What We Mistake for Success: The Hidden Cost of Counting Tumor Shrinkage

Jul 18, 2026

9 min read

1 views

71%

The seductive lie of early victory

What if the most convincing sign that a treatment works is also the least reliable sign that a patient is actually better off?

That is the uncomfortable tension at the heart of modern cancer care. On one side is a clean, immediate metric: objective response rate, or ORR, the percentage of tumors that visibly shrink. It is attractive because it seems direct, measurable, and fast. On the other side is the slower, messier reality of living with a disease over years, where complications accumulate, anatomy changes, infections recur, and the body pays a price that does not show up in the first scan.

This gap matters because medicine is always tempted by the visible win. A tumor that shrinks in a trial looks like a truth. A patient who survives long enough to develop late complications tells a more complicated story. The deeper question is not simply whether a treatment works. It is: works for what, over what horizon, and at what hidden cost?

The problem with measuring the wrong victory

ORR is powerful precisely because it is simple. If a drug makes a tumor visibly recede, that seems like evidence the drug is active. In a single arm trial, it is especially appealing because spontaneous regression is rare, so a response can be plausibly attributed to the therapy itself. For regulators and clinicians, that makes ORR an efficient signal of anti tumor activity, especially when a therapy is being tested for breakthrough potential.

But ORR has a dangerous psychological effect: it can compress time. It turns a complex treatment journey into a single snapshot. That snapshot may be true and still be misleading. A shrinking tumor is like a car dashboard that lights up green while the engine is quietly overheating. You are not wrong to notice the dashboard, but you would be foolish to confuse it with the whole machine.

The issue is not that ORR is useless. It is that ORR answers only one question: does the tumor respond now? It does not answer whether the response lasts, whether the patient experiences fewer symptoms, whether treatment burdens are tolerable, or whether downstream interventions create new problems. In oncology, a metric can be biologically valid and clinically insufficient at the same time.

This is where long term outcomes become humbling. In one clinical context, complications after major urinary diversion can rise from 45 percent at five years to 94 percent in those surviving more than 15 years. That is not a typo, and it is not a minor side note. It is a reminder that the first chapter of success often hides the later chapters of cost.

Time reveals what early metrics conceal

There is a pattern in medicine, and in decision making more broadly: what is easy to measure first is often what matters least in the end.

Short term measures are favored because they are legible. They fit into trial timelines, reimbursement decisions, and public narratives. But biology does not respect our administrative convenience. A treatment may reduce tumor burden quickly and still create a future landscape of stenosis, infection, obstruction, or functional decline. Likewise, a reconstructive or diversion strategy may solve the immediate oncologic problem but generate a new set of chronic complications years later.

The long term data on urinary diversions are instructive because they make visible a general principle. Early complications can occur in nearly half of patients, including urinary tract infections, pyelonephritis, leakage, and stenosis. Over time, stomal complications and upper urinary tract changes remain common, and older experience suggests that some diversion types carry higher rates of ureteral stenosis and ascending infection than others. In other words, the intervention does not end when the operation ends. The cost profile unfolds.

That unfolding matters because human beings experience health in sequence, not in averages. A patient does not live the mean outcome. They live the first month, the third year, the tenth year, and every complication in between. A treatment can look excellent if you stop the clock early, and disappointing if you keep watching long enough.

The central deception in outcome measurement is not that it lies. It is that it tells the truth too early.

This is why ORR and long term complication rates are not just different statistics. They are different philosophies of time. ORR privileges immediate biological signal. Complication curves privilege accumulated lived reality. If we only worship the first, we confuse motion with progress.

A better mental model: the three horizons of treatment value

To connect these ideas, it helps to use a simple framework: every therapy has three horizons of value.

1. The response horizon

This is the first horizon, where ORR lives. Does the tumor shrink? Does a biomarker move? Is there evidence of direct effect? This horizon is essential because without a response, there may be no meaningful anticancer activity at all.

2. The durability horizon

This asks whether the early win persists. Does the response hold? Does the disease stay controlled? Is there a meaningful delay in progression or need for further treatment? A therapy that creates a brief response and then collapses may be impressive in the mirror and useless in the long run.

3. The burden horizon

This is where the deeper truth emerges. What complications accumulate? What does the patient have to live with? How much follow up, maintenance, infection risk, revision surgery, or quality of life erosion accompanies the benefit?

The mistake is treating the first horizon as if it were synonymous with the other two. It is not. A treatment can score high on response, mediocre on durability, and poor on burden. Or it can offer a modest response with a surprisingly favorable burden profile, producing more value than its headline numbers suggest.

