# CAR-T clinical finale — evidence and rendering notes

Prepared 7 September 2026. This is a sourced communication animation, not a predictive patient model or comprehensive prescribing information.

## Story and visual scope

The film follows a representative CAR T cell encounter in a marrow reconstruction. Tisagenlecleucel is the product exemplar, in refractory or second-or-later-relapsed B-cell precursor ALL in patients up to age 25. The animation explains cellular events; it does not model manufacturing yield, trafficking probability, encounter frequency or patient response.

The same scene timeline supports the movie and scroll-through experience. The film contains AI-generated narration and an original synthesized ambient bed. The three figures contain rendered raster scenes with editable SVG titles and captions.

Marrow and cell meshes are original authored assets from the Human Reference project. The rebuilt marrow uses Blender geometry and Three.js lighting and surface shading. AI-generated art direction, a camera-framed distant tissue matte, and a grayscale membrane microtexture inform its appearance; none is microscopy or patient-derived tissue. The moving encounter and nearby vessel remain 3D, while distant tissue is a composited background rather than independently simulated cells. The arrangement, cell density, surface relief and motion are authored for explanation rather than fitted to measured histology. Molecular coordinates are from the research complex [RCSB PDB7URV](https://www.rcsb.org/structure/7URV), a 3.05Å cryo-EM structure of FMC63 recognition fragment with soluble CD19. It is not the full CAR, a complete therapeutic cell or a moving measurement of binding.

## Early remission

Maude et al., NEJM 2018: 92 enrolled, 75 infused; 45 CR and 16 CRi within three months. Thus 61/75 (81%, 95% CI71–89) achieved the composite endpoint. The 14 complementary patients are labeled only as not meeting that endpoint; they are not depicted as deaths. CRi is complete remission with incomplete hematologic recovery, not partial remission. Seventeen enrolled patients did not receive infusion; the film does not assign unverified reasons. Each glyph is one patient.

The original publisher search result was independently retrieved on 7 September 2026 and explicitly supports 92/75/45/16, resolving the previously recorded lack of full-text verification for the subdivisions. Primary source: [Maude et al.](https://www.nejm.org/doi/full/10.1056/NEJMoa1709866). Its accessible abstract also supports the composite 81% remission result. No label-efficacy denominator is substituted.

## Durability

Grupp et al., JCO, online 29 July 2026, PMID42525896, doi10.1200/JCO-25-01471. The primary abstract was independently verified via PubMed search on 7 September 2026: 79 treated patients, median follow-up79.4 months; estimated five-year overall survival 55.0% with SCT included in censoring for new anticancer therapies and 62.4% without SCT included. SCT means stem-cell transplantation. These two figures use different censoring definitions in the same cohort; they are not raw survivor fractions or treatment-comparison arms. No patient-glyph morph or survival curve is used. Full-text confidence intervals were not accessible and are not reconstructed.

Primary: [Grupp et al.](https://pubmed.ncbi.nlm.nih.gov/42525896/).

## Risk

Current FDA-hosted KYMRIAH label, revised June 2025: in its 79-patient pediatric/young-adult ALL safety cohort,61/79 (77%) had CRS; 48% had grade ≥3 CRS using Penn grading. Median onset 3 days. Boxed warnings include CRS, neurological toxicities, and secondary hematological malignancies. Normal CD19-positive B cells may be depleted. Selected risks are shown, with the full label linked; this is not a complete risk presentation.

Primary: [KYMRIAH US prescribing information](https://www.fda.gov/downloads/BiologicsBloodVaccines/CellularGeneTherapyProducts/ApprovedProducts/UCM573941.pdf), boxed warning and sections5.1,6.1,12.1–12.2.

## Review boundary

Population sizes 92, 75, 79 and the label's separate 63-patient evaluable efficacy population must remain distinct. ELIANA is single-arm evidence without a concurrent comparator. Early remission and long-term overall survival are different outcomes. The clinical finale does not derive clinical percentages from the animated killing sequence.

