Ipatasertib

Ipatasertib (GDC-0068, RG7440) is Roche/Genentech's AKT inhibitor that was supposed to be the next big thing in PI3K/AKT pathway oncology. It isn't. After two Phase 3 failures in 2020 - IPATential150 in metastatic castration-resistant prostate cancer (mCRPC) and IPATunity130 in triple-negative breast cancer (TNBC) - Roche effectively shelved it for commercial development [8][9]. AstraZeneca's capivasertib (Truqap) crossed the finish line in November 2023 with FDA approval for hormone receptor-positive, HER2-negative (HR+/HER2-) metastatic breast cancer carrying PIK3CA/AKT1/PTEN alterations, leapfrogging ipatasertib in the same class [3]. Truqap generated $728M in 2025 (+69% YoY per AstraZeneca FY2025 earnings) - meaningful but not yet a top-10 oncology asset. What's left for ipatasertib is largely NCI-sponsored work: MATCH Subprotocol Z1K (NCT06400251) testing it in AKT-altered solid tumors [6], ComboMATCH screening, and head-and-neck cancer combinations with pembrolizumab and chemoradiation [7]. Recent investigator-sponsored Phase 1b/2a work in TAKTIC, SOLTI-1507 IPATHER, and PATHFINDER keeps the biology moving without a clear commercial sponsor pulling it forward [1][2][4]. The drug works at the kinase. The question is whether anyone pays to develop it now that the class has a winner.

Not a novel compound. Ipatasertib has been in clinical development since roughly 2009. Highest historical phase is 3, but the drug failed primary endpoints in two registrational trials in 2020 (IPATential150 in mCRPC, IPATunity130 in TNBC) [8][9]. The DB status field reads phase2; ChEMBL's max_phase=3 captures the historical Phase 3 work. Both are accurate views of the same fact: there is no active Phase 3 program. No active FDA designations (no breakthrough, no orphan, no fast track). No expected regulatory submission, because Roche has not pursued one. Current active trials are all Phase 1/2, mostly NCI-sponsored: MATCH Subprotocol Z1K (NCT06400251, AKT-altered solid tumors, objective response rate (ORR) endpoint, n=35, active not recruiting) [6]; the ComboMATCH screening trial (NCT05564377, n=2,900); HNSCC (head-and-neck squamous cell carcinoma) with chemoradiation Phase 1 (NCT05172245, n=46, recruiting); HNSCC with pembrolizumab Phase 2 (NCT05172258, n=52, active not recruiting) [7]. Recent investigator-sponsored Phase 1b reports refine combination strategies but do not signal a registrational restart: TAKTIC in HR+/HER2- breast with endocrine plus CDK4/6 inhibitor [1], SOLTI-1507 IPATHER in HER2-positive PIK3CA-mutant breast [2], and PATHFINDER Phase 2a in TNBC with non-taxane chemo [4]. Realistic timeline: MATCH Z1K readout 2026-2027. No commercial registration pathway visible. Composition-of-matter patent coverage for ipatasertib dates to filings from the late 2000s; absent a US patent term extension tied to an approved indication (none exists), exclusivity is limited and any out-licensing partner would face a narrow commercial window.

AKT (also called protein kinase B) is a switch that tells cells to grow, survive, and metabolize. When AKT is stuck on, usually because upstream signals from PI3K mutations or loss of PTEN push it there, tumors get a survival and proliferation boost. Roughly 30-40% of breast cancers carry alterations in the PI3K/AKT pathway: activating PIK3CA mutations, AKT1 E17K hotspot mutations, or PTEN loss. Block AKT and you cut off a major growth signal. Ipatasertib is an ATP-competitive small molecule that hits all three AKT isoforms (AKT1, AKT2, AKT3). It sits where ATP would normally bind and shuts the kinase off. Isoform biology matters for both efficacy and toxicity. AKT1 is the dominant driver of proliferation in most epithelial tumors and is the isoform mutated in Proteus syndrome and the AKT1 E17K hotspot. AKT2 is the metabolic isoform - it is required for insulin-stimulated GLUT4 translocation in muscle and adipose, which is why pan-AKT inhibitors produce mechanism-based hyperglycemia. AKT3 is the most brain-enriched isoform with a more limited cancer footprint outside specific amplifications. A pan-isoform inhibitor like ipatasertib accepts the AKT2 metabolic toxicity as the price of broad antitumor coverage. The mechanism is well validated. Genetics: somatic (post-zygotic mosaic) AKT1 E17K mutations cause Proteus syndrome, a gain-of-function genetic disease where overgrowth follows directly from constitutive AKT activation in the affected lineages - the mosaic pattern itself confirms cell-autonomous AKT1 gain-of-function drives the phenotype (germline AKT1 E17K is presumed embryonic-lethal, which is part of why the mutation is only ever seen in mosaic form). Open Targets gives AKT1-Proteus an evidence score of 0.76. Pharmacology: capivasertib's CAPItello-291 Phase 3 in HR+/HER2- breast cancer with PIK3CA/AKT1/PTEN alterations met both PFS endpoints in 2022, leading to FDA approval in November 2023 as Truqap [3]. The biology isn't the problem. Two prior Phase 3 ipatasertib trials hitting this same target failed not because AKT inhibition doesn't work, but because patient selection was wrong: IPATential150's PTEN-loss-by-immunohistochemistry (IHC) strategy in prostate cancer didn't enrich a true biomarker-positive population [8] - IHC under-calls PTEN loss heterogeneity within a tumor and across metastases, and concordance with next-generation sequencing (NGS) calls of genomic PTEN loss is imperfect, so an 'IHC-positive' arm is contaminated with both false positives and tumors with only partial PTEN inactivation. IPATunity130 in TNBC went too broad on selection criteria and the AKT-pathway-altered subgroup did not derive benefit [9].

