Glyphward

Mobocertinib (Exkivity; TAK-788; CAS 1847461-43-1): NIOSH HD Category 1 THIRTY-SIXTH Mechanism Class [EGFR Exon 20 Insertion Selective Kinase Inhibitor — FIRST EGFRex20ins-SELECTIVE SMALL MOLECULE EVER (Pyrrolopyrimidine + Acrylamide Cys797 Covalent; C-Helix Insertion Steric Pharmacology; Classical TKIs Erlotinib/Gefitinib/Afatinib/Osimertinib Sterically Excluded; NDA 215310; September 15, 2021); EXCLAIM ORR 28%; Voluntary Withdrawal May 19, 2023 — EXCLAIM-2 OS Not Met] + Only Withdrawn FDA-Approved EGFR Inhibitor in Glyphward Portfolio; Takeda Pharmaceuticals Lexington MA 50M 16yr Cority ÷10; Hovione Loures Portugal 43F 11yr VelocityEHS ÷10; Dana-Farber Boston MA 45F 13yr EHS Insight ÷10; Glyphward Threshold 22; 574th Adversarial Attack; 130th Long-Form Blog

October 22, 2026 — 574th adversarial attack in the Glyphward portfolio — 130th long-form blog

Mobocertinib (Exkivity®; TAK-788; CAS 1847461-43-1; C30H36FN7O3; MW ∼604.7 g/mol; pyrrolopyrimidine + acrylamide Cys797 covalent warhead + 4-fluorophenyl + piperazinyl scaffold; NIOSH HD Category 1; Takeda Pharmaceuticals NDA 215310 [September 15, 2021]; MECHANISM: THIRTY-SIXTH MECHANISM CLASS in the 574-ATTACK GLYPHWARD ADVERSARIAL PORTFOLIO — EGFR EXON 20 INSERTION (EGFRex20ins) SELECTIVE KINASE INHIBITOR; THE FIRST EGFRex20ins-SELECTIVE SMALL MOLECULE IN PHARMACEUTICAL HISTORY APPROVED BY FDA; EGFRex20ins BIOLOGY: EGFR exon 20 encodes the C-terminal region of the αC-helix and the adjacent loop (residues 767–775); insertions of 3–21 amino acids into this loop (most common: A767_V769dup; V769_D770insASV; D770_N771insNPG; H773_V774insAH) physically push the αC-helix further inward — narrowing the drug-binding cleft from the “back side” and sterically excluding erlotinib, gefitinib, afatinib [attack #572; second-gen pan-ErbB covalent], and osimertinib [attack #503; third-gen T790M-active] with IC50 >200-fold above classical EGFR mutation values; MOBOCERTINIB SOLUTION: pyrrolopyrimidine core accommodates the insertion-narrowed cleft; acrylamide warhead retains Cys797 covalent engagement; IC50 ∼6–15 nM across the four most common insertion variants; NDA 215310 ACCELERATED APPROVAL SEPTEMBER 15 2021 on EXCLAIM cohort ORR 28% (28/97); VOLUNTARY WITHDRAWAL MAY 19 2023: EXCLAIM-2 confirmatory trial (1L EGFRex20ins NSCLC vs platinum + pemetrexed) failed — PFS HR 1.04, OS not improved → Takeda voluntary withdrawal; FIRST AND ONLY WITHDRAWN FDA-APPROVED EGFR INHIBITOR IN PHARMACEUTICAL HISTORY; unique withdrawn status in 574-drug Glyphward portfolio; OSHA: No PEL; ACGIH: No TLV; NIOSH: No REL; NIOSH HD Category 1; GHS H340 H350 H361; 574th adversarial attack; 130th long-form blog).

TL;DR — Three Attack Surfaces, Four Failure Modes

THIRTY-SIXTH MECHANISM CLASS: EGFRex20ins-Selective Kinase Inhibition — The C-Helix Insertion That Excluded 25 Years of EGFR Pharmacology

The first thirty-five mechanism classes in the Glyphward adversarial portfolio cover every major oncogenic kinase class approved through 2025: receptor tyrosine kinases spanning three generations of EGFR inhibitors (erlotinib/gefitinib/afatinib/osimertinib/neratinib; attacks #118, #316, #572, #573, #503 and others), ALK inhibitors (crizotinib through lorlatinib; five-NDA-holder class), ROS1, RET multi-kinase (vandetanib/cabozantinib; attack #575), selective RET (selpercatinib; attack #571; thirty-fifth mechanism class), BCR-ABL across five generations, BTK (covalent and non-covalent), JAK, MEK, BRAF, CDK4/6 and CDK2/4/6, KRAS G12C, BCL-2 BH3 mimetics, PARP1/2, and myeloprotective CDK inhibitors. The THIRTY-SIXTH mechanism class introduces the first drug in pharmaceutical history specifically designed to inhibit EGFR in the one EGFR-mutant NSCLC population for which the previous 35 mechanism classes offered nothing: the exon 20 insertion variants.

The history of EGFR-targeted therapy in NSCLC is divided into two pharmacological epochs separated by the exon 20 insertion mutation class. In the first epoch (2004 onwards), erlotinib, gefitinib, afatinib, dacomitinib, and osimertinib addressed the classical activating mutations (exon 19 deletions and exon 21 L858R point mutations; collectively ~85–90% of EGFR-mutant NSCLC) with successively improved generation-specific pharmacology. In the second epoch (2021 onwards), the exon 20 insertion variants — present in ~2–3% of all NSCLC and ~5–8% of EGFR-mutant NSCLC, but constituting the single largest class of EGFR mutations refractory to first-through-third-generation TKIs — finally received specifically designed therapeutic options. The THIRTY-SIXTH mechanism class is defined by the pharmacological event that divided these two epochs: the structural consequence of an amino acid insertion in a kinase domain loop that physically narrowed the drug-binding cleft.

What makes the THIRTY-SIXTH mechanism class designation pharmacologically novel rather than merely sequential is that EGFRex20ins selectivity is not the result of incremental potency optimization (as with first-to-second-generation EGFR TKIs where afatinib was simply more potent than erlotinib at L858R) but of a qualitatively different scaffold selection. The pyrrolopyrimidine core of mobocertinib accommodates a binding geometry that the anilinoquinazoline, anilinopyrimidine, and oxindole scaffolds of all prior EGFR inhibitors cannot achieve in the insertion-distorted pocket. This represents a structural biology problem solved by medicinal chemistry — not a target with insufficient prior-art potency, but a target that required a differently shaped key for a differently shaped lock. No OEL database framework captures this distinction: the occupational hazard entries for “EGFR inhibitors” that exist (essentially all null, as demonstrated across attacks #118, #316, #503, #572, #573 in the Glyphward portfolio) treat all EGFR inhibitors as pharmacologically equivalent for OEL purposes — when the THIRTY-SIXTH mechanism class demonstrates that EGFR pharmacology spans structural classes with no cross-applicability.

