{"entity":{"id":"b-combination-space","kind":"bottleneck","name":"Too many combinations to test","aka":[],"tldr":"There are thousands of possible drug pairs and sequences. Trials can test a few dozen a year.","summary":"With well over a hundred approved oncology agents and hundreds in development, the space of pairwise combinations runs to tens of thousands, and sequences and schedules multiply it further, while the field can run only a few dozen adequately powered combination trials a year. The combinations that are tested are chosen by commercial ownership and precedent rather than biology: thousands of PD-1/PD-L1 combination trials have been launched, most adding an agent to a checkpoint inhibitor without a predictive biomarker. Analyses of historical combination trials suggest that many 'successes' reflect independent action in different patients rather than synergy, which means better patient selection would achieve the same benefit with fewer drugs. Platform trials, factorial and adaptive designs, ex vivo functional testing, and computational prioritisation from dependency maps and combination screens are the only ways to explore the space at a useful rate.","asOf":"2026-09-08","links":[{"label":"Palmer & Sorger, Combination cancer therapy can confer benefit via patient-to-patient variability without drug additivity or synergy (Cell 2017)","url":"https://doi.org/10.1016/j.cell.2017.11.009"},{"label":"Upadhaya et al., Challenges and opportunities in the PD1/PDL1 inhibitor clinical trial landscape (Nat Rev Drug Discov 2022)","url":"https://doi.org/10.1038/d41573-022-00030-4"},{"label":"Menden et al., Community assessment to advance computational prediction of cancer drug combinations (Nature Communications 2019)","url":"https://doi.org/10.1038/s41467-019-09799-2"}],"tags":[],"related":["depmap","drugbank-chembl","idea-payload-switching","idea-dual-payload-first","idea-organoid-guided-adc","idea-tr2-shared-control-network","idea-tr2-organoid-coclinical-arms","idea-tr2-alternating-vs-concurrent","idea-tr2-rwe-combination-emulation","idea-tr2-factorial-adjuvant-generics","idea-tr2-platform-single-ethics","idea-tr2-resistance-mechanism-baskets","idea-tr2-combination-utility","idea-tr2-combination-patent-pool","idea-tr2-sequence-registry","idea-tr2-bandit-allocation","idea-tr2-qsp-combo-dosing","idea-tr2-combo-readiness-dossier","idea-tr2-rt-drug-platform","idea-tr2-paediatric-combo-prea","idea-tr2-combination-forecast-tournament","idea-tr2-antagonism-surveillance","idea-tr2-contribution-of-components-mandate","idea-tr2-window-of-opportunity-triplets","idea-moon-automated-combination-discovery"],"cancers":["prostate","tnbc","nsclc","melanoma"],"sections":[],"technologies":["ai-drug-design","functional-drug-testing","organoids","bh3-profiling","dual-payload-adc","crispr-screens"],"targets":[],"drugs":[],"companies":[],"institutions":["nci","broad-institute"],"pathways":[],"terms":["basket-umbrella-platform","adc-sequencing","first-line"],"trials":["stampede","myelomatch"],"people":[],"bottlenecks":[],"keyPapers":["paper-palmer-cell","paper-upadhaya-nat-rev-drug-discov","paper-menden-nat-commun"],"journals":[],"dependsOn":[],"notes":[],"stage":"trials","severity":"major","metrics":[{"label":"Clinical trials of PD-1/PD-L1 inhibitors registered by 2021, the large majority as combinations","value":"5,683 trials","source":"Upadhaya et al., Nature Reviews Drug Discovery 2022","url":"https://doi.org/10.1038/d41573-022-00030-4"},{"label":"Approved combination therapies whose clinical benefit is explained by independent drug action (patient-to-patient variability) without synergy","value":"Most of 15 combinations analysed","source":"Palmer & Sorger, Cell 2017","url":"https://doi.org/10.1016/j.cell.2017.11.009"}],"causes":["Combinatorial growth: n drugs give n(n-1)/2 pairs before doses, schedules and sequences are considered.","Companies preferentially