Tissue injury and manageability: evidence for policy
The 1.29.0 central ontology is the current source of tissue and cell calls. The numerical grids below document the earlier tissue-scope screening API; the current CTA cascade is described in CTA selection.
This review concerns significant immune injury, its consequences, persistence and control. Complete organ removal is one evidence source; it is not a model for every pattern of immune injury. Reproductive, Disposable, Manageable, Protected and Vital are project categories. A separate field records evidence for local treatment.
Published classifications
Bradley et al., Nature Communications (2020) provides the closest explicit precedent. Supplementary Table 2 divides human source tissues into four risk groups, plus exceptions. The following is an anatomical summary of that table, not the oncoref assignment:
| Published group | Tissue coverage, condensed from the source labels | Study maximum RNA |
|---|---|---|
| Danger | Aorta, coronary and tibial arteries; sampled brain regions and cervical spinal cord; atrial appendage and left ventricle | 1 TPM |
| Hazard | Lung; tibial nerve; colon; esophageal junction/mucosa/muscularis; renal cortex; liver; small intestine; stomach | 3 TPM |
| Caution | Spleen; pancreas; thyroid; blood; adrenal; pituitary | 10 TPM |
| Non-essential | Mammary tissue; cervix; fallopian tube; ovary; prostate; suprapubic skin; uterus; vagina; subcutaneous adipose; bladder; minor salivary gland; skeletal muscle | 30 TPM |
| Other | Testis and placenta; transformed lymphocyte/fibroblast samples also appear in this source column | No TPM limit in that study |
Source: article and Supplementary Table 2. The table was checked visually in the supplementary PDF, page 25. Its GTEx TPM values are not adopted as HPA nTPM limits. The authors expressly acknowledge that recognition below a proposed threshold is possible. A source classification of skeletal muscle, bladder or pituitary is not clinical validation of tissue loss. Transformed cell lines are not normal-organ absence evidence.
Kaufmann et al., Oncotarget (2019) independently uses brain/heart/lung as an essential group and scores expression in other tissues as predicted adverse events. It provides a precedent for graded penalties rather than treating every nonessential tissue as freely targetable. Its calculations use normalized counts, not our nTPM policy. Primary study
Blagosklonny (2003) supplies a conceptual tissue-selective rationale, including melanocyte and reproductive-organ targeting and thyroid-directed treatment. It does not provide a universal safe-organ whitelist. Paper
What local steroids establish
| Evidence | What it supports | Boundary |
|---|---|---|
| Johnson et al. (2009), antigen-directed TCR therapy | Some eye/ear inflammation following melanocyte targeting was treated with local steroids | This is direct cell-therapy evidence, but not proof that all ocular, retinal or inner-ear cell killing is reversible |
| SITC (2021) guideline | Selected inflammatory skin manifestations can be treated locally; management depends on severity | Severe, blistering or mucosal disease can require escalation; checkpoint toxicity does not validate control of every engineered T-cell product |
| Warner et al. (2019), 20 patients with checkpoint-associated sicca | Symptoms often improved following intervention | Salivary output remained very low; supports a persistent-function-loss concern rather than a reliable local rescue assumption |
Sources: Johnson clinical trial, SITC guideline, Warner clinical series.
Project interpretation: local access and steroid responsiveness are useful attributes, but they must be attached to a specific compartment, manifestation, severity and therapy modality. They cannot by themselves turn a whole organ into Disposable. Suppressing inflammation and restoring cells already lost are different outcomes. No paper above establishes that immune injury can always be stopped locally while preserving antitumor activity.
