A Ki-67 score is a percentage that tells you how many cells in a tumor sample are actively dividing. The number comes from staining a slice of tumor tissue with an antibody that latches onto a protein called Ki-67, which appears only when a cell is gearing up to split or is in the middle of doing so. A tumor with a Ki-67 score of 5% has relatively few cells dividing at any given moment; one scored at 80% is proliferating furiously. The score matters because faster-growing tumors generally behave more aggressively, but they also tend to respond better to treatments that target dividing cells. How oncologists interpret the number depends heavily on the type of cancer, the cutoff values used, and what treatment decisions are on the table.
The Protein Behind the Score
Ki-67 is a nuclear protein found in cells during every active phase of the cell cycle but completely absent from resting cells. That on/off quality makes it a reliable flag for whether a cell is proliferating or sitting quietly.1Journal of Cellular Physiology. The Ki-67 protein: From the known and the unknown The name itself has an unglamorous origin: “Ki” stands for the University of Kiel in Germany, where the antibody was first developed, and “67” was simply the number of the well in the laboratory plate where the clone grew.2PubMed Central. The intrinsically disorderly story of Ki-67 Since then, a refined antibody clone called MIB-1 has become the standard tool used in clinical labs around the world.
Despite decades of use, the exact biological function of Ki-67 remains surprisingly murky. It is not itself driving cell division the way oncogenes do. It sits on the surface of chromosomes during mitosis and seems to play a role in organizing the material around chromosomes, but knocking it out experimentally does not always stop cells from dividing. Researchers treat it more as a bystander that reliably shows up at the scene than as a mastermind. For clinical purposes, though, its reliability as a proliferation flag is what counts.
How the Score Is Produced
After a biopsy or surgical removal, a thin slice of the tumor is embedded in paraffin wax, sliced even thinner, and placed on a glass slide. The slide goes through an automated staining process where the MIB-1 antibody binds to any Ki-67 protein present, and a chemical detection system makes those bound antibodies visible as a brown color under a microscope.3PubMed Central. Clinical and Analytical Validation of Two Methods for Ki-67 Scoring in Formalin Fixed and Paraffin Embedded Tissue Sections of Early Breast Cancer The pathologist then counts how many tumor cell nuclei stain brown versus how many do not, and the result is reported as a percentage.
This sounds straightforward, but the counting step is where things get messy. Some pathologists scan the entire tumor section and estimate an average. Others focus on the tumor’s growing edge, counting a set number of cells in several fields of view.3PubMed Central. Clinical and Analytical Validation of Two Methods for Ki-67 Scoring in Formalin Fixed and Paraffin Embedded Tissue Sections of Early Breast Cancer Different methods can yield different numbers from the same slide. A lightly staining nucleus might be called positive by one reader and negative by another. This observer variability is the biggest practical limitation of Ki-67 scoring and the reason that specific cutoff numbers can vary between institutions and guidelines.
Ki-67 in Breast Cancer
Breast cancer is where Ki-67 sees its heaviest clinical use. The score helps distinguish between subtypes of hormone-receptor-positive disease, particularly Luminal A and Luminal B tumors. Luminal A cancers tend to have lower Ki-67 values and grow slowly, while Luminal B tumors show higher proliferation and more aggressive behavior, including a higher rate of lymph node involvement.4PubMed Central. Difference between Luminal A and Luminal B Subtypes According to Ki-67, Tumor Size, and Progesterone Receptor Negativity Providing Prognostic Information Separating these two groups matters because Luminal B patients are more likely to benefit from chemotherapy on top of hormone therapy, while many Luminal A patients can safely skip it.
The question of where exactly to draw the line has been contentious. An older cutoff of 14% was widely used, but when researchers compared Ki-67-based classification against gene-expression profiling (the PAM50 test), that threshold only correctly classified about 70% of patients. Raising the cutoff to 20% improved accuracy to roughly 76% and bumped sensitivity above 93%, though it came at the cost of some specificity.5EMJ Oncology. Identifying the Best Ki-67 Cut-Off for Determining Luminal Breast Cancer Subtypes Using Immunohistochemical Analysis and PAM50 Genomic Classification Many centers now use 20% as the dividing line, but you may still see 14% referenced in older reports or certain international guidelines. If your pathology report uses a different cutoff than what you read online, that is likely why.