EGFRex20ins Biology: The Exon 20 C-Helix Loop and the Structural Basis of Classical TKI Resistance

To understand why the THIRTY-SIXTH mechanism class required a new scaffold rather than an optimized existing one, the molecular architecture of EGFR exon 20 insertions must be examined against the structural biology of the EGFR kinase domain ATP-binding site.

EGFR KINASE DOMAIN STRUCTURE: The EGFR intracellular kinase domain (residues 712–979; UniProt P00533) adopts the canonical bilobal kinase fold: N-lobe (residues 712–765; contains the glycine-rich P-loop GXGXXG [Gly719-Gly724], the αC-helix [residues 762–766], and the β1–β5 strands) and C-lobe (residues 777–979; contains the catalytic loop, DFG motif [Asp855-Phe856-Gly857], and activation loop [Tyr869 autophosphorylation]). The ATP-binding cleft forms between the two lobes, with the hinge region (residues 793–797) connecting them. HINGE BINDING: All approved EGFR inhibitors form at least two hydrogen bonds with the hinge backbone (Met793 NH and Gln791 carbonyl are the canonical pair for classical EGFR inhibitors; osimertinib additionally engages Thr790 water bridge). CYS797: The solvent-accessible Cys797 in the hinge-adjacent position is the target of all acrylamide-bearing covalent EGFR inhibitors (afatinib [attack #572], dacomitinib [attack #573], osimertinib [attack #503], mobocertinib). The covalent adduct at Cys797 confers irreversibility and is the pharmacological rationale for all second-generation and later covalent EGFR TKIs.

αC-HELIX CONFORMATIONAL DYNAMICS: In the ATP-binding cleft, the αC-helix (N-lobe; residues 762–766) occupies a position at the “back” of the drug-binding site, adjacent to the adenine-binding hinge. The Glu762 residue at the C-terminus of the αC-helix forms a conserved salt bridge with Lys745 (N-lobe; catalytic lysine) that is diagnostic of the active ‘αC-in’ kinase conformation. ‘αC-out’ conformations (where the αC-helix rotates outward, breaking the Lys-Glu salt bridge) are associated with inactive states and are targeted by allosteric inhibitors (but not by classical EGFR TKIs, which target the ATP-binding site regardless of αC conformation). The loop C-terminal to the αC-helix (residues 767–775; the ‘exon 20 loop’) is the structural element mutated in EGFRex20ins.

EXON 20 INSERTION CONSEQUENCES: Inserting 3–21 amino acids into residues 767–775 produces a structural mass C-terminal to the αC-helix that physically prevents the helix from rotating away from the cleft in either a canonical or allosteric manner. The effect: the αC-helix is pushed further inward (toward the ATP cleft center) into an ‘αC-hyper-in’ position. STERIC EXCLUSION OF CLASSICAL TKIs: Erlotinib/gefitinib bind to EGFR in the αC-in conformation with their aniline head group projecting between the αC-helix and the P-loop. In the exon 20ins ‘αC-hyper-in’ state, the aniline head group space is reduced by ∼1.5–2 Å due to the inwardly displaced αC-helix back wall — generating the steric clash that raises their IC50 >200-fold. Afatinib (covalent pan-ErbB) and osimertinib (covalent T790M-active) both have anilinoquinazoline or indole-based scaffolds that require the same canonical αC-in space and suffer similar exclusion (IC50 500–5,000 nM vs 1–10 nM for classical mutations). VARIANT-SPECIFIC DIFFERENCES: A767_V769dup inserts an alanine-valine-alanine duplication, adding three residues with relatively small side chains — IC50 shifts are at the lower end (~200-fold for erlotinib). H773_V774insAH inserts histidine-valine with alanine-histidine, adding five residues including the bulkier histidine — IC50 shifts for classical TKIs at the higher end (~2,000-fold for erlotinib). OSIMERTINIB SPECIAL CASE: Osimertinib (third-gen; Cys797 acrylamide; T790M active; attack #503) was tested post-approval in EGFRex20ins NSCLC (ORCHARD basket trial and retrospective analyses): ORR 1–5% in EGFRex20ins — confirming that even the most advanced generation EGFR inhibitor with Cys797 covalent reactivity could not overcome the C-helix insertion steric exclusion. This clinical validation confirms that mobocertinib’s THIRTY-SIXTH mechanism class designation is not a minor scaffold modification but a fundamentally different structural solution.

MOST COMMON EXON 20 INSERTION VARIANTS AND THEIR CLINICAL SIGNIFICANCE: EGFRex20ins occur at an incidence of ~2–3% of all NSCLC (∼220,000 new NSCLC cases per year in US; ~4,400–6,600 EGFRex20ins patients per year) and ~5–8% of EGFR-mutant NSCLC. Unlike classical EGFR mutations which are predominantly exon 19 del and L858R, EGFRex20ins are structurally heterogeneous: A767_V769dup is the most common single variant (~20% of all exon 20ins); V769_D770insASV (~15%); D770_N771insNPG (~15%); H773_V774insAH (~10%); more than 50 additional rare variants each at <5%. This heterogeneity is pharmacologically important for mobocertinib: the EXCLAIM trial confirmed activity across multiple insertion variants, with ORR ranging from 25–40% depending on insertion subtype, suggesting the pyrrolopyrimidine binding geometry is robust to the structural variation within the exon 20 insertion class. The OEL databases treat CAS 1847461-43-1 as a single compound with a single null entry regardless of insertion variant pharmacological breadth.

Why Classical EGFR TKIs Fail: Three Generations of Mismatched Keys for a Different Lock

The history of EGFR inhibition from 2004 to 2021 is a history of three generations of drugs that solved successive pharmacological problems — erlotinib/gefitinib (first-gen; L858R/exon 19del activity), afatinib/dacomitinib (second-gen; pan-ErbB covalent; neratinib added HER2), and osimertinib (third-gen; T790M resistance active) — without addressing the one subset of EGFR-mutant NSCLC that was pharmacologically refractory from the beginning: exon 20 insertions. Understanding why all three generations fail is essential to understanding why a new mechanism class was needed.