Scientific source verification does not establish physiological accuracy of the rendered mechanism, medical approval, or release readiness. Specialist visual and scientific signoff remains required for a commissioned clinical film.


## Mechanism and tissue references

- [He et al. CD19 CAR antigen engagement mechanisms and affinity tuning (2023)](https://doi.org/10.1126/sciimmunol.adf1426). Experimental FMC63–CD19 recognition complex; not a complete therapeutic CAR.
- [Cazaux et al. Single-cell imaging of CAR T cell activity in vivo reveals extensive functional and anatomical heterogeneity (2019)](https://doi.org/10.1084/jem.20182375). Anti-CD19 CAR T encounters in lymphoma-bearing mice; marrow killing and nonproductive contacts. Research model, not product-specific human kinetics.
- [Montalvo et al. Decoding the mechanisms of CAR T cell-mediated killing of tumors (2024)](https://doi.org/10.1038/s41419-024-06461-8). Single-cell cytotoxicity assays reveal context-dependent contributions of granzymes and Fas; do not treat one route as universal.
- [wwPDB Usage Policy](https://www.wwpdb.org/about/usage-policies). PDB archive data files are CC0 1.0; original author attribution encouraged.
- [Human marrow cell neighborhoods · Bandyopadhyay et al., Cell 2024](https://doi.org/10.1016/j.cell.2024.04.013). Human marrow vascular, stromal, adipocytic and bone-associated compartments; not a pediatric B-ALL spatial reconstruction.
- [B-ALL stromal microenvironment · Ferrao Blanco et al., Leukemia 2025](https://www.nature.com/articles/s41375-025-02734-z). Distinct stromal neighborhoods; spatial dataset is one biopsy with age and subtype unspecified.

## Entry and contact motion — v3

The entry sequence follows one representative activated T cell from a marrow vessel into tissue. Its slowing, protrusion, deformation and rear retraction are authored explanatory motion. The route is not a measured human tisagenlecleucel trajectory; neither successful entry nor killing frequency is inferred from it. The case is B-cell precursor ALL, not a localized solid tumor.

- [Jacobelli et al., 2013](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0075151): activated mouse T cells, endothelial monolayers and mouse tissue entry; supports a deforming, mechanically active transmigration sequence, not human product-specific trafficking kinetics.
- [Cazaux et al., 2019](https://pmc.ncbi.nlm.nih.gov/articles/PMC6504219/): anti-CD19 CAR T cells in lymphoma-bearing mice; variable contacts and killing in marrow. Display timings and event proportions are not imported from this research model.
- [Davenport et al., 2018](https://pmc.ncbi.nlm.nih.gov/articles/PMC5834689/): experimental CAR/TCR system with HER2 targets; supports a nonclassical multifocal CAR contact interface. It does not establish the exact synapse geometry of clinical tisagenlecleucel.

The animation uses a local irregular contact, granule convergence, narrow-cleft release and delayed target blebbing. This is one explanatory cytotoxic route; death-receptor and granule contributions vary with context. Detachment does not guarantee successful serial killing. The generated tissue backgrounds remain artistic plates, while the entry and contact actions use deterministic 3D geometry.

## Molecular visualization upgrade, September 13, 2026

The recognition close-up and molecular inspection use deposited heavy-atom coordinates from [PDB 7URV](https://www.rcsb.org/structure/7URV), showing the FMC63 recognition fragment and CD19. Protein identity colors coexist with conventional element cues. Missing atoms are not reconstructed; surfaces and atom spheres are visualization conventions.

The cell-contact chapter uses simplified, uniformly scaled envelopes derived from the same deposited chains. Their repeated placement and recruitment at the illustrative synapse are authored, not an experimentally resolved synapse or molecular-dynamics trajectory. The envelopes are baked offline and instanced for mobile efficiency.