Canine Mast Cell Tumor Grading: Understanding the Two Systems and What Your Report Is Actually Telling You
Pathology Deep Dive · Canine · Oncology, Dermatopathology
Mast cell tumor grading is one of the most clinically consequential things a pathology report communicates — and one of the most frequently misunderstood. Two grading systems are in active use in veterinary pathology, they use different criteria, and they produce different grade designations that do not map directly onto each other. A Patnaik grade II designation — the most common grade assigned under that system — does not always accurately reflect the degree of malignancy, which is precisely why a second system was developed. Understanding which system your pathologist used, what the criteria are, and how to apply the result to clinical decision-making is essential for interpreting canine MCT reports accurately.
Why two grading systems exist
The Patnaik grading system was published in 1984 and became the standard for canine MCT grading for over two decades. It assigns tumors to one of three grades — I, II, or III — based on a combination of histologic features including cellular differentiation, mitotic index, nuclear morphology, and depth of invasion. For much of its history it was the only validated system in use, and it generated a large body of clinical outcome data linking grade to prognosis and treatment response.
The problem that became increasingly apparent over time was the Patnaik grade II designation. In most published series, grade II tumors account for approximately 40 to 50% of all canine MCTs — a heterogeneous group that included tumors with widely varying clinical behavior. Some Patnaik grade II tumors behaved like grade I (indolent, curable with surgery), while others behaved like grade III (aggressive, high metastatic risk). The grade II designation was not reliably predicting outcome in a substantial proportion of cases, and clinicians were making treatment decisions on an intermediate category that did not accurately reflect degree of malignancy.
The Kiupel grading system, published in 2011, was developed specifically to address this problem. Rather than a three-tier system, Kiupel uses a two-tier system — low grade and high grade — based on two primary criteria: mitotic figures and multinucleated cells, each counted per 10 high-power fields, along with bizarre nuclear morphology. The simplification was intentional: reduce the intermediate category, improve interobserver agreement, and produce a binary result that more cleanly stratifies clinical risk.
The criteria: what each system measures
Understanding how each system assigns a grade helps explain why the same tumor can receive different designations depending on which system is applied.
| Patnaik Grade I | Patnaik Grade II | Patnaik Grade III | |
|---|---|---|---|
| Differentiation | Well differentiated | Intermediate | Poorly differentiated |
| Mitotic figures | 0–1 per 10 HPF | 1–4 per 10 HPF | ≥4 per 10 HPF |
| Nuclear morphology | Round, uniform | Mild to moderate atypia | Marked pleomorphism |
| Depth of invasion | Superficial dermis | Deep dermis | Subcutis and beyond |
| Granulation | Prominent | Moderate | Sparse or absent |
| Kiupel Low Grade | Kiupel High Grade | |
|---|---|---|
| Primary criteria | Fewer than 7 mitoses per 10 HPF AND fewer than 3 multinucleated cells per 10 HPF AND no bizarre nuclei | 7 or more mitoses per 10 HPF OR 3 or more multinucleated cells per 10 HPF OR bizarre nuclear morphology |
| Prognosis | Median survival not reached in most studies; excellent with surgery | Median survival 4–6 months without additional treatment |
How the two systems relate — and where they diverge
In straightforward cases, the two systems agree. Patnaik grade I tumors are almost universally Kiupel low grade. Patnaik grade III tumors are almost universally Kiupel high grade. The divergence occurs in the middle — Patnaik grade II tumors split between Kiupel low grade and Kiupel high grade, and that split is where the clinical value of the Kiupel system lies.
Published studies comparing the two systems consistently show that Kiupel high-grade tumors drawn from the Patnaik grade II pool behave significantly worse than Kiupel low-grade tumors from the same pool. In other words, the two-tier Kiupel system successfully stratifies a group that the three-tier Patnaik system leaves ambiguous. A Patnaik grade II / Kiupel high-grade tumor carries a prognosis much closer to Patnaik grade III than to Patnaik grade I.
This is why it matters which system your report uses. A Patnaik grade II designation tells you the tumor falls in the intermediate category — but it does not tell you which end of that category the tumor is closer to, and it does not accurately reflect degree of malignancy in all cases. A high-grade designation on a Kiupel report tells you the tumor carries significant risk regardless of what its Patnaik grade would have been.
