Towards Realizing the Vision of Precision Medicine: AI Based Prediction of Clinical Drug Response

2020 ◽  
Author(s):  
Johann de Jong ◽  
Ioana Cutcutache ◽  
Matthew Page ◽  
Sami Elmoufti ◽  
Cynthia Dilley ◽  
...  
Keyword(s):  
Precision Medicine ◽  
Drug Response ◽  
Clinical Drug

scholarly journals Tissue-guided LASSO for prediction of clinical drug response using preclinical samples

2019 ◽  
Author(s):  
Edward W Huang ◽  
Ameya Bhope ◽  
Jing Lim ◽  
Saurabh Sinha ◽  
Amin Emad
Keyword(s):  
Machine Learning ◽  
Drug Response ◽  
Drug Sensitivity ◽  
Machine Learning Algorithms ◽  
Learning Models ◽  
Large Databases ◽  
Tissue Of Origin ◽  
Molecular Profiles ◽  
Clinical Drug ◽  
Machine Learning Models

ABSTRACTPrediction of clinical drug response (CDR) of cancer patients, based on their clinical and molecular profiles obtained prior to administration of the drug, can play a significant role in individualized medicine. Machine learning models have the potential to address this issue, but training them requires data from a large number of patients treated with each drug, limiting their feasibility. While large databases of drug response and molecular profiles of preclinical in-vitro cancer cell lines (CCLs) exist for many drugs, it is unclear whether preclinical samples can be used to predict CDR of real patients.We designed a systematic approach to evaluate how well different algorithms, trained on gene expression and drug response of CCLs, can predict CDR of patients. Using data from two large databases, we evaluated various linear and non-linear algorithms, some of which utilized information on gene interactions. Then, we developed a new algorithm called TG-LASSO that explicitly integrates information on samples’ tissue of origin with gene expression profiles to improve prediction performance. Our results showed that regularized regression methods provide significantly accurate prediction. However, including the network information or common methods of including information on the tissue of origin did not improve the results. On the other hand, TG-LASSO improved the predictions and distinguished resistant and sensitive patients for 7 out of 13 drugs. Additionally, TG-LASSO identified genes associated with the drug response, including known targets and pathways involved in the drugs’ mechanism of action. Moreover, genes identified by TG-LASSO for multiple drugs in a tissue were associated with patient survival. In summary, our analysis suggests that preclinical samples can be used to predict CDR of patients and identify biomarkers of drug sensitivity and survival.AUTHOR SUMMARYCancer is among the leading causes of death globally and perdition of the drug response of patients to different treatments based on their clinical and molecular profiles can enable individualized cancer medicine. Machine learning algorithms have the potential to play a significant role in this task; but, these algorithms are designed based the premise that a large number of labeled training samples are available, and these samples are accurate representation of the profiles of real tumors. However, due to ethical and technical reasons, it is not possible to screen humans for many drugs, significantly limiting the size of training data. To overcome this data scarcity problem, machine learning models can be trained using large databases of preclinical samples (e.g. cancer cell line cultures). However, due to the major differences between preclinical samples and real tumors, it is unclear how accurately such preclinical-to-clinical computational models can predict the clinical drug response of cancer patients.Here, first we systematically evaluate a variety of different linear and nonlinear machine learning algorithms for this particular task using two large databases of preclinical (GDSC) and tumor samples (TCGA). Then, we present a novel method called TG-LASSO that utilizes a new approach for explicitly incorporating the tissue of origin of samples in the prediction task. Our results show that TG-LASSO outperforms all other algorithms and can accurately distinguish resistant and sensitive patients for the majority of the tested drugs. Follow-up analysis reveal that this method can also identify biomarkers of drug sensitivity in each cancer type.

A CLIA-certified functional precision medicine assay, the PARIS© test, to predict drug response for ovarian cancer

2020 ◽  
Vol 159 ◽  
pp. 147-148
Author(s):  
A. Richardson ◽  
A. Dullea ◽  
H. Stecher ◽  
S. Pereira ◽  
F. Musa ◽  
...  
Keyword(s):  
Ovarian Cancer ◽  
Precision Medicine ◽  
Drug Response

scholarly journals Rare genetic variability in human drug target genes modulates drug response and can guide precision medicine

