Measuring T-cell engagement
This page is the practical ladder for a deceptively simple question:
Are T cells seeing Diana's tumor, reaching it, and doing useful work?
No single assay answers that end-to-end. H&E can show lymphocytes in the tumor bed. Multiplex imaging can show which immune cells are next to tumor cells. TCR-seq can show whether T-cell clones are expanding in blood or tumor. Single-cell RNA sequencing can show whether those cells look activated, exhausted, suppressive, or irrelevant. ctDNA answers a different but paired question: is tumor-derived DNA clearing while the immune response is changing?
The useful mental model is: location, identity, clonality, function, and tumor-burden response are separate layers. Some require tissue; some are deliberately blood-first.
The two-lane measurement strategy
Do not read this as "tissue first, blood later." Tissue and blood answer different questions and should ideally be collected in parallel. Tissue gives location and state; blood gives trajectory and repeatability.
What each assay can and cannot tell you
| Readout | Sample | What it measures | Best use | Main limitation |
|---|---|---|---|---|
| sTIL score | H&E slide | Percent stromal area occupied by lymphocytes | Cheap baseline pCR / immune-inflamed signal | Does not identify CD8 vs Treg, tumor specificity, or killing |
| IHC | FFPE slide | One protein marker per slide, such as CD8, FOXP3, PD-L1, HLA-I, B2M | Fast targeted confirmation | Marker-by-marker, limited spatial context |
| Multiplex IF / MXIF | FFPE or fresh tissue slide | Many immune and tumor markers on the same slide | Hot vs excluded vs cold; CD8 proximity; Treg rings; HLA patchiness | Requires panel design and image analysis |
| Tumor TCR-seq | Tumor tissue or TIL-rich section | TCR clonotypes physically present in tumor | Identifies clones that reached tumor | Sequence alone does not prove antigen target or cytotoxic function |
| Blood TCR-seq | PBMCs, whole blood, or leukopak | Circulating TCR clonotypes and expansion over time | Longitudinal immune pharmacodynamics during pembro, local priming, or vaccine | Blood clones may not match tumor clones |
| scRNA-seq | Fresh viable cells, sometimes nuclei | Cell-by-cell gene-expression state | Active vs exhausted T cells; Tregs; macrophages; rare populations | Tissue hungry; dissociation can bias cell recovery |
| scTCR-seq | Same single-cell prep | TCR sequence paired to each T cell's state | Links a clone to exhaustion, activation, memory, or Treg identity | Usually needs fresh tissue or high-quality cells |
| Spatial transcriptomics | Tissue section | Gene expression with location preserved | Maps immune programs to tumor/stroma neighborhoods | Platform-specific resolution and QC issues |
| ctDNA | Plasma | Tumor-derived DNA burden over time | MRD, response, recurrence-risk monitoring | Does not say whether T cells caused the tumor response |
P-RAD, PEARL, and where blood fits
P-RAD and PEARL were tissue-first immune-readout studies.
P-RAD TCI was an early post-treatment biopsy readout: CD3/CD8 T-cell access into or around the tumor bed, measured by quantitative / multiplex immunofluorescence and summarized as an Immunoscore-style T-cell infiltration metric. It asked whether T cells were physically present in the tumor architecture soon after pembrolizumab plus radiation.
PEARL TIL reduction was a pathology-style percent TIL readout across serial fresh-frozen tumor biopsies: baseline, after pembrolizumab alone, and after pembrolizumab plus radiation. It asked whether mean TIL scores changed over those tissue timepoints.
Blood repertoire sequencing would not recreate either readout. It cannot tell whether T cells are inside the tumor or parked at the edge. What it can add is the systemic immune-repertoire layer: whether specific clones expand after pembrolizumab, radiation, cryoablation, or vaccination. That is valuable, but it is a companion readout, not a replacement for tissue.
How TCR-seq works
T-cell receptor sequencing reads the hypervariable CDR3 region of T-cell receptors. The CDR3 sequence acts like a clone barcode: T cells with the same TCR sequence belong to the same clonal lineage.
The output is usually:
- Clonotypes: the unique TCR sequences detected.
- Frequency: how common each clone is in the sample.
- Diversity: how many distinct clones are present.
- Clonality: whether the repertoire is broad and flat or dominated by a few expanded clones.
- Overlap: whether the same clones appear across blood, tumor, lymph node, or timepoints.
The most useful longitudinal pattern is not "one high clone." It is baseline -> post-intervention expansion. A clone that is rare before treatment and expands after vaccine, radiation, cryo, or pembrolizumab is a candidate immune-response clone. To prove it is tumor-relevant, pair it with tumor repertoire sequencing, antigen/HLA mapping, ELISpot, polyfunctional ICS, killing assays, or single-cell state data.
Tissue TCR-seq vs blood TCR-seq
| Question | Better readout | Why |
|---|---|---|
| Are T cells physically in the tumor? | Tissue TCR-seq, multiplex IF, spatial biology | Blood does not preserve location |
| Are T cells expanding systemically after treatment? | Blood TCR-seq | Repeatable with serial blood draws |
| Are the expanding clones tumor-specific? | Matched blood + tumor repertoire sequencing plus antigen/function testing | Expansion alone could be viral, bystander, or suppressive |
| Are T cells active or exhausted? | scRNA-seq + scTCR-seq, flow cytometry, ICS | TCR sequence alone does not encode functional state |
| Is tumor burden clearing? | ctDNA, MRI, pathology | Immune measurements do not directly measure tumor burden |
For Diana, the cleanest blood-only add-on is a serial PBMC / whole-blood repertoire plan tied to the same timepoints as tumor-DNA monitoring. The highest-value version adds a tumor or lymph-node tissue repertoire anchor so blood clones can be matched back to the tumor compartment, but the blood series is still useful even before that anchor exists.
