How accurately can the gut microbiome predict immunotherapy success?
The accuracy of microbiome-based prediction varies widely depending on the cancer type and how the prediction model is built. The largest and most comprehensive study here, analyzing fecal samples from 1,359 patients with four different cancers, found that a model using both bacterial and fungal biomarkers could predict immunotherapy response with an area under the curve (AUC) of 72–90%, depending on the cancer [1]. AUC is a measure of accuracy where 0.5 is a coin flip and 1.0 is perfect; so 0.72–0.90 means good to excellent prediction. For example, in renal cell cancer, the model reached 83–90% accuracy [1].
However, when the same approach was tested in a smaller meta-analysis of just four melanoma studies (130 patients total), the predictive model dropped to an AUC of only 0.625 (62.5% accuracy) when applied to a new group of 27 patients [8]. This drop highlights a key problem: models that work well on the data they were trained on often perform worse on new, unseen patients. Similarly, a study on colorectal cancer found that a model built on 63 patients achieved an AUC of 0.89 during training but fell to 0.72 when tested on an external group of 11 patients [9]. These numbers show that while the microbiome is a promising predictor, its accuracy is not yet reliable enough for routine clinical use across all patients.
Which specific gut microbes are linked to better immunotherapy outcomes?
No single microbe is universally predictive, but several bacteria consistently appear in studies of responders. The bacterium Faecalibacterium prausnitzii is one of the most frequently cited: it was enriched in responders across multiple cancer types in the large 2024 study [1], and a separate analysis of non-small cell lung cancer patients found it was significantly more abundant in those who responded to immunotherapy [7]. Another beneficial microbe is Akkermansia muciniphila, which was linked to better outcomes in gastric cancer [4] and was also identified as important for predicting response to CAR-T cell therapy in lymphoma patients [6].
On the other hand, some bacteria are consistently associated with poor response. For example, Hungatella hathewayi was depleted in responders across multiple cancer types [1], and higher levels of Veillonellaceae were linked to worse survival in hepatobiliary cancer patients on immunotherapy [3]. Importantly, the specific 'good' and 'bad' microbes can differ by cancer type: what works for melanoma may not apply to lung cancer [1][8]. This means that a personalized, cancer-specific approach is needed, not a one-size-fits-all microbial test.
Can we improve immunotherapy by changing the gut microbiome?
Yes, early evidence suggests that modifying the microbiome can boost immunotherapy response, but the research is still in its early stages. Fecal microbiota transplantation (FMT) — transferring stool from a healthy donor or a responder into a patient — has shown promise. In one study, patients with anti-PD-1 refractory metastatic melanoma who received FMT from responders experienced improved response rates without added toxicity [5]. Similarly, a mouse study found that FMT from human responders could delay tumor progression and enhance the effect of immunotherapy in lung cancer [7].
Other interventions like probiotics, prebiotics, and dietary changes are also being explored, but the evidence is less direct. A 2024 review noted that these strategies show promise in enhancing immunotherapy responses and reducing side effects, but emphasized that large randomized controlled trials are still needed to confirm their effectiveness and establish guidelines [2]. The key takeaway is that while the microbiome can be altered, we do not yet know the best way to do it for each patient or cancer type.
About These Sources
This answer is built on 9 peer-reviewed studies — published from 2021 to 2024, 4 from 2024 or later, 6 in Q1 journals, collectively cited 610 times — selected as the most relevant from 15 studies that passed quality screening, drawn from 81 papers retrieved from a database of over 500 million.
Sources used in this answer
Effects of gut microbiota on immune checkpoint inhibitors in multi-cancer and as microbial biomarkers for predicting therapeutic response
In the largest study here (1,359 patients, 4 cancers), a model using bacterial and fungal biomarkers predicted immunotherapy response with 72–90% accuracy, depending on the cancer type.
Gut microbiota as a biomarker and modulator of anti-tumor immunotherapy outcomes
A review highlights that specific gut microbes are linked to better immunotherapy outcomes across multiple cancers, but calls for large randomized trials to confirm predictive value.
Gut microbiome is associated with the clinical response to anti-PD-1 based immunotherapy in hepatobiliary cancers
In 65 hepatobiliary cancer patients, higher levels of Lachnospiraceae and Alistipes species were linked to longer survival on anti-PD-1 therapy, while Veillonellaceae was linked to worse outcomes.
Predictable regulation of gut microbiome in immunotherapeutic efficacy of gastric cancer
In 30 gastric cancer patients, a machine learning model using 6 bacterial species predicted immunotherapy response with an AUC of 0.941; Akkermansia muciniphila was a key predictor.
The Gut Microbiome and Cancer Immunotherapy: Can We Use the Gut Microbiome as a Predictive Biomarker for Clinical Response in Cancer Immunotherapy?
A review of 10 clinical studies found that gut microbiome profiles before immunotherapy were consistently related to progression-free and overall survival in melanoma, HCC, and NSCLC.
A non-antibiotic-disrupted gut microbiome is associated with clinical responses to CD19-CAR-T cell cancer immunotherapy
In 172 lymphoma patients receiving CAR-T therapy, a microbiome-based machine learning model trained in Germany and validated in the US robustly predicted long-term responders; Akkermansia was a key feature.
Gut microbiome affects the response to immunotherapy in non‐small cell lung cancer
In 71 NSCLC patients, higher gut microbiome diversity and Faecalibacterium abundance were linked to better immunotherapy response; fecal transplants from responders delayed tumor growth in mice.
Meta-analysis of the gut microbiota in predicting response to cancer immunotherapy in metastatic melanoma.
A meta-analysis of 4 melanoma studies (130 patients) found that a microbiome-based classifier predicted responders in an independent cohort with only 62.5% AUC, showing limited cross-cohort accuracy.
Abstract A030: Gut microbiome as a promising biomarker for colorectal cancer diagnosis and immunotherapy response prediction
In colorectal cancer, a model using gut microbiome data predicted immunotherapy response with an AUC of 0.72 in an external test set; Eubacterium hallii was a key feature.