Skip to main content
Download PDF

Emerging artificial intelligence-driven precision therapies in tumor drug resistance: recent advances, opportunities, and challenges

Molecular Cancer volume 24, Article number: 123 (2025) Cite this article

Abstract

Drug resistance is one of the main reasons for cancer treatment failure, leading to a rapid recurrence/disease progression of the cancer. Recently, artificial intelligence (AI) has empowered physicians to use its powerful data processing and pattern recognition capabilities to extract and mine valuable drug resistance information from large amounts of clinical or omics data, to study drug resistance mechanisms, to evaluate and predict drug resistance, and to develop innovative therapeutic strategies to reduce drug resistance. In this review, we proposed a feasible workflow for incorporating AI into tumor drug resistance research, highlighted current AI-driven tumor drug resistance applications, and discussed the opportunities and challenges encountered in the process. Based on a comprehensive literature analysis, we systematically summarized the role of AI in tumor drug resistance research, including drug development, resistance mechanism elucidation, drug sensitivity prediction, combination therapy optimization, resistance phenotype identification, and clinical biomarker discovery. With the continuous advancement of AI technology and rigorous validation of clinical data, AI models are expected to fuel the development of precision oncology by improving efficacy, guiding therapeutic decisions, and optimizing patient prognosis. In summary, by leveraging clinical and omics data, AI models are expected to pioneer new therapy strategies to mitigate tumor drug resistance, improve efficacy and patient survival, and provide novel perspectives and tools for oncology treatment.

Graphical Abstract

Introduction

Tumor drug resistance refers to the phenomenon of tumor cells evading the effects of anticancer drugs, leading to the failure of treatments such as chemotherapy, targeted therapy, or immunotherapy. Due to the influence of tumor burden, tumor heterogeneity, tumor microenvironment (TME), and other factors [1], the majority of traditional chemotherapy and radiotherapy fail to prevent the development of resistance during treatments. More seriously, current clinical methods for assessing tumor drug resistance have a significant lag effect, leading to poor therapeutic efficacy and serious toxic side effects for patients [2]. Notably, more than 90% of cancer-related deaths have been attributed to drug resistance [3]. Scientists and clinicians have long attempted to address this challenge from multiple dimensions and have developed a variety of methods to predict tumor drug resistance, including in vitro models [4], in vivo preclinical models [5], DNA sequencing technologies [6], immunohistochemistry [7], and liquid biopsies [8]. However, each of these methods has obvious limitations, including high workload, limited predictive accuracy, and difficulty in effectively utilizing the data. In particular, the massive amount of data generated by clinical omics, pathology, and imaging poses a great challenge for direct processing and analysis, thus hindering their effective application in tumor drug resistance practice.

For the large-scale and high-precision multimodal medical oncology data generated by the rapid development of high-throughput sequencing [9], mass spectrometry [10], radiology [11], and testing technologies [12], artificial intelligence (AI) technology has already shown great potential in integrating, analyzing and interpreting multisource tumor drug resistance data [13]. Indeed, by integrating multisource heterogeneous data, including omics data, medical images, and electronic medical records, AI can identify the key resistance features and construct more accurate and comprehensive diagnostic and prognostic models of tumor resistance [14] to facilitate cross-modal information fusion, ultimately guiding clinical precision oncology and personalized therapy. Machine learning (ML), a prominent subset of AI, relies on algorithms that learn from available data to construct models to perform specific tasks [15]. Furthermore, deep learning is a particularly adept form of ML at handling and processing massive data from genomics, transcriptomics, metabolomics, proteomics, and radiomics [16]. For instance, Rathore et al. [17] applied transfer learning using a convolutional neural network pre-trained on 1.2 million ImageNet images to extract resistance features from brain scans of 270 glioblastoma patients. This approach effectively mined resistance-related information linked to O6-methylguanine-DNA methyltransferase promoter methylation status (MGMTpms), achieving robust MGMTpms prediction with cross-validated accuracies of 86.95%, 81.56%, and 82.43% across three independent cohorts.

Artificial intelligence has the potential to significantly advance tumor resistance practice, offering promising avenues for resistance prediction and the development of precision oncology. Given the significance of AI in tumor drug resistance, this review highlights the applications of AI in basic study and clinical practice, mainly including guiding the development of drugs against tumor drug resistance, advancing drug resistance mechanisms discovery, driving drug sensitivity prediction, optimizing combination therapy, facilitating tumor resistant phenotype prediction, and accelerating biomarker discovery. Additionally, we provided a practical workflow of AI-guided tumor resistance practice, applications and discussed the perspectives and challenges associated with its use in tumor drug resistance practice. This review provides novel insights into tumor resistance practice and precision therapy, presents a useful reference for the practice of combating drug resistance in clinical tumors.

Proposed feasible workflow for artificial intelligence-driven tumor resistance practice

A streamlined and practical workflow is crucial to enhance the efficiency and accuracy of tumor drug resistance evaluation and prediction. By summarizing the extensive literature, we propose a feasible and practical workflow (Fig. 1):

Fig. 1
figure 1

A feasible workflow of artificial intelligence driven-tumor drug resistance model in basic studies and clinical practice

Full size image

Tumor drug resistance data collection

Tumor drug resistance-related data collection represents the initial step in the AI-driven workflow, and the acquisition of high-quality data is essential for advancing drug resistance practice. Available clinical data include patient demographic and clinical information [18,19,20], genomic data [21, 22], transcriptomic data [23,24,25], metabolomic data [2], proteomic data [26, 27], imaging data [28,29,30,31], and physiological or biochemical pathology test results [26, 28, 29].

Preprocessing of tumor drug resistance data

Recently, massive tumor data in modern medicine have been rapidly increased and accumulated [32]. Multimodal information, including electronic health records, imaging reports, and genomic data, comprehensively covers the diagnosis and treatment process of cancer patients [33]. However, these data are scattered across different origins and systems, often containing missing values and outliers [34], and remain heterogeneous, posing significant challenges to data integration, analysis and utilize [35]. Therefore, tumor drug resistance-related data must undergo comprehensive preprocessing before being used to train AI models. This process includes several critical steps, such as coding of medical concepts, data cleaning, data standardization and normalization, and feature selection [36].

Tumor drug resistance modeling

Appropriate AI algorithms should be employed to develop diagnostic or predictive models tailored to the specific needs of tumor drug resistance trials. These models should be adept at discerning and interpreting the underlying correlations and patterns within the resistance data. Commonly used drug resistance data mining methods mainly include support vector machines (SVM), random forest (RF), logistic regression (LR), and deep learning [37]. deep learning model HECTOR was established for predicting distant recurrence risk in endometrial cancer, which extracted oncological pathology features from H&E-stained whole-slide images using a Vision Transformer, then integrated these features with image-based molecular classification and anatomical staging through a gating-based attention mechanism to generate prognostic predictions for tumors [29].

Tumor drug resistance model training and validation

Model training and validation are essential steps in applying AI to tumor drug resistance practice, ensuring that the models achieve optimal performance on the training datasets and exhibit robust generalization to unseen and unknown data. Typically, the tumor drug resistance datasets have been partitioned into a training set (often 80% or 70% of the total data) and a validation set (commonly 20% or 30%) [38]. The training set is utilized to train the model, while the validation set is employed to evaluate its performance. Commonly used validation methods include cross-validation, leave-one-out cross-validation, and k-fold cross-validation [39]. This approach can enhance the accuracy and generalization of a model, making it applicable to both basic and clinical studies on tumor drug resistance [40]. Ahn et al. [41] developed a pathology image-based deep learning classifier, PathoRiCH, to predict the response to platinum-based chemotherapy for high-grade serous ovarian cancer, employing pathology images from the SEV cohort for training and initial validation, and then utilizing images from the TCGA and SMC cohorts to further evaluate the generalization of the model.

Interpretation of tumor drug resistance results

The predictive output of AI models requires effective communication with healthcare professionals to ensure understanding and facilitate adoption [42]. To achieve this, it is essential to present model results in an interpretable manner that allows clinicians to understand the basis of the drug resistance prediction. Interpretable machine learning models have emerged as a key tool to address this challenge [43]. Specifically, the calculation of SHAP values can elucidate the biological characteristics or clinical factors with significant impact on tumor drug resistance. Guo et al. [44] constructed a more interpretable prediction model for distant metastasis in ovarian clear cell carcinoma using six different machine learning techniques, and the primary tumor stage (T) was identified as a critical clinical factor influencing metastasis risk through SHAP analysis, which also correlated with drug resistance development.

Validation of tumor drug resistance models in experimental and clinical studies

Once models screen for potential biomarkers or predict tumor resistance, these results must be further validated by molecular biology [45], cell biology [46], and cohort studies [23]. For instance, Cai et al. [47] utilized six machine learning algorithms and successfully identified the core gene RAC3, which is significantly associated with chemoresistance and immune infiltration characteristics of bladder cancer (BCa); Immunohistochemistry (IHC) staining, RT-qPCR, and Western blot were applied to validate the expression of RAC3 in BCa tumor tissues, which successfully provided potential markers for evaluating BCa resistance and addressed a critical gap in the risk assessment of BCa patients.

Application and continuous optimization of tumor drug resistance models

Initial resistance models are difficult to translate directly into clinical practice [48]. Continuous data collection and model optimization are critical to ensure the accuracy of drug resistance prediction.

Artificial intelligence facilitates discovery of tumor resistance mechanism

With the development and application of artificial intelligence, its enormous potential in biomedical and clinical fields is being continuously explored. Specifically, in basic research on tumor drug resistance, AI can (1) identify new effective drugs against tumor resistance via facilitating the design and screening of novel drugs, predicting drug-target interactions [45], and identifying potential targets [49]; (2) elucidate the complex molecular mechanisms underlying tumor resistance in tumor cells through large-scale omics data analysis [50]; (3) construct drug sensitivity prediction models to assist clinicians in assessing the cytotoxicity of various drugs on tumor cells [51]; and (4) optimize drug combination strategies by analyzing the interactions between multiple antitumor agents to mitigate the resistance of monotherapy (Table 1) [52].

