Guide to Use the Atlas

Table of Contents

About the Atlas

The atlas is organized by organ system and is subdivided into chapters, each covering a single organ. Each chapter comprises a set of sections, each describing a single lesion. Each section is meant to be an independent, stand-alone document. The title of each section is the preferred NTP microscopic diagnostic term for that lesion. Each section includes histopathologic descriptions, high-quality images, diagnostic guideline recommendations, and references. Each chapter was reviewed by a recognized expert on the particular organ or organ system, and the final documents have been vetted and approved by senior NTP pathologists.

Back to top

Professional Judgment and Diagnostic Consistency

The study pathologists tasks are to consistently utilize morphologic diagnoses that accurately, clearly, and succinctly reflect the salient tissue changes, and to report and interpret the pathology results in a narrative. The pathology narrative report should not only address the toxicologic relevance of all potential treatment-related effects but should also include detailed histologic descriptions and grading criteria for treatment-related findings. Whatever the diagnostic criteria for a given study, it is of the utmost importance that the study pathologist maintain consistency in application of the criteria to each lesion in each animal. Unwavering diagnostic consistency is by far the most essential factor in light microscopic evaluation of any study.

This atlas lists the preferred terminology for lesions and guidelines for recording the various diagnoses. Some observations associated with predominate processes can be documented in the pathology narrative. For example, inflammatory lesions are often accompanied by cellular changes (e.g., degeneration or necrosis), which can usually be adequately documented in the narrative; however, if a cellular change is unusually or disproportionately severe (relative to the inflammation), then it may be recorded.

Thus, the study pathologist must determine the most appropriate diagnostic approach for each lesion in each study. The approach may vary depending on the study design, study duration, test species and strain, character and extent of the lesions, known class effects of the test article, and many other factors. The atlas diagnostic recommendations are intended to help with this decision-making process.

It is expected that the diagnostic terminology presented in the atlas will be adequate for most studies. But if warranted by the findings, the study pathologist may deviate from the atlas guidelines. In such cases, the study pathologist may request that the NTP add a new diagnostic term to the study lexicon. Unusual terminology or deviations from the atlas recommendations should also be explained in the narrative.

The atlas guidelines can help with maintaining consistency by providing defined criteria and representative images of typical lesions against which actual study findings can be compared. The study pathologist can also employ other methods to ensure consistency and avoid “drift” (cumulative diagnostic inconsistency over the course of a study), including keeping a running handwritten list (diagnostic dictionary) of all new diagnoses as they are created, reexamining representative lesions in high-dose animals before evaluating “cut-down” target tissues in lower-dose animals, and periodically reviewing summary incidence tables and taking necessary steps (including reevaluating target tissues) to address potential discrepancies.

Back to top

Structure of the Light Microscopic Diagnosis

Diagnoses should be constructed in a consistent hierarchical manner. The tissue (site) is listed first, usually followed by a single subsite modifier, followed by a single morphologic diagnosis, followed (if appropriate) by descriptive or distributional modifier(s). For example, suppurative inflammation of the renal pelvis should be recorded as kidney, pelvis – inflammation, suppurative, that is, kidney (site), pelvis (subsite) – inflammation (morphologic diagnosis), suppurative (descriptive modifier). A focus of follicular cell hyperplasia in the thyroid should be recorded as thyroid gland, follicular cells – hyperplasia, focal; that is, thyroid gland (site), follicular cells (subsite), hyperplasia (morphologic diagnosis), focal (distributional modifier). Diagnoses of most nonneoplastic lesions also include severity grades.

Back to top

Multiple Subsites

If multiple subsites in one tissue (site) exhibit a particular lesion, then each subsite should be given a separate diagnosis. For example, if there is acute inflammation in the cornea and sclera of the eye, these should be recorded as eye, cornea – inflammation, acute and eye, sclera – inflammation, acute, not as eye, cornea, sclera – inflammation, acute. The use of a general tissue site without a subsite (e.g., eye, inflammation, acute) is generally not preferred but may be appropriate in some cases.

Back to top

Distributional, Descriptive, and Quantitative Modifiers

Severity grading suffices to describe variations in the extent (degree of overall tissue involvement) of most nonneoplastic lesions. Therefore, distributional modifiers should be used only to distinguish lesions from others with different pathogenesis and toxicologic significance. For example, necrosis of periportal hepatocytes and necrosis of centrilobular hepatocytes are distinct lesions and thus should be diagnosed separately; that is, liver, hepatocyte, periportal – necrosis and liver, hepatocyte – centrilobular, necrosis, not liver, hepatocyte – necrosis.

For most nonneoplastic lesions, the distributional modifiers “focal” and “diffuse” are unnecessary and often obfuscating, since the degree of tissue involvement is adequately conveyed by severity grade. Thus, “focal” and “diffuse” should not be used in most nonneoplastic diagnoses.

The main exceptions are the diagnoses of “hyperplasia” and “hypertrophy, for which “focal” and “diffuse” lesions often have different pathogenesis and toxicologic significance; this is especially true for endocrine organs and gonads. There may also be cases where two distribution patterns (with different toxicologic relevance) may occur in the same tissue (e.g., focal fatty change and diffuse fatty change in the liver). In these situations, the use of “focal” and “diffuse” is appropriate to distinguish lesions with biologically significant differences.

The distributional modifier “multifocal” indicates only the extent of the tissue involvement, which is adequately communicated by severity grade. Thus, the use of “multifocal” is not recommended. The modifier “multiple” should be used only in neoplastic diagnoses.

Anatomic modifiers used in descriptions of clinical or gross necropsy observations (“anterior,” posterior,” “dorsal,” “ventral,” “medial,” “left,” and “right”) should not be used in the corresponding microscopic diagnoses.

Likewise, the use of the modifiers “unilateral” and “bilateral” in paired organs almost always should be avoided for microscopic diagnoses. The routine use of “unilateral” and “bilateral” essentially doubles the number of separate line-entry diagnoses in every paired organ, creating excessively complex summary pathology tables and potentially obfuscating potential treatment effects. Furthermore, for most nonneoplastic lesions in paired organs, the toxicologically salient point is not whether one or both of the pair exhibits the lesion but, rather, the overall extent of tissue involvement, which is virtually always best conveyed by a severity grade collectively assessed for the tissue as a single whole. The rare instances when lesion distribution in one versus both of a pair is toxicologically relevant can be addressed in the narrative.

Back to top

Severity Grades

Most nonneoplastic lesions should be assigned severity grades. The NTP uses an ascending four-level numerical scheme for gradable nonneoplastic lesions: 1 = minimal, 2 = mild, 3 = moderate, and 4 = marked. For a given study, the histologic criteria for each grade for each lesion can be established by the study pathologist, but the criteria for treatment-related findings should be described in the pathology narrative.

As noted in various atlas sections, some morphologic tissue changes (e.g., cysts) are generally not graded but are routinely recorded as “present.” However, such changes sometimes exhibit treatment-related, toxicologically relevant differences in extent or magnitude, and in these instances, they should be assigned severity grades rather than being listed as “present.”

In some studies, the test article is detectable by light microscopy in target tissues. Test articles can be recorded as “present” when the accumulations appear similar from animal to animal. However, when there are toxicologically relevant, dose-related differences in the degree of test article accumulation, these differences should be documented with severity grades. On the other hand, organisms (bacteria, fungi, protozoan and metazoan parasites) in tissues always should be graded as “present” no matter what their numbers.

Back to top

Inflammation versus Infiltration

Accumulation of leukocytes and other cells are common nonneoplastic lesions in many tissues. The term “inflammation” should be used when the cell (leukocyte) accumulations are part of an active inflammatory process (typified by concurrent features such as vascular changes, necrosis, fibrosis, and/or tissue disruption). In contrast, cell “infiltration” may be diagnosed when the cell (e.g., leukocyte) accumulations are present in tissue without other disruption or pathology.

Back to top

Macroscopic Terms and “Disease Entity” Diagnoses

Macroscopic or clinical diagnostic terms such as “abscess” or “hydronephrosis” generally should be avoided in microscopic diagnoses and used only if no alternative is more appropriate.

Microscopic diagnoses ending in “-itis,” “-osis,” and “-opathy” generally also should be avoided. A few exceptions are listed in the atlas. These are mainly complex disease processes that should be recorded using the “disease entity” diagnosis so that the overall process is specifically identified as the particular disease and is not confused with other potentially treatment-related lesions. Examples of such “disease entity” diagnoses include chronic nephropathy, polyarteritis, and eosinophilic crystalline pneumonia.

Back to top

Diagnostic Thresholds

Spontaneous “background” and/or age-related lesions, such as extramedullary hematopoiesis in the spleen or mononuclear cell infiltrates in the Harderian gland, are common in rats and mice. In many studies, such incidental lesions have similar incidences in control and treated groups and lack toxicologic relevance. In such instances, diagnostic “thresholds” can be established, in which only occurrences with severity above a predetermined threshold are diagnosed. However, in some studies, common background or age-related lesions can exhibit treatment-related, toxicologically significant differences in incidence and/or severity. In these situations, the lesions should be diagnosed and graded without a threshold.

Back to top

“Lumping” versus “Splitting”

“Lumping” and “splitting” refer, respectively, to consolidating or dividing morphologic features of a pathologic process into fewer or more individual line-entry diagnoses. There are legitimate differences of opinion among pathologists over the relative importance of “lumping” versus “splitting” as a general practice, as well as which specific lesions are appropriate candidates for either approach.

This is because excessive “splitting” will result in overly complicated summary incidence tables, in which a valid treatment effect may be so “diluted” across numerous line-entry diagnoses as to be imperceptible to the reader. For example, “splitting” the morphologic components of aging-related cardiomyopathy in rats (e.g., inflammation, fibrosis, cardiomyocyte necrosis) into separate line-entry diagnoses may create a false impression that several different pathologic processes are occurring in the heart when only a single specific “disease entity” is present. And of course, the greater the number of separate line-entry diagnoses, the more difficult it is for the pathologist to maintain consistency.

Conversely, excessive “lumping” of pathogenically disparate changes will result in less complex tables but may also mask potential treatment effects. For example, if nose, olfactory epithelium – degeneration has treatment-related increased incidences, but nose, respiratory epithelium – degeneration exhibits only sporadic incidences, “lumping” the olfactory and respiratory epithelial changes under the single line-entry diagnosis “nose, epithelium – degeneration” might obscure the actual treatment effects in the olfactory epithelium.

The optimum balance of “lumping” and “splitting” may also be dictated by the study design and/or duration. In an acute or subchronic study, relatively few aging-related incidental lesions would be expected in both control and treated animals, and the study goal is often to characterize subtle pathologic processes, especially those that might be pertinent in setting doses for longer-term studies. In such shorter-term studies, a greater degree of “splitting” may be most appropriate. In contrast, for many long-term studies (such as two-year bioassays), the primary objective is the detection of carcinogenic effects, and much higher incidences of incidental background lesions typically occur. Here, more “lumping” (especially of incidental aging-related lesions) may be more appropriate.

For these reasons, the study pathologist must employ professional judgment when “lumping” and “splitting” in a given study, so that the summary incidence tables clearly and accurately document all the study results without being overly simplified or complex.

Back to top

Nonneoplastic Lesions within and/or Associated with Neoplasms

Many neoplasms can exhibit internal features (e.g., hemorrhage, necrosis, fibrosis) or can secondarily lead to changes (e.g., inflammation, compression) in adjacent tissue(s). In general, such nonneoplastic features within or secondary to a neoplasm should not be diagnosed separately, as this could potentially obfuscate treatment effects. For example, separately diagnosing “inflammation” within incidental hepatocellular carcinomas could spuriously obscure or exaggerate actual treatment-related incidences of nonneoplastic “inflammation” in regions of the liver distant from the carcinomas.

In a few situations, there may be clinical or gross necropsy observations for which the only appropriate microscopic correlate would be a nonneoplastic change secondary to a neoplasm. In these cases, it may be necessary to add the microscopic diagnosis for a change secondary to a neoplasm.

Back to top

Autolysis and Insufficient Tissue

“Autolysis precludes evaluation” and “insufficient tissue” should be exclusionary and categoric terms used only when autolysis is so extensive or the tissue section so small as to preclude evaluation of any morphologic changes in a given tissue. That is, if even one morphologic diagnosis can be detected in a tissue, concurrent “autolysis precludes evaluation” or “insufficient tissue” should not be recorded. In both cases, the diagnosis should be entered as present with no severity grade. In the case of “insufficient tissue,” efforts should be made to acquire an adequate section of the tissue (e.g., prepare recuts of the original block).

Back to top

Organ Notes

The NTP data entry system permits Organ Notes, which can be used by the study pathologist to add free-text comments under individual tissues in individual animals. Organ Note comments are not a supplement or substitute for line-entry diagnoses or the pathology narrative. They should be utilized very sparingly if at all. Organ Notes should include only information that is not a study result or an interpretation and that therefore cannot be recorded in the pathology data tables or the pathology narrative. An example of an appropriate Organ Note is “Routine pancreas section on slide 6, but additional pancreas sections are also present on slide 7.”

Microscopic diagnoses should never be recorded in Organ Notes. Organ Notes should never be used to record differential diagnoses, histologic descriptions, gross-microscopic correlations, severity grade criteria, cause(s) of death or moribundity, results of special stains or other ancillary procedures, comments about tissue accountability and histotechnique quality, and/or literature references.

Examples of inappropriate Organ Note entries include: “hyperplasia, but could be borderline adenoma”; “inflammation included multiple foci of necrosis, edema, and hemorrhage”; “considered to be acute inflammation but differential could include suppurative or chronic active”; “Fontana-Masson stain negative for melanin”; “spleen is missing”; “extensive autolysis precludes microscopic evaluation”; “pituitary shattered and folded”; “nephropathy may be severe enough have caused early death, but cardiomyopathy may have also been involved”; “skeletal muscle hemorrhage correlates to gross dark red discoloration”; “minimal grade hyperplasia is defined as a layer of 8-10 epithelial cells”; and “renal tubule vacuolization graded based on classification of Doe et al., J Rodent Toxicol 12:3-4, 2012.

Back to top

Citing the Atlas

The NTP Nonneoplastic Lesion Atlas is available to the public worldwide. Anyone may use the atlas; however, we request that the atlas be referenced appropriately. Please use the following citation format so that the individual authors are recognized for their work:

Cora MC, Travlos GS. 2014. Bone Marrow—Fibrosis. In: Cesta MF, Herbert RA, Brix A, Malarkey DE, Sills RC (Eds.), National Toxicology Program Nonneoplastic Lesion Atlas. Available: [accessed 1 February 2014].

The authors’ names can be found under the Authors and Reviewers tabs on each of the lesion pages.

Back to top