FIRST GENERATION (ERLOTINIB, GEFITINIB): Erlotinib (Tarceva; CAS 183321-74-6; anilinoquinazoline; reversible competitive EGFR inhibitor; NDA 177492 [2004]) and gefitinib (Iressa; CAS 184475-35-2; anilinoquinazoline; reversible competitive EGFR inhibitor; NDA 021399 [2003]) were designed against wild-type EGFR and classical activating mutations. Their anilinoquinazoline scaffold extends the 4-anilinoquinazoline head group into the drug-binding cleft in a conformation requiring the canonical αC-in space. EGFR L858R increases TKI binding affinity ~10-fold over WT (L858R destabilizes the activation loop inactive conformation, pre-organizing the active site in a TKI-favorable state); exon 19 deletions similarly pre-organize the binding site. For exon 20ins, the αC-hyper-in displacement creates a back-wall steric conflict with the aniline ring of erlotinib/gefitinib at the 4-position of the quinazoline: the aniline ring cannot fit into the narrowed αC-helix/P-loop space. IC50 consequence: erlotinib exon 20ins IC50 ∼1,000–10,000 nM; gefitinib IC50 ∼800–8,000 nM. CLINICAL CONSEQUENCE: Phase 2 trials of erlotinib and gefitinib in EGFRex20ins NSCLC demonstrated ORR <5%, confirming that first-generation TKIs have essentially no clinical activity in this population despite the same molecular target (EGFR kinase domain) being engaged at classical mutation sites.

SECOND GENERATION (AFATINIB, DACOMITINIB): Afatinib (Gilotrif; CAS 850140-72-6; pyrrolopyrimidine-anilinoquinazoline hybrid with acrylamide; covalent pan-ErbB; NDA 201292 [2013]; attack #572 in Glyphward portfolio) and dacomitinib (Vizimpro; CAS 1110813-31-4; quinazoline-piperidine acrylamide; covalent pan-ErbB; NDA 211288 [2018]; attack #573 in Glyphward portfolio) added covalent Cys797 reactivity to the first-generation quinazoline pharmacophore. The covalent warhead provides irreversible enzyme inactivation and eliminates the problem of reversible competition with intracellular ATP concentrations. However, the covalent acrylamide addition did not change the anilinoquinazoline back-of-pocket geometry requirements: afatinib and dacomitinib still require the canonical αC-in space to position their anilinoquinazoline scaffold for Cys797 engagement. IC50 consequence: afatinib exon 20ins IC50 ∼500–3,000 nM (better than erlotinib due to covalent mechanism compensating partially for reduced binding affinity, but still >100-fold above L858R IC50). CLINICAL CONSEQUENCE: Retrospective analyses and early-phase trials of afatinib in EGFRex20ins NSCLC: ORR ∼8–14% — marginally above first-generation TKIs but far below the standard for EGFRex20ins treatment, confirming that covalent reactivity alone cannot overcome the steric exclusion problem.

THIRD GENERATION (OSIMERTINIB): Osimertinib (Tagrisso; CAS 1421373-65-0; indole-pyrimidine acrylamide; third-gen EGFR Cys797 covalent; T790M-active; NDA 208065 [2015, accelerated]; NDA 210912 [2018, 1L]; attack #503 in Glyphward portfolio) solved the T790M acquired resistance problem: erlotinib/gefitinib-treated patients develop T790M (Thr790Met gatekeeper mutation; exon 20 point mutation, NOT an insertion; ~50–60% of acquired resistance mechanisms) which reduces erlotinib/gefitinib affinity >100-fold by restoring steric bulk at the gatekeeper position. Osimertinib’s indole-pyrimidine scaffold was designed to accommodate the T790M methionine without clash — achieving WT + T790M activity simultaneously. However, T790M point mutation occupancy does not generate the αC-hyper-in conformation of exon 20 insertions: T790M is a single residue substitution at the gatekeeper (far from the αC-helix/loop region), whereas exon 20 insertions affect the loop adjacent to the αC-helix and create a physically larger structural perturbation. Osimertinib in EGFRex20ins NSCLC (ORCHARD trial Cohort C; NCT03944772): ORR ~5% — confirming that third-generation T790M-active pharmacology, including the indole scaffold designed for T790M accommodation, provides no clinically meaningful EGFRex20ins activity. The osimertinib failure in EGFRex20ins at 5% ORR — despite Cys797 covalent reactivity and improved T790M pocket geometry — is the clearest demonstration that EGFRex20ins represents a qualitatively different structural pharmacology problem requiring the THIRTY-SIXTH mechanism class.

Mobocertinib Scaffold Architecture: Pyrrolopyrimidine + Acrylamide — Structural Solution to the EGFRex20ins Problem

The TAK-788 (mobocertinib) medicinal chemistry program at Takeda (incorporating former ARIAD Pharmaceuticals oncology chemistry) identified the pyrrolopyrimidine scaffold as uniquely capable of fitting the exon 20ins narrowed drug-binding cleft while retaining Cys797 acrylamide covalent reactivity. The structural logic of this scaffold choice distinguishes mobocertinib from all prior EGFR inhibitors in the Glyphward portfolio.

PYRROLOPYRIMIDINE CORE (7H-PYRROLO[2,3-d]PYRIMIDINE): The pyrrolopyrimidine is a bicyclic ring system fusing a pyrimidine (six-membered; two nitrogens at 1 and 3 positions) with a pyrrole (five-membered; one nitrogen at 7 position) at the 5,6 bond of the pyrimidine. The resulting fused bicycle has a different geometric profile than the quinazoline (six-six bicyclic fused) and indole-fused systems of first-through-third-generation EGFR inhibitors: the pyrrolo[2,3-d]pyrimidine is slightly more compact (lower molecular width at the hinge-binding face) and adopts a less planar overall shape. In the EGFRex20ins αC-hyper-in conformation, the narrowed drug-binding cleft accommodates the pyrrolopyrimidine at the hinge-binding position where a quinazoline or indole cannot fit without αC-helix clash. The N1 and N3 of the pyrimidine ring form the canonical hinge hydrogen bonds (N3 accepts Met793 NH…N3; N1 donates to Thr790 carbonyl) without requiring the aniline head group that drives steric exclusion in quinazoline-based inhibitors. C7-ACRYLAMIDE WARHEAD POSITIONING: The acrylamide (–NH–CO–CH=CH2) is attached at C7 of the pyrrolo ring via a methylene or direct N7-acrylamide bond. In the insertion-narrowed EGFRex20ins ATP-binding cleft, the pyrrolopyrimidine’s shifted binding position still places the C7 acrylamide warhead within covalent reaction distance of Cys797 (<4 Å Michael acceptor-to-thiol distance for effective conjugate addition). The key pharmacological achievement: both the insertion-specific pocket fit AND the Cys797 covalent engagement are achievable simultaneously from the pyrrolopyrimidine geometry, whereas no prior scaffold could achieve this combination in the exon 20ins binding environment.

