Diagnostic Approach and Prognostic Factors of Cancers

Pathologic diagnosis is still ranked as a gold standard of tumor diagnosis despite remarkable advances in imaging techniques and molecular biology of tumors. As modern medicine is built on the basis of pathology, pathologic terminology has become a common language system between doctors, especially in the oncology related fields. In most instances, treatment of cancer patients starts after pathologic diagnosis; therefore, correct diagnosis is extremely important for timely and appropriate treatment. Although the pathologic diagnosis of tumors comprises a major portion of routine daily practice for the pathologist, it is often problematic for pathology trainees and young pathologists. However, when one follows a systematic approach to make a diagnosis of a lesion, it is relatively easy to render an accurate cancer diagnosis in almost all cases during routine daily practice. In this review, we describe a systematic algorithmic process that may be used by pathologists and trainees during observation of slides to make a correct diagnosis of a lesion and to subsequently provide appropriate grading and staging information. This approach is simple by design and easy to apply, however, as in most issues that we deal with in pathology, there are several exceptions. We believe that if this approach is consistently implemented, the vast majority of lesions in surgical pathology can be successfully analyzed.

SYSTEMATIC APPROACH TO DIAGNOSIS OF TUMORS

The thought process of a pathologist during histopathologic diagnosis can be summarized as follows: Is a lesion identified in the slides, or do the slides show merely normal histologic features? If a lesion is identified, is it neoplastic or non-neoplastic? If the lesion is neoplastic, is it an epithelial neoplasm or a mesenchymal neoplasm in nature? And, finally, is the neoplasm malignant or benign (Fig. 1)? For this process, basic knowledge of normal histology and pertinent clinicoradiologic information are essential. A neoplasm is, in most instances, consisted of a clonal proliferation of 1 lineage of cells (with a few exceptions, of course, such as mixed tumors). Neoplastic proliferations consist of a population of transformed cells having essentially identical morphology; in contrast, non-neoplastic lesions are consisted of mixed cell populations with varying morphologic features (Fig. 2). On the basis of this concept, neoplastic and non-neoplastic lesions are readily distinguished. It is important, however, to understand that neoplastic lesions often contain a background of inflammatory cells, which may confuse trainees when trying to determine if a lesion is consisted of a single clonal population or not. Such inflammatory cells should be “subtracted” or “ignored” during this phase of lesion analysis. If a population of similar cells remains after subtracting out the inflammatory cells, the lesion may be neoplastic.

F1-6FIGURE 1:

Algorithmic approach to diagnosing tumors based on histologic features.

F2-6FIGURE 2:

Neoplastic lesion consisted of single cell population (left). Non-neoplastic (reactive) lesion consisted of mixed cell population (right).

The next important step is distinguishing whether the tumor is epithelial or mesenchymal in nature (Fig. 3). Neoplasms generally have a tendency to recapitulate the morphologic features of their tissue of origin. Epithelial neoplasms generally are consisted of round, oval, or polygonal tumor cells organized in tumor cell nests (or glands), which are separated by non-neoplastic stroma with or without desmoplastic reaction. Feeding vessels are situated in the stroma in between tumor cell nests (Fig. 4). In contrast, mesenchymal neoplasms are consisted of spindle cells, which are not organized in tumor cell nests but rather diffusely arranged. Intervening stroma is not present between tumor cells and desmoplastic reaction is generally absent. Feeding vessels open directly between tumor cells and are not surrounded by any appreciable stroma (Fig. 5). Table 1 contrasts the features of epithelial and mesenchymal neoplasms.

T1-6TABLE 1:

Differential Diagnosis of Epithelial and Mesenchymal Neoplasms

F3-6FIGURE 3:

Normal endometrium, in which epithelial cells having round or oval nuclei are grouped together to form glands surrounded by well-developed stroma (left). Normal muscular layer of uterus (myometrium), in which spindle cells are diffusely arranged and blood vessels open directly between spindle cells (right).

F4-6FIGURE 4:

Epithelial neoplasm. Round, oval, or polygonal nuclei, formation of tumor cell nests, well developed stroma between cell nests and blood vessels within the stroma.

F5-6FIGURE 5:

Mesenchymal neoplasm. Spindle nuclei, diffusely arranged tumor cells without nest formation, no evident intervening stroma and blood vessels opening directly between tumor cells.

