What this is. The engine's global competitive-landscape view of a target — every drug it can find in development/approved, compared by modality, phase, mechanism and differentiation, plus the white space and where a new program should position. This is the analysis a pharma BD/DD team trusts *before* they believe any downstream design output. Generated from the engine's open-data layer (ChEMBL drug-mechanism, ClinicalTrials.gov, patent search). Coverage boundary + how to extend it: see §6.
This landscape + the binder proof case are one case study showing the engine's full loop:
l74_mpnn2).The point isn't the molecule; it's that the engine **judged which market lane the molecule may enter
and which it may not** — competitive-landscape judgement + design + honest positioning, end to end.
PD-L1 (CD274) is a mature, crowded checkpoint target: **5+ globally/regionally approved anti-PD-L1
agents** — note this is *global/regional*, not all FDA: atezolizumab / durvalumab / avelumab / cosibelimab
are FDA-approved, while sugemalimab and the subcutaneous single-domain envafolimab are China/EU-anchored
approvals — and 8+ in active clinical development (ChEMBL drug-mechanism), with the frontier having
moved (8+ clinical-stage agents based on open-source ChEMBL/ClinicalTrials coverage — see §6) from plain mAbs to
**bispecifics, bifunctional fusions, subcutaneous single-domain formats, oral small molecules, conditionally-
active probodies, and imaging polypeptides. A *me-too* anti-PD-L1 blocking antibody has no room**. The
real opportunity is in formats/uses the approved mAbs cannot serve (§4–5).
| Drug | Phase | Modality / mechanism | Differentiation angle |
|---|---|---|---|
| Atezolizumab (Roche) | Approved | anti-PD-L1 IgG1 mAb (blocker) | 1st-in-class, broad labels |
| Durvalumab (AZ) | Approved | anti-PD-L1 mAb (blocker) | lung/biliary leadership |
| Avelumab (Merck/Pfizer) | Approved | anti-PD-L1 mAb (ADCC-competent) | maintenance/Merkel |
| Sugemalimab (CStone) | Approved | anti-PD-L1 mAb | China-anchored |
| Cosibelimab (Checkpoint) | Approved | anti-PD-L1 mAb (ADCC) | cSCC, latecomer |
| Bintrafusp alfa (Merck KGaA) | Ph3 / *setbacks* | PD-L1 × TGF-β trap bifunctional | dual MoA, but major development setbacks (INTR@PID failures); deprioritised in several indications |
| Envafolimab (3D Med/Alphamab) | Ph3 | subcutaneous single-domain (VHH-Fc) | SC dosing, format edge |
| Erfonrilimab (Alphamab) | Ph3 | PD-L1 × CTLA-4 bispecific | dual-checkpoint in one |
| Socazolimab, Adebrelimab | Ph3 | anti-PD-L1 mAbs | China SCLC/regional |
| Pacmilimab (CytomX) | Ph2 | conditionally-active Probody | tumour-localised, safer |
| INCB-086550 (Incyte) | Ph2 | ORAL small-molecule PD-L1 inhibitor | pill, not an antibody |
| Danburstotug | Ph2 | anti-PD-L1 mAb | mainly combo-positioned; limited monotherapy differentiation |
Read: the blocking-mAb lane is saturated (5 approved + regional Ph3s). All the *interesting* recent
entries win on format or mechanism, not on being another blocker — bispecific (CTLA-4), bifunctional
(TGF-β), subcutaneous nanobody, oral small molecule, tumour-conditional probody.
PD-L1 biologics — i.e. the non-blocking / imaging** use is an *active, IP-staked* area, validating
it as a real lane (not a consolation prize). A de-novo, sequence-novel binder there has a cleaner FTO
path than yet another anti-PD-L1 antibody (whose CDR/epitope space is heavily claimed).
*(Representative public titles surfaced by the engine's patent search — "Novel PD-L1 binding
polypeptides for imaging", "Imaging methods using ¹⁸F-radiolabelled biologics"; full patent-family
numbers + legal status are retrievable via the engine's patent module / a premium IP source — see §6.)*
companion imaging. A non-PD-1-competitive binder is a genuine *advantage* here — non-blocking
means it can (i) image PD-L1 without perturbing PD-1/PD-L1 biology, (ii) be used alongside a
therapeutic anti-PD-(L)1 antibody without competing at the same epitope, (iii) select patients,
(iv) dynamically monitor PD-L1 expression over a course of therapy, and (v) serve as the
pre-diagnostic that gates an ADC/RLT delivery program. **This is exactly where our engine's
de-novo binder l74_mpnn2 fits** (validated, selective vs PD-L2, developable, non-blocking — see
ENGINE_PROOF_PDL1.md). Public IP on "PD-L1 binding polypeptides for imaging" confirms the lane is
real, not a consolation prize.
delivery antigen like DLL3.** DLL3 is tumour-high with near-zero normal-periphery expression → it
suits a cytotoxic/radioligand payload. PD-L1 is expressed on **immune cells, inflamed and normal
tissue → imaging is fine, but for a *cytotoxic/radioligand* payload the therapeutic index is the
central risk** and must be proven (normal-tissue uptake + dosimetry) before anything else. Industry
is moving here (e.g. Henlius HLX43, an anti-PD-L1 ADC in global Ph2/3 for pretreated squamous
NSCLC), which shows real interest and the toxicity/window work the lane demands.
the bar is high and the lane is filling.
targeting module** — the lane where non-blocking is a feature, IP is cleaner, and the wet bar is lower
(binding + selectivity + cell binding, not a full efficacy program).
l74_mpnn2 proof-case binder: not a drug, a **targeting modulecandidate** for the imaging/delivery lane.
This landscape is built from OPEN sources — ChEMBL (drug mechanisms/phase), ClinicalTrials.gov
(active trials), and public patents. That is strong for approved + clinical-stage agents and IP
texture. It does not include:
premium databases (Cortellis, Pharmaprojects, Evaluate, GlobalData).
BYOK design (how the engine handles this): each premium source is a bring-your-own-key datasource
slot. **If your organisation configures a Cortellis/Pharmaprojects/Evaluate API key, the engine auto-calls
it and folds preclinical pipeline + deals + forecasts into this table; if not, it falls back to the
open sources above** and flags — right here — *exactly which sections a premium key would deepen*
(pre-clinical pipeline completeness, competitive deal/financial context, sales forecasts). So the report
is honest about its coverage and tells you the precise upgrade that closes the gap.
Engine-generated competitive landscape (open-data layer). Drug list + phases from ChEMBL; trials from
ClinicalTrials.gov; IP texture from patent search. Premium-pipeline coverage available via BYOK keys (§6).