What Does "Complete Response" Actually Mean in Cancer Immunotherapy?

"Complete response" is one of the most hopeful phrases a patient can hear in oncology — and one of the most misunderstood. It has a precise, technical definition rooted in tumor measurement criteria, and it is genuinely good news when it happens. But CR is a statement about what a scan can see on a given day, not a promise about the future. This article explains how complete response is formally defined and measured, how it differs from "cure," why immunotherapy trials sometimes report response patterns that look nothing like chemotherapy trials, and what real CR rates from approved immunotherapies actually look like.

The Four RECIST Categories

Since the early 2000s, most solid-tumor trials have graded response using RECIST — Response Evaluation Criteria in Solid Tumors, now in its 1.1 revision, maintained by the RECIST Working Group. At the start of treatment, radiologists select "target lesions" on imaging (up to five total, up to two per organ) and add up their longest diameters to create a baseline sum. Every follow-up scan is compared back to that baseline. RECIST 1.1 defines four possible outcomes: complete response (CR) — disappearance of all target lesions; partial response (PR) — at least a 30% decrease in the sum of diameters; progressive disease (PD) — at least a 20% increase from the smallest sum recorded during treatment (with an absolute increase of at least 5mm), or the appearance of any new lesion; and stable disease (SD) — shrinkage or growth that doesn't meet either the PR or PD threshold. These categories are the shared language that lets oncologists compare trial results across cancer types and drug classes.

How Doctors Actually Measure a Complete Response

CR isn't just "the tumor looks smaller" — it's a specific, auditable standard. All target lesions must disappear entirely on imaging, and any non-target lesions must resolve or, for lymph nodes, shrink to a non-pathological size. Lymph nodes get their own rulebook because normal nodes are never truly "zero": a node with a short-axis diameter under 10mm is considered normal, nodes between 10 and 14mm are pathological but too small to serve as target lesions, and nodes of 15mm or more can be selected as targets. Critically, for a patient to be scored as a complete response, every pathological lymph node — target or non-target — must shrink below that 10mm short-axis threshold, not just get smaller. This is why a patient can have dramatic tumor shrinkage and still be scored PR rather than CR, if so much as one lingering lymph node sits at 11mm.

Complete Response Is Not the Same Thing as "Cured"

This is the distinction worth sitting with. CR means imaging and exams can no longer detect measurable disease — it does not mean every cancer cell is gone. Standard CT and PET scans can only reliably detect tumor deposits down to roughly a few millimeters; a population of cancer cells too small or too dispersed to image can persist beneath that resolution floor. Oncology calls this minimal residual disease (MRD) — cancer cells that remain detectable only through far more sensitive tools like flow cytometry or next-generation sequencing of blood or bone marrow, capable of finding roughly one tumor cell among 100,000 normal ones. The FDA's own guidance on multiple myeloma treats MRD status as a deeper, more sensitive marker than CR precisely because patients can achieve a "complete response" by imaging and physical exam while still testing MRD-positive underneath. That's the reason oncologists talk about "durable response" and "remission" rather than "cured" the day a scan comes back clean, and why CR patients still follow a defined surveillance schedule rather than walking away from care.

Why Immunotherapy Responses Don't Always Follow the Old Rules

RECIST was built for chemotherapy and targeted drugs, which typically either shrink a tumor directly or don't. Immunotherapies work differently — they activate the immune system, which can take months to fully engage a tumor, and the process of immune cells infiltrating a lesion can temporarily make it look larger on a scan before it shrinks. This phenomenon, pseudoprogression, is uncommon (most published estimates place it at well under 10% of patients on checkpoint inhibitors) but real enough that oncology developed iRECIST, a modified framework built by the same RECIST Working Group specifically for immunotherapy trials. Under iRECIST, apparent progression is initially flagged as "unconfirmed" (iUPD); if a clinically stable patient continues treatment and a follow-up scan 4–8 weeks later confirms genuine growth, it becomes "confirmed progressive disease" (iCPD). This buffer exists because some patients whose scans look worse at first go on to have deep, delayed responses — a pattern chemotherapy essentially never produces.

The Numbers: Real Complete Response Rates From Approved Immunotherapies

When immunotherapy CR does land, the data can be genuinely striking. In the pivotal ZUMA-1 trial of the CAR-T therapy axicabtagene ciloleucel (Yescarta) for refractory large B-cell lymphoma, 83% of patients had an objective response and 58% achieved a complete response — remarkable for a population that had already failed multiple prior treatments. At five-year follow-up, roughly 30% of all treated patients still had an ongoing complete response, with a median CR duration exceeding five years, prompting investigators to describe the results as supporting the drug's "curative potential" in a meaningful subset of patients. In advanced melanoma, the CheckMate 067 trial of nivolumab plus ipilimumab reported a 9% complete response rate for the combination at the initial 9-month analysis (versus 9% for nivolumab alone and 2% for ipilimumab alone) — and because immunotherapy responses deepen over time, longer follow-up of nivolumab monotherapy in the same trial family showed the complete response rate climb to 17% by 10 years, with a median duration of response exceeding eight years. These numbers illustrate both halves of the story this article is built around: CR is a real, durable, life-changing outcome for a meaningful share of patients — and it is measured, reported, and followed over years precisely because "complete" is not the same as "finished."

Bottom Line

If you or someone you love reaches complete response, that is genuinely one of the best outcomes oncology has to offer — but it's worth asking your oncologist three specific questions: what imaging or lab criteria were used to call it a CR, whether any MRD or other sensitive testing is available for your specific cancer type, and what your monitoring schedule looks like going forward and for how long. "Durable" complete response is itself a defined, time-bound term in the trials that establish it (commonly assessed at fixed follow-up marks like one, two, or five years) — so ask where you are on that timeline, and don't let a clean scan today replace the follow-up plan that turns a complete response into a lasting one.