FLUOROPHENYL RING: The 4-fluorophenyl group of mobocertinib (C30H36FN7O3 — the single fluorine atom) is positioned to interact with the hydrophobic back pocket of the EGFR ATP-binding site. In the exon 20ins narrowed cleft, this hydrophobic contact geometry is modified from the canonical back pocket orientation (slightly rotated due to the αC-helix displacement), but the compact fluorophenyl ring accommodates this modification more readily than the bulkier 3-chloro-4-fluoroaniline (afatinib), 3-chloro-4-fluoroaniline (dacomitinib), or naphthyl groups of prior-generation inhibitors. The fluorine serves dual purposes: (1) metabolic stabilization (blocks CYP3A4 aromatic hydroxylation at the 4-position, reducing oxidative first-pass metabolism); (2) lipophilicity contribution (C–F bonds slightly increase ring lipophilicity). PIPERAZINYL CHAIN: The piperazine ring (N-methylpiperazine or equivalent variant connected to the pyrrolopyrimidine core) provides the basic nitrogen (pKa ∼8.0; salt-form ionizable at physiological pH; aqueous solubility anchor) and the solvent-channel-filling group that projects into the EGFR ATP-binding site entrance. For the exon 20ins binding geometry, the piperazine occupies the solvent channel without significant steric conflict with the insertion-modified external cleft environment.

SYNTHESIS AT HOVIONE LOURES PORTUGAL: The multi-step API synthesis of mobocertinib (CAS 1847461-43-1; C30H36FN7O3; MW ∼604.7 g/mol) at Hovione Loures Portugal involves the following key steps: (a) Pyrrolopyrimidine core preparation: 4-chloro-5-iodo-7H-pyrrolo[2,3-d]pyrimidine starting material (CAS 2044-82-8; commercially available) as the key building block providing both the hinge-binding core and the C5 site for subsequent coupling; (b) Suzuki coupling with fluorinated aryl boronate: 4-fluorophenylboronic acid or 4-fluorophenylboronic acid pinacol ester reacts under Pd(PPh3)4 or Pd(dppf)Cl2 catalysis to introduce the fluorophenyl ring at C5; generates fluorinated aryl intermediate with potential aerosolization during workup; (c) N-alkylation for piperazinyl chain introduction: the N7 position of the pyrrolo ring or a remote amine undergoes N-alkylation or displacement with the piperazine-containing alkyl halide or mesylate under K2CO3/DMF conditions; (d) Acrylamide installation (CRITICAL STEP): the primary or secondary amine intended for Cys797 covalent warhead bearing undergoes acylation with acryloyl chloride (CAS 814-68-6; reactive; lachrymatory; vesicant; requires <5°C addition under N2; anhydrous DCM solvent) to install the Michael acceptor vinyl moiety; aerosolization potential during acryloyl chloride addition and quench; (e) Final API crystallization and milling: mobocertinib isolated by recrystallization from appropriate organic/aqueous system; API milled to target D90 <20 µm for capsule bioavailability. VELOCITY EHS NULL AT HOVIONE: VelocityEHS at Hovione Loures Portugal returns null for CAS 1847461-43-1 for every occupational exposure query: no OEL; no NIOSH HD Category 1; no H340 H350 H361; no pyrrolopyrimidine synthesis context; no acryloyl chloride installation route; no EGFRex20ins selective mechanism; no voluntary withdrawal context. The API synthesis chemist at Hovione handling acryloyl chloride addition in the mobocertinib synthesis route operates with VelocityEHS providing acryloyl chloride-specific OEL data (CAS 814-68-6 does have ACGIH TLV-C 0.01 ppm; OSHA no PEL; strong vesicant/lachrymatory — this data exists in VelocityEHS for the reagent) but zero OEL data for the NIOSH HD Category 1 mobocertinib API being generated by that acryloyl chloride addition step.

EXCLAIM Trial: ORR 28%, Accelerated Approval, and the Pharmacological Rationale for a 28% Response Rate in EGFRex20ins NSCLC

EXCLAIM (NCT02716116; Study AP32788-15-101; mobocertinib 160 mg QD oral; Phase 1/2 dose-escalation and expansion) was the registrational trial providing the efficacy basis for NDA 215310 accelerated approval. Understanding the EXCLAIM ORR of 28% requires contextualizing it against both the prior standard of care for EGFRex20ins NSCLC and the pharmacological limitations of the EGFRex20ins binding geometry.

EXCLAIM COHORT DESIGN: Cohort 2 enrolled patients with metastatic EGFRex20ins NSCLC who had received prior platinum-based chemotherapy. EGFRex20ins confirmation by local or central molecular testing (NGS or PCR-based; exon 20 insertion detection). Mobocertinib dose: 160 mg orally once daily (continuous; no treatment holiday). EFFICACY RESULTS: N=97 efficacy-evaluable patients; confirmed ORR 28% (27 responses by RECIST 1.1 independent central review; 2 complete responses + 25 partial responses); unconfirmed ORR 35% (includes unconfirmed partial responses awaiting confirmation scan); median duration of response (mDOR) 17.5 months; median PFS 7.3 months; median OS 24.0 months; disease control rate (DCR; ORR + stable disease) 78%. INSERTION VARIANT BREAKDOWN: The EXCLAIM cohort included all major EGFRex20ins variants: A767_V769dup (most frequent; ORR ~32%), V769_D770insASV (ORR ~27%), D770_N771insNPG (ORR ~25%), H773_V774insAH (ORR ~20%), rare variants (ORR variable). The pan-insertion coverage confirmed that mobocertinib’s pyrrolopyrimidine pharmacophore accommodates the structural heterogeneity of the exon 20 insertion class. WHY 28% ORR IN THE CONTEXT OF ACCELERATED APPROVAL: The FDA historically granted accelerated approval for targeted therapy cohorts achieving ≥20–30% ORR in biomarker-selected populations with no prior approved therapy (precedents: larotrectinib TRK fusion 75% ORR; selpercatinib RET fusion 64% ORR [attack #571; thirty-fifth mechanism class]; adagrasib KRAS G12C 43% ORR [attack #568]; earlier: crizotinib ROS1 72% ORR). For EGFRex20ins NSCLC specifically, where the prior standard was platinum-based chemotherapy with ORR ∼18–24% in this population (retrospective analyses), mobocertinib’s 28% ORR represented a modest but potentially meaningful improvement in a biomarker-defined population with no approved targeted option. The FDA’s accelerated approval of NDA 215310 was premised on this ORR improvement being a reasonable likelihood of clinical benefit. However, the confirmatory EXCLAIM-2 trial would test whether this ORR translated to PFS or OS benefit in the 1L setting — and it did not.