Table 1 Application of artificial intelligence in basic research on tumor drug resistance
Full size table

Artificial intelligence guides anticancer drug development to overcome tumor resistance

AI technologies hold tremendous promise for accelerating drug discovery and development. By leveraging machine learning and deep learning algorithms to analyze large biological and chemical datasets, AI can identify key biomarkers and molecular pathways associated with specific diseases or drug mechanisms of action [45], accelerating the drug discovery process [54], and facilitate rapid screening of potential drug candidates from large numbers of chemical compounds [53]. It is also capable of predicting the biological activity and safety of drug molecules [75], thus increasing the success rate of drug discovery (Fig. 2a).

Fig. 2
figure 2

Artificial intelligence in basic research on tumor drug resistance. a AI can facilitate the design and screening of drugs against tumor drug resistance by predicting molecular properties, screening effective lead compounds from libraries, predicting drug-target interactions, and identifying potential targets. b AI can help elucidate the complex molecular mechanisms underlying drug resistance in tumor cells. c AI can construct drug sensitivity prediction models to assess the inhibitory effects of various drugs on tumor cells. d AI can optimize drug combinations and explore combination strategies by analyzing the interactions between multiple antitumor agents

Full size image
  1. (1)

    AI has facilitated antitumor drug discovery and design: AI can be applied to the design of potentially effective anticancer drugs by constructing chemoinformatics or pharmacoinformatics models to predict the properties (cytotoxicity, safety, metabolicity) of molecules. An advanced deep learning framework POLYGON was constructed via integration of variational autoencoder, reinforcement learning, and random forest regression models to embed chemical spaces and iteratively generate novel molecular structures. Among 32 generated lead compounds targeting MEK1 and mTOR, most significantly inhibited their activity at 1–10 µM and reduced tumor cell viability in vitro experimental validation [53]. Additionally, AI can facilitate the screening of potential antitumor drugs from large compound libraries. For instance, Wen et al. [54] developed an end-to-end deep learning framework combining a self-supervised graph neural network with a Transformer architecture. Fine-tuned on the BindingDB database, it screened 50 candidate clusters from 4,527,000 compounds, with further homogeneous time-resolved fluorescence assays identifying clusters exhibiting IC50 < 200 nM, accelerating cyclin-dependent kinase 12 inhibitor (CDK12i) discovery. Furthermore, AI models have been instrumental in predicting drug-target interactions and assessing drug cytotoxicity, selectivity, and risk profiles of drugs. The BipotentR model, a computational tool integrating linear mixed models and feed-forward neural networks, identified 38 immune-metabolic bifunctional regulators using single-cell data. Integration of experimental, bioinformatics and clinical validation demonstrated that regulator knockdown enhanced metabolic gene suppression and T-cell killing efficacy, with an Area Under the Curve (AUC) of 0.603 [45].

  2. (2)

    AI has promoted the identification and screening of tumor drug resistance targets: AI can assist in transforming known or potential resistance-related genes or proteins into novel therapeutic targets to overcome resistance to existing therapies. For instance, Xiao et al. [75] trained ridge regression models using drug response data from GDSC and transcriptome data to predict drug sensitivity, validated across multiple colorectal cancer cohorts. Two BCL-XL inhibitors, navitoclax and WEHI-539, were identified and demonstrated the sensitivity towards high-chromosomal instability- colorectal cancer cells in vitro pharmacodynamic screening, thereby confirming CIN as a potential therapeutic target for colorectal cancer. Zhang et al. [59] employed a Bayesian model to integrate multi-omics data for predicting patient response to immune checkpoint inhibitors (ICI). They identified a stemness signature (Stem.Sig) negatively correlated with anti-tumor immunity and validated 20 stemness-related genes with immuno-resistance properties through CRISPR screening, suggesting their potential as immunotherapy.

Artificial intelligence advances molecular mechanisms underlying tumor drug resistance

Although the advent of targeted therapies and immunotherapy has significantly improved the survival rates of patients with advanced cancer, tumor resistance remains a major challenge in clinical cancer treatment. Emerging strategies such as genetic testing, liquid biopsy using circulating tumor DNA (ctDNA) technology [8], and single-cell sequencing [76] have elucidated the complex molecular mechanisms underlying tumor drug resistance. However, these approaches have generated massive amounts of data [56], which are challenging to accurately analyze and interpret using conventional statistical analysis methods.

AI can capture the complex nonlinear relationships inherent in tumor drug resistance data and extract characteristics of tumor resistance, including changes in the cell cycle, TME modes [64], modes of cell death [77], abnormal expression of tumor resistance-related proteins [60], gene regulation [66], and the mediation of various signaling pathways, providing valuable insights into the molecular mechanisms of tumor resistance from genomic and proteomic perspectives (Fig. 2b) [78]. Integrating a visible neural network (VNN) with hierarchical structures, backpropagation, AdamW optimizer, BatchNorm, and Dropout, an interpretable deep learning model NeST-VNN was established. Using the Cancer Multi-Protein Complexes Atlas (NeST) and data from GDSC and Cancer Therapeutics Response Portal (CTRP), it analyzed 718 resistance relevant genes with mutations, copy number alterations, and deletions to uncover drug-resistance mechanisms. Based on in vitro cell line screening, patient-derived xenograft (PDX) modeling, clinical data, and CRISPR validation, it revealed that histone regulatory complexes, mediated by KAT6A, TBL1XR1, and RUNX1, promote S-phase entry, driving resistance of cyclin-dependent kinase 4/6 inhibitor (CDK4/6i) [62]. Gerratana et al. [55] employed gradient boosting machines to analyze baseline ctDNA data from 610 hormone receptor-positive/HER2-negative metastatic breast cancer patients, identifying resistance mechanisms of CDK4/6i, including alterations in ER, RTK, and cell cycle pathways.

Artificial intelligence drives drug sensitivity prediction and screening

AI can assist in predicting individual drug responses by rapidly identifying and obtaining specific sets of genes or genetic traits, protein profiles, and metabolic characteristics that correlate with treatment outcomes. Numerous studies have established accurate, efficient, and intuitive drug sensitivity prediction platforms by combining tumor drug sensitivity assessment models [65] with AI algorithms for evaluating the cytotoxicity of various drugs [67], predicting the individual treatment responses, and providing a scientific basis for clinicians to design personalized and precise treatment programs (Fig. 2c).

  1. (1)

    Gene expression profiling: Gene expression profiling has become a widely used method for tumor drug sensitivity screening [79]. However, variations in sequencing depth across different technologies and laboratories, as well as batch effects and heterogeneity, have posed challenges for single-cell RNA sequencing (scRNA-seq) data analysis [80]. The integration of AI technology and gene expression profiling not only enhances data processing and model prediction accuracy but also facilitates a comprehensive understanding of the mechanisms underlying gene expression. TransCell, a two-step deep transfer learning framework integrating autoencoders, transfer learning, and deep feed-forward neural networks, was trained on pan-cancer tumor samples and validated using CellMinerCDB and DepMap 20Q1 data. It predicted drug sensitivities for 124 pediatric cell lines across 4686 drugs, identifying 29 broadly effective and cancer-specific agents [65].

  2. (2)

    Patient-derived xenograft (PDX) model: The PDX model is widely considered to more accurately reflect tumor heterogeneity, with efficacy evaluation results closely resembling those observed in clinical patients [81]. However, the use of the PDX models has been constrained by several factors: tumor implantation and expansion, as well as extensive in vivo drug sensitivity testing, which typically requires 10–15 months [82], and has hindered the application of these models in tumor resistance studies. AI can be subjected to assist in identifying factors that influence the success rate of PDX model establishment, such as tumor tissue processing methods and transplantation site selection, to optimize the establishment process [83]. Furthermore, AI can analyze genomic and transcriptomic data from PDX models to reduce time consumption. For instance, Cotler et al. [69] developed a platform to accelerate ovarian cancer drug sensitivity testing, predicted outcomes of intraperitoneal injections for three second-line cytotoxic therapies, achieving an average AUC of 0.91, based on machine learning classifier with linear regression and forward–backward stepwise feature selection,

  3. (3)

    Single-cell drug susceptibility testing: Single-cell drug susceptibility testing, a method developed in recent years, can evaluate the sensitivity of single cells to various antitumor drugs [84]. Cellular parameters obtained from single-cell cytotoxicity assays can be processed as input files and further analyzed by AI algorithms, providing deeper insights for clinical decision-making [85]. Additionally, integrating tumor sensitivity testing with AI to analyze various omics data can facilitate more personalized drug selection for cancer patients in the era of precision oncology. Based on RF, SVM, and k-nearest neighbors algorithms, the first single-cell transcriptome-based AI model, SCATTome, was developed predict individual cell responses to proteasome inhibitors, validated by in vitro pharmacodynamic screening, addressing the challenge of drug sensitivity heterogeneity in multiple myeloma [86].

  4. (4)

    Organoid models: The use of organoids in tumor drug resistance studies has gained significant attention recently [87]. The integration of AI-driven data analysis can optimize quality control processes and culture conditions, alleviating the financial and technical hurdles inherently associated with organoid culture [88], while simultaneously enhancing the efficiency and accuracy of drug sensitivity predictions. Kong et al. [67] integrated ridge regression with linear regression, support vector regression, and deep neural networks to train a model using transcriptomic and pharmacovigilance data from colorectal and bladder cancer organoids. Based on TCGA patient transcriptomic data, the model predicted drug responses for 114 colorectal and 77 bladder cancer patients, with survival analysis significantly supporting the predictions (P = 0.014 and P = 0.01, respectively).

Artificial intelligence assists in optimizing combination therapy development

AI can integrate data from multiple biomedical sources [70], organize drug combination datasets [72], predict drug combination sensitivity, and save valuable time in drug combination screening [89], thereby overcoming tumor resistance, addressing combinatorial explosion challenges, and enhancing cost-effectiveness in oncology drug screening (Fig. 2d). scTherapy, a machine learning model based on LightGBM, integrated single-cell transcriptome and drug response data to predict combination therapies for metastatic/refractory tumors. Following validation of flow cytometry and bulk cell viability assays in acute myeloid leukemia patient samples, 96% of predicted combinations demonstrated selective efficacy or synergy [90]. ComboPred combining RF, gradient boosting, and XGBoost, identified synergistic drug combinations with high selectivity against ovarian cancer. It prioritized candidates like the mTOR inhibitor vistusertib and BCL2L1 inhibitor A1155463 (HSA score 9.7) for single-cell validation, guiding preclinical oncology drug testing [72].

