Gut microbiome functional pathways outperform taxonomic profiles in predicting immune checkpoint inhibitor response in non-small cell lung cancer: an interpretable machine learning approach with SHAP - Report - MDSpire

Gut microbiome functional pathways outperform taxonomic profiles in predicting immune checkpoint inhibitor response in non-small cell lung cancer: an interpretable machine learning approach with SHAP

  • By

  • Feifei Wei

  • Yoshiro Nakahara

  • Junya Isobe

  • Yuka Igarashi

  • Haruhiro Saito

  • Shuji Murakami

  • Tetsuro Kondo

  • Hidetomo Himuro

  • Taku Kouro

  • Tomoya Matsui

  • Satoshi Wada

  • Takuya Tsunoda

  • Kiyoshi Yoshimura

  • Tetsuro Sasada

  • May 15, 2026

Share

Clinical Report: Functional Pathways of the Gut Microbiome in NSCLC

Overview

This study demonstrates that functional profiles of the gut microbiome are superior to taxonomic profiles in predicting responses to immune checkpoint inhibitors (ICIs) in non-small cell lung cancer (NSCLC). A specific signature of four metabolic pathways was identified as a strong predictor of treatment response.

Background

Non-small cell lung cancer (NSCLC) is the leading cause of cancer-related mortality worldwide, and immune checkpoint inhibitors (ICIs) have transformed its treatment. However, the variability in patient responses to ICIs remains a significant challenge. Understanding the role of the gut microbiome, particularly its functional characteristics, may provide insights into improving treatment outcomes.

Data Highlights

Feature SetCorrelation with Response
MetaCyc PathwaysStrongest
Taxonomic ProfilesWeaker

Key Findings

  • Functional profiles from MetaCyc pathways correlated significantly with RECIST-defined response.
  • A signature of four pathways was identified: urea cycle, adenosine nucleotide degradation, O-antigen biosynthesis, and L-glutamate degradation.
  • Pathways related to nitrogen metabolism and short-chain fatty acid biosynthesis were key factors in responder classification.
  • Taxonomic characteristics alone were less predictive of treatment outcomes compared to functional profiles.
  • Machine learning models based on functional data showed robust predictive capabilities.

Clinical Implications

The findings suggest that assessing gut microbiome functional profiles may enhance the ability to predict patient responses to ICIs in NSCLC. This approach could lead to more personalized treatment strategies, focusing on metabolic pathways as potential therapeutic targets.

Conclusion

Functional characteristics of the gut microbiome provide a more accurate prediction of ICI responses in NSCLC than taxonomic profiles. This highlights the potential for metabolic pathway signatures to inform precision medicine approaches in cancer treatment.

Related Resources & Content

  1. The ASCO Post, 2017 -- Gut Bacteria May Enhance, or Hamper, Response to Anti–PD-1 Agents
  2. The New Gastroenterologist, 2025 -- Insights into Colorectal Cancer Prognosis Through Computational Pathology Analysis
  3. The ASCO Post, 2026 -- Metastatic NSCLC: Deep Learning Pathomics Platform May Help Predict Response to Immunotherapy
  4. ASCO AI in Oncology, 2026 -- Multimodal Model Uses Pathology Data to Predict Immunotherapy Response in NSCLC
  5. PubMed, 2026 -- Therapy for Stage IV Non-Small Cell Lung Cancer Without Driver Alterations: ASCO Living Guideline
  6. PubMed, 2026 -- Gut microbial signatures and immunotherapy outcomes in NSCLC and melanoma: a systematic review and meta-analysis
  7. Frontiers, 2026 -- Gut Microbiome Functional Pathways Outperform Taxonomic Profiles in Predicting Immune Checkpoint Inhibitor Response in Non-Small Cell Lung Cancer
  8. Therapy for Stage IV Non-Small Cell Lung Cancer Without Driver Alterations: ASCO Living Guideline, 2026.3.0 - PubMed
  9. Gut microbial signatures and immunotherapy outcomes in NSCLC and melanoma: a systematic review and meta-analysis - PubMed
  10. Frontiers | Gut Microbiome Functional Pathways Outperform Taxonomic Profiles in Predicting Immune Checkpoint Inhibitor Response in Non-Small Cell Lung Cancer: An Interpretable Machine Learning Approach with SHAP

Original Source(s)

Gut microbiome functional pathways outperform taxonomic profiles in predicting immune checkpoint inhibitor response in non-small cell lung cancer: an interpretable machine learning approach with SHAP

  • by Feifei Wei, Yoshiro Nakahara, Junya Isobe, Yuka Igarashi, Haruhiro Saito, Shuji Murakami, Tetsuro Kondo, Hidetomo Himuro, Taku Kouro, Tomoya Matsui, Satoshi Wada, Takuya Tsunoda, Kiyoshi Yoshimura, Tetsuro Sasada

  • May 15, 2026

Related Content

Does the Timing of Immunotherapy Administration Affect Outcomes in Advanced Cancers?

A systematic review and meta-analysis found that early time-of-day immunotherapy was associated with improved survival outcomes in patients with advanced cancers, with the association appearing to be ...

Combination Enzalutamide and Radium-223 Extends Overall Survival in Bone-Dominant mCRPC

In the final analysis of the phase III EORTC 1333/PEACE-3 trial, with a median follow-up of 58 months, the addition of radium-223 to the androgen receptor inhibitor enzalutamide significantly prolonge...

Lack of Research in Understanding the Global Population of Childhood Cancer Survivors Leads to Gaps in Care and Cures

Each year, globally, about 400,000 children and adolescents aged between 0 and 19 years are diagnosed with cancer; over 100,000 die from the disease, with most of those cases, over 80%, and deaths occ...