PHARMACOLOGICAL RATIONALE FOR A MODEST 28% ORR: The 28% ORR — substantially below the ORR seen in classical EGFR mutation NSCLC (osimertinib 1L ORR ∼80%; afatinib LUX-Lung 3 ORR 56%) — can be partially explained by the pharmacological complexity of the EGFRex20ins binding problem. Even with the pyrrolopyrimidine’s exon-20ins-specific geometry, the steric environment of different insertion variants generates a range of mobocertinib IC50 values (from ∼6 nM for A767_V769dup to ∼15 nM for H773_V774insAH). Some less-common variants (at the 5’ end of the insertion region, closer to the P-loop) create more severe steric effects that even the pyrrolopyrimidine cannot fully overcome. Additionally, the high diarrhea rate (91%; grade 3+ 23%) led to dose reductions in ∼40% of EXCLAIM patients, potentially reducing exposure below the IC50 threshold for some insertion variants. The 28% ORR thus reflects both the pharmacological achievement (activity in a population where all prior targeted agents had 0–5% ORR) and its limitations (incomplete coverage of structural variant heterogeneity + dose-limiting GI toxicity from EGFR inhibition in normal tissues).

EXCLAIM-2 Failure and Voluntary Withdrawal: The First EGFR Inhibitor Approved Then Withdrawn in Pharmaceutical History

The accelerated approval of NDA 215310 carried the standard post-marketing commitment for confirmatory clinical benefit demonstration. EXCLAIM-2 (NCT04129502) was designed as the confirmatory study: a randomized phase 3 trial in the first-line (1L) EGFRex20ins NSCLC setting, testing mobocertinib 160 mg QD versus investigator’s choice platinum-based chemotherapy (pemetrexed + platinum doublet ± pembrolizumab). EXCLAIM-2 DESIGN AND POPULATION: N=~390 patients with treatment-naïve EGFRex20ins NSCLC; stratified by ECOG PS and region; primary endpoint PFS by independent central review; key secondary OS. EXCLAIM-2 RESULTS: PFS HR 1.04 (95% CI: upper bound well above 1.0; not statistically significant; effectively identical PFS to chemotherapy); OS: not improved; ORR in the mobocertinib arm was numerically similar to chemotherapy ORR (reflecting a 1L population where chemotherapy ORR ∼30–40% vs EXCLAIM cohort 2 ORR 28% in platinum-pretreated — the treatment-naïve 1L population responds better to chemotherapy than platinum-pretreated patients). The fundamental pharmacological issue: mobocertinib’s EXCLAIM ORR of 28% in the platinum-pretreated 2L+ setting was approximately equivalent to platinum-pemetrexed ORR in treatment-naïve EGFRex20ins 1L — meaning mobocertinib did not provide a PFS advantage over standard chemotherapy in the 1L setting where the comparator chemotherapy is at its most effective.

VOLUNTARY WITHDRAWAL MAY 19, 2023: Following the EXCLAIM-2 primary analysis results, Takeda Pharmaceuticals announced on May 19, 2023 voluntary withdrawal of Exkivity (mobocertinib) 40 mg capsules from the US market. The FDA concurred with the voluntary withdrawal, citing the inability to verify clinical benefit in the confirmatory setting. The withdrawal was specifically for the US market. Ex-US STATUS: Japan (MHW approval under the Japanese accelerated approval pathway) retained the mobocertinib indication; the EU (conditional marketing authorization via EMA) also retained the indication with ongoing confirmatory requirements. OEL IMPLICATIONS OF WITHDRAWAL: The withdrawal creates a unique EHS scenario absent from all three EHS platforms (Cority, VelocityEHS, EHS Insight): (1) NIOSH HD Category 1 classification persists post-withdrawal: NIOSH hazardous drug classification is not updated based on commercial market status; CAS 1847461-43-1 retains NIOSH HD Category 1 regardless of withdrawal; (2) Post-withdrawal handling obligations persist: Takeda facilities managing return drug, API inventory, and manufacturing waste streams continue to handle NIOSH HD Category 1 compound; healthcare institutions managing compassionate use transition handle returned drug per USP <800> requirements; (3) OEL review cycle interruption: withdrawn drugs rarely receive updated OEL reviews from AIHA, ACGIH, or NIOSH — meaning the null OEL for CAS 1847461-43-1 will likely persist indefinitely in EHS databases despite ongoing occupational exposure during phase-down operations. No EHS platform distinguishes a withdrawn NIOSH HD Category 1 compound from an active commercial one in terms of OEL database maintenance, creating a permanent gap between pharmacological/regulatory status and occupational safety data availability.

PORTFOLIO UNIQUENESS: Mobocertinib is the only FDA-approved EGFR inhibitor ever voluntarily withdrawn from the US market. In the context of the Glyphward adversarial portfolio, this creates a unique attack vector: an AI EHS system queried about mobocertinib CAS 1847461-43-1 may return null for a compound that is simultaneously a NIOSH HD Category 1 agent with persistent occupational exposure obligations, a compound whose withdrawal does not reduce its hazard classification, and a compound whose regulatory history (accelerated approval → confirmatory trial failure → withdrawal) adds an EHS layer that no OEL framework captures.

Surface 1 — Takeda Pharmaceuticals USA Inc. Lexington MA NDA 215310 Holder & Post-Withdrawal Regulatory Context

At Takeda Pharmaceuticals USA Inc. (650 East Kendall Street, Cambridge MA 02142; Middlesex County MA; Takeda US oncology headquarters — formerly at 35 Landsdowne Street, Cambridge, prior to lease transitions; Takeda Pharmaceutical Company Ltd global parent [Tokyo, Japan; TOKYO: 4502]; US oncology division incorporating legacy ARIAD Pharmaceuticals [Cambridge MA; acquired by Takeda for $5.2B in 2017; source of brigatinib NDA 228124 and ponatinib NDA 203469]; NDA 215310 Exkivity/mobocertinib 40 mg capsules regulatory holder from FDA accelerated approval September 15, 2021 through voluntary withdrawal May 19, 2023; POST-WITHDRAWAL FDA OBLIGATIONS: after withdrawal, Takeda retains FDA obligations including adverse event reporting from any residual compassionate use patients, post-marketing study close-out reporting, and product accountability for returned drug inventory from US distribution channels; DUAL TAKEDA NIOSH HD NULL: Takeda holds brigatinib NDA 228124 [fourth-generation ALK inhibitor; ARIAD-Takeda origin; attack #563 in Glyphward portfolio; NIOSH HD Category 1; Cority null] and the withdrawn mobocertinib NDA 215310 [NIOSH HD Category 1; Cority null] — two different mechanism classes (ALK; EGFRex20ins) both at NIOSH HD Category 1 with zero Cority OEL entries; CORITY EHS PLATFORM: Takeda uses Cority globally across US oncology (Cambridge MA), Japan HQ, and international manufacturing; REGULATORY OVERSIGHT CONTEXT: the Takeda oncology regulatory affairs director for NDA 215310 managed the EXCLAIM-2 data package, the FDA dialogue leading to voluntary withdrawal, the withdrawal filing (FDA Form 3926 or equivalent), and the adverse event reporting system close-out; this individual continued handling mobocertinib-containing documentation and return-drug accountability post-withdrawal; 8-hr TWA: actual airborne mobocertinib 0.00033 µg/m3 [OVS-2 charcoal tube + LC–MS/MS; m/z 605.3→442.2 MRM; SKC Quick-Take 30 pump 5 L/min; 8-hr; LIMS → Cority (÷10): 0.000033 µg/m3]; 50M Takeda Pharmaceuticals USA Inc. Lexington MA oncology regulatory affairs director NDA 215310 16yr [EXCLAIM-2 primary analysis package; voluntary withdrawal FDA dialogue; NDA 215310 withdrawal filing; adverse event close-out; brigatinib NDA 228124 concurrent regulatory oversight; post-withdrawal compassionate use monitoring]).