Meanwhile, AI models can analyze the synergistic mechanism of antitumor agents, thereby optimizing drug combinations, facilitating the identification of the optimal therapeutic regimen, and maximizing the benefits for oncology patients. For instance, Zhou et al. [73] developed prediction models using RF, XGBoost, and CatBoost to analyze drug combination datasets from DrugComb, DrugCombDB, and SYNERGxDB, and identified the combination of lapatinib (RTK inhibitor) and pazopanib (multi-kinase inhibitor) as potent against breast cancer by blocking the PI3K/AKT/mTOR pathway, validated by in vitro cytotoxicity screening towards MDA-MB-231 and HCC1937 cell. Davis et al. [71] trained an AI model combining neural networks and Bayesian network propagation on IC50 values and RNA sequencing (RNA-seq) data from six DLBCL cell lines, predicting strong synergy between histone deacetylase inhibitors and JAK inhibitors. Experimental validation confirmed significant synergistic effects of these combination therapy (p < 0.01), establishing a computational-experimental closed-loop framework for cancer combination therapy development.

Artificial intelligence assists reduce tumor drug resistance in clinical oncology

Tumor drug resistance poses a significant challenge in clinical oncology, making studies into this phenomenon essential for enhancing the effectiveness of cancer therapies [91], improving patient outcomes [92], reducing healthcare costs, and advancing medical science. Recently, AI has been increasingly applied to clinical trials of tumor drug resistance, offering promising solutions to address resistance issues. By integrating diverse sets, including genomic, transcriptomic, proteomic, imaging, and clinical data, machine learning models have been developed to predict patient responses to specific oncology drugs (Table 2) [93].

Table 2 Application of artificial intelligence in clinical research on tumor drug resistance
Full size table

Artificial intelligence facilitates the prediction of tumor-resistant phenotypes

Undeniably, AI is a powerful tool for processing and analyzing patient genomic data [22] and clinical information [113]. Its introduction could assist physicians in identifying patients at higher risk of developing tumor resistance. Tumor drug resistance is influenced by multiple factors, such as gene mutations [101] and alterations in gene expression [75], which contribute to the tumor's increased tolerance to therapeutic agents. AI can predict drug resistance phenotypes by systematically analyzing large-scale clinical information and genomic data (Fig. 3a). Furthermore, factors such as tumor patients' genetic background and mutations, weight, gender, age, and lifestyle habits may influence treatment outcomes [114], and AI contributed to clinicians systematically evaluating patients' drug resistance and predicting resistance-prone populations or cohorts [115].

Fig. 3
figure 3

Artificial intelligence-guided tumor drug clinical practice. a AI models predict patient responses to specific drugs and identify and differentiate individuals with distinct drug-resistant phenotypes. b AI models screen predictive and prognostic biomarkers associated with drug resistance

Full size image

AI can fully mine tumor resistance-related data from vast amounts of data in single-cell omics to guide tumor drug resistance. A drug response prediction model PERCEPTION using transfer learning and elastic net regularization, predicted responses of transcriptional clones within tumors, with the most resistant clone indicating overall patient response. It stratified responders and non-responders in multiple myeloma (AUC = 0.83) and breast cancer (AUC = 0.776) experiments [116]. Liu et al. [2] mined and analyzed HCT-116 colon cancer single-cell mass spectrometry metabolomic data, then used RF, artificial neural networks, and penalized LR to predict drug resistance in individual cells, achieving 86.5% accuracy in validation, highlighting its clinical potential. Goldstein et al. [94] developed an AI model with SVM, RF, and XGBoost, classify drug resistance and metastatic potential with > 95% accuracy using lung cancer cell features (size, granularity, fluorescence intensity).

Additionally, liquid biopsy samples can provide genomic, transcriptomic, epigenomic, proteomic, and metabolomic information about tumor resistance through omics analysis [8]. AI can efficiently integrate and process these multidimensional datasets, enhancing assay precision and supporting more informed diagnostic and therapeutic decisions. Based on plasma proteomics data of 184 non-small cell lung cancer patients, Shaked et al. used supervised learning, and identified resistance-associated proteins and combined key clinical parameters to predict response of ICB therapy. Survival analysis revealed risk ratios of 4.5 (CI 2.07–9.77; p < 0.0001) for overall survival and 2.27 (CI 1.7–4.03; p = 0.004) for progression-free survival, enabling patient stratification and dynamic monitoring [117].

Artificial intelligence accelerates discovery of tumor resistance biomarkers

In addition to tumor drug resistance phenotypes, there is an urgent clinical need for specific and highly sensitive biomarkers to monitor tumor progression and resistance [118]. By leveraging AI-based bioinformatics tools and computational biology models, the gene expression levels of proto-oncogenes and oncogenes [99], enzyme activity, and metabolic reprogramming [103] in patients' tumors can be thoroughly analyzed to screen for molecular markers that are closely related to tumor drug resistance, providing reliable evidence for the progression, prediction [109], and prognosis [111], while also providing scientific rationale for the development of novel drug resistance detection technologies and therapeutic strategies (Fig. 3b).

The discovery of predictive biomarkers is critical for stratifying patients into distinct susceptibility subtypes and enabling personalized treatment [119]. AI can assist physicians in identifying resistance genes [106] or proteins [107] through minimally invasive manipulation, thereby enhancing patient risk stratification and clinical trial selection. Based on tissue information normalization and deep learning with a fully connected neural network, TINDL was established and trained on RMA-normalized data from 958 GDSC cancer cell lines, and tested on TCGA primary tumor RNA-seq data. It identified key resistant genes (such as SLFN11, RPS6, RPL13) and pathways, validated by siRNA knockdown experiments [95]. Lee et al. [25] used LR and network propagation to analyze TME interactions, predicting ICI responses in melanoma, lung, bladder, and gastric cancers, and achieved a median AUC of 0.79 across 11 ICI cohorts, with single-cell experiments and enrichment analyses identifying resistance-associated pathways as potential combination therapy targets.

Prognostic biomarkers are valuable tools for monitoring cancer progression and treatment efficacy, allowing physicians to make timely adjustments to treatment regimens [120]. By integrating single-cell and multi-patient sample sequencing [100] and leveraging AI to explore the relationship between specific tumor cell subpopulations and patient prognosis, disease progression and outcome can be predicted, providing novel insights for clinical diagnosis and treatment of cancer [105]. An AI model named AE-SDN, combining autoencoders and deep neural networks, extracted key features from tumor RNA-seq data into a Cox regression layer to output patient risk scores and identify immune-, oncogenic-, and tumor suppressor-related genes. Compared to CD3+/CD8+ T-cell-density-based immune scores, AE-SDN improved predictive power by > 20% [111]. Guan et al. [103] used support vector machine-recursive feature elimination to screen PYGL, a prognostic metabolic gene, from 858 KEGG pathway genes and single-cell data. A xenograft model confirmed that PYGL knockdown inhibited tumor growth, supporting PYGL as a potential metabolic therapy target.

Available online tumor drug resistance databases or servers

Available, easy-to-use tumor resistance databases shorten the gap between basic research and clinical application, allowing physicians to quickly obtain therapeutic references and guidance based on laboratory data, contributing to precision oncology decision-making (Table 3) [121]. In addition to The Cancer Genome Atlas (TCGA) (https://www.cancer.gov/tcga) [122], based on the aggregation of drug sensitivity data from nearly 75,000 experiments, Yang et al. [123] developed the GDSC database (https://www.cancerrxgene.org/), which can identify molecular biomarkers of drug sensitivity by querying for specific anticancer drugs or cancer genes, facilitating the discovery of novel biomarkers for cancer therapy. The DRMref database (https://ccsm.uth.edu/DRMref/) [124], which analyzed tumor cell composition, intra-tumor heterogeneity, and epithelial-mesenchymal transition scores, provides a comprehensive characterization of drug resistance mechanisms and supports the development of drug combinations and innovative therapeutic targets. The Cancer Therapeutics Response Portal (CTRP) (http://portals.broadinstitute.org/ctrp/) [125], links genetic and cellular characteristics of 860 cancer cell lines to their sensitivity to 481 small molecule probes and drugs, accelerating the discovery of patient-matched therapies. ncRNADrug (http://www.jianglab.cn/ncRNADrug) [126] catalogs non-coding RNAs (ncRNAs) associated with drug resistance and targets, and predicts drug-ncRNA interactions to support drug development and cancer treatment. ScDrugAct (http://bio-bigdata.hrbmu.edu.cn/scDrugAct) [127] compiles 17,274 drug-related genes and 276,559 associations between over 10,000 drugs and 53 cell types, linking drugs, genes, and cells to support cell type-specific therapies and the identification of therapeutic biomarkers.

Table 3 Available databases on tumor drug resistance
Full size table

Current challenges and future perspectives

AI models outperform traditional methods in data integration, handling complex data, and adaptability, offering deeper insights into biomedical data for clinical decision-making and drug development [137, 138]. However, their effectiveness depends on high-quality data [139], and their "black-box" nature poses interpretability challenges, particularly with complex omics data and resistance mechanisms [61]. Therefore, standardized data management and acquirement is essential to ensure data quality and consistency [140], while efforts should prioritize enhancing model interpretability and visualization [141]. Moving forward, multimodal AI models should be utilized to integrate diverse data sources, emphasizing key interactions between oncological data modalities to boost predictive accuracy for resistance. Moreover, strengthening collaboration among computer scientists [142], biologists [143], clinicians, and pharmacologists [144] will be vital for translating research into practical applications and advancing precision oncology.

Quality and standardization of tumor drug resistance data

The scarcity of high-quality and suitable datasets is a major challenge for the application of AI algorithms to tumor drug resistance [145]. Currently, AI models based on in vitro cancer cell lines show limited translational potential in forecasting clinical drug responses in real-world scenarios [146]. Although many studies have moved towards AI models using clinical data, appropriate resistance datasets remain limited [139]. Even in large databases such as TCGA, clinical drug response data are typically sparse [95].

The integrity and comprehensiveness of clinical data are fundamental to clinical research and evidence-based decision-making, necessitating rigorous quality control and validation protocols [147]. Effective integration and processing of heterogeneous clinical data are critical for ensuring the reliability of AI models [148]. In particular, the National Cancer Institute Genomic Data Commons (GDC) dataset, which integrated data from multiple cancer genome programs, provided comprehensive clinical drug response data alongside multi-omics profiles. By processing patient-derived molecular data through standardized GDC workflows, researchers can easily achieve data normalization, thereby enhancing the quality of tumor resistance modeling [149].

