The rising interest for precise characterization of the tumour immune contexture has recently brought forward the high potential of RNA sequencing (RNA-seq) in identifying molecular mechanisms engaged in the response to immunotherapy. In this review, we provide an overview of the major principles of single-cell and conventional (bulk) RNA-seq applied to onco-immunology. We describe standard preprocessing and statistical analyses of data obtained from such techniques and highlight some computational challenges relative to the sequencing of individual cells. We notably provide examples of gene expression analyses such as differential expression analysis, dimensionality reduction, clustering and enrichment analysis. Additionally, we used public data sets to exemplify how deconvolution algorithms can identify and quantify multiple immune subpopulations from either bulk or single-cell RNA-seq. We give examples of machine and deep learning models used to predict patient outcomes and treatment effect from high-dimensional data. Finally, we balance the strengths and weaknesses of single-cell and bulk RNA-seq regarding their applications in the clinic.
6 September 2024 in The Lancet Regional Health - Europe
In this Personal View, we address the latest advancements in automatic text analysis with artificial intelligence (AI) in medicine, with a focus on its implications in aiding treatment decisions in medical oncology. Acknowledging that a majority of hospital medical content is embedded in narrative format, natural language processing has become one of the most dynamic research fields for developing clinical decision support tools. In addition, large language models have recently reached unprecedented performance, notably when answering medical questions. Emerging applications include prognosis estimation, treatment recommendations, multidisciplinary tumor board recommendations and matching patients to recruiting clinical trials. Altogether, we advocate for a forward-looking approach in which the community efficiently initiates global prospective clinical evaluations of promising AI-based decision support systems. Such assessments will be essential to validate and evaluate potential biases, ensuring these innovations can be effectively and safely translated into practical tools for oncological practice. We are at a pivotal moment, where continued advancements in patient care must be pursued with scientific rigor.
Large language models (LLMs) are rapidly being integrated into health care, with substantial implications for oncology practice. The European Society for Medical Oncology (ESMO) developed the ESMO guidance on the use of Large Language Models in Clinical Practice (ELCAP) to provide a structured framework and basic guidance for their safe and effective application in oncology. Between November 2024 and February 2025, a multidisciplinary group of 20 experts convened under the ESMO Real World Data and Digital Health Task Force. Using literature review and a Delphi consensus process, the panel defined three categories of LLM use in oncology: type 1 (patient-facing applications), type 2 [health care professional (HCP)-facing applications], and type 3 (background institutional systems). Consensus statements were developed for each type to provide basic practical guidance. ELCAP highlights opportunities such as improved patient education and symptom management, streamlined clinical workflows, and enhanced data processing. At the same time, it addresses challenges including data privacy, algorithmic bias, regulatory compliance, and the risk of unsupervised use. The framework emphasises human oversight, protection of patient privacy, and alignment with clinical and ethical standards. Patient-facing tools should complement, not replace, professional advice and should be embedded in supervised care pathways. HCP-facing and background systems may improve efficiency and decision support but require systematic validation, transparency, and continuous monitoring. ELCAP provides a three-tier framework and basic practical guidance for LLM use in oncology. ESMO supports efforts to use this framework to improve patient care, but warns against unsupervised or unvalidated use.
Metastatic relapse after treatment is the leading cause of cancer mortality, and known resistance mechanisms are missing for most treatments administered to patients. To bridge this gap, we analyze a pan-cancer cohort (META-PRISM) of 1,031 refractory metastatic tumors profiled via whole-exome and transcriptome sequencing. META-PRISM tumors, particularly prostate, bladder, and pancreatic types, displayed the most transformed genomes compared with primary untreated tumors. Standard-of-care resistance biomarkers were identified only in lung and colon cancers—9.6% of META-PRISM tumors, indicating that too few resistance mechanisms have received clinical validation. In contrast, we verified the enrichment of multiple investigational and hypothetical resistance mechanisms in treated compared with nontreated patients, thereby confirming their putative role in treatment resistance. Additionally, we demonstrated that molecular markers improve 6-month survival prediction, particularly in patients with advanced breast cancer. Our analysis establishes the utility of the META-PRISM cohort for investigating resistance mechanisms and performing predictive analyses in cancer. Significance: This study highlights the paucity of standard-of-care markers that explain treatment resistance and the promise of investigational and hypothetical markers awaiting further validation. It also demonstrates the utility of molecular profiling in advanced-stage cancers, particularly breast cancer, to improve the survival prediction and assess eligibility to phase I clinical trials. This article is highlighted in the In This Issue feature, p. 1027