Is Tumor Shrinking a Good Sign During Cancer Treatment?

Tumor shrinkage during cancer treatment is generally a positive signal, and across many cancer types it correlates with longer survival. But the relationship between a tumor getting smaller on a scan and what actually happens to a patient over the following months or years is more complicated than it first appears. A shrinking tumor can sometimes mask resistant cells lurking underneath, and in certain treatment scenarios, a tumor that appears to grow initially may actually be responding. The size of a tumor on imaging is one of the most important pieces of information oncologists use, but it is far from the only one that matters.

What Large Studies Show About Shrinkage and Survival

Oncologists have two ways of talking about tumor shrinkage. “Early tumor shrinkage” refers to a measurable reduction within the first several weeks of treatment, while “depth of response” captures the maximum reduction a tumor reaches at any point during therapy. Both have been studied extensively in colorectal cancer, where some of the strongest data exists. A large analysis drawing on the ARCAD database found that early tumor shrinkage substantially stratified survival outcomes across different treatment groups, with patients whose tumors shrank early living longer on average regardless of whether they received anti-EGFR or bevacizumab-based therapies.¹ A separate analysis from the phase III TRIBE trial confirmed the same pattern: both early shrinkage and depth of response predicted overall survival, progression-free survival, and post-progression survival even after adjusting for other prognostic factors.²

These findings extend beyond colorectal cancer. A broad review of early tumor shrinkage and depth of response across solid tumors found that both measures hold predictive value for progression-free and overall survival in multiple cancer types, though the exact cutoffs used to define “meaningful” shrinkage vary between studies, typically ranging from about a 20% to 30% reduction.³ So at a population level, across thousands of patients, the trend is clear: tumors that shrink early and deeply tend to forecast better outcomes.

Why Shrinkage Alone Can Be Misleading

The population-level trend hides a deceptive detail about individual tumors. A fast-growing tumor can shrink dramatically under treatment and still carry a worse prognosis than a slow-growing tumor that barely budges. This is not just theoretical. An analysis published in the Journal of Nuclear Medicine walked through a straightforward example: imagine two tumors diagnosed at the same size. One doubles in volume every 90 days and the other every 200 days. After three months of therapy, the fast grower shrinks by a third of its diameter while the slow grower stays the same. Under standard response criteria, the fast grower looks like a success and the slow grower looks like a failure. But because the fast grower rebounds quickly once treatment pressure lifts, it surpasses the slow grower in size within about a year.⁴ The point is that shrinkage reflects what happened to the tumor during treatment, but it says less about the biology driving the tumor’s long-term behavior.

A classic review of growth rate patterns reinforced this idea decades ago: rapidly growing tumors often respond well to therapy and can achieve complete responses that are durable, but rapidly growing tumors in advanced stages behave more like slowly growing tumors, with responses that tend not to last. Slowly growing tumors, meanwhile, respond less favorably overall.⁵ This creates an awkward situation for patients and doctors alike: a tumor can look like it is melting away on scans while the underlying biology is setting the stage for a comeback.

The Cancer Stem Cell Problem

One reason tumor shrinkage sometimes fails to translate into a cure involves a subpopulation of cells within tumors that behave differently from the bulk. These cells, known as cancer stem cells, can self-renew and restart tumor growth even after the vast majority of the tumor has been killed. If treatment destroys the ordinary cancer cells that make up most of a tumor’s mass but leaves cancer stem cells intact, scans will show an impressive reduction in size. The patient and the oncologist see encouraging results. But the surviving cells can regrow the tumor months or years later.⁶

Cancer stem cells are particularly resistant to chemotherapy and radiation, which is part of why they survive when other cancer cells do not. Their high adaptive abilities and capacity for self-renewal make them a key driver of treatment failure and tumor relapse.⁷ This does not mean that tumor shrinkage is meaningless in the presence of cancer stem cells. It does mean that shrinkage on a scan is measuring the loss of the treatment-sensitive population, not necessarily the complete elimination of every dangerous cell.

