. Prospective validation required
• Fingerprint classification isn’t standardized across labs
• A commercial diagnostic for clinicians is 5–10 years out

What it changes now:

• Trial design — patient stratification using microbial fingerprint
• FMT trial patient selection
• Peri-operative diet and antibiotic guidance (protecting the microbiome around surgery)

What this means if you don’t have melanoma

You don’t need a cancer diagnosis to take this seriously. The gut microbiome runs the immune system more than it runs digestion. Antibiotic overuse, low-fiber diets, and chronic stress flatten the microbiome — and the effects propagate to skin and systemic immunity.

Basics to protect microbial diversity:

• 25–38g fiber per day; aim for 30 different plants per week
• Regular fermented foods (kimchi, yogurt, kefir)
• Avoid unnecessary antibiotics
• Sleep, exercise, stress management — all feed back into the microbiome

Tetrapod’s editorial position

The gut-skin axis is no longer a hypothesis. It’s the variable that splits immunotherapy responders from non-responders in the clinic. For chronic inflammation and skin problems, looking only at topicals misses the larger upstream system. Commercial probiotics, however, have not yet shown the precision-matched effects this Cell paper describes. Fiber diversity and whole plant food variety remain the most reliable microbiome strategy.