Blood-first TCR strategy
If tissue is scarce, the blood-first version is:
- Find or create a baseline. Confirm whether April 2 or another pre-treatment blood sample has usable PBMC / DNA substrate.
- Repeat at biologically meaningful timepoints. Pair blood repertoire sequencing with ctDNA at end-taxane, pre-surgery, post-surgery, vaccine dosing, and any local-priming intervention.
- Track clone expansion, contraction, and overlap. The main signal is not whether one clone is large; it is whether clones change after an intervention.
- Add function when possible. ELISpot, ICS, flow cytometry, or antigen stimulation turns "clone expanded" into "clone responds to this target."
- Back-map to tissue when possible. If surgery or residual disease provides tissue, compare blood clones against tumor/TIL clones.
This plan does not require spending scarce tumor tissue up front. It creates a longitudinal immune record that can later be connected to tissue, vaccine antigens, and tumor-DNA response.
Single-cell RNA sequencing and scTCR-seq
Single-cell RNA sequencing asks what each cell is doing. In immune monitoring, the important distinction is not just "T cell present" but which T-cell state is present:
| Cell state | Why it matters |
|---|---|
| Cytotoxic CD8 effector | The desired killing population |
| Stem-like / memory-like T cell | Durable response potential; useful for vaccine and TCR-T thinking |
| Exhausted T cell | May be rescuable by checkpoint blockade, depending on depth of exhaustion |
| Terminally exhausted T cell | Often less rescuable |
| FOXP3+ Treg | Suppressive; can expand with treatment and blunt effector response |
| Bystander T cell | Present but not necessarily tumor-reactive |
Pairing single-cell RNA with scTCR-seq is powerful because it links clone identity to cell state. The same expanded TCR clone can mean very different things if the cells are cytotoxic effectors, exhausted CD8 cells, or suppressive Tregs.
How to pair immune readouts with tumor-DNA monitoring
Tumor-DNA and immune assays answer different sides of the same story:
Use the combination this way:
| Pattern | Interpretation |
|---|---|
| Tumor DNA falling/clearing + effector TCR expansion | Best signal that treatment is reducing tumor burden while immune learning is happening |
| Tumor DNA falling/clearing + no TCR expansion | Tumor response may be chemo-dominant; still clinically good, but less evidence of immune education |
| Tumor DNA persistent/rising + TCR expansion | Immune response may be present but insufficient, mistargeted, exhausted, or blocked from tumor access |
| Tumor DNA persistent/rising + no useful TCR expansion | Stronger resistance concern; prioritize tissue/spatial/function workup |
| TCR expansion dominated by Tregs | Possible suppressive immune response; flow/scRNA confirmation needed |
This is why repertoire sequencing should not be interpreted alone. It becomes much more meaningful when read beside tumor-DNA monitoring, MRI, pathology, and tissue/spatial immune context.
Practical Diana measurement plan
If we were designing the cleanest monitoring stack without overusing tissue:
- Baseline H&E / sTIL score on existing diagnostic tissue.
- Baseline blood repertoire sequencing if April 2 or another pre-treatment blood aliquot is usable.
- Serial blood repertoire sequencing around major intervention points: early pembrolizumab, end of taxane phase, local priming if ever done, vaccine dosing, and any tumor-DNA recurrence.
- Paired tumor-DNA monitoring at the same blood draws so immune-repertoire changes can be read against tumor-burden changes.
- Multiplex IF or spatial biology on surgery tissue, especially if residual disease remains.
- Single-cell RNA + scTCR-seq only when tissue quality and decision value justify it, because it is powerful but tissue- and logistics-heavy.
The practical question for PHM / Echo / Invocata / Stanford is whether a staged blood plan is enough now: smaller PBMC draws for discovery and TCR tracking, with a full leukopak reserved for the vaccine / TCR-T optionality decision.
Cross-links
- reading-a-tumor -- how this fits the broader diagnostic layer stack.
- immune-hot-vs-cold -- the state call these assays build toward.
- 04-spatial-biology -- spatial and multiplex IF details.
- tcr-and-tcr-sequencing -- TCR biology, age-20 repertoire, and vaccine monitoring.
- is-pembro-working -- Diana-specific pembro interpretation and timing.
- ctdna-mrd -- operational ctDNA / MRD monitoring.
- should-we-bank-leukopak -- PBMC vs leukopak optionality.
- pre-op-radiation-nodal-response -- P-RAD / PEARL tissue readouts.
Sources
- pre-op-radiation-nodal-response -- P-RAD TCI and PEARL TIL measurement distinction.
- p-rad-nct04443348 -- P-RAD T-cell infiltration / Immunoscore endpoint details.
- ho-2024-pearl-radiation-pembrolizumab -- PEARL TIL scoring and serial biopsy methods.
- is-pembro-working -- blood TCR-seq vs ctDNA companion-readout framing.
- tcr-and-tcr-sequencing -- TCR-seq mechanics.
- Adaptive Biotechnologies TCR assay methods -- CDR3-based clone tracking.
- MedlinePlus tumor-DNA overview -- tumor-DNA definition and monitoring frame.
- NCI liquid biopsy overview -- liquid biopsy as repeatable tumor-material monitoring.