combine their own assets and rarely cross-license for early testing.","Conventional two-arm trials test one combination each and take years.","Preclinical synergy poorly predicts clinical benefit, so prioritisation is weak.","Regulatory paths for combinations of two unapproved agents are complex."],"currentEfforts":["I-SPY 2 (Quantum Leap Healthcare Collaborative) and STAMPEDE run adaptive platform trials that add and graduate combination arms continuously.","NCI ComboMATCH tests biomarker-directed combinations across cooperative groups.","DREAM Challenges and the NCI-DREAM drug combination prediction challenge benchmark computational prioritisation of pairs.","The AstraZeneca-Sanger drug combination screen and DepMap provide public combination and dependency data for hypothesis generation.","Ex vivo functional testing (organoids, explants, BH3 profiling) is used to select combinations for individual patients.","Payload-class switching and dual-payload ADCs address the sequencing problem structurally instead of trial by trial."],"successLooksLike":"Platform trials in each major cancer test dozens of combination arms a year against shared controls, and computational or ex vivo prioritisation has a demonstrated positive predictive value for which combinations succeed in the clinic."},"route":"/bottlenecks/b-combination-space/","neighbours":{"collection":[{"id":"depmap","kind":"collection","name":"DepMap (Cancer Dependency Map)","route":"/collections/depmap/"},{"id":"drugbank-chembl","kind":"collection","name":"DrugBank & ChEMBL","route":"/collections/drugbank-chembl/"}],"idea":[{"id":"idea-reg-combination-price-attribution","kind":"idea","name":"A combination pricing rule so two-drug regimens are not priced as two monopolies","route":"/ideas/idea-reg-combination-price-attribution/"},{"id":"idea-fund-compulsory-combination-access","kind":"idea","name":"A legal right to obtain marketed cancer drugs at cost for combination trials","route":"/ideas/idea-fund-compulsory-combination-access/"},{"id":"idea-fund-rt-io-platform","kind":"idea","name":"A neutral platform trial for radiotherapy plus immunotherapy combinations","route":"/ideas/idea-fund-rt-io-platform/"},{"id":"idea-tr2-combination-utility","kind":"idea","name":"A non-profit phase 1b combination unit that any drug owner can use","route":"/ideas/idea-tr2-combination-utility/"},{"id":"idea-tr2-combination-patent-pool","kind":"idea","name":"A patent pool for combination method-of-use claims","route":"/ideas/idea-tr2-combination-patent-pool/"},{"id":"idea-fund-neutral-platform-sponsor","kind":"idea","name":"A permanent neutral non-profit sponsor for multi-company platform trials","route":"/ideas/idea-fund-neutral-platform-sponsor/"},{"id":"idea-tr2-perpetual-platforms","kind":"idea","name":"A perpetual platform trial in every major cancer, funded as infrastructure","route":"/ideas/idea-tr2-perpetual-platforms/"},{"id":"idea-tr2-organoid-matrix-atlas","kind":"idea","name":"A public atlas of drug-pair responses across a thousand patient-derived organoids","route":"/ideas/idea-tr2-organoid-matrix-atlas/"},{"id":"idea-tr2-public-combination-formulary","kind":"idea","name":"A publicly held library of investigational drugs available for academic combination trials","route":"/ideas/idea-tr2-public-combination-formulary/"},{"id":"idea-tnbc-adc-sequencing-trial","kind":"idea","name":"A randomised trial of antibody-drug conjugate sequence in metastatic triple-negative breast cancer","route":"/ideas/idea-tnbc-adc-sequencing-trial/"},{"id":"idea-data-sequencing-analysis-per-approval","kind":"idea","name":"A real-world sequencing analysis within a year of every new approval","route":"/ideas/idea-data-sequencing-analysis-per-approval/"},{"id":"idea-tr2-sequence-registry","kind":"idea","name":"A registry of every treatment sequence patients actually receive, with