2021 ◽  
Vol 7 (36) ◽  
Author(s):  
Yitian Zhou ◽  
Gabriel Herras Arribas ◽  
Ainoleena Turku ◽  
Tuuli Jürgenson ◽  
Souren Mkrtchian ◽  
...  
Keyword(s):  
Genetic Variability ◽  
Precision Medicine ◽  
Drug Target ◽  
Drug Response ◽  
Target Genes

scholarly journals Functional Precision Medicine in AML: Technical Performance Evaluation for in Vitro Diagnostics Using High-Throughput Image-Based Screening of Primary Patient Cells

Blood ◽  
2019 ◽  
Vol 134 (Supplement_1) ◽  
pp. 3366-3366 ◽  
Author(s):  
Nikolaus Krall ◽  
Noemi Meszaros ◽  
Karin Lind ◽  
Bojan Vilagos ◽  
Heinz Sill ◽  
...  
Keyword(s):  
Bone Marrow ◽  
Precision Medicine ◽  
Drug Response ◽  
Ex Vivo ◽  
Advisory Board ◽  
Small Molecule Drugs ◽  
Cd34 Cells ◽  
Equity Ownership ◽  
The Right

Introduction: Identification of the right drug, for the right patient, at the right time is the defining goal for precision and personalized medicine programs. Many tools are currently being pursued for this including next-gen sequencing and other -omics approaches to determine target availability (Moscow et al, 2018 Nat Rev Clin Oncol.), functional drug response profiling in viable patient samples (Snijder et al, 2017 Lancet Haem), or a combination of both (Schmidl et al, 2019 Nat Chem Bio). Of those, functional precision medicine approaches are only just beginning to gain clinical interest, and because of this, nearly all programs lack systematic testing in prospective clinical studies (Letai, 2017 Nat Rev Med). One of the first functional precision medicine approaches to be tested in a small prospective clinical study was the use of high-throughput image-based screening and single cell analysis of drug response in primary tissues ("Pharmacoscopy") as a tool to prospectively rank drug options for patients with late-stage hematological indications (Snijder et al 2017, Lancet Haem, NCT03096821). In this study, taking Pharmacoscopy data into account for treatment selection led to a higher overall response rate (88% vs. 24%) and longer progression free survival (22.6 vs. 5.7 weeks) compared to the previous line of therapy. In order for functional precision medicine programs to reach widespread routine use, ultimately the development of fully-certified in vitro diagnostic test products will be required. This requires amongst others a systematic understanding of i) the technical robustness of functional drug testing approaches and ii) a detailed understanding of how pre-analytical sample handling influences assay results. Here, we set out to investigate two crucial questions that arose in the course of systematic assay development for AML: a) do CD34+ cells derived from peripheral blood and bone marrow AML patients respond differently to short term ex vivo treatment with small molecule drugs and b) does the overall response of these cells change upon cryopreservation? These issues are of high relevance even outside the setting of functional precision medicine development programs given worldwide biobanking efforts, and the lack of robust systematic model comparison. Methods: Fresh patient blood and bone marrow samples of newly diagnosed as well as relapsed and refractory patients with AML were collected at Medical University of Graz under appropriate ethics approval. Samples were divided into two parts and one part used immediately, the other cryopreserved as viable cells. The response of CD34+ cells against 140 different small molecule drugs (two concentrations and three technical replicates) was evaluated using single-cell high content microscopy (Allcyte's "Pharmacoscopy" platform) in both peripheral blood and bone marrow as well as freshly used and previously viably frozen samples. Drug response was determined by fitting to generalised linear and mixed models. Further, RNA is isolated from additional patient samples before and after biobanking for Nanostring analysis. Results: Overall, across all tested patient samples, we observed a high correlation between drug response tested in blood and bone marrow samples from this group of patient samples, indicating, at least for AML, no strong niche dependence for the drugs tested. Additionally, we also observed a high correlation between drug response tested in fresh and frozen tissues (Figure 1). The differences that did appear were, in particular, in drugs targeting metabolic and cell-growth dependent functions, amongst others. Conclusions: Understanding how biobanking affects the ex vivo drug response of primary human tissues is paramount to being able to understand how far translational efforts can go using tissues and functional drug response assays. We provide systematic data supporting the use of cryopreserved samples for the functional analysis of drug response in AML. This is a fundamental question associated with the use of biobanks for investigations of primary patient material. Figure 1 Disclosures Krall: Allcyte: Employment, Equity Ownership. Meszaros:Allcyte: Employment. Vilagos:Allcyte: Employment. Sill:Astellas: Other: Advisory board; Novartis: Other: Advisory board; Astex: Other: Advisory board; AbbVie: Other: Advisory board. Vladimer:Allcyte: Employment, Equity Ownership, Patents & Royalties: EP3198276A1.