Which system should you expect on your report?
Both systems remain in active use, and practice varies between pathologists and institutions. Some reports provide only one grade; some provide both. The Kiupel system has gained significant adoption since its publication and is now recommended by many veterinary oncology guidelines as the preferred system given its stronger prognostic stratification and improved interobserver agreement.
Regardless of which system is used, the grade should always be accompanied by the specific criteria on which it was based — at minimum the mitotic count, and ideally the other features used in the assessment. A grade designation without supporting data is less useful than one that tells you the mitotic index was 8 per 10 HPF, because that number has direct meaning beyond whatever category it places the tumor in.
If a report provides only a Patnaik grade and the tumor is grade II, it is entirely reasonable to contact the pathologist and ask whether Kiupel criteria were assessed and what the result would be. That conversation may materially change the prognosis and treatment conversation with the owner.
Additional prognostic factors beyond grade
Grade is the most commonly cited prognostic factor in canine MCT, but it is not the only one, and it should not be read in isolation.
Mitotic index is the single most powerful independent prognostic variable in most studies — more powerful than grade designation itself in some analyses. A mitotic count of 5 or more per 10 HPF is consistently associated with significantly shorter survival times regardless of which grading system is applied. The mitotic count in the report is worth noting and carrying into the clinical conversation.
c-Kit mutation status (PCR for exon 11 and exon 8 mutations) identifies a subset of MCTs with dysregulated receptor tyrosine kinase signaling that are candidates for tyrosine kinase inhibitor therapy (toceranib, masitinib). This is a molecular test performed on tissue, not a histologic finding, but it is directly relevant to treatment planning in higher-grade tumors and should be considered as part of the complete workup.
c-Kit IHC staining pattern — membranous versus cytoplasmic versus diffuse — has been associated with mutation status and clinical behavior and is sometimes included in reports as a surrogate marker, though PCR remains the definitive method for mutation detection.
Location matters independently of grade. MCTs on the muzzle, mucocutaneous junctions, nail bed, prepuce, and perineum have historically been associated with more aggressive behavior regardless of histologic grade. Subcutaneous MCTs, conversely, tend to behave more favorably than their cutaneous counterparts even when grade designations are similar.
Surgical margin status is the most actionable histopathologic finding after grade. Complete excision with clear margins is the single strongest predictor of local recurrence risk and should be evaluated and reported for every excised MCT.
AgNOR staining and the MSU prognostic panel deserve specific mention because they are familiar to many clinicians from reference laboratory reports and represent an established approach to prognostic refinement beyond histologic grade alone. AgNOR — argyrophilic nucleolar organizer region — staining is a silver-based technique that identifies nucleolar organizer regions, which reflect ribosomal RNA gene activity and correlate with cellular proliferative capacity. The AgNOR count per nucleus provides a measure of proliferative activity that is independent of, but complementary to, the mitotic index.
The Michigan State University (MSU) MCT prognostic panel, developed from work by Kiupel and colleagues, combines AgNOR count with Ki-67 immunohistochemistry to produce a composite proliferation score. The AgNOR x Ki-67 index — calculated by multiplying the mean AgNOR count by the percentage of Ki-67 positive cells — has been validated as a prognostic marker in canine MCT and outperforms either marker alone in stratifying clinical outcome. Published threshold values separate low-risk from high-risk tumors, and the combined index has been shown to predict disease-free interval and survival time independently of Patnaik grade.
In practical terms, clinicians familiar with MSU panel results will recognize reports that include AgNOR count, Ki-67 percentage, and the composite index alongside histologic grade. These findings add prognostic resolution — particularly within the Patnaik grade II and Kiupel low-grade populations where biologic behavior is most variable. A Kiupel low-grade tumor with a high AgNOR x Ki-67 index warrants closer monitoring than one with a low index, even though both fall within the same grade category. This is precisely the kind of additional stratification that the AgNOR x Ki-67 panel was designed to provide — and it is one of the clearest current examples of quantitative ancillary testing meaningfully extending the prognostic value of histologic grading.