Think of this like buying a house. A beautiful kitchen is a genuine asset, but you would not buy the house without inspecting the roof, plumbing, and wiring. ORR is the renovated kitchen. Long term complications are the roof leak, the basement flooding, and the electrical issues that appear after the move. Smart buyers do not reject aesthetics, but they also do not confuse aesthetics with structural integrity.

The same logic applies to cancer care. A fast radiographic response is an attractive kitchen. The patient still has to live in the house.

Why regulators and clinicians keep falling for the same trap

There is a reason ORR persists as a favored endpoint: medicine is under pressure to move fast. Patients with aggressive cancers cannot wait years for definitive survival data before promising therapies are made available. Regulators also need endpoints that are practical, especially for single arm trials where randomized evidence is unavailable or difficult to obtain.

So the problem is not that ORR is irrational. The problem is that speed changes epistemology. When the system is optimized for quick approval, quick publication, and quick optimism, it will naturally privilege measures that arrive quickly. That is a structural bias, not just an analytical mistake.

But there is a second bias at work: the human preference for clean stories. A drug that shrinks tumors sounds like a breakthrough. A surgery that cures the cancer but leads to years of infection risk and anatomical complications sounds less glamorous, even if the latter outcome matters more to patients in daily life.

This creates a recurring mismatch. Scientific success is often measured by what can be counted soonest. Patient success is often determined by what can be endured longest.

The challenge, then, is not to eliminate ORR or to romanticize long term complications as if they were the only truth. It is to build a richer scorecard that respects the full arc of treatment. The right question is not “Did it work?” but “What did it cost to work, and how long did the bill take to arrive?”

From endpoint to ecology: how to think like a patient

One of the most useful shifts in medicine is to stop thinking of treatment as a single event and start thinking of it as an ecology.

In an ecology, one intervention changes the habitat for everything else. A diversion alters urinary flow, infection risk, maintenance needs, and the anatomy of future care. A response inducing therapy may shrink tumors but also create toxicities, monitoring burdens, or downstream surgical consequences. The body is not a static target. It is a living system that adapts, compensates, and sometimes breaks in places the initial metric cannot see.

This is why patient centered thinking often looks less like chasing the largest early effect and more like balancing competing risks. A patient may prefer a slightly less dramatic early response if it comes with fewer late complications, fewer hospitalizations, or better function. Another patient may gladly accept higher burden for the chance of a major oncologic response. The right choice depends on values, prognosis, anatomy, and time horizon.

A useful question for clinicians and researchers is this: what outcome would still matter if the first scan were hidden? If the answer is “not much,” then the endpoint may be too narrow. If the answer remains meaningful over years of living, then you have found something closer to true benefit.

This framework also explains why older long term data are so valuable. They act like slow motion footage. Early results show the ball leaving the bat. Long term follow up shows whether it was a home run, a foul ball, or a hit that bounced into the stands but caused damage on the way out.

Key Takeaways

Do not confuse early response with durable benefit. Tumor shrinkage is important, but it is only the first horizon of value.
Always ask what happens after the initial win. Late complications, maintenance burden, and quality of life can overwhelm early gains.
Use a three horizon lens. Evaluate treatment by response, durability, and burden, not just by one headline metric.
Think in patient time, not trial time. Patients live the full sequence of consequences, while studies often stop at convenient milestones.
When choosing between options, compare the total life cycle cost of benefit. The best therapy is not always the one with the biggest early signal.

The real lesson: medicine should measure what it is willing to live with

The deepest connection between these two ideas is that both expose the same intellectual temptation: to treat a proxy as if it were the thing itself. ORR is a proxy for anticancer activity. Early postoperative success is a proxy for durable recovery. But proxies are only useful if we remember what they hide.

Cancer care teaches this with unusual force because the stakes are so high and the timeline so unforgiving. A treatment that impresses in the short term can impose a large, delayed burden. A surgery that appears definitive can generate a chronic maintenance problem. A trial can celebrate response while the patient’s life is still being rewritten by complications that only emerge years later.

The right conclusion is not cynical. It is more demanding. We should celebrate real response, but only when we can place it inside a longer story about survival, function, and lived burden. Otherwise, we are measuring victory at the wrong altitude.

The next time a treatment is called a success because it shrank a tumor, ask a harder question: What kind of success still looks like success ten years later? That is the question that separates a promising signal from a truly meaningful advance.