The most relevant active trial is NCI-MATCH Subprotocol Z1K (NCT06400251), a single-arm Phase 2 in AKT-altered solid tumors. Enrollment target n=35, primary endpoint is objective response rate (ORR). Patients are selected from the larger MATCH/ComboMATCH biomarker-screening framework (NCT05564377, n=2,900 screening cohort) [6]. This is a tissue-agnostic basket design: AKT pathway alterations regardless of tumor histology. The trial is active not recruiting, which suggests the n=35 target is essentially met. Strengths: clean biomarker selection (AKT1/AKT2/AKT3 activating mutations or amplifications) confirmed by NGS rather than IHC, academic infrastructure that runs efficiently for early-signal questions, low cost. Weaknesses: single-arm ORR in n=35 doesn't move regulatory needles for a drug with two prior Phase 3 failures. A 25% ORR could mean real signal or noise in a small basket. The threshold for 'interesting' here is roughly >20% ORR with durability >6 months; anything less is not a re-engagement trigger. NCT05172258 (HNSCC plus pembrolizumab, Phase 2, n=52) is the more commercially relevant trial, with progression-free survival (PFS) as primary endpoint and a checkpoint-inhibitor partner that has differentiation potential [7]. The combination rationale is mechanistic: PI3K/AKT signaling in tumor-infiltrating T cells dampens effector function, and tumor-intrinsic PI3K/AKT activation upregulates immunosuppressive features (PD-L1 expression, regulatory T-cell recruitment). Suppressing AKT in this context may both relieve T-cell exhaustion and increase tumor sensitivity to checkpoint blockade, which is the hypothesis that makes a positive readout interpretable. NCT05172245 is a Phase 1 dose-finding study in HNSCC with chemoradiation. None of these is a registrational program. Roche-sponsored Phase 3 work has been absent since 2020.

Our model gives this drug an 8% chance of eventually reaching approval. That estimate starts from the historical approval rate for Phase 2 drugs in this area - about 13% - then adjusts based on ten facts about the trial and its sponsor. Two factors push the number up: the trial's blinding design and the sponsor's strong record of getting drugs approved; two factors pull it down: weak earlier-phase results and a randomized trial design. The remaining facts fall close to average for this stage, so they leave the final estimate roughly where the historical base rate put it.

Efficacy risk is no longer theoretical. Two Phase 3 trials failed in 2020: IPATential150 in PTEN-loss mCRPC missed its radiographic progression-free survival (rPFS) primary endpoint in the intention-to-treat population [8]; IPATunity130 in TNBC also missed PFS [9]. The PATHFINDER Phase 2a in pretreated TNBC with non-taxane chemo (PMID 40770648) showed modest activity, not transformative [4]. TAKTIC and SOLTI-1507 IPATHER Phase 1b combination data from 2026 are interesting but small and uncontrolled [1][2]. Safety risk: AKT inhibitors carry mechanism-based hyperglycemia (AKT2 is the isoform required for insulin-stimulated glucose uptake in skeletal muscle and adipose, so a pan-isoform inhibitor blunts insulin signaling), rash, and diarrhea. Capivasertib's label carries the same warnings, so the class is manageable. QT prolongation is the specific ipatasertib concern: Poon et al. 2025 (PMID 40671326) quantified delayed QT prolongation risk in preclinical and clinical data [5]. Not a deal-breaker, but it requires monitoring and dose attention. Commercial risk is the killer. Capivasertib (Truqap) is FDA-approved and on-market for the most lucrative AKT-pathway indication (HR+/HER2- metastatic breast cancer with biomarker-positive disease), with AstraZeneca holding the label and sales motion [3]. Truqap reported $728M in FY2025 sales (+69% YoY) [10] - established and growing. Ipatasertib would need to show meaningful differentiation: better efficacy, better safety, or a distinct biomarker niche capivasertib doesn't cover. No data so far supports any of those claims. Execution risk: NCI-sponsored trials run on academic timelines and academic priorities. Without Roche's drug-supply muscle and registrational urgency, the data that emerges will inform biology, not labels.

Watch for biology, not commerce. Ipatasertib is functionally a research tool now. The only readout that would change the picture is NCI-MATCH Z1K (NCT06400251) producing an ORR of 30%+ with durable responses in a biomarker-defined population that capivasertib's label doesn't cover [6]: for example, AKT1 E17K mutant tumors outside breast cancer, or AKT3-amplified or AKT2-mutant tumors. That kind of signal could re-engage Roche or trigger a partnered development effort. Without it, ipatasertib stays a footnote to the AKT inhibitor story that capivasertib finished. Investment angle: zero direct exposure. Roche has effectively written off the R&D investment by walking away after the 2020 Phase 3 failures. Roche's 2024 group revenue was CHF 60.5B (approximately $68B at 2024 average rates) per the Roche 2024 annual report - this drug is rounding-error optionality at that scale. The commercial story for AKT inhibition belongs to AstraZeneca and Truqap, which delivered $728M in FY2025 sales (+69% YoY) [10] - meaningful and growing, but not yet a top-10 oncology asset. When to check back: Q2-Q3 2027 for MATCH Z1K readout. If no positive biomarker signal, this becomes a closed chapter and a useful case study in why second-mover AKT inhibitors needed sharper biomarker strategies - NGS-based pathway calls, not IHC - from the start.