Interpretable Machine Learning Models to Predict the Resistance of Breast Cancer Patients to Doxorubicin from Their microRNA Profiles

2023 Beijing Health Data Science Summit

The 5th annual Beijing Health Data Science Summit, organized by the National Institute of Health Data Science at Peking University, recently concluded with resounding success. This year, the summit aimed to foster collaboration among researchers, practitioners, and stakeholders in the field of health data science to advance the use of data for better health outcomes. One significant highlight of this year's summit was the introduction of the Abstract Competition, organized by Health Data Science, a Science Partner Journal, which focused on the use of cutting-edge data science methodologies, particularly the application of artificial intelligence in the healthcare scenarios. The competition provided a platform for researchers to showcase their groundbreaking work and innovations. In total, the summit received 61 abstract submissions. Following a rigorous evaluation process by the Abstract Review Committee, eight exceptional abstracts were selected to compete in the final round and give presentations in the Abstract Competition. The winners of the Abstract Competition are as follows:•First Prize: "Interpretable Machine Learning for Predicting Outcomes of Childhood Kawasaki Disease: Electronic Health Record Analysis" presented by researchers from the Chinese Academy of Medical Sciences, Peking Union Medical College, and Chongqing Medical University (presenter Yifan Duan).•Second Prize: "Survival Disparities among Mobility Patterns of Patients with Cancer: A Population-Based Study" presented by a team from Peking University (presenter Fengyu Wen).•Third Prize: "Deep Learning-Based Real-Time Predictive Model for the Development of Acute Stroke" presented by researchers from Beijing Tiantan Hospital (presenter Lan Lan). We extend our heartfelt gratitude to the esteemed panel of judges whose expertise and dedication ensured the fairness and quality of the competition. The judging panel included Jiebo Luo from the University of Rochester (chair), Shenda Hong from Peking University, Xiaozhong Liu from Worcester Polytechnic Institute, Liu Yang from Hong Kong Baptist University, Ma Jianzhu from Tsinghua University, Ting Ma from Harbin Institute of Technology, and Jian Tang from Mila-Quebec Artificial Intelligence Institute. We wish to convey our deep appreciation to Zixuan He and Haoyang Hong for their invaluable assistance in the meticulous planning and execution of the event. As the 2023 Beijing Health Data Science Summit comes to a close, we look forward to welcoming all participants to join us in 2024. Together, we will continue to advance the frontiers of health data science and work toward a healthier future for all.

Early Breast Cancer Risk Assessment: Integrating Histopathology with Artificial Intelligence

Effective risk assessment in early breast cancer is essential for informed clinical decision-making, yet consensus on defining risk categories remains challenging. This paper explores evolving approaches in risk stratification, encompassing histopathological, immunohistochemical, and molecular biomarkers alongside cutting-edge artificial intelligence (AI) techniques. Leveraging machine learning, deep learning, and convolutional neural networks, AI is reshaping predictive algorithms for recurrence risk, thereby revolutionizing diagnostic accuracy and treatment planning. Beyond detection, AI applications extend to histological subtyping, grading, lymph node assessment, and molecular feature identification, fostering personalized therapy decisions. With rising cancer rates, it is crucial to implement AI to accelerate breakthroughs in clinical practice, benefiting both patients and healthcare providers. However, it is important to recognize that while AI offers powerful automation and analysis tools, it lacks the nuanced understanding, clinical context, and ethical considerations inherent to human pathologists in patient care. Hence, the successful integration of AI into clinical practice demands collaborative efforts between medical experts and computational pathologists to optimize patient outcomes.

