Page 88 - Withrow and MacEwen's Small Animal Clinical Oncology, 6th Edition

CHAPTER 3 The Pathology of Neoplasia 67

of tissue area evaluated. 41,42 The clinician should remember that subpopulations and genetic instability that characterize many
the designation “clean” or “negative” node is based only on the advanced tumors. If heterogeneity is present, the most malig-
46
nant-appearing areas (usually defined by cellularity, degree of dif-
tissue section (4–5 μm thick) or sections evaluated; on a routine
VetBooks.ir basis, this is often less than 1% of the total tissue area. One study ferentiation, and mitotic activity) should be assessed for grading
purposes. Grading of specific tumor subpopulations, such as those
of 20 retrospectively evaluated LNs found that 25% of previously
“negative” LNs could be reclassified after more comprehensive at the “invasive front,” might also help to better prognosticate
47
sectioning and staining. Gross examination of the tissue by a behavior. Human biases have the potential to amplify natural
41
pathologist or trained histotechnologist may assist with trimming biologic variation in tumors. In small biopsy samples, the sam-
in more suspicious areas based on color and/or texture, but an pling procedure can affect the representation of different compo-
understanding of the total tissue area evaluated microscopically nents in a heterogeneous tumor. Some studies have demonstrated
relative to the total area of the tissue submitted remains important. that pretreatment incisional biopsies are a reasonably accurate esti-
mation of grade in the excisional sample; however, results might
vary between tumor types. 48,49 If necrosis is considered directly in
Histologic Grading and Clinical Staging of the grade, clinicians and/or pathologists might avoid these areas
Neoplasia for diagnostic sampling. Therefore incisional biopsies could have a
tendency to underestimate tumor grade. 48,49
In certain tumors, histologic grading, which is based on micro- Guidelines for evaluating the prognostic utility of grading
scopic evaluation of defined histopathologic criteria (many of schemes have been described. 43,50 An awareness of study design,
which have been presented previously) is predictive of biologic assessment of clinical outcomes and endpoints, statistical analy-
behavior. 2,5,14 Grading can be a powerful estimate of expected sis, methods of marker evaluation, and specific study conclusions
clinical progression or clinical outcome, such as metastatic poten- should all influence decisions to assign clinical relevance to spe-
tial, disease-free interval, or overall survival. For some tumors, cific histopathologic features and/or grade. Table 3.4 lists tumors
such as human and canine soft tissue sarcoma (STS), the histo- for which strong data suggest that grade and/or specific histologic
logic grade is more important than the tumor type. 2,22,32 Grade features are prognostic in dogs. Other canine tumors in which
frequently affects clinical decisions, such as recommendations for histomorphology has been linked to the prognosis include appen-
adjuvant therapy. Although this grade can be a valuable compo- dicular OSA 51,52 ; multilobular osteochondrosarcoma (MLO) 53,54 ;
56
57
55
nent of the clinical picture, no grading scheme accurately predicts mandibular OSA ; chondrosarcoma ; hemangiosarcoma (HSA) ;
behavior in 100% of patients. This introduces the critical dif- urothelial carcinoma ; SCC of the tongue ; splenic sarcoma
58
43
10
60
59
ference between histologic grading and clinical staging; grading (nonangiomatous/nonlymphomatous) ; primary lung tumors ;
62
63
is based almost exclusively on histomorphologic characteristics ocular melanoma ; splenic liposarcoma ; and lingual HSA. In
61
of the primary tumor and does not typically incorporate clini- cats, histologic features may be relevant in cutaneous MCTs 64,65 ;
66
cal staging components, such as the presence of metastases or LN primary lung tumors ; mammary carcinoma 67,68 ; and uveal mel-
involvement. anoma. 30,69 Independent validation of many veterinary grading
Histomorphologic features used to determine grade may schemes is limited, and some have been questioned in the recent
include cellular differentiation, number of MFs, invasiveness, literature. 43,70,71 In canine malignant mammary tumors, the his-
necrosis, cellular or nuclear pleomorphism or atypia, nucleolar tologic subtype may be more relevant than conventional grad-
72
size and number, stromal reaction, and overall cellularity. For ing. As with MFs, grade is ultimately an estimate of biologic
many routinely applied grading schemes, histologic features are activity to be taken into account with other parameters and the
described categorically (i.e., range of MFs, invasion pattern, per- overall clinical picture.
centage necrosis, degree of differentiation) and assigned a score. For tumors where grading systems are not well established, the
These grading schemes are time efficient, cost-effective, involve no pathologist may still estimate presumed biologic behavior based
new technology, and easily can be applied to routine diagnostic on the overall degree of tumor differentiation. This might be
specimens. To ensure valid clinical application, grading must use convenient for those tumors where there are insufficient data to
the same methods and criteria as described in the relevant publi- directly associate histomorphology and prognosis. In these cases,
cation; clinicians should be familiar with these studies and ensure the terms well differentiated, moderately differentiated, or poorly
that accurate methods were applied by the pathologist in report- differentiated may infer a low-grade, medium-grade, and high-
73
ing grade. grade malignancy, respectively. Pathologists who advocate for
Because of the nature of grading criteria, there is potential for evidence-based medicine may chose not to include this in the final
subjectivity, and reproducibility between pathologists can be vari- diagnosis unless clearly defined criteria and clinically validated
able. 14,43,44 Despite this limitation, one human study of 440,000 data justify its prognostic utility; however, the degree of differen-
cases found that interobserver variation did not alter the relation- tiation, should be reported in the histopathologic description. 74
ship between grade and outcome. A recent international study Many histomorphologic grading criteria are likely proxies for
14
that applied the World Health Organization (WHO) system of underlying molecular mechanisms that define most malignant
lymphoma classification to canine lymphomas demonstrated an tumors (Table 3.5). In human breast cancer morphologic phe-
overall accuracy of 83% among 17 pathologists. With prepara- notypes and grading criteria are associated with molecular signa-
45
75
tive training and careful application to well-described criteria, rea- tures. Tumoral necrosis, measured in multiple different grading
sonable accuracy can be achieved. If the criteria are more loosely schemes, might be related partly to inadequate perfusion, hypoxic
defined, pathologist concordance likely will be reduced. 44 injury, and vascularization. Tumoral neoangiogenesis is a hallmark
Some degree of grading variation is inevitable because of known of most malignant neoplasms and can be estimated by microvessel
biologic variation that defines tumor types and even individual density. Important mediators in this process include the vascu-
tumors. Tumors often have considerable morphologic heteroge- lar endothelial growth factor (VEGF), thrombospondin-1, and
neity – this might directly reflect the expansion of differing clonal hypoxia-inducible factor-1α. Underlying molecular mechanisms