scholarly journals TMOD-04. IMAGE-BASED DRUG RESPONSE PROFILING FROM PEDIATRIC TUMOR CELL SPHEROIDS USING PATIENT-BY-PATIENT DEEP TRANSFER LEARNING

2021 ◽  
Vol 23 (Supplement_1) ◽  
pp. i36-i36
Author(s):  
Yannick Berker ◽  
Dina ElHarouni ◽  
Heike Peterziel ◽  
Ina Oehme ◽  
Matthias Schlesner ◽  
...  
Keyword(s):  
Drug Discovery ◽  
Tumor Cell ◽  
Precision Medicine ◽  
Cell Line ◽  
Transfer Learning ◽  
Cell Cultures ◽  
Drug Response ◽  
Fine Tuning ◽  
Braf V600e ◽  
Drug Concentrations

Abstract Introduction Image-based phenotypic drug profiling is receiving increasing attention in drug discovery and precision medicine. Compared to classical end-point measurements quantifying drug response, image-based profiling enables both the quantification of drug response and characterization of disease entities and drug mechanisms of actions. In pediatric precision oncology, we aim to study drug response in patient-derived 3D spheroid tumor cell cultures and tackle the challenges of a lack of image-segmentation methods and limited patient-derived material. Methods We investigate deep transfer learning with patient-by-patient fine-tuning for cell-viability quantification. We fine-tune a convolutional neural network (pre-trained on ImageNet) with many cell-line-specific and few patient-specific assay controls. The method is validated using 3D cell cultures in 384-well microplates derived from cell lines with known drug sensitivities and tested with primary patient-derived samples. Network outputs at different drug concentrations are used for drug-sensitivity scoring; dense-layer activations are used in t-distributed stochastic neighbor embedding and clustering of drugs. Results Cell-line experiments confirm expected hits, such as effective treatment with BRAF inhibitors in a BRAF V600E mutated brain tumor model and NTRK inhibitors in a cell line harboring an NTRK-fusion, indicating the predictive power of deep learning to identify drug-hit candidates for individual patients. In patient-derived samples, clustering of drugs further confirms phenotypic similarity according to their mechanisms of actions. Combining drug scoring with phenotypic clustering may provide opportunities for complementary combination treatments. Conclusion Deep transfer learning with patient-by-patient fine-tuning is a promising, segmentation-free image-analysis approach for precision medicine and drug discovery based on 3D spheroid cell cultures.

EIT PACMAN Study preliminary results: OncoSignal pathway analysis identifies actionable cancer targets.

2020 ◽  
Vol 38 (15_suppl) ◽  
pp. e15606-e15606
Author(s):  
Patricia Martin ◽  
Sigi Neerken ◽  
Anja Van De Stolpe ◽  
Eveline den Biezen-Timmermans ◽  
Martijn Akse ◽  
...  
Keyword(s):  
Precision Medicine ◽  
Pathway Analysis ◽  
Drug Response ◽  
Tumor Tissue ◽  
Ovarian Tissue ◽  
Metastatic Tumor ◽  
Pi3k Pathway ◽  
Expression Array ◽  
Tissue Samples ◽  
Targeted Drug

e15606 Background: Precision medicine refers to tailoring of treatment to each individual patient, although identifying tumor driving signaling pathways (SP) that are functionally active is still a challenge. OncoSignal pathway tests quantitatively measure activity of SP such as estrogen receptor (ER), androgen receptor (AR), PI3K, MAPK, Hedgehog (HH), TGF-β, Notch on fresh frozen and formalin-fixed paraffin-embedded (FFPE) tissue samples. OncoSignal pathway analysis aimed at assessing clinically actionable SP and retrospectively predicting targeted drug response on a series of patients’ (pts) samples from the MOSCATO trial run at Gustave Roussy. Methods: OncoSignal pathway analysis (ER, AR, PI3K, MAPK, HH, Notch, TGF-β) was performed blinded by Molecular Pathway Dx (Philips, Eindhoven) on metastatic tumor tissue samples from breast cancer (BC), prostate (PC), and high grade serous ovarian cancers (OC). Using Affymetrix expression array data from public GEO datasets, SP activity was analyzed in healthy prostate, breast, and ovarian tissue to define abnormal SP activity thresholds for tumor tissue pathway analysis. For each individual sample, SP alterations were considered tumor driving SP if sample SP activity exceeded the 95th percentile of SP activity within healthy tissue. Results by OncoSignal were also combined with clinical characteristics and molecular alterations identified in the MOSCATO trial. Results: Identified tumor driving SP were ER, AR, MAPK-AP1, HH, PI3K pathway in BC (n = 5), AR in PC (n = 30); AP1, Notch, TGFβ in OC (n = 17). OncoSignal identified clinically actionable tumor driving pathways in all BC samples (median tumor cellularity [MTC]: 40%, range 15-80%); 30/31 PC samples (MTC: 62%, range 25-90%), 16/17 OC samples (MTC 62%, range 15-80%). Actionable mutations were previously identified in 4/5 BC; 13/31 PC; 6/17 OC. Seven pts with BC and PC were treated with targeted therapy. OncoSignal pathway analysis correctly predicted response/resistance in 4 of these pts (57%). Conclusions: OncoSignal pathway analysis correctly identified SP activity alterations and predicted targeted drug response in this series of patients. OncoSignal will be further validated prospectively in precision medicine studies at Gustave Roussy in which patients are stratified for targeted treatment by mutation analysis.