A gray box framework that optimizes a white box logical model using a black box optimizer for simulating cellular responses to perturbations

Predicting cellular responses to perturbations requires interpretable insights into molecular regulatory dynamics to perform reliable cell fate control, despite the confounding non-linearity of the underlying interactions. There is a growing interest in developing machine learning-based perturbation response prediction models to handle the non-linearity of perturbation data, but their interpretation in terms of molecular regulatory dynamics remains a challenge. Alternatively, for meaningful biological interpretation, logical network models such as Boolean networks are widely used in systems biology to represent intracellular molecular regulation. However, determining the appropriate regulatory logic of large-scale networks remains an obstacle due to the high-dimensional and discontinuous search space. To tackle these challenges, we present a scalable derivative-free optimizer trained by meta-reinforcement learning for Boolean network models. The logical network model optimized by the trained optimizer successfully predicts anti-cancer drug responses of cancer cell lines, while simultaneously providing insight into their underlying molecular regulatory mechanisms.

A Comprehensive Investigation of Active Learning Strategies for Conducting Anti-Cancer Drug Screening

It is well-known that cancers of the same histology type can respond differently to a treatment. Thus, computational drug response prediction is of paramount importance for both preclinical drug screening studies and clinical treatment design. To build drug response prediction models, treatment response data need to be generated through screening experiments and used as input to train the prediction models. In this study, we investigate various active learning strategies of selecting experiments to generate response data for the purposes of (1) improving the performance of drug response prediction models built on the data and (2) identifying effective treatments. Here, we focus on constructing drug-specific response prediction models for cancer cell lines. Various approaches have been designed and applied to select cell lines for screening, including a random, greedy, uncertainty, diversity, combination of greedy and uncertainty, sampling-based hybrid, and iteration-based hybrid approach. All of these approaches are evaluated and compared using two criteria: (1) the number of identified hits that are selected experiments validated to be responsive, and (2) the performance of the response prediction model trained on the data of selected experiments. The analysis was conducted for 57 drugs and the results show a significant improvement on identifying hits using active learning approaches compared with the random and greedy sampling method. Active learning approaches also show an improvement on response prediction performance for some of the drugs and analysis runs compared with the greedy sampling method.

Large-Scale Machine Learning Analysis Reveals DNA Methylation and Gene Expression Response Signatures for Gemcitabine-Treated Pancreatic Cancer

Background: Gemcitabine is a first-line chemotherapy for pancreatic adenocarcinoma (PAAD), but many PAAD patients do not respond to gemcitabine-containing treatments. Being able to predict such nonresponders would hence permit the undelayed administration of more promising treatments while sparing gemcitabine life-threatening side effects for those patients. Unfortunately, the few predictors of PAAD patient response to this drug are weak, none of them exploiting yet the power of machine learning (ML). Methods: Here, we applied ML to predict the response of PAAD patients to gemcitabine from the molecular profiles of their tumors. More concretely, we collected diverse molecular profiles of PAAD patient tumors along with the corresponding clinical data (gemcitabine responses and clinical features) from the Genomic Data Commons resource. From systematically combining 8 tumor profiles with 16 classification algorithms, each of the resulting 128 ML models was evaluated by multiple 10-fold cross-validations. Results: Only 7 of these 128 models were predictive, which underlines the importance of carrying out such a large-scale analysis to avoid missing the most predictive models. These were here random forest using 4 selected mRNAs [0.44 Matthews correlation coefficient (MCC), 0.785 receiver operating characteristic-area under the curve (ROC-AUC)] and XGBoost combining 12 DNA methylation probes (0.32 MCC, 0.697 ROC-AUC). By contrast, the hENT1 marker obtained much worse random-level performance (practically 0 MCC, 0.5 ROC-AUC). Despite not being trained to predict prognosis (overall and progression-free survival), these ML models were also able to anticipate this patient outcome. Conclusions: We release these promising ML models so that they can be evaluated prospectively on other gemcitabine-treated PAAD patients.