Predicting Response to Chemotherapy Before Surgery
When breast cancer patients receive chemotherapy before surgery to shrink a tumor, Ki-67 can help predict whether the tumor will disappear entirely, a result called pathological complete response. A meta-analysis pooling data from over 10,000 patients found that those with high Ki-67 values were roughly three times more likely to achieve a complete response compared to those with lower scores.6PubMed. The predictive value of Ki-67 before neoadjuvant chemotherapy for breast cancer: a systematic review and meta-analysis That held true regardless of whether the tumors were hormone-receptor-positive, HER2-positive, or triple-negative.
The paradox here is important to understand: a high Ki-67 score means the cancer is growing fast, which sounds bad, and it often is in terms of long-term prognosis if left untreated. But chemotherapy works by targeting cells that are dividing. A rapidly proliferating tumor presents more targets. In one study, Ki-67 was the single strongest independent predictor of complete response to pre-surgical chemotherapy, outperforming estrogen receptor status and HER2 status in multivariate analysis, with a cutoff around 25% best separating responders from non-responders.7PubMed Central. Ki-67 as a predictor of response to neoadjuvant chemotherapy in breast cancer patients Another study placed the optimal threshold even higher, at 35%, for predicting complete response.8PubMed Central. Assessment of the Predictive Role of Ki-67 in Breast Cancer Patients’ Responses to Neoadjuvant Chemotherapy The variation in cutoffs across studies reflects the reality that Ki-67 is a continuous variable being forced into binary decisions.
Tracking Ki-67 Changes During Treatment
Ki-67 is not only useful as a one-time snapshot. Changes in the score during a short course of treatment can reveal whether a therapy is hitting the brakes on the tumor’s growth. In hormone-receptor-positive breast cancers treated with a brief course of endocrine therapy before surgery, the median Ki-67 dropped from 12% to 3%. The vast majority of patients saw their score fall to 10% or below, and among those who started with elevated Ki-67, about three-quarters achieved a meaningful response.9ESMO Open. Early proliferation dynamics and long-term outcomes of short-term preoperative endocrine therapy in early estrogen receptor-positive/HER2-negative breast cancer
This approach, sometimes called a “window-of-opportunity” trial, can be completed in as little as one to two weeks. Research with the aromatase inhibitors anastrozole and tamoxifen showed that most of the Ki-67 suppression happened within the first two weeks and held steady at twelve weeks.10Clinical Cancer Research. Short-Term Changes in Ki-67 during Neoadjuvant Treatment of Primary Breast Cancer with Anastrozole or Tamoxifen Alone or Combined Correlate with Recurrence-Free Survival Even seven days of letrozole significantly reduced Ki-67 in postmenopausal women with hormone-positive tumors.11PubMed Central. Effective Ki-67 Reduction by Pre-Surgery Short-Term Administration of Letrozole in (Hormone-Positive) Breast Cancer The idea is that if a tumor’s Ki-67 drops sharply in response to a drug, that drug is likely to work well for that patient long term. If it does not budge, the oncologist has early evidence that a different strategy may be needed.
How Ki-67 Relates to Genomic Tests
Patients with early hormone-receptor-positive breast cancer may be offered multigene assays like Oncotype DX, which analyze the activity of a panel of genes, including proliferation genes, to produce a recurrence score that guides chemotherapy decisions. A natural question is whether Ki-67 and these expensive genomic tests are measuring the same thing. The answer is: partly, but not closely enough to be interchangeable.
A systematic review of 18 studies found that most reported a positive correlation between Ki-67 and the Oncotype DX recurrence score, and the combined result was statistically significant.12PubMed Central. Association Between Ki-67 Proliferative Index and Oncotype-Dx Recurrence Score in Hormone Receptor-Positive, HER2-Negative Early Breast Cancers: A Systematic Review of the Literature A population-based study found a moderate correlation, with about 71% of patients who had high recurrence scores also showing Ki-67 at or above 20%.13PubMed Central. Population-based estimate for the correlation of the Oncotype Dx Breast Recurrence Score® result and Ki-67 IHC MIB-1 pharmDx in HR+, HER2-, node-positive early breast cancer But at the individual level the agreement can break down. One small study found essentially zero correlation between Ki-67 and recurrence score.14PubMed Central. Correlating Ki67 and other prognostic markers with Oncotype DX recurrence score in early estrogen receptor-positive breast cancer
The disconnect makes sense when you consider that genomic tests measure the activity of many genes beyond just proliferation, including genes involved in invasion, hormone signaling, and immune response. Ki-67 captures one axis of tumor behavior; genomic tests capture several. For patients in regions where genomic testing is unavailable or unaffordable, Ki-67 remains a useful and widely accessible surrogate, but where genomic tests are available, they provide a more complete picture.
