Returning to “Same Stage, Different Destiny”
In my previous post, I wrote that TNM staging is an excellent method of assessment, but it cannot fully describe the blueprint of a tumor—namely, the biology of the tumor microenvironment (TME). That blueprint often cannot be seen through stage alone, or through one or two single biomarkers. Yet what makes clinical practice and drug development truly difficult is this: those invisible differences can change the direction of treatment response and recurrence/resistance.
Today, I want to begin with two numbers—the grammar we need to read that blueprint.
20 µm: a working distance within which immune cells can actually attack tumor cells
150 µm: a gate at the boundary where tumor and stroma meet—where treatment response can diverge
From “Immune Cells Are Present” to “Immune Cells Can Reach”
As the era of immunotherapy opened, we began counting immune cells within tumors, calculating TILs, and reading PD‑L1. This was a major step forward in evaluating a patient’s treatment direction, but in real-world clinical practice we often face paradoxical situations:
There are many immune cells, but there is no response. Immune cells are present, yet tumor cells survive.
These paradoxes are usually not a matter of quantity, but of architecture. If immune cells are only circling around the tumor, or if they are blocked at the tumor–stroma boundary, then high numbers alone may not translate into function. This is exactly why spatial biology can be so powerful for drug development. Spatial data goes beyond prediction—it transforms “why there was no response” into something that can be explained.
The 20 µm Distance: Response Begins with “Proximity,” Not “Presence”
A recent neoadjuvant therapy cohort study in breast cancer quantified how close CD8+ T cells are to tumor cells, stratified by distance. The study compared radii from 5–50 µm around tumor cells and reported that, among these, the fraction/metric of CD8+ T cells located within 20 µm of tumor cells was strongly associated with pCR, DFS, and OS.
(Liang et al., “Spatial proximity of CD8+ T cells to tumor cells predicts neoadjuvant therapy efficacy in breast cancer,” npj Breast Cancer, 2025)
The most important point here is not the number 20 µm itself, but the concept of spatial proximity.
The core of treatment response is not “whether cells are present/absent,” but whether cells are positioned at a distance where they can perform their function.
Even among CD8 cells, those that penetrate deep into the tumor core and can contact tumor cells have a completely different clinical meaning from CD8 cells that remain outside the boundary and cannot make contact.
From a drug development perspective, this concept is highly attractive—because it is information we could not know from previous data types (bulk, single-cell, etc.).
This is because “proximity” is often not a simple correlation; it frequently connects directly to the mechanism of action. The process by which immune cells approach tumor cells involves intertwined networks—vasculature, stroma, chemokines, and immunosuppressive pathways—and proximity metrics can offer clues as to where the bottleneck lies.
The 150 µm Boundary: The Tumor–Stroma Boundary Becomes an Immune Gate
Now to the second number: 150 µm.
In colorectal cancer, there is a study that defined the tumor–stroma interface as a functional compartment—the tumor–stroma boundary (0 ± 150 µm)—and showed that the boundary’s spatial organization and immune state are associated with response to immunotherapy (anti‑PD‑1, etc.).
(Feng et al., “Spatially organized tumor-stroma boundary determines the efficacy of immunotherapy in colorectal cancer patients,” Nature Communications, 2025)
At the boundary, the study reported that:
In responders, interactions favorable for immune activation (e.g., proximity/interaction between LAMP3+ DCs and CXCL13+ T cells) may be observed;
In non-responders, CXCL14+ CAFs were reported to remodel the ECM and form a structural barrier.
Here, the boundary is not a simple anatomical line—it is a physical and biological gate that treatment must pass through. Stroma surrounding tumor cells, fibrosis, collagen, vascular structure, and the positioning of immunosuppressive cells can all strengthen or weaken this gate. In other words, by reading the boundary we can see why immune cells cannot enter—and that insight can directly lead to what combination strategies are logical.
Let’s assume there are 100 immune cells. A test that analyzes tissue by grinding and averaging might predict, “There are many immune cells, so the prognosis will be good.” But if we look into the actual spatial context, those immune cells may be drifting on the outside, never approaching cancer cells. What matters more than the count is the question: “Why can they not attack the cancer cells?” And the answer is often hidden somewhere other than the immune cells themselves.
Spatial Biology with ADC / RPT / TCE: The Same Numbers Mean Different Things Across Modalities
The grammar of space (distance and boundary) is not limited to IO. In fact, the more modalities diversify, the more valuable spatial data becomes.
ADC has substantial variability that cannot be explained by target expression level alone (e.g., H‑score). Delivery and efficacy can differ depending not only on how much target there is, but also where and how it is distributed (patchy vs diffuse), what the surrounding stromal barrier looks like, and how the vasculature relates to the target.
RPT cannot be fully explained by mean uptake alone when it comes to true intratumoral microdosimetry and biological response. Boundary/vascular/hypoxic niches can alter delivery and response, and spatial data makes the question “where was the energy delivered?” more precise.
TCE is literally a “cell-bridging drug.” The key, ultimately, is the distance/barrier conditions under which cell–cell bridging can form. Even with the same target and the same T cells, if the boundary is blocked, the bridge fails.
In short, even when the modality changes, the questions remain the same:
Has the ‘distance’ required for a drug to work been secured?
What is the ‘boundary/barrier’ that blocks function?
Conclusion: 20 µm and 150 µm Are Not Numbers—They Are the Questions of Drug Development
The 20 µm and 150 µm I presented today are not absolute cutoffs for drug development. What matters is that the spatial distances between the cells that constitute a tumor can be critically important for drug response and resistance—and the questions these numbers raise are powerful:
Can immune cells reach tumor cells? Are blood vessels close to the target?
Is the boundary a gate—or a barrier?
Is resistance a problem of one cell type—or a problem of structure and spatial distance?
What spatial biology shows us is not a new “test,” but the conditions of treatment response that we have been missing. And the faster and more clearly we can organize those conditions on human data, the more we reduce uncertainty in drug development.
Portrai’s interest lies exactly there: turning the space beyond the microscope into a language that enables decisions.
