Some patients on immunotherapy appear to get dramatically worse before they get better because their immune system is flooding the tumor with attack cells, causing inflammation, swelling, and temporary enlargement that mimics cancer progression on scans. This phenomenon, known as pseudoprogression, occurs in roughly 6% of solid tumor patients receiving immune checkpoint inhibitors and is actually one of the most encouraging signs a patient can receive — pseudoprogressors in one study had a median overall survival of 29.9 months compared to just 8.0 months for those with true progression. The problem is that this temporary worsening looks nearly identical on imaging to genuine treatment failure, and in the worst cases, to a dangerous acceleration of disease called hyperprogression. Consider a patient with advanced melanoma who begins pembrolizumab and returns for a follow-up scan eight weeks later.
The images show the primary tumor has grown by 25% and a new lesion has appeared in the liver. The oncologist faces a decision that could determine whether this patient lives or dies: stop treatment because it appears to be failing, or continue because the immune system may be doing exactly what it should. In melanoma specifically, pseudoprogression rates run between 6.4% and 15.8%, making this dilemma far from rare. This article examines why pseudoprogression happens at the cellular level, how common it really is across different cancer types, and the critical distinction between pseudoprogression and its dangerous counterpart, hyperprogression. We also cover the diagnostic tools oncologists now use to tell the difference — from liquid biopsies that can detect true responses in as few as 24 days to machine learning models that may predict outcomes before treatment even begins — and what all of this means for patients and caregivers navigating the anxiety of early immunotherapy scans.
Table of Contents
- What Causes Some Immunotherapy Patients to Get Worse Before Getting Better?
- Pseudoprogression vs. Hyperprogression — Why the Distinction Is Life or Death
- How Common Is Pseudoprogression Across Different Cancer Types?
- The Diagnostic Tools Oncologists Use to Tell the Difference
- The Real Danger of Stopping Immunotherapy Too Soon
- What Pseudoprogression Means for Brain Health and Dementia Caregivers
- Where Pseudoprogression Research Is Headed
- Conclusion
- Frequently Asked Questions
What Causes Some Immunotherapy Patients to Get Worse Before Getting Better?
When immune checkpoint inhibitors like nivolumab or pembrolizumab release the brakes on the immune system, T cells and other lymphocytes surge into tumor tissue in large numbers. this infiltration causes local inflammation that physically enlarges the tumor mass. On a CT or MRI scan, this swollen, immune-cell-packed tumor looks bigger — sometimes substantially bigger — than it did before treatment started. But a biopsy of these pseudoprogressing tumors tells a completely different story. Instead of thriving cancer cells, pathologists find extensive lymphocyte infiltration and tumor necrosis, confirming that the immune system is actively destroying the cancer from within even as the tumor appears to grow on the outside. This is fundamentally different from how traditional chemotherapy works, and that difference is the source of enormous confusion. Chemotherapy directly poisons cancer cells, so tumor shrinkage on imaging reliably indicates treatment success.
Immunotherapy works indirectly — it activates the patient’s own immune system, which then attacks the cancer. That activation phase, with its accompanying inflammation and cellular infiltration, creates a window where imaging cannot distinguish between a tumor being destroyed and a tumor that is growing. Roughly 77% of pseudoprogression cases are detected at the very first follow-up scan after baseline, meaning most patients who will experience this phenomenon encounter it early, during peak immune activation. The timeline matters enormously. The average time from apparent progression to confirmed tumor response in pseudoprogression cases is approximately six months. That is six months of scans that may continue to look concerning, six months of uncertainty for patients and families, and six months during which a physician might be tempted to abandon a therapy that is actually working. For caregivers of patients with brain tumors or brain metastases — populations already dealing with cognitive and neurological challenges — this waiting period adds a layer of emotional difficulty that cannot be overstated.
Pseudoprogression vs. Hyperprogression — Why the Distinction Is Life or Death
The existence of hyperprogression is what makes pseudoprogression so dangerous to mismanage. Hyperprogression is the genuine acceleration of tumor growth triggered by immunotherapy, occurring in an estimated 4% to 29% of patients depending on the study and cancer type, with a combined incidence of roughly 13.4%. Patients who experience hyperprogression have a median overall survival of only 3.4 months, compared to 6.2 months for patients whose disease progresses at a normal rate without immunotherapy-driven acceleration. These two phenomena — one a sign the treatment is working brilliantly, the other a sign it is making the disease actively worse — can look nearly identical on early imaging. This is one of the most urgent problems in modern oncology. A patient showing tumor growth at eight weeks could be in the 6% experiencing pseudoprogression, with excellent long-term prospects, or in the roughly 13% experiencing hyperprogression, with a life expectancy measured in weeks.