Beyond Breast Cancer
Ki-67 is not exclusive to breast cancer. It shows up on pathology reports across a wide range of tumor types, though its role and the cutoff values change with each disease.
Neuroendocrine Tumors
In neuroendocrine tumors of the gut and pancreas, Ki-67 has its most formally established diagnostic role. These tumors are graded directly by their Ki-67 index: Grade 1 tumors have a Ki-67 below 3%, Grade 2 falls between 3% and 20%, and Grade 3 exceeds 20%. The World Health Organization classification system requires Ki-67 evaluation in the diagnostic workup of these tumors, making it one of the few cancers where Ki-67 is not optional but mandatory for classification.15PubMed Central. Diagnostic, Prognostic, and Predictive Role of Ki67 Proliferative Index in Neuroendocrine and Endocrine Neoplasms: Past, Present, and Future The grade directly influences treatment decisions, distinguishing tumors that can be watched or treated with targeted agents from those requiring aggressive chemotherapy.
Prostate Cancer
In prostate cancer, Ki-67 is not yet part of routine pathology reports at most institutions, but a growing body of evidence suggests it probably should be. A systematic review and meta-analysis found that high Ki-67 expression in localized prostate cancer predicted worse outcomes across multiple endpoints, including disease-specific survival, biochemical recurrence, and overall survival.16PubMed. Prognostic role of Ki-67 score in localized prostate cancer: A systematic review and meta-analysis A study from the Mayo Clinic reported that after adjusting for other known risk factors, each 1% increase in Ki-67 expression was associated with a 12% increased risk of cancer-specific death.17PubMed. Prostate cancer Ki-67 (MIB-1) expression, perineural invasion, and gleason score as biopsy-based predictors of prostate cancer mortality: the Mayo model More recent work has confirmed a positive correlation between Ki-67 and both higher-grade disease and worse risk stratification.18PubMed Central. Can we use Ki67 expression to predict prostate cancer aggressiveness? The biomarker could help identify which localized prostate cancers truly need treatment versus which can be safely monitored, a distinction that matters enormously given the side effects of prostate surgery and radiation.
Lymphomas
In blood cancers, Ki-67 helps distinguish between different types of lymphoma that can look similar under the microscope. Diffuse large B-cell lymphoma, the most common aggressive lymphoma, typically shows Ki-67 values ranging from 40% to 90%. Burkitt lymphoma, an even more aggressively proliferating cancer, drives Ki-67 close to 100%.19PubMed Central. Ki-67 as a Marker to Differentiate Burkitt Lymphoma and Diffuse Large B-cell Lymphoma: A Literature Review Using a cutoff of roughly 98% for strongly positive cells, researchers in one multicenter study were able to distinguish the two with near-perfect sensitivity and specificity.20PubMed Central. Comparison of Ki-67 Labeling Index Patterns of Diffuse Large B-Cell Lymphomas and Burkitt Lymphomas Using Image Analysis: A Multicenter Study Getting the diagnosis right matters because these two lymphomas require different chemotherapy regimens.
The Reproducibility Problem and AI Solutions
The elephant in the room with Ki-67 is reproducibility. Two pathologists looking at the same slide can come up with noticeably different percentages, especially when the staining is uneven or the tumor has a mix of hot spots and quiet zones. This inter-observer variability is one reason Ki-67 has not been adopted into formal staging systems for most cancers the way it has for neuroendocrine tumors. A biomarker is only as good as the consistency with which it is measured.