outcomes","route":"/ideas/idea-tr2-sequence-registry/"},{"id":"idea-fund-standard-combination-agreement","kind":"idea","name":"A regulator-endorsed standard contract for inter-company combination trials","route":"/ideas/idea-fund-standard-combination-agreement/"},{"id":"idea-tr2-combination-template-agreement","kind":"idea","name":"A standard cross-company combination agreement that takes weeks, not years, to sign","route":"/ideas/idea-tr2-combination-template-agreement/"},{"id":"idea-tr2-rt-drug-platform","kind":"idea","name":"A standing platform for testing new drugs with radiotherapy","route":"/ideas/idea-tr2-rt-drug-platform/"},{"id":"idea-tr1-standing-platform-per-cancer","kind":"idea","name":"A standing platform trial for every major cancer, funded as infrastructure","route":"/ideas/idea-tr1-standing-platform-per-cancer/"},{"id":"idea-bio1-virtual-cell-perturbation","kind":"idea","name":"A virtual cancer cell that predicts what a drug will do before you test it","route":"/ideas/idea-bio1-virtual-cell-perturbation/"},{"id":"idea-bio1-upfront-bypass-combination","kind":"idea","name":"Add the second drug on day one when the escape route is predictable","route":"/ideas/idea-bio1-upfront-bypass-combination/"},{"id":"idea-fund-combination-patent-pool","kind":"idea","name":"An oncology patent pool for combination trials across companies","route":"/ideas/idea-fund-combination-patent-pool/"},{"id":"idea-bio1-collateral-sensitivity-atlas","kind":"idea","name":"An open atlas of collateral sensitivity for every approved targeted drug","route":"/ideas/idea-bio1-collateral-sensitivity-atlas/"},{"id":"idea-tr2-synergy-ranking-engine","kind":"idea","name":"An open engine that ranks every drug pair by predicted synergy before anyone runs a trial","route":"/ideas/idea-tr2-synergy-ranking-engine/"},{"id":"idea-tr2-combination-forecast-tournament","kind":"idea","name":"An open forecasting tournament on which combination trials will succeed","route":"/ideas/idea-tr2-combination-forecast-tournament/"},{"id":"idea-data-open-cell-foundation-model","kind":"idea","name":"An open foundation model of the cancer cell trained on perturbation data","route":"/ideas/idea-data-open-cell-foundation-model/"},{"id":"idea-moon-automated-combination-discovery","kind":"idea","name":"Automated combination discovery: patient-sample screens feeding Bayesian platform trials","route":"/ideas/idea-moon-automated-combination-discovery/"},{"id":"idea-tr2-resistance-mechanism-baskets","kind":"idea","name":"Combination baskets defined by resistance mechanism rather than by cancer type","route":"/ideas/idea-tr2-resistance-mechanism-baskets/"},{"id":"idea-bio1-evolutionary-double-bind","kind":"idea","name":"Design drug pairs where resisting one makes you vulnerable to the other","route":"/ideas/idea-bio1-evolutionary-double-bind/"},{"id":"idea-dual-payload-first","kind":"idea","name":"Dual-payload ADCs in first line to prevent resistance","route":"/ideas/idea-dual-payload-first/"},{"id":"idea-tr2-rwe-combination-emulation","kind":"idea","name":"Emulate combination trials from real-world data to triage which ones to run","route":"/ideas/idea-tr2-rwe-combination-emulation/"},{"id":"idea-tr2-combo-readiness-dossier","kind":"idea","name":"Every approved cancer drug ships with a public combination-readiness data pack","route":"/ideas/idea-tr2-combo-readiness-dossier/"},{"id":"idea-tr1-factorial-dose-finding-for-combinations","kind":"idea","name":"Factorial dose finding for drug combinations instead of full dose of everything","route":"/ideas/idea-tr1-factorial-dose-finding-for-combinations/"},{"id":"idea-tr2-factorial-adjuvant-generics","kind":"idea","name":"Factorial trials that test several cheap generics at once in the adjuvant setting","route":"/ideas/idea-tr2-factorial-adjuvant-generics/"},{"id":"idea-tr2-combo-dose-matrix","kind":"idea","name":"Find the lowest effective doses of both drugs in a combination, not