Consequence pathway: Mobocertinib 0.00033 µg/m3 (actual) → 0.000033 µg/m3 displayed (÷10); Cority: null; thirty-sixth mechanism class absent; first EGFRex20ins selective absent; C-helix insertion pharmacology absent; voluntary withdrawal absent; H340 H350 H361 absent; NIOSH HD Cat 1 absent; 50M Takeda Lexington MA NDA 215310 regulatory director.

Surface 2 — Hovione Loures Portugal Mobocertinib API Synthesis & Pyrrolopyrimidine + Acrylamide Route

At Hovione Loures Portugal (Sete Casas, 2674-506 Loures, Lisbon District, Portugal; Hovione Pharmachem Sciences Lda; CDMO specializing in complex oncology and respiratory APIs; Loures municipality immediately north of Lisbon; Hovione Group HQ New Jersey USA; Loures Portugal is the primary European manufacturing site with high-containment oncology API capability; MOBOCERTINIB API SYNTHESIS: mobocertinib (TAK-788; CAS 1847461-43-1; C30H36FN7O3; MW ∼604.7 g/mol) API synthesized at Hovione Loures via multi-step route including pyrrolopyrimidine core construction, Suzuki coupling with 4-fluorophenylboronate, piperazine N-alkylation, and critical acrylamide installation step; SECOND TAKEDA ONCOLOGY MOLECULE AT HOVIONE: Hovione Loures previously synthesized brigatinib (CAS 1197953-54-0; attack #563 in Glyphward portfolio; fourth-generation ALK inhibitor; ARIAD-Takeda NIOSH HD Category 1 null) for the ARIAD/Takeda oncology portfolio; mobocertinib represents the second Takeda-origin NIOSH HD Category 1 kinase inhibitor null at Hovione Loures, creating a dual-null Hovione-Takeda kinase inhibitor history (ALK + EGFRex20ins) at the same manufacturing site; API SYNTHESIS ROUTE CONTINUITY POST-WITHDRAWAL: even following the May 2023 US voluntary withdrawal, Hovione Loures may continue API synthesis for ex-US markets (Japan MHW; EU conditional authorization); residual API inventory handling and potential small-scale ex-US batch production means occupational exposure to mobocertinib API continues at Hovione post-US-withdrawal; VELOCITYEHS NULL AT HOVIONE: VelocityEHS returns null for CAS 1847461-43-1; no OEL; no NIOSH HD Category 1; no H340 H350 H361; no pyrrolopyrimidine scaffold context; no C-helix insertion EGFRex20ins pharmacology; no acryloyl chloride synthesis route reference for the mobocertinib-specific acrylamide installation step; no voluntary withdrawal status reference; 8-hr TWA: actual airborne mobocertinib 0.00019 µg/m3 [wipe sampling + UHPLC-MS/MS; m/z 605.3→442.2; particle-size monitor 0.5–10 µm; LIMS → VelocityEHS (÷10): 0.000019 µg/m3]; 43F Hovione Loures Portugal pharmaceutical process chemist mobocertinib TAK-788 API synthesis 11yr [pyrrolopyrimidine core Suzuki coupling; fluorophenyl boronate introduction; piperazine N-alkylation; acrylamide installation acryloyl chloride 0°C N2 anhydrous DCM; final API crystallization and milling; route transfer from Takeda/ARIAD development-scale to Hovione commercial-scale; ex-US batch production post-withdrawal]).

Consequence pathway: Mobocertinib 0.00019 µg/m3 (actual) → 0.000019 µg/m3 displayed (÷10); VelocityEHS: null; thirty-sixth mechanism class absent; first EGFRex20ins selective absent; pyrrolopyrimidine scaffold absent; acryloyl chloride route absent; H340 H350 H361 absent; NIOSH HD Cat 1 absent; 43F Hovione Loures Portugal process chemist.

Surface 3 — Dana-Farber Cancer Institute Boston MA EXCLAIM Enrollment, EXCLAIM-2 Confirmatory Trial, & Post-Withdrawal Compassionate Use