Interpretable and transparent AI models urgently needed in clinical oncology

Tumor drug resistance prediction models based on AI algorithms are often considered “black box” models due to the difficulty of explaining how these models actually arrive at their decisions [43]. When faced with the challenge of balancing performance and interpretability, scientists often prioritize performance metrics such as accuracy, precision, and recall. However, healthcare decisions require weighing complex, sometimes conflicting data, and clinicians value the interpretability and practical applicability of tumor resistance models [150].

Enhancing the interpretability of models can assist physicians in gaining a deeper understanding of the molecular basis of drug resistance toward tumor and developing more effective oncotherapy accordingly [43], potentially improving the feasibility and practicability of drug resistance models in clinical [62]. Zhao et al. [61] developed a series of “visible” neural network (VNN) models that linked genetic alterations to drug responses, utilizing knowledge maps of biological components and functions to guide the internal architecture of the model. Unlike traditional “black box” neural networks, VNN predictions of biomedical outcomes could be mapped to changes in molecular mechanisms and pathways, thereby enhancing the interpretability of clinical decisions. Ogunleye et al. [139] constructed a patient-interpretable machine learning model, where expression levels of selected miRNAs were nonlinearly combined by the CART algorithm in a correlated manner, supporting the model's predictive outcomes.

Emerging multimodal artificial intelligence models for enhanced robustness and accuracy in tumor drug resistance

Clinical data sources are massive and diverse, encompassing a wide range of data types and variables, such as patient charts, hospital records, laboratory test results, radiological imaging, histologic and histopathologic analyses, genomic profiling, and electronic health records [110]. These sources contain structured data, including clinical tests [151], semi-structured data like patient questionnaires [152], and unstructured data, such as physician's medical records [153].

The integration of such multimodal data has significantly enhanced the robustness and accuracy of diagnostic or prognostic models, driving advancements in AI applications within clinical settings [154]. MOMLN, an advanced multimodal and multi-omics machine learning integration framework, has demonstrated exceptional predictive performance by incorporating comprehensive input data, including clinical characteristics, DNA mutation profiles, gene expression signatures, TME features, and molecular pathway information. This framework achieved a remarkable mean AUC of 0.989 in classifying drug response types among 147 breast cancer patients [97].

Concluding remarks

Artificial intelligence (AI), with its powerful data processing and analysis capabilities, has shown significant potential in both basic and clinical studies on tumor resistance. By analyzing clinical data and omics data, AI provides innovative perspectives and tools to understand the onset and progression of tumor drug resistance, driving advances in cancer prediction, treatment, and prognosis. Its successful application not only underscores the potential of AI in the medical field, but also points to new development directions for precision oncology.

However, the application of AI technology in tumor resistance practice still faces several major challenges, particularly the incompleteness and bias of medical data, model interpretability, and robustness. To tackle these recent challenges, it is essential to implement standardized protocols for data collection, integration, processing, analyzing, modelling and validation, and to focus on developing robust, interpretable AI systems. The adoption of emerging technologies, particularly multimodal AI models, can greatly advance tumor drug resistance research by enabling more effective synthesis and analysis of diverse data types. Clinical validation of AI models is also crucial to ensure their reliable application in real-world studies.

Despite these above challenges, AI is poised to play an increasingly pivotal role in mitigating tumor drug resistance in clinical practice as technology continues to advance and more comprehensive clinical data becomes available. In the near future, AI is expected to predict and combat tumor drug resistance with higher efficiency and precision and become an integral part of every stage of tumor screening strategy, patient management, and prognosis, thus realizing personalized treatment and precision oncology.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

AI:

Artificial Intelligence

AUC:

Area Under the Curve

BCa:

Bladder Cancer

CDK4/6i:

Cyclin-Dependent Kinase 4/6 Inhibitor

CDK12i:

Cyclin-Dependent Kinase 12 Inhibitor

CIN:

Chromosomal Instability

CRC:

Colorectal Cancer

ctDNA:

Circulating tumor DNA

CTRP:

Cancer Therapeutics Response Portal

GDSC:

Genomics of Drug Sensitivity in Cancer

ICB:

Immune Checkpoint Blockade

ICI:

Immune Checkpoint Inhibitors

PDX:

Patient-Derived Xenograft

RF:

Random Forest

RNA-seq:

RNA sequence

scRNA-seq:

Single-cell RNA sequencing

SVM:

Support Vector Machine

TCGA:

The Cancer Genome Atlas

TME:

Tumor Microenvironment

VNN:

Visible neural network

References

  1. Vasan N, Baselga J, Hyman DM. A view on drug resistance in cancer. Nature. 2019;575:299–309.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Liu R, Sun M, Zhang G, Lan Y, Yang Z. Towards early monitoring of chemotherapy-induced drug resistance based on single cell metabolomics: combining single-probe mass spectrometry with machine learning. Anal Chim Acta. 2019;1092:42–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Bukowski K, Kciuk M, Kontek R. Mechanisms of multidrug resistance in cancer chemotherapy. Int J Mol Sci. 2020;21:3233.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Rahmanian M, Seyfoori A, Ghasemi M, Shamsi M, Kolahchi AR, Modarres HP, Sanati-Nezhad A, Majidzadeh-A K. In-vitro tumor microenvironment models containing physical and biological barriers for modelling multidrug resistance mechanisms and multidrug delivery strategies. J Control Release. 2021;334:164–77.

    Article  CAS  PubMed  Google Scholar 

  5. Kang JH, Heo MH, Kim HK, Cho J, Kim Y, Lee H, Lee S-C, Park SH. Development of a novel patient-derived preclinical model from malignant effusions in patients with tyrosine kinase inhibitor resistant clear cell renal cell carcinoma. J Clin Oncol. 2017;35:445–445.

    Article  Google Scholar 

  6. Spalato Ceruso M, Toulmonde M, Blouin L, Italiano A. Circulating tumour DNA sequencing reveal mechanisms of resistance to BRAF-targeted therapies in BRAF-mutated gastrointestinal stromal tumour. Eur J Cancer. 2023;179:25–7.

    Article  CAS  PubMed  Google Scholar 

  7. Lin Q, Wu Z, Yue X, Yu X, Wang Z, Song X, Xu L, He Y, Ge Y, Tan S, et al. ZHX2 restricts hepatocellular carcinoma by suppressing stem cell-like traits through KDM2A-mediated H3K36 demethylation. EBioMedicine. 2020;53:102676.

    Article  PubMed  PubMed Central  Google Scholar 

  8. López López V, Serrano A, Fuentes A, Ferrer Lores B, Chaves JF, Solano C, Terol MJ. Liquid biopsy and lymphoma monitoring in clinical practice. Blood. 2021;138:4483–4483.

    Article  Google Scholar 

  9. Jia Q, Chu H, Jin Z, Long H, Zhu B. High-throughput single-cell sequencing in cancer research. Signal Transduct Target Ther. 2022;7:145.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Beekhof R, Bertotti A, Böttger F, Vurchio V, Cottino F, Zanella ER, Migliardi G, Viviani M, Grassi E, Lupo B, et al. Phosphoproteomics of patient-derived xenografts identifies targets and markers associated with sensitivity and resistance to EGFR blockade in colorectal cancer. Sci Transl Med. 2023;15:eabm3687.

    Article  CAS  PubMed  Google Scholar 

  11. Rudqvist N-P, Charpentier M, Lhuillier C, Wennerberg E, Spada S, Sheridan C, Zhou XK, Zhang T, Formenti SC, Sims JS, et al. Immunotherapy targeting different immune compartments in combination with radiation therapy induces regression of resistant tumors. Nat Commun. 2023;14:5146.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Bury D, Martin-Hirsch PL, Martin FL, Dawson TP. Are new technologies translatable to point-of-care testing? The Lancet. 2017;390:2765–6.

    Article  Google Scholar 

  13. Bueschbell B, Caniceiro AB, Suzano PMS, Machuqueiro M, Rosário-Ferreira N, Moreira IS. Network biology and artificial intelligence drive the understanding of the multidrug resistance phenotype in cancer. Drug Resist Updates. 2022;60:100811.

    Article  CAS  Google Scholar 

  14. Fraunhoffer N, Hammel P, Conroy T, Nicolle R, Bachet JB, Harlé A, Rebours V, Turpin A, Ben Abdelghani M, Mitry E, et al. Development and validation of AI-assisted transcriptomic signatures to personalize adjuvant chemotherapy in patients with pancreatic ductal adenocarcinoma. Ann Oncol. 2024;35:780–91.

    Article  CAS  PubMed  Google Scholar 

  15. Hajjo R, Sabbah DA, Bardaweel SK, Tropsha A. Identification of tumor-specific MRI biomarkers using Machine Learning (ML). Diagnostics. 2021;11:742.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Ballard Jenna L, Wang Z, Li W, Shen L, Long Q. Deep learning-based approaches for multi-omics data integration and analysis. BioData Mining. 2024;17:38.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Rathore S, Nasrallah M, Akbari H, Shukla G, Bagley S, Watt C, Min Ha S, Mamourian E, Sako C, Binder Z, et al. TMOD-40. in vivo evaluation of o6-methylguanine-dna-methyltransferase (mgmt) promoter methylation status for de novo glioblastoma patients using deep learning features. Neuro Oncol. 2019;21:vi271–2.

    Article  PubMed Central  Google Scholar 

  18. Salahi A, Honrado C, Moore J, Adair S, Bauer TW, Swami NS. Supervised learning on impedance cytometry data for label-free biophysical distinction of pancreatic cancer cells versus their associated fibroblasts under gemcitabine treatment. Biosens Bioelectron. 2023;231:115262.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Mu Q, Chai R, Yang Y, Jiang T, Wang J. EPCO-49. Myc amplification at diagnosis regulates therapy-induced hypermutation at recurrent glioma. Neuro Oncol. 2023;25:v135–v135.

    Article  PubMed Central  Google Scholar 

  20. Mu Q, Chai R, Pang B, Yang Y, Liu H, Zhao Z, Bao Z, Song D, Zhu Z, Yan M, et al. Identifying predictors of glioma evolution from longitudinal sequencing. Sci Transl Med. 2023;15:eadh4181.