A note on feline MCT
Feline mast cell tumors are biologically and histologically distinct from their canine counterparts and are not graded using the Patnaik or Kiupel systems. The most common form in cats — cutaneous MCT — is typically the mastocytic type, which tends to behave more favorably than canine MCTs of equivalent appearance. A separate two-tier grading system for feline cutaneous MCT has been proposed and studied (Sabattini et al.), but it has not achieved the same level of clinical adoption as the Kiupel system has in dogs. Feline visceral MCT — splenic, intestinal — carries a different prognosis and is not graded histologically in the same way. When a feline MCT report does not include a grade, this reflects the less standardized state of feline MCT grading rather than an omission.
The future: the limitations of current grading and the hope for AI
It is worth being honest about what even the best current grading systems can and cannot do. Both Patnaik and Kiupel represent meaningful improvements over ungraded diagnosis, and the Kiupel system's ability to stratify the Patnaik grade II pool is a genuine clinical advance. But grading is still an imperfect predictor of biologic behavior. Grade I MCTs can and do metastasize. Kiupel low-grade tumors occasionally behave aggressively. The histologic features we measure today — mitotic index, nuclear morphology, multinucleated cells — capture important signals but do not capture everything that determines how a given tumor will behave in a given patient. This is not a failure of the pathologist or the system. It reflects the limits of what light microscopy of fixed tissue can reveal about a tumor's biology.
The hope for AI is not simply that it will count mitoses more consistently — it is that it will identify morphologic features that human eyes cannot reliably quantify, patterns within the tissue that correlate with outcome in ways that existing criteria do not, and ultimately provide a more accurate and more complete picture of tumor behavior than current grading systems alone can offer. Identifying mitotic figures on H&E requires distinguishing true mitotic figures from apoptotic cells, pyknotic nuclei, and other mimics, and the count is sensitive to section thickness, staining quality, and the specific area of the tumor chosen for evaluation.
AI-assisted mitotic detection is one of the most mature applications of computational pathology in veterinary medicine, and canine MCT has been one of the primary development targets. Published algorithms have demonstrated mitotic detection accuracy approaching interobserver agreement between pathologists, and in some studies exceeding the consistency of manual counting. The MIDOG (Mitosis Domain Generalization) challenge datasets, which include canine MCT cases, have been used to develop and validate models that generalize across different tissue preparation and scanning conditions.
The practical implication is significant: automated, reproducible mitotic counts applied consistently across an entire tumor section — not a selected hotspot — may ultimately provide more accurate and more standardized prognostic data than manual counting in a diagnostic workup. AI grading of this kind does not replace the pathologist's interpretive role; it provides a more reliable quantitative foundation for the grade designation the pathologist assigns.
Full AI-assisted MCT grading — incorporating mitotic count, nuclear morphology, and other Kiupel criteria — is an active area of development that has not yet reached commercial deployment in veterinary diagnostics. But the longer-term potential goes beyond reproducible counting. Machine learning models trained on large MCT cohorts with known outcomes have the capacity to identify morphologic features and spatial patterns within the tumor microenvironment that pathologists cannot reliably quantify manually — and that may predict biologic behavior more accurately than current grading criteria. A grade I MCT that metastasizes, or a Kiupel low-grade tumor that recurs aggressively, may harbor histologic signals that current criteria miss but that an AI system trained on outcomes could learn to recognize. That kind of discovery — identifying what we are not yet measuring — is where the most meaningful advance in MCT prognostication is likely to come from.
Reading the report: a practical summary
When a canine MCT report arrives, the following questions are worth applying systematically. Which grading system was used — Patnaik or Kiupel? If Patnaik grade II, has Kiupel been applied or can it be requested? What is the mitotic count, and is it above or below 7 per 10 HPF? Are margins complete, and how close is the closest margin? Has c-Kit mutation status been assessed or recommended? And does the anatomic location place this tumor in a higher-risk category regardless of grade?
A grade designation is the summary. The data behind it — mitotic count, nuclear features, margin status, location — is where the clinical decision-making lives.
Eric Snook, DVM, PhD, DACVP — Vetopathy. MCT grading reports include mitotic count, Kiupel and Patnaik criteria, margin assessment, and c-Kit IHC pattern as standard. Questions about a specific report? Direct pathologist contact is available.