the highest tolerated","route":"/ideas/idea-tr2-combo-dose-matrix/"},{"id":"idea-tr2-organoid-coclinical-arms","kind":"idea","name":"Grow each trial patient's tumour as organoids to decide which platform arm opens next","route":"/ideas/idea-tr2-organoid-coclinical-arms/"},{"id":"idea-bio2-implantable-microdevice-screen","kind":"idea","name":"Implant a tiny device that tests twenty drugs inside the patient's own tumour","route":"/ideas/idea-bio2-implantable-microdevice-screen/"},{"id":"idea-data-in-silico-trials-calibrated","kind":"idea","name":"In silico trials to prioritise combinations, scored against later real trials","route":"/ideas/idea-data-in-silico-trials-calibrated/"},{"id":"idea-tr2-ctdna-futility-gates","kind":"idea","name":"Kill combination arms early using circulating tumour DNA, before waiting for scans","route":"/ideas/idea-tr2-ctdna-futility-gates/"},{"id":"idea-tr2-bandit-allocation","kind":"idea","name":"Let the trial learn: response-adaptive allocation across many combination arms","route":"/ideas/idea-tr2-bandit-allocation/"},{"id":"idea-tr2-paediatric-combo-prea","kind":"idea","name":"Make paediatric combination studies part of every relevant adult cancer drug approval","route":"/ideas/idea-tr2-paediatric-combo-prea/"},{"id":"idea-tr2-qsp-combo-dosing","kind":"idea","name":"Mechanistic computer models to pick combination doses before dosing patients","route":"/ideas/idea-tr2-qsp-combo-dosing/"},{"id":"idea-tr2-contribution-of-components-mandate","kind":"idea","name":"No accelerated approval for a combination without proof each part contributes","route":"/ideas/idea-tr2-contribution-of-components-mandate/"},{"id":"idea-tr2-platform-single-ethics","kind":"idea","name":"One ethics approval and one consent form for a platform trial across countries","route":"/ideas/idea-tr2-platform-single-ethics/"},{"id":"idea-organoid-guided-adc","kind":"idea","name":"Patient-derived organoids to pick ADC payloads","route":"/ideas/idea-organoid-guided-adc/"},{"id":"idea-tr2-payer-combo-cwe","kind":"idea","name":"Payers cover off-label combinations only inside registry-randomised trials","route":"/ideas/idea-tr2-payer-combo-cwe/"},{"id":"idea-payload-switching","kind":"idea","name":"Payload-class switching as the rule for ADC sequencing","route":"/ideas/idea-payload-switching/"},{"id":"idea-tr2-neoadjuvant-combo-platform","kind":"idea","name":"Pre-surgery platform trials that test combinations on pathological response in months","route":"/ideas/idea-tr2-neoadjuvant-combo-platform/"},{"id":"idea-prostate-randomise-the-sequence-not-only-the-drugs","kind":"idea","name":"Randomise the order of treatment, not only the drugs: a strategy platform for metastatic prostate cancer","route":"/ideas/idea-prostate-randomise-the-sequence-not-only-the-drugs/"},{"id":"idea-pdac-ras-inhibitor-combinations-and-sequencing","kind":"idea","name":"RAS inhibitor combinations and sequence: pan-RAS plus G12D-selective, plus chemotherapy, and what to give after progression","route":"/ideas/idea-pdac-ras-inhibitor-combinations-and-sequencing/"},{"id":"idea-tr2-pragmatic-sequence-randomisation","kind":"idea","name":"Registry-embedded randomisation of treatment order in routine care","route":"/ideas/idea-tr2-pragmatic-sequence-randomisation/"},{"id":"idea-bio1-alternating-schedules","kind":"idea","name":"Rotate between drugs on a fixed schedule instead of waiting for failure","route":"/ideas/idea-bio1-alternating-schedules/"},{"id":"idea-tr2-smart-sequencing-adc","kind":"idea","name":"Sequential multiple-assignment randomised trials to find the best order of ADCs","route":"/ideas/idea-tr2-smart-sequencing-adc/"},{"id":"idea-tr2-shared-control-network","kind":"idea","name":"Shared concurrent control arms across sponsors' trials in the same setting","route":"/ideas/idea-tr2-shared-control-network/"},{"id":"idea-tr2-alternating-vs-concurrent","kind":"idea","name":"Test