BMS-986205

Linrodostat (BMS-986205) is Bristol-Myers Squibb's selective oral IDO1 inhibitor, designed to lift a metabolic brake that tumors put on T cells by stopping an enzyme that chews up tryptophan in the tumor microenvironment. BMS picked the molecule up in the 2015 acquisition of Flexus Biosciences for $800M upfront plus milestones, and it has been tested in combination with nivolumab across head and neck squamous cell carcinoma (Phase 2, NCT03854032) [1], muscle-invasive bladder cancer (Phase 3 ENERGIZE, NCT03661320) [2], endometrial cancer, and glioblastoma. The whole IDO1 class took a near-fatal hit when Incyte's epacadostat plus pembrolizumab failed Phase 3 in metastatic melanoma in 2018 (ECHO-301) [3], and linrodostat now carries the job of proving the IDO1 thesis isn't dead. BMS has effectively stopped marketing this asset in pipeline communications, though it is letting committed trials run to readout. The program-level phase is 3 (ENERGIZE is the most advanced active study); the featured trial here is the Phase 2 HNSCC investigator-initiated study.

Linrodostat is a small-molecule, selective IDO1 inhibitor that BMS acquired by buying Flexus Biosciences in 2015 [4]. The most advanced active study is ENERGIZE (NCT03661320), a Phase 3 in muscle-invasive bladder cancer testing neoadjuvant (given before surgery to shrink the tumor) chemotherapy alone versus chemotherapy plus nivolumab versus chemotherapy plus nivolumab plus linrodostat in 855 patients [2]. Per ClinicalTrials.gov as of May 2026, ENERGIZE status is active, not recruiting, with actual primary completion recorded as February 16, 2026 and estimated study completion December 30, 2027 - meaning a primary endpoint readout is plausible in late 2026 or 2027, though sponsor communication on timing has been sparse. The featured Phase 2 here (NCT03854032) is a Thomas Jefferson-led investigator-initiated study in stage II-IV HNSCC, n=45 [1]. There are no FDA breakthrough therapy, fast track, orphan, or RMAT designations on file. The Phase 1/2 combination study with nivolumab or nivolumab plus ipilimumab in advanced solid tumors was published by Luke et al. in Clinical Cancer Research in 2025 (n=590 across three study parts) and reported manageable safety with responses primarily in immunotherapy-naive patients [5]. There is no announced near-term regulatory submission. BMS, with roughly $48B in annual revenue per its most recent 10-Q [6], has not flagged linrodostat as a near-term commercial asset in earnings calls for several years.

Cancer cells are good at convincing the immune system to stand down. One trick they use: an enzyme called IDO1 (indoleamine 2,3-dioxygenase 1), which lives inside cells and breaks down tryptophan, an essential amino acid that T cells need to function. When IDO1 is overactive in the tumor microenvironment, tryptophan gets depleted and a metabolite called kynurenine builds up. Both effects starve and suppress the T cells trying to kill the cancer. Linrodostat blocks IDO1, in theory letting T cells get their fuel back and pairing with checkpoint inhibitors like nivolumab (which releases a different brake, PD-1) to restart anti-tumor immunity. The biology is well-mapped: IDO1 is genuinely upregulated in many tumors, kynurenine is genuinely immunosuppressive in T cell assays, and mouse models showed clean synergy with PD-1 blockade. But the human data has been brutal. Incyte's epacadostat combined with Merck's pembrolizumab failed Phase 3 in advanced melanoma (ECHO-301) in 2018 with no PFS or OS benefit, despite a glowing single-arm Phase 2 [3]. That failure prompted hard questions: was epacadostat hitting IDO1 hard enough in the tumor, was IDO1 actually the bottleneck or just one of many redundant pathways, and was tryptophan depletion the right target node at all. The leading hypothesis for ECHO-301's failure has been insufficient IDO1 inhibition at clinically tolerated epacadostat doses. Luke et al. reported that linrodostat plus nivolumab decreased plasma kynurenine across patients regardless of response - confirming target engagement but undermining the 'deeper inhibition fixes it' rescue story, since the drug suppresses the pathway and patients still mostly don't respond [5]. Mechanism is biologically validated. Clinical translation is not.

