Abstract:Oncology therapeutic development continues to be plagued by high failure rates leading to substantial costs with only incremental improvements in overall benefit and survival. Advances in technology including the molecular characterisation of cancer and computational power provide the opportunity to better model therapeutic response and resistance. Here we use a novel approach which utilises Bayesian statistical principles used by astrophysicists to measure the mass of dark matter to predict therapeutic response. We construct “Digital Twins” of individual cancer patients and predict response for cancer treatments. We validate the approach by predicting the results of clinical trials. Better prediction of therapeutic response would improve current clinical decision-making and oncology therapeutic development.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the accuracy of predicting tumor treatment response in order to improve clinical decision - making and the development of tumor treatment drugs. Specifically, researchers use advanced Bayesian statistical methods and "digital twin" technology to construct computational models that can simulate the treatment response of individual cancer patients. These models can predict the efficacy of a single drug or a combination of drugs, and can also predict the performance of specific treatment regimens in individual patients and virtual clinical trials. ### Main problems and goals: 1. **Improve the accuracy of treatment response prediction**: By constructing "digital twins", that is, virtual replicas of cancer patients, to more accurately predict the effects of different treatment methods, thereby reducing the "trial - and - error" process in clinical trials. 2. **Optimize clinical decision - making**: Provide more accurate treatment response predictions to help doctors select the most appropriate treatment plan and improve the treatment effect of patients. 3. **Accelerate drug development**: Predict the efficacy of new drugs through virtual clinical trials, reduce the cost and time of drug development, and improve the success rate of drug development. 4. **Verify the validity of the model**: Verify the accuracy and reliability of the model prediction by comparing with the results of historical clinical trials. ### Method overview: - **Data sources**: Multiple large - scale databases are used, such as TCGA (The Cancer Genome Atlas), CTRP (Cancer Therapeutic Response Portal) and CCLE (Cancer Cell Line Encyclopedia), which provide rich molecular characteristics and clinical data. - **Model construction**: Three main models are constructed: - **Drug Efficacy Model**: Based on molecular characteristics and drug fingerprints, predict the IC50 value of drugs in vitro and the drug synergy score. - **Treatment Response Model**: Predict patients' response to treatment, divided into response (partial response or complete response) and no response (stable disease or progressive disease). - **Overall Survival Model**: Integrate the clinical data, treatment response and pathological information of individual patients to predict the overall survival time of patients. - **Verification method**: By simulating historical clinical trials, compare the treatment response rate (ORR) predicted by the model with the actual clinical trial results to verify the accuracy of the model. ### Result verification: - **Single - drug chemotherapy arm**: Multiple historical clinical trials are simulated, such as the advanced pancreatic cancer trial conducted by Burris et al. in 1997 and the metastatic breast cancer trial conducted by Chan et al. in 2023. The model successfully predicted the treatment effects of different drugs. - **Single drug vs. combined chemotherapy**: For example, in the recurrent ovarian cancer trial conducted by Cantù et al. in 2022, the model predicted that combined chemotherapy (CAP) had a better treatment response rate than the single - drug paclitaxel (Paclitaxel). - **Blinded verification**: Without knowing the actual clinical trial results, the model also successfully predicted the results of multiple clinical trials, further verifying its effectiveness. In conclusion, this paper aims to improve the accuracy of predicting tumor treatment response by constructing and verifying "digital twin" models, thereby optimizing clinical decision - making and accelerating drug development.

Computational prediction of therapeutic response and cancer outcomes

Prediction of therapeutic response and cancer outcomes in solid tumours via in silico clinical trials.

Personalized Cancer Therapy Using a Patient-Derived Tumor Tissue Xenograft Model: a Translational Field Worthy of Exploring Further?

Machine learning and data mining frameworks for predicting drug response in cancer: An overview and a novel in silico screening process based on association rule mining

Improving personalized tumor growth predictions using a Bayesian combination of mechanistic modeling and machine learning

Developing Anticancer Drug Response System Using Deep Learning System with Hybrid Genomic and Chemical Features

Exploring approaches for predictive cancer patient digital twins: Opportunities for collaboration and innovation

Immunotherapy Efficacy Prediction in Cancer: An Artificial Intelligence Approach with Unannotated H&E Whole-Slide Images

Evaluating the Molecule-Based Prediction of Clinical Drug Responses in Cancer.

Leveraging State-of-the-Art AI Algorithms in Personalized Oncology: From Transcriptomics to Treatment

Predicting tumour response to anti-PD-1 immunotherapy with computational modelling

Current status and future developments in predicting outcomes in radiation oncology

Mathematical prediction of clinical outcomes in advanced cancer patients treated with checkpoint inhibitor immunotherapy

Personalized Anti-Tumor Drug Efficacy Prediction Based on Clinical Data

Machine learning model to predict oncologic outcomes for drugs in randomized clinical trials

DeepTTA: a transformer-based model for predicting cancer drug response

Drug response prediction by inferring pathway-response associations with Kernelized Bayesian Matrix Factorization

Biology-guided deep learning predicts prognosis and cancer immunotherapy response

Predicting patient treatment response and resistance via single-cell transcriptomics of their tumors.

Abstract IA021: Digital patient avatars predict cancer evolution and personalised therapeutic strategies

Novel computational biology modeling system can accurately forecast response to neoadjuvant therapy in early breast cancer