alternating drug schedules against giving both drugs at once","route":"/ideas/idea-tr2-alternating-vs-concurrent/"},{"id":"idea-bio1-immunopeptidome-timing","kind":"idea","name":"Time immunotherapy to the moment targeted drugs make tumours visible","route":"/ideas/idea-bio1-immunopeptidome-timing/"},{"id":"idea-tr2-window-of-opportunity-triplets","kind":"idea","name":"Two-week pre-operative windows to compare combination biology head to head","route":"/ideas/idea-tr2-window-of-opportunity-triplets/"},{"id":"idea-bio2-epigenetic-priming-cold-tumours","kind":"idea","name":"Unmask hidden antigens with a short epigenetic course before immunotherapy","route":"/ideas/idea-bio2-epigenetic-priming-cold-tumours/"},{"id":"idea-tr2-antagonism-surveillance","kind":"idea","name":"Watch routine care for drug combinations that quietly make cancer treatment worse","route":"/ideas/idea-tr2-antagonism-surveillance/"},{"id":"idea-moon-validated-digital-twins","kind":"idea","name":"Whole-patient digital twins validated in prospective randomised trials","route":"/ideas/idea-moon-validated-digital-twins/"}],"cancer":[{"id":"melanoma","kind":"cancer","name":"Melanoma","route":"/cancers/melanoma/"},{"id":"nsclc","kind":"cancer","name":"Non-small-cell lung cancer","route":"/cancers/nsclc/"},{"id":"prostate","kind":"cancer","name":"Prostate cancer","route":"/cancers/prostate/"},{"id":"tnbc","kind":"cancer","name":"Triple-negative breast cancer (TNBC)","route":"/cancers/tnbc/"}],"technology":[{"id":"ai-drug-design","kind":"technology","name":"AI-driven drug & target discovery","route":"/technologies/ai-drug-design/"},{"id":"bh3-profiling","kind":"technology","name":"BH3 profiling (functional apoptosis testing)","route":"/technologies/bh3-profiling/"},{"id":"crispr-screens","kind":"technology","name":"CRISPR functional genomics","route":"/technologies/crispr-screens/"},{"id":"dual-payload-adc","kind":"technology","name":"Dual-payload ADC","route":"/technologies/dual-payload-adc/"},{"id":"functional-drug-testing","kind":"technology","name":"Functional (ex vivo) drug testing","route":"/technologies/functional-drug-testing/"},{"id":"organoids","kind":"technology","name":"Patient-derived organoids","route":"/technologies/organoids/"}],"institution":[{"id":"broad-institute","kind":"institution","name":"Broad Institute of MIT and Harvard","route":"/institutions/broad-institute/"},{"id":"nci","kind":"institution","name":"National Cancer Institute (NIH)","route":"/institutions/nci/"}],"term":[{"id":"adc-sequencing","kind":"term","name":"ADC sequencing","route":"/terms/adc-sequencing/"},{"id":"basket-umbrella-platform","kind":"term","name":"Basket, umbrella, and platform trials","route":"/terms/basket-umbrella-platform/"},{"id":"first-line","kind":"term","name":"Lines of therapy","route":"/terms/first-line/"}],"trial":[{"id":"myelomatch","kind":"trial","name":"myeloMATCH","route":"/trials/myelomatch/"},{"id":"rosewood","kind":"trial","name":"ROSEWOOD","route":"/trials/rosewood/"},{"id":"stampede","kind":"trial","name":"STAMPEDE","route":"/trials/stampede/"}],"paper":[{"id":"paper-amplify-acalabrutinib-venetoclax-nejm-2025","kind":"paper","name":"AMPLIFY: fixed-duration acalabrutinib plus venetoclax, with or without obinutuzumab, versus chemo-immunotherapy in fit CLL patients","route":"/key-papers/paper-amplify-acalabrutinib-venetoclax-nejm-2025/"},{"id":"paper-breakwater-nejm-2025","kind":"paper","name":"BREAKWATER: encorafenib plus cetuximab with chemotherapy as first treatment for BRAF V600E-mutated colorectal cancer","route":"/key-papers/paper-breakwater-nejm-2025/"},{"id":"paper-upadhaya-nat-rev-drug-discov","kind":"paper","name":"Challenges and opportunities in the PD1/PDL1 inhibitor clinical trial landscape","route":"/key-papers/paper-upadhaya-nat-rev-drug-discov/"},{"id":"paper-clear-nejm-2021","kind":"paper","name":"CLEAR: lenvatinib plus pembrolizumab versus