    Article  CAS  PubMed  Google Scholar 

  21. Huang H, Li Q, Tu X, Yu D, Zhou Y, Ma L, Wei K, Gao Y, Zhao G, Han R, et al. DNA hypomethylation patterns and their impact on the tumor microenvironment in colorectal cancer. Cell Oncol. 2024;47:1375–89.

    Article  CAS  Google Scholar 

  22. Thomas M, Theophanous S, Whittle I, Finetti M, Pollock S, Ahmed N, Tanner G, Chakrabarty A, Ismail A, Bulpitt A, Stead L. Predicting glioblastoma gene expression therapy response with machine learning. Neuro Oncol. 2023;25:iii13–4.

    Article  PubMed Central  Google Scholar 

  23. Suh Y-S, Lee J, George J, Seol D, Jeong K, Oh S-Y, Bang C, Jun Y, Kong S-H, Lee H-J, et al. RNA expression of 6 genes from metastatic mucosal gastric cancer serves as the global prognostic marker for gastric cancer with functional validation. Br J Cancer. 2024;130:1571–84.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Tang X, Thompson KJ, Kalari KR, Sinnwell JP, Suman VJ, Vedell PT, McLaughlin SA, Northfelt DW, Aspitia AM, Gray RJ, et al. Integration of multiomics data shows down regulation of mismatch repair and tubulin pathways in triple-negative chemotherapy-resistant breast tumors. Breast Cancer Res. 2023;25:57.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Lee J, Kim D, Kong J, Ha D, Kim I, Park M, Lee K, Im S-H, Kim S. Cell-cell communication network-based interpretable machine learning predicts cancer patient response to immune checkpoint inhibitors. Sci Adv. 2024;10:eadj0785.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Hu D, Shen X, Gao P, Mao T, Chen Y, Li X, Shen W, Zhuang Y, Ding J. Multi-omic profiling reveals potential biomarkers of hepatocellular carcinoma prognosis and therapy response among mitochondria-associated cell death genes in the context of 3P medicine. EPMA Journal. 2024;15:321–43.

    Article  PubMed  Google Scholar 

  27. Harel M, Lahav C, Jacob E, Issler E, Bar H, Dicker A, Sharon O, Bacchiocchi A, Halaban R, Sznol M, Shaked Y. 21 Plasma-based proteomic profiling as a tool for predicting response to immunotherapy in melanoma patients. J Immunother Cancer. 2020;8:A11–2.

    Google Scholar 

  28. Hwangbo S, Kim SI, Kim J-H, Eoh KJ, Lee C, Kim YT, Suh D-S, Park T, Song YS. Development of machine learning models to predict platinum sensitivity of high-grade serous ovarian carcinoma. Cancers. 1875;2021:13.

    Google Scholar 

  29. Volinsky-Fremond S, Horeweg N, Andani S, Barkey Wolf J, Lafarge MW, de Kroon CD, Ørtoft G, Høgdall E, Dijkstra J, Jobsen JJ, et al. Prediction of recurrence risk in endometrial cancer with multimodal deep learning. Nat Med. 2024;30:1962–73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Pour AF, Wu TC, Martinek J, Kumar S, Anzeneder T, Torkler S, Sommermeyer F, Palucka K, Chuang J: Abstract P1–02–01: Deep learning identifies morphological changes in whole slide images of treatment-resistant TNBC. Cancer Research 2022, 82:P1–02–01-P01–02–01.

  31. Barrera C, Corredor G, Viswanathan VS, Ding R, Toro P, Fu P, Buzzy C, Lu C, Velu P, Zens P, et al. Deep computational image analysis of immune cell niches reveals treatment-specific outcome associations in lung cancer. npj Precis Oncol. 2023;7:52.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Jiang P, Sinha S, Aldape K, Hannenhalli S, Sahinalp C, Ruppin E. Big data in basic and translational cancer research. Nat Rev Cancer. 2022;22:625–39.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Kann BH, Hosny A, Aerts HJWL. Artificial intelligence for clinical oncology. Cancer Cell. 2021;39:916–27.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Liu M, Li S, Yuan H, Ong MEH, Ning Y, Xie F, Saffari SE, Shang Y, Volovici V, Chakraborty B, Liu N. Handling missing values in healthcare data: a systematic review of deep learning-based imputation techniques. Artif Intell Med. 2023;142:102587.

    Article  PubMed  Google Scholar 

  35. Boehm KM, Khosravi P, Vanguri R, Gao J, Shah SP. Harnessing multimodal data integration to advance precision oncology. Nat Rev Cancer. 2022;22:114–26.

    Article  CAS  PubMed  Google Scholar 

  36. Albahra S, Gorbett T, Robertson S, D’Aleo G, Kumar SVS, Ockunzzi S, Lallo D, Hu B, Rashidi HH. Artificial intelligence and machine learning overview in pathology & laboratory medicine: a general review of data preprocessing and basic supervised concepts. Semin Diagn Pathol. 2023;40:71–87.

    Article  PubMed  Google Scholar 

  37. Firoozbakht F, Yousefi B, Schwikowski B. An overview of machine learning methods for monotherapy drug response prediction. Brief Bioinform. 2021;23:bbab408.

    Article  PubMed Central  Google Scholar 

  38. Yao Y, Lv Y, Tong L, Liang Y, Xi S, Ji B, Zhang G, Li L, Tian G, Tang M, et al. ICSDA: a multi-modal deep learning model to predict breast cancer recurrence and metastasis risk by integrating pathological, clinical and gene expression data. Brief Bioinform. 2022;23:bbab408.

    Article  Google Scholar 

  39. Álvez MB, Edfors F, von Feilitzen K, Zwahlen M, Mardinoglu A, Edqvist P-H, Sjöblom T, Lundin E, Rameika N, Enblad G, et al. Next generation pan-cancer blood proteome profiling using proximity extension assay. Nat Commun. 2023;14:4308.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Xia X, Zhu C, Zhong F, Liu L. TransCDR: a deep learning model for enhancing the generalizability of drug activity prediction through transfer learning and multimodal data fusion. BMC Biol. 2024;22:227.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Ahn B, Moon D, Kim HS, Lee C, Cho NH, Choi HK, Kim D, Lee JY, Nam EJ, Won D, et al. Histopathologic image–based deep learning classifier for predicting platinum-based treatment responses in high-grade serous ovarian cancer. Nat Commun. 2024;15:4253.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Gunning D, Stefik M, Choi J, Miller T, Stumpf S, Yang GZ. XAI—Explainable artificial intelligence. Sci Robot. 2019;4:eaay7120.

    Article  PubMed  Google Scholar 

  43. Tang S, Zhang H, Liang J, Tang S, Li L, Li Y, Xu Y, Wang D, Zhou Y. Prostate cancer treatment recommendation study based on machine learning and SHAP interpreter. Cancer Sci. 2024;115:3755–66.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Guo Q-H, Xie F-C, Zhong F-M, Wen W, Zhang X-R, Yu X-J, Wang X-L, Huang B, Li L-P, Wang X-Z. Application of interpretable machine learning algorithms to predict distant metastasis in ovarian clear cell carcinoma. Cancer Med. 2024;13:e7161.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Sahu A, Wang X, Munson P, Klomp JPG, Wang X, Gu SS, Han Y, Qian G, Nicol P, Zeng Z, et al. Discovery of targets for immune-metabolic antitumor drugs identifies estrogen-related receptor alpha. Cancer Discov. 2023;13:672–701.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Sui Q, Chen Z, Hu Z, Huang Y, Liang J, Bi G, Bian Y, Zhao M, Zhan C, Lin Z, et al. Cisplatin resistance-related multi-omics differences and the establishment of machine learning models. J Transl Med. 2022;20:171.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Cai T, Feng T, Li G, Wang J, Jin S, Ye D, Zhu Y. Deciphering the prognostic features of bladder cancer through gemcitabine resistance and immune-related gene analysis and identifying potential small molecular drug PIK-75. Cancer Cell Int. 2024;24:125.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Yujia X, Zhangsheng Y. Thorny but rosy: prosperities and difficulties in â□˜AI plus medicineâ□™ concerning data collection, model construction and clinical deployment. Gen Psychiatry. 2024;37:e101436.

    Article  Google Scholar 

  49. Kadioglu O, Efferth T. A machine learning-based prediction platform for P-glycoprotein modulators and its validation by molecular docking. Cells. 2019;8:1286.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Xie F, Zhang J, Wang J, Reuben A, Xu W, Yi X, Varn FS, Ye Y, Cheng J, Yu M, et al. Multifactorial deep learning reveals pan-cancer genomic tumor clusters with distinct immunogenomic landscape and response to immunotherapy. Clin Cancer Res. 2020;26:2908–20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Ren S, Cooper GF, Chen L, Lu X. An interpretable deep learning framework for genome-informed precision oncology. Nat Machine Intellig. 2024;6:864–75.

    Article  Google Scholar 

  52. Corridore S, Verreault M, Martin H, Delobel T, Carrère C, Idbaih A, Ballesta A: Circumventing glioblastoma resistance to temozolomide through optimal drug combinations designed by quantitative systems pharmacology and machine learning. bioRxiv 2024:2024.2005.2031.596811.

  53. Munson BP, Chen M, Bogosian A, Kreisberg JF, Licon K, Abagyan R, Kuenzi BM, Ideker T. De novo generation of multi-target compounds using deep generative chemistry. Nat Commun. 2024;15:3636.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Wen T, Li L, Li Y, Wang J, Gao P, Xie G, Ma F. A deep learning approach to discover cyclin-dependent kinases 12 (CDK12) inhibitors in breast cancer. J Clin Oncol. 2022;40:e15086–e15086.

    Article  Google Scholar 

  55. Gerratana L, Reduzzi C, Davis AA, Velimirovic M, Clifton K, Hensing WL, Shah AN, Dai CS, D’Amico P, Donahue J, et al. Defining resistance mechanisms to CDK4/6 inhibition in hormone receptor-positive HER2-negative metastatic breast cancer (MBC) through a machine learning approach applied to circulating tumor DNA (ctDNA). J Clin Oncol. 2022;40:3055–3055.

    Article  Google Scholar 

  56. Misek SA, Appleton KM, Dexheimer TS, Lisabeth EM, Lo RS, Larsen SD, Gallo KA, Neubig RR. Rho-mediated signaling promotes BRAF inhibitor resistance in de-differentiated melanoma cells. Oncogene. 2020;39:1466–83.