When a Growing Tumor Is Actually Good News

Immunotherapy has introduced a phenomenon that can confuse patients and clinicians in the opposite direction. In pseudoprogression, a tumor initially appears to grow larger on imaging, or new lesions seem to appear, before the tumor eventually starts shrinking. This happens because the immune system, newly activated by the therapy, floods the tumor with inflammatory cells. The influx of immune cells temporarily inflates the measured size of the mass, even though the treatment is working.⁸

The practical challenge is that pseudoprogression looks identical to true progression on a standard scan. Patients who experience it can benefit from continuing immunotherapy, but the false appearance of growth often leads to premature discontinuation of treatment because doctors and patients assume the drug has failed.⁹ Pseudoprogression is relatively uncommon, occurring in a minority of immunotherapy patients, but it is common enough that oncologists now use modified response criteria and sometimes repeat scans several weeks later before making the call to stop treatment. For patients on immunotherapy, a single scan showing growth is not necessarily bad news.

Pathological Complete Response and What It Means

If tumor shrinkage on imaging is an imperfect measure, what does the evidence say about shrinkage confirmed under a microscope? When chemotherapy or immunotherapy is given before surgery (called neoadjuvant therapy), surgeons remove the treated area, and pathologists examine the tissue for any remaining cancer cells. A pathological complete response means no viable cancer cells are found. This is as close as medicine gets to confirming that treatment truly wiped out the local tumor.

A meta-analysis of patients who received neoadjuvant immunotherapy across multiple cancer types found that achieving a pathological complete response was associated with roughly half the risk of disease progression or death compared to patients who did not achieve it.¹⁰ In non-small cell lung cancer specifically, patients who achieved a pathological response to neoadjuvant treatment survived an average of 75 months, compared to 30 months for unresponsive patients. Recurrence was strikingly rare in the responsive group: only 1 out of 30 patients who recurred during follow-up came from the group that had achieved a pathological response.¹¹

Pathological complete response is not available for every patient, since it requires surgery. But where it can be assessed, it provides a much stronger signal than imaging alone. A tumor that appears smaller on a CT scan may still harbor pockets of living cancer cells. A tumor confirmed to be entirely gone under the microscope is a substantially better indicator.

The Scar Tissue Puzzle

After chemotherapy, some patients are left with a visible mass on imaging even though the treatment has been effective. This is especially well studied in Hodgkin lymphoma, where a residual mass after chemotherapy is common and does not necessarily contain active disease. PET scans, which measure metabolic activity rather than just physical size, help distinguish scar tissue from active cancer. In Hodgkin lymphoma, a negative PET scan after treatment carries a very high negative predictive value, ranging from about 86% to 94% depending on treatment intensity, meaning that when PET says there is no active disease, it is almost always right. But when PET flags something as potentially active, the positive predictive value is much lower, around 25% to 40%.¹²

This asymmetry matters in two directions. A patient who has a residual lump after treatment but a clean PET scan can usually be reassured. But a patient whose PET lights up in a treated area may not actually have recurrent cancer. The leftover tissue may just be inflammation or healing. Doctors are often cautious about restarting aggressive treatment based on a positive PET alone, preferring to biopsy or re-scan before committing.

When Tumors Shrink Too Fast

Rapid tumor destruction is not always physically safe. When large numbers of cancer cells die at once, they dump their internal contents into the bloodstream. This can cause a condition called tumor lysis syndrome, which leads to dangerously high levels of uric acid, potassium, and phosphorus, along with low calcium and potential kidney damage.¹³ Tumor lysis syndrome is most common in blood cancers with high tumor burden, though it can occur in solid tumors as well. Oncologists monitor for it during the first cycles of treatment and often give preventive medications to patients at high risk. The irony is that the better the treatment works, the more dangerous this particular side effect becomes.

Measuring Response Beyond Size

Oncologists increasingly recognize that tumor size on imaging is only part of the picture. Circulating tumor DNA, fragments of tumor genetic material shed into the blood, has emerged as a complementary tool for monitoring treatment response. Because ctDNA levels can change faster than tumor size, a drop in ctDNA can signal that treatment is working before any change shows up on a scan. Conversely, rising ctDNA may warn of resistance or recurrence while images still look stable.¹⁴

In breast cancer, researchers have also worked on predicting the pattern of shrinkage before surgery. The way a tumor shrinks matters for surgical planning: a tumor that contracts uniformly into a smaller ball is easier to remove conservatively than one that fragments into scattered islands. Predictive models are being developed to help identify which patients are good candidates for less invasive surgery based on how their tumor is expected to respond.¹⁵ So even when shrinkage is clearly happening, the shape and pattern of that shrinkage carry additional clinical information.