Current categories and the intermediate Manageable tier
The current policy separates Manageable (lower protection priority) from Protected (higher protection priority). Bradley's Caution/Hazard distinction is a published anchor. Our exact assignments remain project inferences:
| Category | Assignment relevant to this review | Evidence and interpretation |
|---|---|---|
| Disposable | Cutaneous melanocytes, hair follicles, tonsil, appendix, gallbladder | Selected compartment-loss or organ-removal rationale; acute immune injury still needs review |
| Manageable | Thyroid, adipose; resolved mature B cells | Still screened; any positive IHC excludes. This does not imply easy local control or harmless loss |
| Protected | Whole skin, spleen, thymus, salivary glands, cartilage, esophagus and other protected GI tissues, lymph nodes and remaining listed protected organs | Persistent dysfunction, structural damage or important mucosal/distributed function |
| Vital | Existing vital and critical sensory/neurological organs | Highest protection priority; local access does not remove irreversible-loss concerns |
Reproductive remains the separate anatomical/developmental union. See the complete root lists and IHC-dependent RNA grid.
Two additional primary clinical studies argue against a default spleen/thymus allowance. Kristinsson et al. (2014) followed 8,149 splenectomized veterans and reported elevated infection and thromboembolic risks. Kooshesh et al. (2023) found associations between adult thymectomy and increased mortality/cancer. These are observational surgical cohorts, not causal estimates of antigen-directed immune injury. They support retaining protection and reviewing age/immune reserve; they do not prove equivalence between thymus, spleen and skin.
For cartilage, Heinemeier et al. (2016) found minimal replacement of mature articular collagen. Braaten et al. (2020) reported persistent inflammatory arthritis after checkpoint treatment cessation. These support concern about lasting joint injury, but are indirect evidence for a cartilage category: matrix turnover is not identical to repair capacity, arthritis is not selective chondrocyte injury, and a generic cartilage sample does not locate the joint or exclude an airway site.
The category grid is a research prioritization tool. It does not establish that every esophageal injury is worse than every skin injury. Tissue compartment, extent, therapy modality, persistence and treatment response remain separate considerations. Checkpoint-toxicity experience cannot establish the controllability of a persistent engineered T-cell response.
Machine-readable representation
The taxonomy's context_reviews records:
allowance_basis: removal tolerance, an explicit narrow exception, or no default allowance.local_steroid_control: the supported manifestation or absence of established control.references,rationaleandflags: evidence, consequences and unresolved questions.
The RNA ceilings are explicit project starting parameters: Vital 2/4/6/12, Protected 6/8/12/24, Manageable 10/20/30/60 nTPM for missing/uncertain, Approved, Supported and Enhanced negative IHC. The user requested that every absolute RNA ceiling be doubled, giving a minimum ceiling of 2 nTPM. The Manageable name refers to potentially manageable or compensated injury, subject to context review. No GTEx-to-HPA conversion or clinical outcome calibration is claimed. HPA reliability labels do not express a probability that a negative stain rules out antigen or peptide-HLA presentation.
Review annotations do not themselves change thresholds. Complete routine negative IHC is required to earn relaxation, and any positive IHC outside Reproductive or Disposable excludes. Whole skin/salivary/thymus/spleen allowances from policy 2026-10-20.3 are retained only in immutable historical reports.
Human clinical rationale is not transferred to dogs or mice. The shared vocabulary maps anatomy; tolerability evidence, numerical policies, age and patient context remain separate. The historical CTA report separates the effects of reclassification from the new RNA grid.
Combined absolute and selectivity gates
The same local IHC evidence now sets both an absolute RNA ceiling and minimum pairwise selectivity: highest Reproductive RNA divided by that reference plus the screened tissue's RNA. Disposable is excluded from this comparison and cannot supply the required reproductive-positive evidence. Missing or zero reproductive RNA cannot establish a selectivity score.
The approved percentage ladders are Vital 75/70/65/60%, Protected 70/65/60/55%, and Manageable 65/60/55/50%, in order of missing/uncertain, Approved, Supported, and Enhanced negative IHC. Admission requires valid reproductive RNA >0 or positive reproductive IHC. These are target-informed project parameters with a different meaning from the historical global deflated fraction. They do not estimate clinical risk. See the complete matrix and formula and the candidate comparison.