    Article  CAS  PubMed  Google Scholar 

  57. Chapple RH, Liu X, Natarajan S, Alexander MIM, Kim Y, Patel AG, LaFlamme CW, Pan M, Wright WC, Lee H-M, et al. An integrated single-cell RNA-seq map of human neuroblastoma tumors and preclinical models uncovers divergent mesenchymal-like gene expression programs. Genome Biol. 2024;25:161.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Wiecek AJ, Cutty SJ, Kornai D, Parreno-Centeno M, Gourmet LE, Tagliazucchi GM, Jacobson DH, Zhang P, Xiong L, Bond GL, et al. Genomic hallmarks and therapeutic implications of G0 cell cycle arrest in cancer. Genome Biol. 2023;24:128.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Zhang Z, Wang Z-X, Chen Y-X, Wu H-X, Yin L, Zhao Q, Luo H-Y, Zeng Z-L, Qiu M-Z, Xu R-H. Integrated analysis of single-cell and bulk RNA sequencing data reveals a pan-cancer stemness signature predicting immunotherapy response. Genome Med. 2022;14:45.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Yang ZY, Ye ZF, Xiao YJ, Hsieh CY, Zhang SY. SPLDExtraTrees: robust machine learning approach for predicting kinase inhibitor resistance. Brief Bioinform. 2022;23:bbac050.

    Article  PubMed  Google Scholar 

  61. Zhao X, Singhal A, Park S, Kong J, Bachelder R, Ideker T. Cancer mutations converge on a collection of protein assemblies to predict resistance to replication stress. Cancer Discov. 2024;14:508–23.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Park S, Silva E, Singhal A, Kelly MR, Licon K, Panagiotou I, Fogg C, Fong S, Lee JJY, Zhao X, et al. A deep learning model of tumor cell architecture elucidates response and resistance to CDK4/6 inhibitors. Nat Cancer. 2024;5:996–1009.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  63. Zhou Q-n, Lei R-e, Liang Y-x, Li S-q, Guo X-w, Hu B-l. Oxaliplatin related lncRNAs prognostic models predict the prognosis of patients given oxaliplatin-based chemotherapy. Cancer Cell Int. 2023;23:103.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  64. Chen D, Liu J, Zang L, Xiao T, Zhang X, Li Z, Zhu H, Gao W, Yu X. Integrated machine learning and bioinformatic analyses constructed a novel stemness-related classifier to predict prognosis and immunotherapy responses for hepatocellular carcinoma patients. Int J Biol Sci. 2022;18:360–73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  65. Yeh S-J, Paithankar S, Chen R, Xing J, Sun M, Liu K, Zhou J, Chen B: TransCell: In silico characterization of genomic landscape and cellular responses by deep transfer learning. Gen Proteo Bioinform 2024, 22.

  66. Nasimian A, Ahmed M, Hedenfalk I, Kazi JU. A deep tabular data learning model predicting cisplatin sensitivity identifies BCL2L1 dependency in cancer. Comput Struct Biotechnol J. 2023;21:956–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. Kong J, Lee H, Kim D, Han SK, Ha D, Shin K, Kim S. Network-based machine learning in colorectal and bladder organoid models predicts anti-cancer drug efficacy in patients. Nat Commun. 2020;11:5485.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. Zeng Z, Gu SS, Wong CJ, Yang L, Ouardaoui N, Li D, Zhang W, Brown M, Liu XS. Machine learning on syngeneic mouse tumor profiles to model clinical immunotherapy response. Sci Adv. 2022;8:eabm8564.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  69. Cotler MJ, Ramadi KB, Hou X, Christodoulopoulos E, Ahn S, Bashyam A, Ding H, Larson M, Oberg AL, Whittaker C, et al. Machine-learning aided in situ drug sensitivity screening predicts treatment outcomes in ovarian PDX tumors. Transl Oncol. 2022;21:101427.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Ji G, Liu J, Zhao Z, Lan J, Yang Y, Wang Z, Feng H, Ji K, Jiang X, Xia H, et al: Polyamine Anabolism Promotes Chemotherapy-Induced Breast Cancer Stem Cell Enrichment. Advanced Science, n/a:2404853.

  71. Davis N, McKinney MS, Reddy A, Love C, Smith E, Happ L, Dave S. Novel mechanisms for resistance to targeted therapy identified through machine learning approaches in 1167 RNA-seq drug exposure profiles in lymphoma. Blood. 2018;132:1370–1370.

    Article  Google Scholar 

  72. He L, Bulanova D, Oikkonen J, Häkkinen A, Zhang K, Zheng S, Wang W, Erkan EP, Carpén O, Joutsiniemi T, et al. Network-guided identification of cancer-selective combinatorial therapies in ovarian cancer. Brief Bioinform. 2021;22:bbab272.

    Article  PubMed  PubMed Central  Google Scholar 

  73. Zhou J-B, Tang D, He L, Lin S, Lei JH, Sun H, Xu X, Deng C-X. Machine learning model for anti-cancer drug combinations: analysis, prediction, and validation. Pharmacol Res. 2023;194:106830.

    Article  CAS  PubMed  Google Scholar 

  74. Lamballe F, Ahmad F, Vinik Y, Castellanet O, Daian F, Müller A-K, Köhler UA, Bailly A-L, Josselin E, Castellano R, et al. Modeling heterogeneity of triple-negative breast cancer uncovers a novel combinatorial treatment overcoming primary drug resistance. Adv Sci. 2021;8:2003049.

    Article  CAS  Google Scholar 

  75. Fang X, Yu WY, Zhu CM, Zhao N, Zhao W, Xie TT, Wei LJ, Sun XR, Xie J, Zhao Y. Chromosome instability functions as a potential therapeutic reference by enhancing chemosensitivity to BCL-XL inhibitors in colorectal carcinoma. Acta Pharmacol Sinic. 2024;45(11):2420–31.

    Article  CAS  Google Scholar 

  76. Asada K, Takasawa K, Machino H, Takahashi S, Shinkai N, Bolatkan A, Kobayashi K, Komatsu M, Kaneko S, Okamoto K, Hamamoto R. Single-cell analysis using machine learning techniques and its application to medical research. Biomedicines. 2021;9:1513.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  77. Zou Y, Xie J, Zheng S, Liu W, Tang Y, Tian W, Deng X, Wu L, Zhang Y, Wong C-W, et al. Leveraging diverse cell-death patterns to predict the prognosis and drug sensitivity of triple-negative breast cancer patients after surgery. Int J Surg. 2022;107:106936.

    Article  PubMed  Google Scholar 

  78. Gu H, Yin X, Peng T, Pan Y, Cui H, Li Z, Sun W, Ding B, Hu X, Zhang Z, Liu Z. Geographical origin identification and chemical markers screening of Chinese green tea using two-dimensional fingerprints technique coupled with multivariate chemometric methods. Food Control. 2022;135:108795.

    Article  CAS  Google Scholar 

  79. Ran D, Moharil J, Lu J, Gustafson H, Culm-Merdek K, Strand-Tibbitts K, Benjamin L, Navratil M. Platform comparison of HTG EdgeSeq and RNA-Seq for gene expression profiling of tumor tissue specimens. J Clin Oncol. 2020;38:3566–3566.

    Article  Google Scholar 

  80. Hafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. 2019;20:296.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  81. Garcia PL, Miller AL, Yoon KJ. Patient-derived xenograft models of pancreatic cancer: overview and comparison with other types of models. Cancers. 2020;12:1327.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  82. Liu Y, Wu W, Cai C, Zhang H, Shen H, Han Y. Patient-derived xenograft models in cancer therapy: technologies and applications. Signal Transduct Target Ther. 2023;8:160.

    Article  PubMed  PubMed Central  Google Scholar 

  83. Lee J, Lee G, Park HS, Jeong B-K, Gong G, Jeong JH, Lee HJ. Factors associated with engraftment success of patient-derived xenografts of breast cancer. Breast Cancer Res. 2024;26:49.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  84. Pellecchia S, Viscido G, Franchini M, Gambardella G. Predicting drug response from single-cell expression profiles of tumours. BMC Med. 2023;21:476.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  85. Yang C, Yang C, Yarden Y, To KKW, Fu L. The prospects of tumor chemosensitivity testing at the single-cell level. Drug Resist Updates. 2021;54:100741.

    Article  CAS  Google Scholar 

  86. Mitra AK, Mukherjee U, Harding T, Stessman H, Li Y, Jin J, Kumar SK, Rajkumar SV, Van Ness BG. Scattome: a single-cell analysis of targeted transcriptome program to predict drug sensitivity of single cells within human myeloma tumors. Blood. 2015;126:4249–4249.

    Article  Google Scholar 

  87. Zhao Z, Chen X, Dowbaj AM, Sljukic A, Bratlie K, Lin L, Fong ELS, Balachander GM, Chen Z, Soragni A, et al. Organoids. Nat Rev Methods Prim. 2022;2:94.

    Article  CAS  Google Scholar 

  88. Bai L, Wu Y, Li G, Zhang W, Zhang H, Su J. AI-enabled organoids: Construction, analysis, and application. Bioactive Mater. 2024;31:525–48.

    Article  Google Scholar 

  89. Güvenç Paltun B, Kaski S, Mamitsuka H. Machine learning approaches for drug combination therapies. Brief Bioinform. 2021;22:bbab293.

    Article  PubMed  PubMed Central  Google Scholar 

  90. Ianevski A, Nader K, Driva K, Senkowski W, Bulanova D, Moyano-Galceran L, Ruokoranta T, Kuusanmäki H, Ikonen N, Sergeev P, et al. Single-cell transcriptomes identify patient-tailored therapies for selective co-inhibition of cancer clones. Nat Commun. 2024;15:8579.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  91. Ianevski A, Lahtela J, Javarappa KK, Sergeev P, Ghimire BR, Gautam P, Vähä-Koskela M, Turunen L, Linnavirta N, Kuusanmäki H, et al. Patient-tailored design for selective co-inhibition of leukemic cell subpopulations. Sci Adv. 2021;7:eabe4038.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  92. Wang H, Wang R, Luo L, Hong J, Chen X, Shen K, Wang Y, Huang R, Wang Z. An exosome-based specific transcriptomic signature for profiling regulation patterns and modifying tumor immune microenvironment infiltration in triple-negative breast cancer. Front Immunol. 2023;14:1295558.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  93. Wang Z, Boretto M, Millen R, Natesh N, Reckzeh ES, Hsu C, Negrete M, Yao H, Quayle W, Heaton BE, et al. Rapid tissue prototyping with micro-organospheres. Stem Cell Reports. 2022;17:1959–75.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  94. Goldstein Y, Cohen OT, Wald O, Bavli D, Kaplan T, Benny O. Particle uptake in cancer cells can predict malignancy and drug resistance using machine learning. Sci Adv. 2024;10:eadj4370.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  95. Hostallero DE, Wei L, Wang L, Cairns J, Emad A. Preclinical-to-clinical anti-cancer drug response prediction and biomarker identification using TINDL. Gen Proteom Bioinform. 2023;21:535–50.