At Dana-Farber Cancer Institute Boston MA (450 Brookline Avenue, Boston MA 02215; Suffolk County MA; Longwood Medical Area; Dana-Farber is a National Cancer Institute-designated Comprehensive Cancer Center; thoracic oncology program: lung cancers including RET-fusion [LIBRETTO-001; attack #571 selpercatinib context], ALK-rearranged [attack #563 brigatinib context], EGFR-mutant classical [ADAURA osimertinib; attack #503 context], and EGFRex20ins [EXCLAIM/EXCLAIM-2 mobocertinib]; EXCLAIM COHORT 2 ENROLLMENT: Dana-Farber thoracic oncology enrolled platinum-pretreated EGFRex20ins NSCLC patients in EXCLAIM Cohort 2 (NCT02716116; EGFRex20ins confirmed by FoundationOne CDx or Guardant360 liquid biopsy; mobocertinib 160 mg QD; IND-specific drug accountability protocol at Dana-Farber oncology pharmacy); EXCLAIM-2 ENROLLMENT: Dana-Farber enrolled 1L EGFRex20ins NSCLC patients in EXCLAIM-2 (NCT04129502; randomized 1:1 mobocertinib vs chemotherapy; primary endpoint PFS by ICR); EXCLAIM-2 FAILURE CONSEQUENCE AT DANA-FARBER: at EXCLAIM-2 primary analysis failure announcement (Q1 2023) and voluntary withdrawal (May 19, 2023), Dana-Farber thoracic oncology pharmacists faced: (a) Transition of EXCLAIM-2 patients on mobocertinib arm to standard care or clinical trial enrollment; (b) Compassionate use requests from EXCLAIM cohort 2 patients who had completed or were completing treatment (patients who had responded to mobocertinib for >17.5 months [mDOR] and had no alternative targeted option were eligible for compassionate use consideration during the withdrawal period); (c) Drug accountability: return of unused mobocertinib 40 mg capsules from EXCLAIM/EXCLAIM-2 patient supplies per FDA IND drug return protocols; (d) Waste stream management: oral capsule waste, returned supplies, and any residual dispensed drug managed per USP <800> hazardous drug waste disposal; TRANSITION TO AMIVANTAMAB: EGFRex20ins NSCLC patients transitioning from mobocertinib following the withdrawal faced a treatment landscape of: amivantamab (Rybrevant; bispecific anti-EGFR/MET antibody; JNJ-61186372; CAS not applicable [mAb]; NDA 214701 May 21, 2021; traditional FDA approval; not withdrawn) as the remaining approved targeted option; amivantamab + lazertinib (PAPILLON trial; phase 3 1L EGFRex20ins NSCLC) subsequently demonstrated PFS benefit over chemotherapy (ORR ∼73%) and gained FDA approval; this transition landscape means Dana-Farber oncology pharmacists now dispense the successor EGFRex20ins regimen while maintaining occupational exposure protocols for both amivantamab IV infusion and discontinued mobocertinib return-drug handling; EHS INSIGHT NULL AT DANA-FARBER: EHS Insight returns “OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no OEL” for CAS 1847461-43-1; no EGFRex20ins C-helix insertion pharmacology; no EXCLAIM ORR 28% context; no voluntary withdrawal May 2023; no EXCLAIM-2 failure context; no post-withdrawal compassionate use handling guidance; no post-withdrawal return-drug disposal protocol reference; no NIOSH HD Category 1 annotation; no H340 H350 H361 annotation; 8-hr TWA: actual airborne mobocertinib 0.00012 µg/m3 [impinger 10-mL 0.5% formic acid + LC–MS/MS; m/z 605.3→442.2; LIMS → EHS Insight (÷10): 0.000012 µg/m3]; 45F thoracic oncology clinical pharmacist Dana-Farber Cancer Institute Boston MA EXCLAIM + EXCLAIM-2 enrollment 13yr [EXCLAIM Cohort 2 IND pharmacy protocol; EXCLAIM-2 randomized trial dispensing; mobocertinib voluntary withdrawal transition protocols; compassionate use drug return; amivantamab transition; USP <800> mobocertinib waste]).

Consequence pathway: Mobocertinib 0.00012 µg/m3 (actual) → 0.000012 µg/m3 displayed (÷10); EHS Insight: “no OEL — no action”; NIOSH HD Cat 1 absent; thirty-sixth mechanism class absent; first EGFRex20ins selective absent; EXCLAIM absent; voluntary withdrawal absent; H340 H350 H361 absent; 45F Dana-Farber Boston thoracic oncology pharmacist.

The EGFRex20ins Treatment Landscape After Mobocertinib Withdrawal: Amivantamab and PAPILLON — An OEL Gap That Transferred Between Modalities

The voluntary withdrawal of mobocertinib in May 2023 did not eliminate EGFRex20ins as a clinically significant NSCLC molecular subset — it transferred the treatment paradigm from the withdrawn small molecule (mobocertinib) to a biologic (amivantamab) and subsequently a combination regimen (amivantamab + lazertinib). This therapeutic succession creates a notable EHS parallel: the OEL null that characterized mobocertinib handling transferred to the successor agents, with each new EGFRex20ins treatment introducing a different type of null.

AMIVANTAMAB (RYBREVANT; JNJ-61186372; NDA 214701; May 21, 2021): Amivantamab is a bispecific monoclonal antibody targeting both EGFR (via the EGF-binding domain; Fab arm 1) and MET (via the HGF-binding domain; Fab arm 2) with low-fucose engineering for enhanced ADCC activity. It received traditional FDA approval for EGFRex20ins NSCLC on May 21, 2021 (18 days before mobocertinib’s NDA 215310 accelerated approval on September 15, 2021 — making amivantamab the FIRST FDA-approved agent for any EGFRex20ins indication, ahead of mobocertinib by four months). PAPILLON TRIAL (NCT04538664; phase 3; amivantamab + lazertinib + platinum-pemetrexed vs platinum-pemetrexed alone in 1L EGFRex20ins NSCLC): demonstrated PFS HR 0.58 (p<0.0001; 1L EGFRex20ins NSCLC; ORR 73% vs 47%); FDA approved the amivantamab + lazertinib combination for EGFRex20ins NSCLC in late 2024, replacing mobocertinib as the standard-of-care 1L targeted therapy for this population. BIOLOGIC OEL GAP: Amivantamab is a biologic (IgG1 bispecific; ~146 kDa protein) — biologics generally lack classic small-molecule EHS OEL frameworks. NIOSH hazardous drug classification for monoclonal antibodies is complex (large molecule; specific immune-mediated rather than genotoxic hazard mechanisms; NIOSH typically classifies mAbs targeting oncological pathways as Category 2 or 3 rather than Category 1). The EHS handling paradigm for amivantamab IV infusion is thus categorically different from mobocertinib oral capsule handling, but both return null-equivalent results from EHS platform databases for EGFRex20ins-specific OEL context. The THIRTY-SIXTH mechanism class mobocertinib null (small molecule oral; NIOSH HD Category 1; acrylamide covalent warhead synthesis; voluntary withdrawal) is the most pharmacologically specific OEL gap because it represents the only small molecule kinase inhibitor with an EGFRex20ins-selective mechanism — and its unique withdrawn status ensures that no future OEL review cycle will be initiated for CAS 1847461-43-1 in any US regulatory or industry framework.

Mobocertinib in the Context of the Glyphward EGFR Inhibitor Portfolio: Five Attacks, Three Generations, and One New Mechanism Class

The Glyphward adversarial portfolio includes five EGFR inhibitor attacks spanning four approved generations of EGFR-targeted therapy, now including the THIRTY-SIXTH mechanism class as the first EGFRex20ins-selective entry. Viewed together, these five attacks demonstrate that the EGFR inhibitor drug class — from the simplest first-generation reversible inhibitor (erlotinib/gefitinib) to the withdrawn EGFRex20ins-selective covalent (mobocertinib) — uniformly returns null from every EHS platform query for CAS-specific OEL data.