Artificial intelligence tools are starting to close this gap. In a study where pathologists scored Ki-67 with and without AI assistance, the average scoring error dropped from about 6 percentage points without AI to about 2 percentage points with it.21Scientific Reports. AI improves accuracy, agreement and efficiency of pathologists for Ki67 assessments in breast cancer Another study found that when pathologists used standardized reference charts alongside an AI system, inter-observer agreement reached an intraclass correlation above 0.99, essentially eliminating meaningful disagreement.22PubMed Central. Artificial intelligence-assisted interpretation of Ki-67 expression and repeatability in breast cancer Digital image analysis on its own showed moderate correlation with manual pathologist scoring and, importantly, the digitally derived values still predicted outcomes like relapse-free survival.23PubMed. Comparison of Ki-67 labeling index measurements using digital image analysis and scoring by pathologists
These tools are not replacing pathologists. They are acting more like spell-checkers: flagging areas of the slide the human might have underweighted, providing a numerical count to anchor the subjective estimate, and smoothing out the variation between readers. As adoption grows, Ki-67 scores should become more comparable across laboratories, which in turn makes the cutoff values used in treatment guidelines more reliable.
What a High or Low Score Actually Means for You
If you have received a Ki-67 result on your pathology report, the first thing to understand is that the number means nothing in isolation. A Ki-67 of 30% in breast cancer, neuroendocrine tumors, and lymphoma implies very different things because the baseline proliferation rate and the relevant cutoffs vary across diseases. Even within breast cancer, the same Ki-67 value carries different implications depending on hormone receptor status, HER2 status, tumor size, and lymph node involvement. In triple-negative breast cancer, for instance, a cutoff of 30% showed prognostic value only in tumors that were small or had not spread to lymph nodes.24Scientific Reports. The prognostic and predictive potential of Ki-67 in triple-negative breast cancer
A few practical points are worth knowing:
- Higher is not always worse in treatment terms. A high Ki-67 may mean the tumor is more likely to respond to chemotherapy, which can be good news if chemotherapy is the planned treatment.
- The score can change over time. A biopsy taken before treatment and a surgical specimen taken after a few weeks of therapy can show dramatically different Ki-67 values. The direction of that change itself carries information about prognosis.
- Cutoffs are not universal. If one source says 14% is the threshold and another says 20%, neither is wrong. Guidelines evolve, and different institutions adopt different values based on their validation data.
- Ki-67 is one piece of a larger puzzle. Oncologists weigh it alongside tumor grade, stage, receptor status, and in many cases genomic test results. No single biomarker dictates treatment on its own.
Ki-67 in Pediatric Cancers
Ki-67 is not limited to adult cancers. In children, it serves a similar prognostic role across several solid tumor types. In neuroblastoma, a childhood cancer arising from nerve tissue, Ki-67 expression correlated negatively with prognosis, meaning children whose tumors had higher proliferation tended to have more advanced disease and worse outcomes.25Asian Pacific Journal of Cancer Prevention. Expression of Ki-67, p53 and VEGF in Pediatric Neuroblastoma In Wilms tumor, the most common kidney cancer in children, Ki-67 emerged as one of seven key predictors of recurrence in a machine-learning model, alongside tumor stage, age, and histological subtype.26PubMed Central. Prediction of pediatric Wilms tumor recurrence using interpretable machine learning models: insights from a 20-year real-world study and the prognostic value of Ki-67 Across a broader range of childhood solid tumors, Ki-67 has been studied as an adjunct for predicting biological behavior, though it has not been formally incorporated into pediatric staging systems the way it has for adult neuroendocrine tumors.27Indian Journal of Pathology and Oncology. Comparative Immunohistochemical Analysis of Ki -67 in a Spectrum of Pediatric Solid Tumours
Ki-67 in Veterinary Medicine
The same protein shows up in animal tumors and is measured in essentially the same way. In veterinary oncology, Ki-67 has been studied most extensively in canine mast cell tumors, one of the most common skin cancers in dogs. Researchers found that Ki-67 scored from simple needle-aspirate samples could accurately distinguish low-grade from high-grade tumors, with a cutoff value of about 6.5 providing a useful balance of sensitivity and specificity.28PubMed Central. Cutaneous Canine Mast Cell Tumor: The Use of Proliferative Markers (Ki-67 and Ki-67 × AgNOR) in Cytological Samples for Diagnosis and Prognosis This matters for pet owners and veterinarians deciding between watch-and-wait approaches and aggressive surgery or chemotherapy. The underlying biology is the same: cells that are dividing fast stain for Ki-67, and more staining generally means a more aggressive tumor. The principle translates across species even when the specific cutoff numbers do not.