ABP · AI as a Biological Primitive

Builds a model of one patient from that patient's own history, then runs interventions against it in reverse: specify the state you want and it returns the intervention directions that patient's record supports. No population cohort supplies the answer.

Discovery · target-directed interrogation

Interrogate the primitive, then show what moved.

Necrotic core—

Enhancing rim—

Infiltrative margin—

Move an intervention control. ABP will report the compartment-specific response instead of collapsing it to one number.

Discovery searches admissible intervention directions against the declared target while preserving unsupported regions as unresolved.

ABP workflow

Discovery

Interrogate the primitive against a declared target.

Evidence context

Glioblastoma intervention design

The rotating GBM primitive is a literature-grounded demonstrator of target-directed inversion, intervention design, MGMT response modification, infiltration and treatment-associated harm.

Drug-development mode is defined as compound × dose × schedule × combination. Exact TARGET-NBL candidate drug names will be inserted only from the frozen ranked drug-regulator result table; this review build does not invent them.

Compartments

necrotic core contoured

enhancing rim contoured

infiltrative margin diffuse, no boundary

This tumour

Intervention coordinates

Cytoreduction / resection levelBiopsy only Radiation0.0 Gy-eq ~60 Gy conventional-course reference; higher values illustrate escalation beyond the conventional course. Alkylating agent0.00 Weeks since start12

Cyan = enhancing tumour · purple = infiltrative margin · orange = necrotic core · red glow = treatment injury / necrosis risk.

Target-directed inverse interrogation

No optimized intervention selected yet.

Press Optimize Intervention. ABP will compare supported intervention routes and return the best route and levels for the declared target.

Unconstrained numerical optimum

—

ABP supported optimum

—

Route comparison will appear after optimization.

The optimizer is support-constrained: numerically attractive settings outside the declared support envelope remain unresolved.

Response

Cytoreduction effect—

Radiation kill\u2014

Agent kill\u2014

Combined\u2014

Enhancing volume—

Infiltrative extent—

Total volume—

Enhancing-volume response—

Clinical course—

Radiation necrosis—

Induced invasion—

Untreated baseline

Move a control to perturb the primitive.

Why radiation can worsen the margin. Sublethal irradiation has been shown repeatedly to increase glioma cell migration and invasiveness, through αvβ3 integrin and MMP-2/9 (Wild-Bode, Cancer Res 2001), p53–MMP-2 (2015), HIF1α/JMY in glioma stem-like cells at doses as low as 0.5 Gy (2020), and CXCL12/KCa3.1 (2019). Irradiated glioma produces more tumour satellites in vivo, and most relapses arise adjacent to the irradiated field. The effect is consistently reported across cell lines, stem-like cells and in vivo models, with at least one co-culture study reporting the opposite.

Literature-grounded ABP demonstrator. The primitive encodes published qualitative relationships among glioblastoma growth, radiation exposure, alkylating therapy, MGMT status, treatment-associated necrosis, pseudoprogression, and infiltrative response. Numerical parameterization is illustrative and is not fitted to an individual patient. The demonstrator illustrates biological-primitive perturbation, target-directed inverse interrogation, intervention design, and response-modifier exploration; it does not provide treatment recommendations. These are not patient images and no restricted data are distributed here.