Clinically oriented prediction of patient response to targeted and immunotherapies from the tumor transcriptome

Abstract

Precision oncology is gradually advancing into mainstream clinical practice, demonstrating significant survival benefits. However, eligibility and response rates remain limited in many cases, calling for better predictive biomarkers. Here we present ENLIGHT, a transcriptomics-based computational approach that identifies clinically relevant genomic interactions and uses them to predict patients' response to a variety of therapies in multiple cancer types without training on previous response data. We study ENLIGHT in two translationally oriented scenarios, Personalized Medicine (PM), aimed at prioritizing treatments for a single patient, and Clinical Trial Design (CTD), selecting the most likely responders in a patient cohort. Evaluating ENLIGHT's PM performance on 21 blinded clinical trial datasets, we show that it can effectively predict treatment response across multiple therapies and cancer types (obtaining an odds ratio of 2.59), substantially improving upon SELECT, a previously published transcriptomics-based approach, and performing as well as supervised predictors developed for specific indications and drugs, but on a much broader array of therapies and indications. In the CTD scenario, ENLIGHT can markedly enhance the success of clinical trials by excluding non-responders while achieving more than 90% of the response rate attainable under an optimal exclusion strategy. In sum, ENLIGHT is one of the first approaches to demonstrably predict therapeutic response across multiple cancer types by leveraging the transcriptome.

Prediction of cancer treatment response from histopathology images through imputed transcriptomics

“…First, we applied DeepPT to predict the patient transcriptomics from their H&E slides. Second, based on this predicted gene expression, we used the precision oncology algorithm, ENLIGHT [33], to predict the patients' response from the predicted expression.…”

“…Data utilized in this analysis included the response to therapy by residual cancer burden criteria and the fresh frozen H&E-stained primary tumor slides. were greater/equal (lesser) than a decision threshold value of 0.54, a threshold that was fixed and determined already in [33], again, on completely independent data. The same threshold was used here, without any training or modification, both for ENLIGHT-DeepPT and ENLIGHTactual.…”

“…First, we used the DeepPT model trained beforehand on the 1,043 TCGA-Breast patients, without any changes and with no further training on this test dataset, to predict the gene expression values from the H&E slides of each patient's tumor. Second, we applied ENLIGHT to these predicted gene expression values to produce EMS scores based on the same SL/SR partners of ERBB2 that were already inferred in[33]. Importantly, we do not restrict the model to genes with high correlations between the actual and predicted expression values; ENLIGHT considers the combined effect of a large set of genes, mitigating the noise of individual expression prediction, which leads to robust response predictions as we show below.…”

Drug combination prioritization for cancer treatment using single-cell RNA-seq based transfer learning

“…Several research groups have recently attempted to incorporate multiple sources of information into cutting-edge machine learning approaches in order to create clinically oriented predictions of patient responses to medications. I-PREDICT (7), NCI-MATCH (8), MI-ONCOSEQ (9), WINTHER (10), ONCO-TARGET/ONCOTREAT (11), SELECT (12), and ENLIGHT (13) are studies that aim to match patients with therapies and predict their impact on treatment outcome using DNA biomarkers, genomic and transcriptomic information, protein-protein interaction networks, synthetic lethal and/or synthetic rescue interactions among other factors (7–13). However, developing such studies requires data from a large number of patients as well as extensive knowledge of the exact molecular targets for each of the drugs under investigation, limiting their applicability (14).…”

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

“…However, a large fraction of cancer patients still do not bene t from such targeted therapies, and efforts are hence much needed to nd ways to analyze other molecular omics data types to bene t more patients. Addressing this challenge, recent studies have begun to explore the bene t of collecting and analyzing bulk tumor transcriptomics data to guide cancer patient treatment (Beaubier et al, 2019;Hayashi et al, 2020;Rodon et al, 2019;Tanioka et al, 2018;Vaske et al, 2019;Wong et al, 2020, Lee et al, 2021, Dinstag et al, 2022. These studies have demonstrated the potential of such approaches to complement DNA sequencing approaches in increasing the bene t of omics-guided precision treatments to patients.…”