    Article  CAS  Google Scholar 

  96. Mitra AK, Mukherjee UK, Harding T, Jang JS, Stessman H, Li Y, Abyzov A, Jen J, Kumar S, Rajkumar V, Van Ness B. Single-cell analysis of targeted transcriptome predicts drug sensitivity of single cells within human myeloma tumors. Leukemia. 2016;30:1094–102.

    Article  CAS  PubMed  Google Scholar 

  97. Rashid MM, Selvarajoo K. Advancing drug-response prediction using multi-modal and -omics machine learning integration (MOMLIN): a case study on breast cancer clinical data. Brief Bioinform. 2024;25:bbae300.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  98. Che G, Yin J, Wang W, Luo Y, Chen Y, Yu X, Wang H, Liu X, Chen Z, Wang X, et al. Circumventing drug resistance in gastric cancer: a spatial multi-omics exploration of chemo and immuno-therapeutic response dynamics. Drug Resist Updates. 2024;74:101080.

    Article  CAS  Google Scholar 

  99. Blatti C, de la Fuente Js, Gao H, Marín-Goñi I, Chen Z, Zhao SD, Tan W, Weinshilboum R, Kalari KR, Wang L, Hernaez M: Bayesian machine learning enables identification of transcriptional network disruptions associated with drug-resistant prostate cancer. Cancer Res 2023, 83:1361-1380.

  100. Zhong F, Zhang X, Wang Z, Li X, Huang B, Kong G, Wang X. The therapeutic and biomarker significance of ferroptosis in chronic myeloid leukemia. Front Immunol. 2024;15:1402669.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  101. Shi Y, Li C, Zhang X, Peng C, Sun P, Zhang Q, Wu L, Ding Y, Xie D, Xu Z, Zhu W. D3EGFR: a webserver for deep learning-guided drug sensitivity prediction and drug response information retrieval for EGFR mutation-driven lung cancer. Brief Bioinform. 2024;25:bbae121.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  102. Zhou Y, Liu Z, Gong C, Zhang J, Zhao J, Zhang X, Liu X, Li B, Li R, Shi Z, et al. Targeting treatment resistance: unveiling the potential of RNA methylation regulators and TG-101,209 in pan-cancer neoadjuvant therapy. J Exp Clin Cancer Res. 2024;43:232.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  103. Guan J, Xu X, Qiu G, He C, Lu X, Wang K, Liu X, Li Y, Ling Z, Tang X, et al. Cellular hierarchy framework based on single-cell/multi-patient sample sequencing reveals metabolic biomarker PYGL as a therapeutic target for HNSCC. J Exp Clin Cancer Res. 2023;42:162.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  104. Cardoso MR, Silva AAR, Talarico MCR, Sanches PHG, Sforça ML, Rocco SA, Rezende LM, Quintero M, Costa TBBC, Viana LR, et al. Metabolomics by NMR combined with machine learning to predict neoadjuvant chemotherapy response for breast cancer. Cancers. 2022;14:5055.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  105. Chen Q, Gao F, Wu J, Zhang K, Du T, Chen Y, Cai R, Zhao D, Deng R, Tang J. Comprehensive pan-cancer analysis of mitochondrial outer membrane permeabilisation activity reveals positive immunomodulation and assists in identifying potential therapeutic targets for immunotherapy resistance. Clin Transl Med. 2024;14:e1735.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  106. Shi R, Sun J, Zhou H, Hu T, Gao Z, Wang X, Li M, Zhou Z, Shu Y. Hypoxia within tumor microenvironment characterizes distinct genomic patterns and aids molecular subtyping for guiding individualized immunotherapy. J Big Data. 2024;11:81.

    Article  Google Scholar 

  107. Qian L, Zhu J, Xue Z, Zhou Y, Xiang N, Xu H, Sun R, Gong W, Cai X, Sun L, et al. Proteomic landscape of epithelial ovarian cancer. Nat Commun. 2024;15:6462.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  108. Bonstingl L, Zinnegger M, Sallinger K, Pankratz K, Pritz E, Odar C, Skofler C, Ulz C, Oberauner-Wappis L, Borrás-Cherrier A, et al: Circulating tumor cell characterization and classification by novel combinatorial dual-color (CoDuCo) <em>in situ</em> hybridization and supervised machine learning. bioRxiv 2024:2024.2005.2008.592946.

  109. Wang Z, Wang Y, Yang T, Xing H, Wang Y, Gao L, Guo X, Xing B, Wang Y, Ma W. Machine learning revealed stemness features and a novel stemness-based classification with appealing implications in discriminating the prognosis, immunotherapy and temozolomide responses of 906 glioblastoma patients. Brief Bioinform. 2021;22:bbab032.

    Article  PubMed  PubMed Central  Google Scholar 

  110. Shaked Y, Harel M, Lahav C, Jacob E, Sela I, Yahalom G, Elon Y, Sharon O, Kamer I, Bar H, et al. Integration of proteomic and clinical data for the prediction of response to immune checkpoint inhibitor therapy in non-small cell lung cancer. J Clin Oncol. 2021;39:e21110–e21110.

    Article  Google Scholar 

  111. Saghand PG, Naqa IE, Tan AC, Xie M, Dai D, Chen JL, Ratan A, McCarter M, Carpten JD, Shah H, et al. A deep learning approach utilizing clinical and molecular data for identifying prognostic biomarkers in patients treated with immune checkpoint inhibitors: An ORIEN pan-cancer study. J Clin Oncol. 2022;40:2619–2619.

    Article  Google Scholar 

  112. Harel M, Lahav C, Jacob E, Dahan N, Sela I, Elon Y, Raveh Shoval S, Yahalom G, Kamer I, Zer A, et al. Longitudinal plasma proteomic profiling of patients with non-small cell lung cancer undergoing immune checkpoint blockade. J Immunother Cancer. 2022;10:e004582.

    Article  PubMed  PubMed Central  Google Scholar 

  113. Scanlon E, Lavery A, Stevenson L, Kennedy C, Byrne R, Walker A, Eatock M, Middleton M, Thomas A, Turkington R: P-OGC08 Translational Insights from the Dual ErbB Inhibition in Oesophago-gastric Cancer (DEBIOC) Clinical Trial - a bioinformatic analysis. British J Surg 2021, 108.

  114. Alnuhait MA, Shahbar AN, Alrumaih I, Alzahrani T, Alzahrani A, alanizi A, Alrashed MA, Elrggal M, Alhuthali A, Alsuhebany N: Advancing cancer care: How artificial intelligence is transforming oncology pharmacy. Inform Med Unlocked 2024, 50:101529.

  115. Lipkova J, Chen RJ, Chen B, Lu MY, Barbieri M, Shao D, Vaidya AJ, Chen C, Zhuang L, Williamson DFK, et al. Artificial intelligence for multimodal data integration in oncology. Cancer Cell. 2022;40:1095–110.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  116. Sinha S, Vegesna R, Mukherjee S, Kammula AV, Dhruba SR, Wu W, Kerr DL, Nair NU, Jones MG, Yosef N, et al. PERCEPTION predicts patient response and resistance to treatment using single-cell transcriptomics of their tumors. Nature Cancer. 2024;5:938–52.

    Article  PubMed  Google Scholar 

  117. Shaked Y, Harel M, Lahav C, Yellini B, Tepper E, Wolf I, Harkovsky T, Leibowitz R, Gottfried M, Abu-Amana M, et al. Personalized approach for response prediction and treatment management for non-small cell lung cancer patients based on a liquid biopsy. J Clin Oncol. 2022;40:e21132–e21132.

    Article  Google Scholar 

  118. Zhou Y, Tao L, Qiu J, Xu J, Yang X, Zhang Y, Tian X, Guan X, Cen X, Zhao Y. Tumor biomarkers for diagnosis, prognosis and targeted therapy. Signal Transduct Target Ther. 2024;9:132.

    Article  PubMed  PubMed Central  Google Scholar 

  119. Lou E, Vogel RI, Hoostal S, Klein M, Linden MA, Teoh D, Geller MA. Tumor-stroma proportion as a predictive biomarker of resistance to platinum-based chemotherapy in patients with ovarian cancer. JAMA Oncol. 2019;5:1222–4.

    Article  PubMed  PubMed Central  Google Scholar 

  120. Long L, Assaraf YG, Lei Z-N, Peng H, Yang L, Chen Z-S, Ren S. Genetic biomarkers of drug resistance: a compass of prognosis and targeted therapy in acute myeloid leukemia. Drug Resist Updates. 2020;52:100703.

    Article  Google Scholar 

  121. Aaltonen LA, Abascal F, Abeshouse A, Aburatani H, Adams DJ, Agrawal N, Ahn KS, Ahn S-M, Aikata H, Akbani R, et al. Pan-cancer analysis of whole genomes. Nature. 2020;578:82–93.

    Article  Google Scholar 

  122. McLendon R, Friedman A, Bigner D, Van Meir EG, Brat DJ, M. Mastrogianakis G, Olson JJ, Mikkelsen T, Lehman N, Aldape K, et al. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455:1061–8.

    Article  CAS  Google Scholar 

  123. Yang W, Soares J, Greninger P, Edelman EJ, Lightfoot H, Forbes S, Bindal N, Beare D, Smith JA, Thompson IR, et al. Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. Nucleic Acids Res. 2012;41:D955–61.

    Article  PubMed  PubMed Central  Google Scholar 

  124. Liu X, Yi J, Li T, Wen J, Huang K, Liu J, Wang G, Kim P, Song Q, Zhou X. DRMref: comprehensive reference map of drug resistance mechanisms in human cancer. Nucleic Acids Res. 2023;52:D1253–64.

    Article  PubMed Central  Google Scholar 

  125. Seashore-Ludlow B, Rees MG, Cheah JH, Cokol M, Price EV, Coletti ME, Jones V, Bodycombe NE, Soule CK, Gould J, et al. Harnessing connectivity in a large-scale small-molecule sensitivity dataset. Cancer Discov. 2015;5:1210–23.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  126. Cao X, Zhou X, Hou F, Huang YE, Yuan M, Long M, Chen S, Lei W, Zhu J, Chen J, et al. ncRNADrug: a database for validated and predicted ncRNAs associated with drug resistance and targeted by drugs. Nuc Acids Res. 2023;52:D1393–9.

    Article  Google Scholar 

  127. Xu Y, Zhang Y, Song K, Liu J, Zhao R, Zhang X, Pei L, Li M, Chen Z, Zhang C, et al. ScDrugAct: a comprehensive database to dissect tumor microenvironment cell heterogeneity contributing to drug action and resistance across human cancers. Nucleic Acids Res. 2024;53(D1):D1536–46.

    Article  PubMed Central  Google Scholar 

  128. Sun X, Zhang Y, Li H, Zhou Y, Shi S, Chen Z, He X, Zhang H, Li F, Yin J, et al. DRESIS: the first comprehensive landscape of drug resistance information. Nucleic Acids Res. 2023;51:D1263-d1275.

    Article  CAS  PubMed  Google Scholar 

  129. Jiang J, Ma Y, Yang L, Ma S, Yu Z, Ren X, Kong X, Zhang X, Li D, Liu Z. CTR-DB 2.0: an updated cancer clinical transcriptome resource, expanding primary drug resistance and newly adding acquired resistance datasets and enhancing the discovery and validation of predictive biomarkers. Nuc Acids Res. 2024;53(D1):D1335–47.

    Article  Google Scholar 

  130. Kumar R, Chaudhary K, Gupta S, Singh H, Kumar S, Gautam A, Kapoor P, Raghava GPS. CancerDR: cancer drug resistance database. Sci Rep. 2013;3:1445.

    Article  PubMed  PubMed Central  Google Scholar 

  131. Yang Z, Ye Z, Qiu J, Feng R, Li D, Hsieh C, Allcock J, Zhang S. A mutation-induced drug resistance database (MdrDB). Commun Chemist. 2023;6:123.

    Article  CAS  Google Scholar 

  132. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, Wilson CJ, Lehár J, Kryukov GV, Sonkin D, et al. The cancer cell line encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483:603–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  133. Whirl-Carrillo M, Huddart R, Gong L, Sangkuhl K, Thorn CF, Whaley R, Klein TE. An evidence-based framework for evaluating pharmacogenomics knowledge for personalized medicine. Clin Pharmacol Ther. 2021;110:563–72.

    Article  PubMed  PubMed Central  Google Scholar 

  134. Sondka Z, Dhir NB, Carvalho-Silva D, Jupe S, Madhumita, McLaren K, Starkey M, Ward S, Wilding J, Ahmed M, et al: COSMIC: a curated database of somatic variants and clinical data for cancer. Nucleic Acids Research 2023, 52:D1210-D1217.

  135. Suehnholz SP, Nissan MH, Zhang H, Kundra R, Nandakumar S, Lu C, Carrero S, Dhaneshwar A, Fernandez N, Xu BW, et al. Quantifying the expanding landscape of clinical actionability for patients with cancer. Cancer Discov. 2024;14:49–65.

    Article  CAS  PubMed  Google Scholar 

  136. Wang C, Yang J, Luo H, Wang K, Wang Y, Xiao Z-X, Tao X, Jiang H, Cai H. CancerTracer: a curated database for intrapatient tumor heterogeneity. Nucleic Acids Res. 2019;48:D797–806.

    PubMed Central  Google Scholar 

  137. Arango-Argoty G, Kipkogei E, Stewart R, Sun GJ, Patra A, Kagiampakis I, Jacob E. Pretrained transformers applied to clinical studies improve predictions of treatment efficacy and associated biomarkers. Nat Commun. 2025;16:2101.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  138. Zhang H, Yang F, Xu Y, Zhao S, Jiang YZ, Shao ZM, Xiao Y. Multimodal integration using a machine learning approach facilitates risk stratification in HR+/HER2− breast cancer. Cell Rep Med. 2025;6:101924.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  139. Ogunleye AZ, Piyawajanusorn C, Gonçalves A, Ghislat G, Ballester PJ. Interpretable machine learning models to predict the resistance of breast cancer patients to doxorubicin from their microRNA profiles. Adv Sci. 2022;9:2201501.

    Article  CAS  Google Scholar 

  140. Rivera D, Lee JJ, Royce ME, Kluetz PG. FDA oncology center of excellence landscape analysis of real-world data submissions for oncology drugs. J Clin Oncol. 2021;39:e18787–e18787.

    Article  Google Scholar 

  141. Tavolara TE, Su Z, Gurcan MN, Niazi MKK. One label is all you need: interpretable AI-enhanced histopathology for oncology. Semin Cancer Biol. 2023;97:70–85.

    Article  CAS  PubMed  Google Scholar 

  142. Allen GM, Lim WA. Rethinking cancer targeting strategies in the era of smart cell therapeutics. Nat Rev Cancer. 2022;22:693–702.

    Article  CAS  PubMed  Google Scholar 

  143. Tebon PJ, Wang B, Markowitz AL, Davarifar A, Tsai BL, Krawczuk P, Gonzalez AE, Sartini S, Murray GF, Nguyen HTL, et al. Drug screening at single-organoid resolution via bioprinting and interferometry. Nat Commun. 2023;14:3168.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  144. Chan JY. Special issue “cancer immunotherapy: tumor microenvironment, biomarker discovery and immune resistance.” Int J Mol Sci. 2024;25:5113.

    Article  PubMed  PubMed Central  Google Scholar 

  145. Viswanathan VS, Parmar V, Madabhushi A. Towards equitable AI in oncology. Nat Rev Clin Oncol. 2024;21:628–37.

    Article  PubMed  Google Scholar 

  146. Yang J, Li A, Li Y, Guo X, Wang M. A novel approach for drug response prediction in cancer cell lines via network representation learning. Bioinformatics. 2018;35:1527–35.

    Article  Google Scholar 

  147. You SC, Krumholz HM. the evolution of evidence-based medicine: when the magic of the randomized clinical trial meets real-world data. Circulation. 2022;145:107–9.

    Article  PubMed  Google Scholar 

  148. Elkhader J, Elemento O. Artificial intelligence in oncology: from bench to clinic. Semin Cancer Biol. 2022;84:113–28.

    Article  CAS  PubMed  Google Scholar 

  149. Zhang Z, Hernandez K, Savage J, Li S, Miller D, Agrawal S, Ortuno F, Staudt LM, Heath A, Grossman RL. Uniform genomic data analysis in the NCI genomic data commons. Nat Commun. 2021;12:1226.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  150. Karim MR, Islam T, Shajalal M, Beyan O, Lange C, Cochez M, Rebholz-Schuhmann D, Decker S. Explainable ai for bioinformatics: methods, tools and applications. Brief Bioinform. 2023;24:bbad236.

    Article  PubMed  Google Scholar 

  151. Folstad M, Augustine A, Hollnagel F, Carroll C, Saner A, Pier J, Lavitschke M, Kleinschmidt P, Twedt H, Tevaarwerk AJ, Emamekhoo H. Improving quality of oncology (onc) documentation and enhancing structured data collection using a standardized onc note template. J Clin Oncol. 2024;42:11128–11128.

    Article  Google Scholar 

  152. Le D, Brain A, Shenkier TN, Ingledew P-A. Virtual health in cancer care: results from a semi-structured interview-survey of oncology health care providers. J Clin Oncol. 2021;39:e13618–e13618.

    Article  Google Scholar 

  153. Eldridge EH, Trowbridge E, Zandt MV, Adam A, Mack CD. Unstructured EMR data hold the key to oncology studies: methods for validating NLP-extracted phenotypes from EMR notes. J Clin Oncol. 2024;42:11133–11133.

    Article  Google Scholar 

  154. Acosta JN, Falcone GJ, Rajpurkar P, Topol EJ. Multimodal biomedical AI. Nat Med. 2022;28:1773–84.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

Yes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

This work was supported by the National Natural Science Foundation of China [No. 82003929], Provincial Natural Science Foundation of Hunan [No. 2023JJ40857] and Scientific Research Program of Hunan Provincial Department of education [No. 23A0665].

Author information

Author notes

  1. Dangang Shangguan

    Present address: Hunan Cancer Hospital/The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, Hunan, China

Authors and Affiliations

  1. Hunan Cancer Hospital/The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, Hunan, China

    Yuan Mao, Dongsheng Cao & Yi-Kun Wang

  2. Xiangya School of Pharmaceutical Sciences, Central South University, Changsha, China

    Dangang Shangguan

  3. Department of Pharmacy, Xiangya Hospital, Central South University, Changsha, Hunan, China

    Qi Huang

  4. Department of Histology and Embryology of Xiangya School of Medicine, Central South University, Changsha, Hunan, People’s Republic of China

    Ling Xiao

  5. Department of Lymphoma and Hematology, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, Hunan, People’s Republic of China

    Yuan Mao & Hui Zhou

  6. Department of Lymphoma and Hematology, Hunan Cancer Hospital, Changsha, Hunan, People’s Republic of China

    Hui Zhou

Authors
  1. Yuan Mao
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Dangang Shangguan
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Qi Huang
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Ling Xiao
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Dongsheng Cao
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Hui Zhou
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Yi-Kun Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Y.M. and DG.SG.: Literature search,wrote the main manuscript text. Q.H. L.X. and DS.C: Literature search. H.Z and YK.W: Conceptualization, Funding and revision of manuscripts.

Corresponding authors

Correspondence to Hui Zhou or Yi-Kun Wang.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Mao, Y., Shangguan, D., Huang, Q. et al. Emerging artificial intelligence-driven precision therapies in tumor drug resistance: recent advances, opportunities, and challenges. Mol Cancer 24, 123 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12943-025-02321-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12943-025-02321-x

Keywords

Molecular Cancer

ISSN: 1476-4598