Clinical drug response prediction from preclinical cancer cell lines by logistic matrix factorization approach

2021 ◽  
Author(s):  
Akram Emdadi ◽  
Changiz Eslahchi
Keyword(s):  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Matrix Factorization ◽  
Drug Response ◽  
Drug Sensitivity ◽  
Cancer Cell Lines ◽  
Tissue Type ◽  
Factorization Approach ◽  
Clinical Drug

Predicting tumor drug response using cancer cell line drug response values for a large number of anti-cancer drugs is a significant challenge in personalized medicine. Predicting patient response to drugs from data obtained from preclinical models is made easier by the availability of different knowledge on cell lines and drugs. This paper proposes the TCLMF method, a predictive model for predicting drug response in tumor samples that was trained on preclinical samples and is based on the logistic matrix factorization approach. The TCLMF model is designed based on gene expression profiles, tissue type information, the chemical structure of drugs and drug sensitivity (IC 50) data from cancer cell lines. We use preclinical data from the Genomics of Drug Sensitivity in Cancer dataset (GDSC) to train the proposed drug response model, which we then use to predict drug sensitivity of samples from the Cancer Genome Atlas (TCGA) dataset. The TCLMF approach focuses on identifying successful features of cell lines and drugs in order to calculate the probability of the tumor samples being sensitive to drugs. The closest cell line neighbours for each tumor sample are calculated using a description of similarity between tumor samples and cell lines in this study. The drug response for a new tumor is then calculated by averaging the low-rank features obtained from its neighboring cell lines. We compare the results of the TCLMF model with the results of the previously proposed methods using two databases and two approaches to test the model’s performance. In the first approach, 12 drugs with enough known clinical drug response, considered in previous methods, are studied. For 7 drugs out of 12, the TCLMF can significantly distinguish between patients that are resistance to these drugs and the patients that are sensitive to them. These approaches are converted to classification models using a threshold in the second approach, and the results are compared. The results demonstrate that the TCLMF method provides accurate predictions across the results of the other algorithms. Finally, we accurately classify tumor tissue type using the latent vectors obtained from TCLMF’s logistic matrix factorization process. These findings demonstrate that the TCLMF approach produces effective latent vectors for tumor samples. The source code of the TCLMF method is available in https://github.com/emdadi/TCLMF.

scholarly journals Plasmodium vivax Polymorphism in a Clinical Drug Trial

2001 ◽  
Vol 8 (5) ◽  
pp. 891-894 ◽  
Author(s):  
T. Adak ◽  
Neena Valecha ◽  
V. P. Sharma
Keyword(s):  
Plasmodium Vivax ◽  
Drug Response ◽  
Drug Trial ◽  
Short Term ◽  
Double Blind ◽  
Response Data ◽  
Clinical Drug Trial ◽  
Clinical Drug

ABSTRACT Data from a double-blind randomized clinical drug trial were analyzed to find the comparative responses of two antirelapse drugs, bulaquine and primaquine, against different relapsing forms ofPlasmodium vivax infection. A 1-year follow-up study strongly suggests that the duration of preerythrocytic development ofP. vivax is a polymorphic characteristic, exhibited by two strains of hypnozoites responsible for early and late manifestations after primary infection. Short-term relapses were significantly higher in the first half year than long-term relapses, and the reverse was true in the second half year. Clinical drug response data showed that the hypnozoites characterized for short-term relapse were not susceptible to either of the antirelapse drugs in the currently administered dose, whereas hypnozoites characterized for long incubation were significantly susceptible.

Sign in / Sign up

Export Citation Format

Share Document