Clinically Explainable Prediction of Immunotherapy Response Integrating Radiomics and Clinico-Pathological Information in Non-Small Cell Lung Cancer †.

Simple Summary: Only 20% patients with non-small cell lung cancer respond to immunotherapy alone and 40% to immunotherapy in combination with chemotherapy. The PD-L1 value cutoff from immunohistochemistry that is used to select patients who would respond to immunotherapy lacks accuracy. A combinatio...

Descripción completa

Detalles Bibliográficos
Publicado en:Cancers Vol. 17; no. 16; pp. 2679 - 2700
Autores principales: Mitra, Jhimli, Ghose, Soumya, Thawani, Rajat
Formato: diagnostic images pictorial research tables/charts Journal Article
Publicado: MDPI Aug2025
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=187558568&site=ehost-live
header:
  @attributes:
    shortDbName: ccm
    uiTerm: 187558568
    longDbName: CINAHL Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    dissinfo:
    jinfo:
      jid:
        20726694
        B74B
      jtl: Cancers
      issn: 20726694
      maglogo: N
    pubinfo:
      dt: Aug2025
      vid: 17
      iid: 16
      pid: 97109
      pub: MDPI
    artinfo:
      ui:
        187558568
        187558568
        187558568
        10.3390/cancers17162679
        187558568
      ppf: 2679
      ppct: 21
      formats:
      tig:
        atl: Clinically Explainable Prediction of Immunotherapy Response Integrating Radiomics and Clinico-Pathological Information in Non-Small Cell Lung Cancer †.
      aug:
        au:
          Mitra, Jhimli
          Ghose, Soumya
          Thawani, Rajat
        affil: GE HealthCare, Niskayuna, NY 12309, USA
      sug:
        subj:
          Carcinoma, Non-Small-Cell Lung Pathology
          Carcinoma, Non-Small-Cell Lung Drug Therapy
          Immunotherapy
          Radiomics
          Machine Learning
          Prediction Models
          Natural Language Processing
          Random Forest
          Treatment Outcomes
          Human
          Male
          Female
          Adult
          Middle Age
          Aged
          Aged, 80 and Over
          Cancer Patients
          Retrospective Design
          Comparative Studies
          Descriptive Statistics
          Prospective Studies
          Tumor Markers, Biological
          Health Care Delivery, Integrated
          Adult: 19-44 years
          Middle Aged: 45-64 years
          Aged: 65+ years
          Aged, 80 & over
          Male
          Female
      ab: Simple Summary: Only 20% patients with non-small cell lung cancer respond to immunotherapy alone and 40% to immunotherapy in combination with chemotherapy. The PD-L1 value cutoff from immunohistochemistry that is used to select patients who would respond to immunotherapy lacks accuracy. A combination of other clinical biomarkers and radiomic biomarkers from CT should be analyzed for the selection of patients who would benefit from immunotherapy. The aim of our retrospective study was to develop a machine learning model that predicted treatment response from multimodal data (clinical, peritumoral and tumoral radiomics features). This combination of features outperfomed using radiomics or clinical variables alone. A large language model was further used to explain the response predictions in natural-language using the biomarkers that are comprehensible to clinicians. Background/Objectives: Immunotherapy is a viable therapeutic approach for non-small cell lung cancer (NSCLC). Despite the significant survival benefit of immune checkpoint inhibitors PD-1/PD-L1, on average; the objective response rate is around 20% as monotherapy and around 50% in combination with chemotherapy. While PD-L1 IHC is used as a predictive biomarker, its accuracy is subpar. Methods: In this work, we develop a machine learning (ML) method to predict response to immunotherapy in NSCLC from multimodal clinicopathological biomarkers, tumor and peritumoral radiomic biomarkers from CT images. We further learn a graph structure to understand the associations between biomarkers and treatment response. The graph is then used to create sentences with clinical hypotheses that are finally used in a Large Language Model (LLM) that explains the treatment response predicated on the biomarkers that are comprehensible to clinicians. From a retrospective study, a training dataset of NSCLC with n = 248 tumors from 140 subjects was used for feature selection, ML model training, learning the graph structure, and fine-tuning LLM. Results: An AUC = 0.83 was achieved for prediction of treatment response on a separate test dataset of n = 84 tumors from 47 subjects. Conclusions: Our study therefore not only improves the prediction of immunotherapy response in patients with NSCLC from multimodal data but also assists the clinicians in making clinically interpretable predictions by providing language-based explanations.
      pubtype: Academic Journal
      doctype:
        diagnostic images
        pictorial
        research
        tables/charts
        Journal Article
      ougenre: Article
    language: English
    refInfo:
    holdings:
      @attributes:
        islocal: N