sunitinib as first treatment for advanced kidney cancer","route":"/key-papers/paper-clear-nejm-2021/"},{"id":"paper-codebreak-200-lancet-2023","kind":"paper","name":"CodeBreaK 200: sotorasib versus docetaxel in KRAS G12C-mutated lung cancer, a modest win for the first KRAS drug","route":"/key-papers/paper-codebreak-200-lancet-2023/"},{"id":"paper-palmer-cell","kind":"paper","name":"Combination Cancer Therapy Can Confer Benefit via Patient-to-Patient Variability without Drug Additivity or Synergy","route":"/key-papers/paper-palmer-cell/"},{"id":"paper-menden-nat-commun","kind":"paper","name":"Community assessment to advance computational prediction of cancer drug combinations in a pharmacogenomic screen","route":"/key-papers/paper-menden-nat-commun/"},{"id":"paper-ev-302-nejm-2024","kind":"paper","name":"EV-302: enfortumab vedotin plus pembrolizumab replaces chemotherapy as first treatment for advanced bladder cancer","route":"/key-papers/paper-ev-302-nejm-2024/"},{"id":"paper-imbrave150-nejm-2020","kind":"paper","name":"IMbrave150: atezolizumab plus bevacizumab replaces sorafenib as first treatment for advanced liver cancer","route":"/key-papers/paper-imbrave150-nejm-2020/"},{"id":"paper-mariposa-nejm-2024","kind":"paper","name":"MARIPOSA: amivantamab plus lazertinib versus osimertinib as first treatment for EGFR-mutated lung cancer","route":"/key-papers/paper-mariposa-nejm-2024/"},{"id":"paper-planchard-flaura2-osimertinib-chemotherapy-nejm-2023","kind":"paper","name":"Osimertinib with or without chemotherapy in EGFR-mutated advanced NSCLC","route":"/key-papers/paper-planchard-flaura2-osimertinib-chemotherapy-nejm-2023/"},{"id":"paper-beer-prevail-enzalutamide-nejm-2014","kind":"paper","name":"PREVAIL: enzalutamide in metastatic prostate cancer before chemotherapy","route":"/key-papers/paper-beer-prevail-enzalutamide-nejm-2014/"},{"id":"paper-relativity-047-nejm-2022","kind":"paper","name":"RELATIVITY-047: relatlimab plus nivolumab, the first LAG-3 checkpoint combination, in untreated advanced melanoma","route":"/key-papers/paper-relativity-047-nejm-2022/"},{"id":"paper-teply-restore-bipolar-androgen-therapy-lancet-oncol-2018","kind":"paper","name":"RESTORE: bipolar androgen therapy after progression on enzalutamide in metastatic castration-resistant prostate cancer","route":"/key-papers/paper-teply-restore-bipolar-androgen-therapy-lancet-oncol-2018/"},{"id":"paper-rosewood-zanubrutinib-obinutuzumab-follicular-jco-2023","kind":"paper","name":"ROSEWOOD: a phase II randomized study of zanubrutinib plus obinutuzumab versus obinutuzumab monotherapy in patients with relapsed or refractory follicular lymphoma","route":"/key-papers/paper-rosewood-zanubrutinib-obinutuzumab-follicular-jco-2023/"},{"id":"paper-spotlight-lancet-2023","kind":"paper","name":"SPOTLIGHT: zolbetuximab, the first Claudin 18.2 antibody, added to chemotherapy in gastric cancer","route":"/key-papers/paper-spotlight-lancet-2023/"},{"id":"paper-inmind-tafasitamab-lenalidomide-rituximab-follicular-lancet-2026","kind":"paper","name":"Tafasitamab, lenalidomide, and rituximab in relapsed or refractory follicular lymphoma (inMIND): a global, phase 3, randomised controlled trial","route":"/key-papers/paper-inmind-tafasitamab-lenalidomide-rituximab-follicular-lancet-2026/"},{"id":"paper-hallmarks-of-cancer-cell-2000","kind":"paper","name":"The Hallmarks of Cancer: six capabilities every tumour must acquire","route":"/key-papers/paper-hallmarks-of-cancer-cell-2000/"}],"person":[{"id":"emil-frei","kind":"person","name":"Emil \"Tom\" Frei III","route":"/people/emil-frei/"},{"id":"emil-freireich","kind":"person","name":"Emil J Freireich","route":"/people/emil-freireich/"}],"roadmap":[{"id":"targeted-therapy-roadmap","kind":"roadmap","name":"Targeted therapy roadmap: imatinib → designed for resistance → the undruggable drivers fall","route":"/roadmaps/targeted-therapy-roadmap/"}]}}