The clinical consequences of guessing wrong run in both directions. Stopping immunotherapy in a pseudoprogressor deprives them of a treatment that may have given them years of additional life. Continuing immunotherapy in a hyperprogressor wastes precious time that could be spent on alternative therapies while the disease accelerates. However, it is important to note that current tools for distinguishing between these two outcomes remain imperfect. No single biomarker reliably separates pseudoprogression from hyperprogression at the time of that first ambiguous scan. Oncologists must rely on a combination of clinical judgment, serial imaging, and emerging liquid biopsy technologies — and even with all of these, certainty often comes only in retrospect. Patients and caregivers should understand that when a physician recommends continuing immunotherapy through apparent progression, they are making a calculated bet based on probabilities, not a guarantee.
How Common Is Pseudoprogression Across Different Cancer Types?
A systematic review and meta-analysis published in Radiology found that the pooled incidence of pseudoprogression across all solid tumors treated with immune checkpoint inhibitors is approximately 6%, with a 95% confidence interval of 5% to 7%. But this average obscures meaningful variation by cancer type. Melanoma has the highest reported rates, ranging from 6.4% to 15.8%, likely because melanoma is among the most immunogenic cancers and tends to provoke strong immune responses to checkpoint blockade. Non-small cell lung cancer shows notably lower rates, between 1.5% and 5%, while genitourinary cancers come in at roughly 7%. For patients with brain metastases or primary brain tumors — a population of particular relevance to dementia and brain health — pseudoprogression presents unique challenges. Brain imaging is already difficult to interpret after radiation therapy, which can cause its own form of pseudoprogression unrelated to immunotherapy.
When immunotherapy is added to the mix, distinguishing true progression from immune-mediated swelling from radiation-induced changes becomes extraordinarily complex. A growing lesion in the brain also carries immediate clinical consequences — seizures, cognitive decline, personality changes — that a growing lesion in the lung or liver may not, making the stakes of the wait-and-see approach considerably higher. These numbers also highlight an uncomfortable reality: pseudoprogression is relatively uncommon. When a patient on immunotherapy shows tumor growth on a scan, true progression is still the most likely explanation in the vast majority of cases. The challenge lies in identifying the minority of patients for whom apparent progression is actually a good sign, without subjecting the majority to prolonged ineffective treatment. This is why the development of reliable biomarkers for pseudoprogression has become such a research priority.
The Diagnostic Tools Oncologists Use to Tell the Difference
The introduction of iRECIST criteria in 2017 represented the first systematic attempt to address immunotherapy’s unusual response patterns. Under traditional RECIST guidelines, new lesions or significant tumor growth automatically classified a patient as having progressive disease. iRECIST changed this by requiring that progression be confirmed on two consecutive scans taken at least four weeks apart before classifying it as true failure. New lesions alone no longer automatically mean the treatment has stopped working. This gives pseudoprogression time to declare itself without immediately triggering a change in therapy. However, waiting four or more weeks for a confirmatory scan is not always practical or safe, particularly for patients with symptomatic progression. This is where liquid biopsy technologies are beginning to change the calculus.
Circulating tumor DNA, or ctDNA, can detect a true treatment response within approximately 24.5 days — compared to 72.5 days for conventional imaging. When ctDNA levels are declining even as tumors appear to grow on scans, this discordance strongly suggests pseudoprogression rather than true progression. Serum IL-8 levels offer another window: early decreases in IL-8 correlate with longer overall survival even in patients showing initial tumor growth on imaging, providing a biochemical signal that the immune response is headed in the right direction. Perhaps the most promising development came in a 2024 study demonstrating that CT-based radiomic biomarkers, analyzed by a machine learning model using intratumoral and peritumoral features, could predict pseudoprogression versus hyperprogression in non-small cell lung cancer patients before treatment even begins. If validated in larger studies and across cancer types, this kind of pre-treatment risk stratification could fundamentally change how oncologists counsel patients and plan follow-up. The tradeoff is that these tools remain largely confined to research settings and academic medical centers. A community oncologist managing a patient in a rural clinic may have access to ctDNA testing but almost certainly does not yet have access to radiomic prediction models.
The Real Danger of Stopping Immunotherapy Too Soon
The survival data on premature treatment discontinuation tells a stark story. Patients who continued immunotherapy through pseudoprogression achieved a median overall survival of 16.6 months, compared to only 7.4 months for patients who stopped treatment when scans looked worse. That is more than a doubling of survival time, lost because of a misinterpretation of imaging. In one study, 100% of patients in the pseudoprogression group were alive at one year, an extraordinary outcome for patients with advanced solid tumors. These numbers carry an important caveat. The decision to continue immunotherapy through apparent progression should never be made casually. Some patients experience cytokine release syndrome during immunotherapy — a systemic inflammatory response featuring fever, dangerously low blood pressure, and difficulty breathing, typically occurring 3 to 14 days after T-cell-based immunotherapy.
In a case series, 40% of patients with cytokine release syndrome experienced multiple flare episodes with a median duration of 15 days. Most patients recover within one to two weeks without long-term complications, but the experience can be frightening and physically taxing, especially for older patients or those with existing neurological conditions. The warning here is for patients and caregivers who may hear about pseudoprogression and interpret it as a reason to push through any and all treatment-related deterioration. Not every worsening is pseudoprogression. In fact, most are not. The decision to continue therapy should be a carefully reasoned clinical judgment involving imaging trends, biomarker data, the patient’s functional status, and an honest conversation about probabilities. Patients who are rapidly declining physically are generally not good candidates for a wait-and-see approach, regardless of what the statistics say about pseudoprogression rates.