Preclinical Anticipation of On- and Off-Target Resistance Mechanisms to Anti-Cancer Drugs: A Systematic Review

The advent of targeted therapies has led to tremendous improvements in treatment options and their outcomes in the field of oncology. Yet, many cancers outsmart precision drugs by developing on-target or off-target resistance mechanisms. Gaining the ability to resist treatment is the rule rather than the exception in tumors, and it remains a major healthcare challenge to achieve long-lasting remission in most cancer patients. Here, we discuss emerging strategies that take advantage of innovative high-throughput screening technologies to anticipate on- and off-target resistance mechanisms before they occur in treated cancer patients. We divide the methods into non-systematic approaches, such as random mutagenesis or long-term drug treatment, and systematic approaches, relying on the clustered regularly interspaced short palindromic repeats (CRISPR) system, saturated mutagenesis, or computational methods. All these new developments, especially genome-wide CRISPR-based screening platforms, have significantly accelerated the processes for identification of the mechanisms responsible for cancer drug resistance and opened up new avenues for future treatments.

FINDER: A Fluidly Confined CRISPR-Based DNA Reporter on Living Cell Membranes for Rapid and Sensitive Cancer Cell Identification

The accurate, rapid, and sensitive identification of cancer cells in complex physiological environments is significant in biological studies, personalized medicine, and biomedical engineering. Inspired by the naturally confined enzymes on fluid cell membranes, a fluidly confined CRISPR-based DNA reporter (FINDER) was developed on living cell membranes, which was successfully applied for rapid and sensitive cancer cell identification in clinical blood samples. Benefiting from the spatial confinement effect for improved local concentration, and membrane fluidity for higher collision efficiency, the activity of CRISPR-Cas12a was, for the first time, found to be significantly enhanced on living cell membranes. This new phenomenon was then combined with multiple aptamer-based DNA logic gate for cell recognition, thus a FINDER system capable of accurate, rapid and sensitive cancer cell identification was constructed. The FINDER rapidly identified target cells in only 20 min, and achieved over 80 % recognition efficiency with only 0.1 % of target cells presented in clinical blood samples, indicating its potential application in biological studies, personalized medicine, and biomedical engineering.

A Boolean-based machine learning framework identifies predictive biomarkers of HSP90-targeted therapy response in prostate cancer

Precision medicine has emerged as an important paradigm in oncology, driven by the significant heterogeneity of individual patients' tumour. A key prerequisite for effective implementation of precision oncology is the development of companion biomarkers that can predict response to anti-cancer therapies and guide patient selection for clinical trials and/or treatment. However, reliable predictive biomarkers are currently lacking for many anti-cancer therapies, hampering their clinical application. Here, we developed a novel machine learning-based framework to derive predictive multi-gene biomarker panels and associated expression signatures that accurately predict cancer drug sensitivity. We demonstrated the power of the approach by applying it to identify response biomarker panels for an Hsp90-based therapy in prostate cancer, using proteomic data profiled from prostate cancer patient-derived explants. Our approach employs a rational feature section strategy to maximise model performance, and innovatively utilizes Boolean algebra methods to derive specific expression signatures of the marker proteins. Given suitable data for model training, the approach is also applicable to other cancer drug agents in different tumour settings.

NeRD: a multichannel neural network to predict cellular response of drugs by integrating multidimensional data

BACKGROUND

Considering the heterogeneity of tumors, it is a key issue in precision medicine to predict the drug response of each individual. The accumulation of various types of drug informatics and multi-omics data facilitates the development of efficient models for drug response prediction. However, the selection of high-quality data sources and the design of suitable methods remain a challenge.

METHODS

In this paper, we design NeRD, a multidimensional data integration model based on the PRISM drug response database, to predict the cellular response of drugs. Four feature extractors, including drug structure extractor (DSE), molecular fingerprint extractor (MFE), miRNA expression extractor (mEE), and copy number extractor (CNE), are designed for different types and dimensions of data. A fully connected network is used to fuse all features and make predictions.

RESULTS

Experimental results demonstrate the effective integration of the global and local structural features of drugs, as well as the features of cell lines from different omics data. For all metrics tested on the PRISM database, NeRD surpassed previous approaches. We also verified that NeRD has strong reliability in the prediction results of new samples. Moreover, unlike other algorithms, when the amount of training data was reduced, NeRD maintained stable performance.

CONCLUSIONS

NeRD's feature fusion provides a new idea for drug response prediction, which is of great significance for precise cancer treatment.