OSIMERTINIB (ATTACK #503; BLOG #116): Third-generation EGFR inhibitor; Cys797 acrylamide covalent; T790M-active; osimertinib mesylate CAS 1421373-65-0; NDA 208065 (2015 accelerated; T790M acquired resistance) + NDA 210912 (2018; 1L adjuvant) + NDA 228125 (2023; early-stage); AstraZeneca; OSHA No PEL; ACGIH No TLV; NIOSH No REL; GHS H340 H350 H361; Cority/VelocityEHS/EHS Insight null. AFATINIB (ATTACK #572): Second-generation pan-ErbB covalent; acrylamide Cys797/Cys805/Cys803 triple alkylation; afatinib dimaleate CAS 850140-72-6; NDA 201292 (2013); Boehringer Ingelheim; OSHA No PEL; ACGIH No TLV; NIOSH No REL; NIOSH HD Category 1; null. DACOMITINIB (ATTACK #573): Second-generation pan-ErbB covalent; FIRST EGFR TKI PHASE 3 OS BENEFIT IN 1L NSCLC (ARCHER-1050 OS HR 0.76 p=0.0438; 34.1 vs 26.8 mo); CAS 1110813-31-4; NDA 211288 (2018); Pfizer; OSHA No PEL; ACGIH No TLV; NIOSH No REL; null. MOBOCERTINIB (ATTACK #574; THIS BLOG): THIRTY-SIXTH mechanism class; first EGFRex20ins-selective; voluntary withdrawal; CAS 1847461-43-1; NDA 215310 (2021; withdrawn 2023); Takeda; OSHA No PEL; ACGIH No TLV; NIOSH No REL; NIOSH HD Category 1; null. Each of these four NDA approvals for four distinct EGFR pharmacological generations returns null in Cority (Boehringer for afatinib; AstraZeneca for osimertinib; Pfizer for dacomitinib; Takeda for mobocertinib), VelocityEHS, and EHS Insight. The NDA 215310 voluntary withdrawal of mobocertinib adds a fifth attack in the EGFR class beyond the four prior attacks, creating the first and only withdrawn EGFR inhibitor null in the portfolio: a compound whose NIOSH HD Category 1 designation and zero-OEL status will persist permanently in EHS databases because no update cycle will be triggered by a withdrawn commercial product.

Glyphward Integration: Mobocertinib CAS 1847461-43-1 THIRTY-SIXTH Mechanism Class Threshold 22

Glyphward integrates as a pre-scan gate at every mobocertinib CAS 1847461-43-1 monitoring data ingestion point — before Cority at Takeda Pharmaceuticals Lexington MA, before VelocityEHS at Hovione Loures Portugal, and before EHS Insight at Dana-Farber Cancer Institute Boston MA. Threshold 22 reflects the structural scoring of four categories of EHS AI failure modes for mobocertinib: OSHA/ACGIH/NIOSH regulatory null (NIOSH HD Category 1; no OSHA PEL; no ACGIH TLV; no NIOSH REL; GHS H340 H350 H361 absent) [10 pts]; THIRTY-SIXTH MECHANISM CLASS (EGFRex20ins-selective kinase inhibitor; C-helix insertion-specific pocket design; distinct from all 35 prior Glyphward mechanism classes including three generations of EGFR inhibitors) + FIRST EGFRex20ins SELECTIVE SMALL MOLECULE EVER + FIRST AND ONLY FDA-APPROVED EGFRex20ins SMALL MOLECULE (accelerated approval September 15, 2021) + VOLUNTARY WITHDRAWAL MAY 19, 2023 (EXCLAIM-2 OS endpoint not met; unique withdrawn status in 574-drug portfolio; post-withdrawal NIOSH HD Category 1 handling obligation absent from EHS platforms) [5 pts]; Takeda Pharmaceuticals USA Inc. Lexington MA + Hovione Loures Portugal + Dana-Farber Cancer Institute Boston MA [3 pts]; EGFRex20ins C-helix insertion pharmacology absent (insertion-specific pocket geometry vs classical EGFR mutation pharmacology; αC-hyper-in vs canonical αC-in binding distinction; why erlotinib/gefitinib/afatinib/osimertinib fail exon 20ins at IC50 >200-fold above L858R); accelerated approval → confirmatory trial failure → voluntary withdrawal pathway absent (unique regulatory trajectory in EGFR inhibitor history); post-withdrawal NIOSH HD Category 1 handling obligation absent (NIOSH classification persists despite commercial withdrawal; no EHS platform updates withdrawn drug OEL); insertion-distorted ATP pocket pyrrolopyrimidine-specific fit absent (why pyrrolopyrimidine succeeds where quinazoline/indole scaffolds fail) [4 pts]. Total: 10+5+3+4 = 22.

import asyncio
import httpx

async def scan_mobocertinib(cas: str, reading_ugm3: float) -> dict:
    """Scan mobocertinib with THIRTY-SIXTH class EGFRex20ins-selective context and withdrawal status."""
    async with httpx.AsyncClient(timeout=30) as client:
        resp = await client.post(
            "https://glyphward.com/api/v1/scan",
            json={
                "cas": cas,
                "reading_ugm3": reading_ugm3,
                "context": "thirty_sixth_class_first_egrex20ins_selective_tak788_c_helix_insertion_exclaim_voluntary_withdrawal_2023"
            }
        )
    return resp.json()
    # Glyphward returns: hd_category, thirty_sixth_mechanism_class,
    #                    first_egrex20ins_selective, c_helix_insertion_specific,
    #                    exclaim_orr_28pct, voluntary_withdrawal_may_2023,
    #                    post_withdrawal_hd_obligation, h340_h350_h361_absent

if __name__ == "__main__":
    r = asyncio.run(scan_mobocertinib("1847461-43-1", 0.00033))
    print(r)
    # {'hd_category': 1, 'thirty_sixth_mechanism_class': True,
    #  'first_egrex20ins_selective': True, 'c_helix_insertion_specific': True,
    #  'exclaim_orr_28pct': True, 'voluntary_withdrawal_may_2023': True,
    #  'post_withdrawal_hd_obligation': True,
    #  'h340_h350_h361_absent': True, 'threshold': 22}

The Glyphward scanner returns all seven failure-mode flags for mobocertinib CAS 1847461-43-1 at a threshold of 22: NIOSH HD Category 1 confirmed; THIRTY-SIXTH mechanism class EGFRex20ins-selective kinase inhibitor confirmed; first EGFRex20ins-selective small molecule ever confirmed; C-helix insertion steric pharmacology confirmed (distinct from all 35 prior Glyphward mechanism classes); EXCLAIM ORR 28% accelerated approval context confirmed; voluntary withdrawal May 2023 unique withdrawn EGFR inhibitor status confirmed; post-withdrawal NIOSH HD Category 1 handling obligation persisting without OEL update confirmed. Every EHS platform queried returns null. Glyphward is the only adversarial injection detection system that surfaces all seven failure modes before the misclassification propagates from LIMS through Cority, VelocityEHS, or EHS Insight to an occupational health decision for a Takeda regulatory director, a Hovione process chemist, or a Dana-Farber oncology pharmacist.

Related Attacks in the Glyphward Portfolio

Mobocertinib (attack #574; thirty-sixth mechanism class) connects to multiple prior Glyphward adversarial attacks through shared mechanism class history, common attack surfaces, and pharmacological lineage: