Nomenclature and Diagnostic Resources in Anatomic Toxicologic Pathology


Acknowledgments

The authors acknowledge previous authors of this chapter, Drs. Amy Brix and Ron Herbert. This work was supported in part by the NIH, National Institute of Environmental Health Sciences.

Introduction

In general nonclinical toxicology studies conducted for hazard evaluation and/or safety assessment, toxicologic pathologists document nonneoplastic and neoplastic morphological changes from laboratory animals to identify potential treatment-related effects of a test article (see ; Pathology in Nonclinical Safety , Vol 2, Chap 4). The primary means of communicating such morphological changes is through a system of nomenclature, i.e., the terminology of diagnoses. Such terminology of diagnoses is often amenable to statistical analysis (especially for neoplasms), which is a necessary requirement in carcinogenicity assessments (see Experimental Design and Statistical Analysis for Toxicologic Pathologists , Vol 1, Chap 16 and Carcinogenicity Assessment , Vol 2, Chap 5). Since histopathological diagnoses are primarily based on subjective observations, terminology used to classify morphological changes must be standardized so that valid analysis and interpretation of data derived from nonclinical toxicology studies can be made.

Many factors can affect the application of specific diagnostic terms to morphological changes by pathologists, and such factors can contribute to the inconsistencies observed within and between toxicological studies. This chapter reviews the factors and challenges that commonly affect terminology usage in toxicologic pathology and proposes recommended practices to help minimize common challenges. Efforts undertaken to harmonize nomenclature worldwide include the Standards for Exchange of Nonclinical Data (SEND), International Harmonization of Nomenclature and Diagnostic Criteria for Lesions in Rats and Mice (INHAND), National Toxicology Program (NTP) Nonneoplastic Lesion Atlas (NNLA), and global open Registry Nomenclature Information System (goRENI).

The Need for Standardized Nomenclature

In toxicologic pathology, the pathologist wears two hats: the diagnostic pathologist (morphological diagnosis) and the scientist (generating data). Due to inherent complexities of evaluating histopathological changes, the training and experience of the toxicologic pathologist is of primary importance, because the diagnoses provided by the pathologist are the substrate for statistical analyses. Regulatory authorities often require statistical inferences of the relationship of adverse morphological changes, toxicity, or carcinogenicity with a test article. These adverse morphological changes are often presented as a morphological diagnosis, e.g., hepatocellular carcinoma. Such morphological diagnoses (variables) must be independent to comply with the assumptions required for valid statistical analysis and inferences drawn with respect to the morphological changes and the treatment.

Standardized nomenclature is essential not only to communicate results to other colleagues but also to ensure the independence of the named variable. Standardized nomenclature should aim to provide categories that are mutually exclusive (i.e., only those morphological diagnoses placed in a specific diagnosis category, e.g., hepatocellular carcinoma) as well as mutually exhaustive, such that all cases of the category in question are included (e.g., all cases of hepatocellular carcinoma are within the named category) (see Experimental Design and Statistical Analysis for Toxicologic Pathologists , Vol 1, Chap 16 ).

When separate categories are generated from overlapping diagnostic criteria, the toxicologic pathologist should determine which categories are relevant to the study at hand before subjecting diagnoses to statistical testing. Conducting statistical tests across multiple correlated categories can lead to increased false positive rates. Subset relations between different diagnoses also challenge the independence assumption of statistical testing and further complicate the interpretation of the results. The toxicologic pathologist must be well trained and experienced in order to select the categories best suited to the scientific goals of the investigation.

Reproducibility is a growing concern in scientific research and is indispensable to regulatory practices that must compare results across studies. Nomenclature may be updated to accommodate the expansion of knowledge. However, such revisions should only be done when necessary, since even small modifications in nomenclature may complicate data management as well as the interpretation and reporting of results. Moreover, the diverse interests and focus areas of various laboratories often produce inconsistent diagnostic criteria and terminology. Basing diagnoses on published and accepted nomenclature standards will mitigate observer bias and promote consistency and objectivity in toxicology studies. The previous lack of a universally accepted and standardized system of nomenclature within toxicologic pathology has, in the past, resulted in inconsistent terminology, controversy, additional costs, and/or delays in the product review and approval by regulatory authorities. Under such circumstances, regulators are unable to evaluate the dose–response effects, and the statistical analyses from such data may violate assumptions of the test. Moreover, inaccuracies or inconsistencies in the diagnostic terminology used for morphological changes impact safety assessment and/or hazard identification, and ultimately risk assessment (see Risk Assessment , Vol 2, Chap 16). Diagnostic terminology should therefore be precise, reliable, consistent, and convey a clear picture of the important morphological changes to facilitate unambiguous interpretation of the pathological effects of chemicals, drugs, biologics, and/or medical devices, as well as to aid in understanding underlying mechanisms of toxicity.

Components in Nomenclature

In toxicologic pathology, the morphologic diagnosis is the foundation of data collection (see Pathology and GLPs, Quality Control and Quality Assurance , Vol 1, Chap 27 ). The approach to making a morphological diagnosis is hierarchical, resulting in flexibility and variation in the range of diagnostic terminology used. Essentially, a morphologic diagnosis is made by sequentially designating the topography first (organ/tissue affected), with or without a subsite modifier, followed by a single morphological diagnosis (predominant pathological change or type of lesion), followed, where applicable, by descriptive qualifier(s) that include duration, distribution, and/or severity. Topographic designations may include as many as three hierarchical terms: (1) organ, (2) site within an organ, and, if necessary, (3) a subset of increasing specificity within sites. An example would be “brain, cerebral cortex, cingulate.” Morphological diagnoses are used to describe the major pathological processes or abnormalities occurring in an organ or tissue (neoplastic or nonneoplastic). Although both nonneoplastic and neoplastic lesions may have site modifiers, most types of qualifiers (e.g., duration, distribution, and severity) are predominantly used for nonneoplastic lesions. Further, each diagnosis typically allows the inclusion of a comment to capture the intricacies of a specific lesion.

Terminology for Nonneoplastic Lesions

For nonneoplastic lesions, morphologic changes often allow the use of qualifiers that specify important characteristics of the morphological disease process (e.g., distribution, severity, and duration) to effectively convey additional information about morphologic changes. For example, “liver, infiltrate, mononuclear cells, focal, moderate” allow some degree of perspective on the distribution and severity of the morphological change noted. It should be noted that each category of qualifiers typically has a large pool of potential terms lending several possible permutations and combinations available for the capture of the final diagnosis. For example, the qualifier for distribution may include such terms as focal, multifocal, locally extensive, or diffuse.

Organ-specific site qualifiers should be used to delineate lesions based on different regional anatomical differences, tissue-specific responses, or biological significance. The choice of site qualifiers can vary with the anatomic complexity of an organ or tissue and the prevalence of treatment-related effects in an organ. For example, at different “levels” along its length, the nasal cavity is lined by squamous, transitional, respiratory, or olfactory surface epithelia. The response(s) of these epithelia to inhaled toxicants may differ by location, and the use of a site qualifier helps characterize site-specific effects. Likewise, chemically induced lesions in the larynx often have site specificity at the base of the epiglottis; thus, the designation “larynx, epiglottis” is appropriate. Other possible examples of site qualifiers include “liver, bile duct”; “nerve, sciatic”; and “kidney, renal tubule.” Alternatively, in some situations, it may be best to refrain from using site qualifiers. As an example, for inflammation in hollow organs or tissues that have surface epithelia (e.g., the nose, stomach, gastrointestinal tract, or urinary bladder), the use of qualifiers that indicate specific sites (e.g., lumen, epithelium, mucosa, submucosa) may not be appropriate, because the presence of small numbers of inflammatory cells around organs associated with points of entry and/or external environment is generally considered to be within normal limits and has no biological relevance. Therefore, unless a site-specific toxicity is suspected, in such instances, characterization of a background lesion is unnecessary, and designating the overall severity of inflammation may be best used to convey differences in lesion distribution.

Distribution qualifiers are used to indicate the pattern of an effect (e.g., diffuse, locally extensive, multifocal, or focal). However, focal and multifocal qualifiers may also be addressed by the severity grade. Distribution modifiers can provide additional site specificity and indicate an inherent feature of a lesion that has a particular biological significance or reflects a specific pathogenesis. For example, centrilobular hepatic necrosis rarely has the same pathogenesis as multifocal, randomly distributed hepatic necrosis; therefore, it is important to include these types of distribution modifiers. Likewise, it may be toxicologically relevant to the interpretation of a study to know whether hyperplasia of the forestomach is focal versus diffuse. However, the use of distribution qualifiers is usually not warranted for common background, spontaneous, and/or age-related lesions specific for certain species and strains, unless an exacerbation of such lesions is suspected. Because the objective of the pathologist is to identify test article–related effects, such common background, spontaneous, and/or age-related findings may typically be cited via references in the pathology report. Severity grading of common background lesions is particularly relevant when such lesions are dose response related, where grading may be the most reliable means of effectively documenting differences between test groups and controls and among test groups, i.e., dose response. For example, certain classes of chemicals can potentiate chronic progressive nephropathy (CPN) as well as alpha 2u -globulin nephropathy in male rats. In such cases, assigning severity grades can provide important information.

Although distribution qualifiers depict the pattern of a lesion with additional specificity, the intensity or magnitude of the lesion is reflected by the use of severity grades. Severity grading is the designation of a semiquantitative score to a lesion or process (e.g., often a four- or five-point scale, depending on the software program used for data entry). It is used to reflect a combination of the amount and distribution of tissue involvement, as well as the actual degree of tissue damage. Typically, these include minimal (1), mild (2), moderate (3), and marked (4), and if a five-point scale is used, severe (5). Regardless of whether a four- or five-point scale is used, general criteria used to define each grade should be included in the pathology report. Severity grading of nonneoplastic lesions can provide useful dose–response information in that a test article may alter incidence of a lesion or its severity, or both. Severity grades are continuous rather than discrete data representing degrees of changes relative to similar lesions in controls and other animals in a study, and thresholds for grading often vary between individual pathologists and or study types.

Severity grading (ordinal data) is reserved for nonneoplastic lesions, whereas neoplasms are considered either present or not present (nominal [diagnosis] or quantal [malignant vs. benign] data). Severity grading adds little valuable information in some nonneoplastic lesions. Examples include calculus in the urinary tract or cysts of the pituitary pars distalis . Severity grades are most useful when making comparisons of qualitative changes, for example, when average severity increases with increasing dose, or when comparing the relative toxicity of structurally related compounds. The generation of data allows statistical analysis and, in some instances, may be the only indication of treatment-related differences among test groups.

Duration qualifiers are more commonly used in association with particular types of inflammatory changes. The commonly used notation of “-itis” in diagnostic pathology is discouraged in toxicologic pathology and replaced with more descriptive terminology. Commonly used duration qualifiers include acute (neutrophils), subacute (lymphocytes), chronic (macrophages or mononuclear cells with or without fibrosis), and chronic active (a mixture of macrophages or mononuclear cells and neutrophils). In histopathological evaluation of toxicological studies, it is recommended to describe the cell type within the morphological description in the pathology narrative rather than duration qualifiers.

Terminology for Neoplastic Lesions

A few rules exist for the terminology applied to neoplasms. By convention, it is assumed that neoplasms not designated as metastatic are primary or arising at that site. For neoplasms in tissues distant from the primary site, the qualifier “metastatic” is added to the diagnosis, and the site of the primary lesion can be indicated parenthetically (e.g., “lung—hepatocellular carcinoma, metastatic (liver)”). For some neoplasms, the diagnosis should indicate whether the lesion is benign or malignant (e.g., “adrenal gland, medulla—pheochromocytoma, benign” and “adrenal gland, medulla—pheochromocytoma, malignant”). The occurrence of multiple and bilateral neoplasms should be indicated, especially when this occurrence appears to be a treatment-related phenomenon. Examples include “liver—hepatocellular adenoma, multiple” and “kidney, renal tubule—adenoma, bilateral.”

Diagnoses for neoplasms should not contain site qualifiers unless such qualifiers serve to distinguish the histogenesis of one neoplasm from another. For example, the site qualifiers “C-cell” or “follicular cell” should be used to distinguish carcinomas of the thyroid gland.

Finally, inflammation and other types of secondary changes that may result from tissue damage associated with the presence of neoplasms (e.g., a histiocytic infiltrate in the alveoli associated with an alveolar bronchiolar carcinoma) are usually not meaningful to the study and therefore should not be recorded as separate diagnoses, but, if deemed important, they can be described in the narrative section of the pathology report (see Preparation of the Anatomic Pathology Report for a Toxic ity Study , Vol 2, Chap 13).

Challenges in Standardizing Nomenclature

Histopathological evaluation of the tissues is one of the most resource-intensive phases of toxicity and carcinogenicity studies. Given that histopathology is a descriptive and interpretative science, the inherent complexity underlying a biological response compounded with the spectrum of subjectivity in determining selective diagnostic criteria between pathologists often underscores the lack of consistency in diagnoses within and between studies. No two pathologists independently evaluating the same study could be expected to arrive at identical findings for every tissue, organ, and animal. Factors influencing standardization of histopathological diagnoses include the training and experience level of the pathologist to select diagnostic criteria within the spectrum of available terminology, thresholds, diagnostic drift, severity grading, lesion complexity, and involvement of multiple pathologists for a single test article. Moreover, in some pathology data collection systems, variations in the use of topographical, morphological, and/or qualifier designations may result in separate and distinct diagnosis categories when tabulated, thereby generating more than one morphologic diagnosis for the same lesion. For more information, see Practices to Optimize Generation, Interpretation, and Reporting of Pathology Data from Toxicity Studies , Vol 1, Chap 28 .

Training

Most anatomic toxicologic pathologists are trained in a veterinary diagnostic pathology setting, usually at a college of veterinary medicine or, much less frequently, a medical college. The primary objective in diagnostic pathology is to identify the cause of disease or death for an individual animal and communicate the findings in a detailed fashion (i.e., each disease process is identified and recorded as a separate entity) to the veterinarian or clinician for that single case. In contrast to toxicologic pathology, consistency in morphologic diagnosis terminology across cases is not a major concern.

The emphasis in toxicologic pathology is on identifying changes in relation to a treatment group rather than in an individual animal—dose and response. Specifically, the goal of the toxicologic study pathologist is to determine if there are differences in the incidences and/or severities of lesions in treated groups compared to the control animals (positive and negative controls, where applicable) in order to provide dose–response data, and occasionally, where applicable, to delineate effects of test article from formulation. Study pathologists must identify the various lesions in tissues, but generally have less freedom than a veterinary or medical diagnostic pathologist in the type and number of diagnoses that they may record. For the study pathologist, it is especially important to use consistency and brevity in recording diagnoses in order to facilitate the statistical evaluation (e.g., carcinogenicity studies) and interpretation of the findings.

Diagnostic terminology is influenced by the training and experience in laboratory animal pathology that the toxicologic pathologist may have received. A histopathological diagnosis is primarily a qualitative judgment of the nature of a specific lesion and its apparent or expected biological behavior. Each diagnosis is a subjective observation, the accuracy of which depends on the pathologist's training and experience, the state of knowledge of the specific disease process, and the generally accepted diagnostic criteria and nomenclature within the profession.

The training, qualifications, and experience of the pathologist can influence how lesions are interpreted, and hence, the selection of diagnostic terminology. The majority of veterinary anatomic pathology training occurs in domestic and companion animal settings, with occasional exposure to the common background lesions of laboratory animals; trainees are less often exposed to toxicologic pathology. This may lead to selection of inappropriate terminology by pathologists with less experience evaluating rodent studies. Common issues encountered in evaluations by novice pathologists include the use of multiple morphological descriptive terms or synonyms for a lesion or disease process; inconsistencies in the designation of topography, sites, and subsites; and duplication of diagnoses, any of which may result in misinterpretation of toxicologic data.

Among experienced pathologists, nomenclature is still an issue. It is influenced by their training, as well as by individual and philosophical biases. One example includes terminology to describe inflammation composed primarily of mononuclear cells and fewer neutrophils. Some pathologists prefer the term “chronic active,” while others prefer the term “chronic.” Still others may use the term “cellular infiltrate” rather than “inflammation” with or without cell type modifiers.

Regardless of the chosen terminology, it is important for the study pathologist to identify such diagnostic differences and either resolve the difference, explain the reason for the use of alternate terminology in the pathology narrative, or state that the two different diagnoses reflect the same pathological process. Pathologists with experience evaluating rodents trained in a “basic” experimental research setting may be accustomed to nomenclature used for genetically engineered rodents, which may differ from that used in toxicologic pathology ( ).

Some customized Good Laboratory Practice–compliant computerized systems allow study pathologists to build a selective diagnostic vocabulary. For experienced and novice pathologists alike, the availability of a wide range of possible diagnostic terms can be a source of inconsistency due to overlapping criteria in defining subjective lesions, thus violating the assumptions for the statistical tests used. Additionally, multiple studies conducted with different pathology data capture systems for the same test article often generate nomenclature issues based on the differences in lexicon available within each data capture system. Often, this is mitigated by inclusion of a second-tier evaluation by a peer review pathologist. The peer review process allows the peer review pathologist to evaluate and compare diagnoses with the study pathologist, which results in better standardized nomenclature for the study (see Pathology Peer Review , Vol 1, Chap 26 ). Occasionally, unresolved diagnoses between the study and peer review pathologist are further examined by a Pathology Working Group (PWG) to reach a consensus. Such a process allows for consistency in diagnoses between pathologists and between studies for the same test article.

Thresholds

Training and/or experience of the pathologist can play a role in the applications of “thresholds” by individual pathologists from a diagnostic perspective. The use of “thresholds” should not be confused with use of this term by toxicologists as relevant in carcinogenicity risk assessment. Histopathology evaluation of subtle lesions poses problems when one pathologist considers a tissue to show variation within normal limits of the species/strain, while another pathologist considers the variation to be above the threshold and, therefore, a lesion. Such differences impede the dose–response assessment.

In general, there exists an inherent spontaneous biological variation in a given tissue that is expected for the species/strain, which may change with age. Pathologists use knowledge of this variation in species/strain, as well as incidence in concurrent control group animals (when appropriate), to ascertain diagnostic criteria to determine thresholds for selecting and differentiating diagnoses. Exacerbation may occur as increased incidences or severities or both. Some pathologists may choose to diagnose all changes considered to be background at the lowest severity grade, and any treatment-related changes with an increased severity grade (where applicable). Alternatively, a diagnosis may be made only when the level of variation is determined to be above a diagnostic “threshold” of the background finding when it exceeds that considered to be within normal limits. For example, peripheral nerve fiber degeneration is a common age-related finding in older rats (especially males). Because an actual treatment-related effect is not diagnostically different from background findings, an increase in severity across dose groups may be the only parameter that captures potential exacerbation at the nomenclature level. Therefore, such triggers during histopathological evaluation should spur reevaluation of the diagnostic threshold initially set for this finding to ensure that exacerbation of the background lesion due to treatment is indeed considered. Further, the pathologist must address this specific finding in the narrative within the context of what is known about background nerve fiber degeneration in rodents, potential differences between sexes, age (a one-month repeat-dose study vs. a carcinogenicity study), and other influencing factors.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here