NCT03854032 is a Phase 2, n=45, investigator-initiated study at Thomas Jefferson University testing nivolumab plus linrodostat in stage II-IV squamous cell head and neck cancer, with objective response rate as the primary endpoint [1]. Status: active, not recruiting. This is a small single-arm signal-finding study, not a registrational design. No comparator arm, no statistical power for survival endpoints, and a 45-patient ceiling means confidence intervals on ORR will be wide. The bigger program-defining trial is ENERGIZE (NCT03661320), a three-arm Phase 3 in muscle-invasive bladder cancer (n=855): neoadjuvant (pre-surgery) chemotherapy alone, plus nivolumab, or plus nivolumab plus linrodostat, with pathological complete response rate (pCR - no visible tumor at the time of surgery) as the primary endpoint [2][7]. ENERGIZE directly tests the question epacadostat could not answer: does adding IDO1 inhibition to a checkpoint inhibitor beat the checkpoint inhibitor alone. Both trials were enrolled before the IDO1 thesis cratered, which is partly why they are still running. The HNSCC Phase 2 has an obvious calibration problem: nivolumab monotherapy ORR in HNSCC sits around 13-20%, so anything under roughly 35% in the combination arm cannot be cleanly distinguished from nivolumab alone in a single-arm study. The completed Phase 1 dose-finding study (NCT03459222) and the Luke et al. 2025 Phase 1/2 expansion (n=590 across parts) selected 200 mg once daily as the maximum tolerated dose; responses were concentrated in immunotherapy-naive patients, and a composite biomarker of low TDO2 (a second tryptophan-degrading enzyme that can compensate when IDO1 is blocked) expression plus high IFN-γ gene signature was associated with response in non-melanoma cohorts [5]. That biomarker is exploratory and was not used to prospectively enrich ENERGIZE or NCT03854032. Important context: the standard of care in advanced bladder cancer has evolved since ENERGIZE was designed. Enfortumab vedotin plus pembrolizumab (EV-pembro) became the first-line standard in metastatic urothelial carcinoma after EV-302, and it is rapidly being studied in the neoadjuvant MIBC setting. ENERGIZE's chemotherapy-only control arm now looks anachronistic against that backdrop - a positive ENERGIZE readout would need to clear not just the chemo control but also justify a position in a treatment landscape that is moving toward EV-pembro-based regimens.

Our model gives this drug a 5% chance of eventually being approved. That starts from a historical baseline - about 13% of Phase 2 drugs in this area make it through - then adjusts based on ten facts about the trial and its sponsor. The factors pulling the estimate up are the trial's blinding design, more secondary endpoints than usual, and the sponsor's strong approval record; the main factor pulling it down is weak earlier-phase results. Everything else came in near average for this stage, leaving the final number well below the baseline.

Efficacy is the dominant risk. The IDO1-plus-checkpoint hypothesis has exactly one Phase 3 readout, and it was negative [3]. Linrodostat could repeat epacadostat's failure if the mechanism simply does not translate from mouse to human. Plausible reasons: redundancy with TDO2 (a second tryptophan-degrading enzyme that can compensate when IDO1 is blocked); alternative kynurenine production routes that bypass IDO1; or the tumor microenvironment finding other ways to suppress T cells (PD-L2, TIM-3, LAG-3, regulatory T cells, MDSCs - immune cells that suppress anti-tumor responses). Patient selection is unsolved. There is no validated prospective biomarker for IDO1-driven immune evasion; the Luke et al. TDO2-low/IFN-γ-high composite is exploratory and post-hoc, and ENERGIZE and NCT03854032 are enrolling unselected populations [5]. In HNSCC, nivolumab monotherapy ORR is already 13-20%, so the Phase 2 needs to land well above that to mean anything. Safety is the lower-tier risk: the Luke et al. Phase 1/2 reported Grade 3/4 adverse events in 50.1-63.4% of patients across cohorts, with dose-limiting toxicities primarily immune-related and hepatotoxicity/rash as the dominant on-target effects [5]. Commercial risk is elevated by competitive evolution: enfortumab vedotin plus pembrolizumab (EV-pembro) is now the first-line standard in metastatic urothelial carcinoma and is migrating into the neoadjuvant MIBC setting. ENERGIZE was designed against a chemotherapy-only control that is becoming obsolete, so even a statistically positive pCR readout would face the question of whether IDO1 plus nivolumab plus chemo beats EV-pembro-based regimens - a comparison ENERGIZE will not deliver. HNSCC and bladder both already have approved checkpoint inhibitors, and linrodostat would need to deliver meaningful incremental survival or response benefit to justify a combination price tag in a cost-sensitive oncology market. Execution risk is low - BMS knows how to run these trials - but BMS's silence on this program in earnings calls is itself a signal about internal expectations.

Mostly noise, with one signal worth tracking. The IDO1 thesis is on life support after epacadostat's failure. BMS is letting committed Phase 2/3 trials run to completion because costs were already sunk and the data has academic value, but the company has effectively stopped describing linrodostat as a near-term commercial asset. The HNSCC Phase 2 (NCT03854032) is investigator-driven, small, and single-arm, and is unlikely to move the needle in either direction. The one data point worth watching is ENERGIZE (NCT03661320) - a clean Phase 3 readout in muscle-invasive bladder cancer that will either rescue the IDO1 thesis or close it. Primary completion was logged on ClinicalTrials.gov as February 16, 2026; topline data is plausible in late 2026 or 2027, though sponsor communication has been notably silent. Even a positive ENERGIZE readout faces a commercial-relevance question because EV-pembro has moved the MIBC standard of care since the trial was designed. For BMS as a company, this asset is rounding error against ~$48B in annual revenue [6]; it does not change the BMS thesis either way. The more interesting question for BioCosm readers is not 'will linrodostat work' but 'what does an IDO1 rescue or failure say about how we should weight the other genetically validated oncology targets that have not yet had their Phase 3 reality check.'