What Pseudoprogression Means for Brain Health and Dementia Caregivers
For families already managing cognitive decline in a loved one, the possibility of pseudoprogression during immunotherapy adds a deeply unsettling layer of uncertainty. A patient with brain metastases who develops worsening confusion, memory problems, or behavioral changes after starting immunotherapy may be experiencing immune-mediated inflammation in the brain — a potentially encouraging sign — or genuine tumor growth. The symptoms feel the same to the patient and the family regardless of the underlying cause.
Caregivers in this situation should insist on clear communication with the oncology team about the specific criteria being used to evaluate treatment response, whether ctDNA or other biomarker testing is available, and what the concrete plan is for the next scan and the decision points it will trigger. Documenting day-to-day cognitive and functional changes in a journal can provide the care team with granular information that a scan taken weeks apart cannot capture. These observations — when the patient is sharper versus more confused, whether symptoms are stable or worsening over days — become data points that help the medical team interpret ambiguous imaging in the broader clinical context.
Where Pseudoprogression Research Is Headed
The central goal of current research is to move from reactive assessment — waiting to see whether apparent progression resolves — to predictive assessment that can classify patients before or immediately after treatment begins. The 2024 radiomic biomarker study represents one frontier, using machine learning to extract predictive signals from standard CT scans that human radiologists cannot perceive. If these models prove generalizable across tumor types and treatment regimens, they could eliminate much of the diagnostic uncertainty that currently forces oncologists into educated guesses.
Simultaneously, the integration of multiple biomarker streams — ctDNA kinetics, serum cytokine profiles, and imaging-based features — into composite prediction models may eventually give physicians a reliable dashboard for distinguishing pseudoprogression from hyperprogression in real time. The clinical impact would be profound. Rather than the current approach of treating all apparent progressors the same and hoping for the best, oncologists could stratify patients into high-confidence categories and tailor their management accordingly. For patients with brain involvement, where the consequences of both undertreating and overtreating are severe, this precision cannot come soon enough.
Conclusion
Pseudoprogression remains one of the most counterintuitive phenomena in modern medicine — a scenario where getting worse is actually the clearest sign that treatment is working. The roughly 6% of solid tumor patients who experience it go on to have dramatically better outcomes than those with true progression, with median survival differences measured in years rather than months. But the existence of hyperprogression, which can look nearly identical on early scans and carries a devastating prognosis, means that optimism must always be tempered by careful diagnostic work.
For patients, caregivers, and families navigating immunotherapy — particularly those already dealing with brain health challenges — the practical takeaway is this: ask your oncology team specifically about pseudoprogression, request information about available biomarker testing including ctDNA, and understand the timeline and criteria they will use to evaluate treatment response. Continuing effective immunotherapy through a period of pseudoprogression can more than double survival time compared to stopping prematurely. That knowledge, paired with the right diagnostic tools and an honest conversation about probabilities, is the difference between abandoning a treatment that is saving your life and staying the course through a storm that will pass.
Frequently Asked Questions
What percentage of immunotherapy patients experience pseudoprogression?
Approximately 6% of all solid tumor patients on immune checkpoint inhibitors experience pseudoprogression, though rates vary by cancer type. Melanoma has the highest rates at 6.4% to 15.8%, while non-small cell lung cancer is lower at 1.5% to 5%.
How long does pseudoprogression last before the tumor actually shrinks?
The average time from apparent progression to confirmed treatment response is approximately six months. About 77% of pseudoprogression cases are first detected at the very first follow-up scan after starting treatment.
What is the difference between pseudoprogression and hyperprogression?
Pseudoprogression is temporary tumor swelling caused by immune cells infiltrating and attacking the cancer — it is a positive sign. Hyperprogression is a genuine acceleration of cancer growth triggered by immunotherapy, occurring in roughly 13.4% of patients, with a median survival of only 3.4 months. They can look similar on early scans but have opposite outcomes.
Can blood tests help distinguish pseudoprogression from true progression?
Yes. Circulating tumor DNA testing can detect a true response within about 24.5 days, compared to 72.5 days for imaging. Declining ctDNA levels during apparent radiographic progression strongly suggest pseudoprogression. Decreasing serum IL-8 levels also correlate with better outcomes despite initial tumor growth on scans.
Should immunotherapy be continued if scans show the tumor is growing?
This decision must be made carefully with the oncology team. Data shows patients who continued immunotherapy through pseudoprogression had a median overall survival of 16.6 months versus 7.4 months for those who stopped. However, not all apparent progression is pseudoprogression, and patients who are rapidly declining physically may not be candidates for continued treatment.
What are the iRECIST criteria?
Introduced in 2017, iRECIST are modified response evaluation guidelines designed specifically for immunotherapy. They require that progression be confirmed on two consecutive scans at least four weeks apart before classifying it as true treatment failure, giving pseudoprogression time to reveal itself.