How Shrinkage Affects How Patients Actually Feel

For patients with symptomatic tumors, the practical question is often less about survival statistics and more about whether treatment makes them feel better. A study of patients with metastatic colorectal cancer found that those who experienced early tumor shrinkage had meaningful improvements in quality of life scores within the first eight weeks. Among patients who had symptoms at baseline, those whose tumors shrank early saw their global health and quality-of-life scores improve by about 6 points, while patients without early shrinkage actually declined by about 5 points. Physical functioning and social functioning showed similar gaps.¹⁶ This makes intuitive sense: a tumor pressing on an organ or nerve that gets smaller relieves pressure, reduces pain, and restores function. For patients dealing with day-to-day symptoms, early shrinkage is a genuinely positive experience beyond its statistical implications.

The Expectation Gap

One of the more sobering findings in cancer communication research is how patients interpret treatment response. A study of patients with advanced cancer found that, compared to their oncologists, patients consistently overestimated the chances of their tumor responding, their likelihood of being cured, how much symptom relief they would experience, and how much longer treatment would help them survive. The median gap between patient and oncologist estimates was about 20 percentage points for curability, tumor response, and survival benefit.¹⁷

This matters directly to how patients interpret tumor shrinkage. A patient who sees their tumor get smaller on a scan may interpret that as strong evidence of cure, while their oncologist understands it as one encouraging data point in a complex picture. Neither interpretation is wrong exactly, but the gap can lead to significant emotional whiplash if the disease later progresses despite an initially positive imaging response. Having a realistic framework for what shrinkage does and does not guarantee is one of the most important things a patient can take from understanding this topic.

Acquired Resistance and the Limits of Initial Response

Even when a targeted therapy produces a dramatic initial response, tumors can develop resistance mutations that allow them to escape the drug. This has been documented across many cancer types and drug classes. In one case, a pancreatic cancer patient achieved a partial response to a third-generation EGFR inhibitor, only for the tumor to develop a specific resistance mutation that rendered the drug ineffective.¹⁸ The tumor shrank, which was a good sign at the time, but the shrinkage ultimately did not prevent the cancer from finding a way around the treatment.

Kinetic modeling of tumors has attempted to tease apart which features of a treatment response actually predict long-term outcomes. In gastric cancer, one analysis found that neither the rate of tumor shrinkage nor the fraction of resistant cells (which relates to how deeply the tumor responds) was associated with increased survival among tumors that showed a response to treatment.¹⁹ In other words, two patients whose tumors shrank at the same rate and to the same degree could still have very different outcomes depending on other biological factors that scans do not capture.

The Abscopal Effect and Radiation

Radiation therapy is designed to shrink or destroy tumors in a targeted area, but occasionally it triggers something far more interesting. In the abscopal effect, radiation directed at one tumor site causes distant, untreated tumors to shrink as well. This happens because local radiation can release tumor-specific signals that activate the immune system broadly, essentially teaching the body to recognize and attack cancer cells elsewhere.²⁰ The abscopal effect was historically considered rare, but it has been observed more frequently in patients who receive radiation alongside immunotherapy.²¹

For patients, the abscopal effect represents one of the most unambiguously good versions of tumor shrinkage: not only is the treated tumor responding, but the immune system has been primed to hunt down disease in other parts of the body. Researchers are actively studying how to make this effect more reliable rather than sporadic, including by combining specific radiation doses and schedules with checkpoint inhibitors.

Why Drugs Get Approved Based on Tumor Shrinkage

The FDA’s accelerated approval pathway allows cancer drugs to reach patients based on surrogate measures like tumor shrinkage rather than waiting for long-term survival data. The logic is straightforward: if patients with deadly cancers have no good options, waiting years for survival data costs lives. Tumor response rate, where the drug is shown to shrink tumors, serves as a stand-in for clinical benefit.²²

The catch is that companies are required to run confirmatory trials afterward to prove the drug actually improves survival or quality of life, not just tumor size. These confirmatory studies frequently take a long time to complete, and in some cases they show that a drug that shrank tumors did not ultimately help patients live longer or feel better.²³ This regulatory dynamic underscores the broader scientific point: tumor shrinkage is a reasonable bet, and it is often the right bet, but it is not a guarantee. The gap between “the tumor got smaller” and “the patient lived longer” is exactly the space that confirmatory trials are supposed to fill, and the fact that filling it sometimes reveals no benefit is a reminder that imaging response and meaningful clinical outcome are related but distinct.