When following the above-mentioned systematic approach, approximately 80% to 90% of tumors can be classified into either epithelial or mesenchymal categories. Some tumors, however, may show overlapping features with both epithelial and mesenchymal characteristics. Sarcomatoid carcinomas, sarcomas with epithelioid features, lymphomas, melanomas, neuroendocrine tumors, and small round cell tumors are included in this category. These tumors may have round, oval, or polygonal cells mimicking an epithelial neoplasm but show a diffuse growth pattern with minimal or absent desmoplastic stroma. Vessels may open directly between tumor cells, as well. These overlapping features may cause confusion when attempting to discern the origin of the tumor cells (Figs. 6A, B, upper). In these cases, histochemical stains, immunohistochemical stains, electron microscopic analysis may be required to further classify the lesion (Figs. 6C, D, lower). Intelligent, well planned use of immunohistochemical stains with appropriate antibody selection is especially important in such overlapping lesions. For example, cytokeratin for epithelial tumors, S-100 protein and HMB45 for melanoma, leukocyte common antigen (CD45) for lymphoma, smooth muscle actin and desmin (muscle-related markers), and synaptophysin and chromogranin for neuroendocrine tumors are all extremely useful for the diagnosis of confusing cases. In the near future (and even in the present!), molecular diagnostic pathology, particularly in soft tissue sarcomas, will give clues and vital information for the diagnosis.

F6-6FIGURE 6:

Some tumors may show epithelial cellular morphology with mesenchymal growth pattern. This tumor is consisted of round, oval, or polygonal shaped epithelioid cells, but the tumor cells are arranged in diffuse sheets without tumor nest formation or intervening stroma. Blood vessels open directly between the tumor cells (A, low power; B, high power). The tumor cells show positive immunostaining for cytokeratin, confirming the diagnosis of sarcomatoid carcinoma (C, low power; D, high power).

The next step is determination of the behavioral nature of the tumor (ie, whether the tumor is benign or malignant). Pathologic characteristics of benign and malignant tumors are summarized in Table 2.1,2 In most clinical settings, the diagnostic approach is started with a less invasive small biopsy of the tumor early before metastasis presents; the presence of locally invasive growth pattern or metastases cannot always be assessed in such a biopsy or early lesion. Therefore, the biological nature of a tumor often must be determined based on morphologic features representing growth rate and on nuclear characteristics. Malignant tumors generally grow rapidly, and, accordingly, show frequent mitoses, higher cellularity (cellular density), tumor cell necrosis, and apoptosis (Fig. 7).

T2-6TABLE 2:

Differential Diagnosis of Benign and Malignant Neoplasms (I)

F7-6FIGURE 7:

Malignant neoplasm showing high cellularity (left), and high cellularity in conjunction with necrosis (right).

Nuclear characteristics are important parameters to determine the malignant potential of a tumor. Malignant tumors usually have large nuclei with higher nuclear-cytoplasmic ratio, nuclear hyperchromasia, prominent nucleoli, nuclear pleomorphism, and increased number of mitoses. When biopsies do not include normal cells to compare nuclear morphology, nuclear size and chromasia cannot be determined easily (Fig. 8). In this case, nuclei of endothelial cells or fibroblasts adjacent to the tumor can be used as reference cells with which to compare tumor cell nuclei to determine nuclear size and hyperchromasia. Prominent nucleoli are also usually helpful to favor a diagnosis of malignancy. However, the observation of “prominent” nucleoli may be subjective according to the individual pathologist. Thus, we recommend usage of the Fuhrman nuclear grade criteria, which is widely used in renal cell carcinoma and as a reference for grading in general.3 In the Fuhrman system, nucleoli can be considered as prominent when nucleoli are identified on medium power (100× magnification, 10× objective). Nuclear pleomorphism is also an important parameter for determining malignant potential in tumors, yet it too may be subjective. To increase objectivity of the observation of nuclear pleomorphism, the standard of the National Wilms' Tumor Study for determination of anaplasia of Wilms' tumor is applicable.4 When tumor cells have variable sized nuclei and the size difference between the largest nucleus and the smallest nucleus is more than 3 times, the tumor nuclei can then be considered pleomorphic. If mitotic figures are 10 or more per 10 high power fields (HPFs; 400×) and atypical mitotic figures are observed, the tumor is most probably malignant (Fig. 9). When mitotic figures are absent or <1 per 10 HPFs a benign tumor is favored. When mitotic figures are more than 1 but <10 per 10 HPFs in a tumor, other nuclear features, as mentioned above, are considered together with the mitotic rate to determine malignancy (Table 3). However, we have to keep in mind that some types of tumors may require a specific mitotic rate to determine malignant potential (eg, in gastrointestinal stromal tumor, 5 or more mitoses/50 HPFs is a sign of malignancy).

T3-6TABLE 3:

Differential Diagnosis of Benign and Malignant Neoplasms (II)

F8-6FIGURE 8:

Malignant cells have large nuclei (A), hyperchromatic nuclei (B), with prominent nucleoli (C), and nuclear pleomorphism (D).

F9-6FIGURE 9:

Malignant neoplasm showing frequent mitoses (left) with occasional atypical mitotic figures (right).

Benign tumors usually grow slowly, and therefore, show lower cellularity, no or minimal necrosis or apoptosis, and rare or absent mitoses. Nuclear atypia is usually minimal. Therefore, nuclear enlargement, hyperchromasia, pleomorphism, prominent nucleoli, and mitoses are usually absent in benign tumors.

One important consideration is that certain lesions may be unclassifiable; that is, they cannot be categorized easily into (1) neoplastic condition versus non-neoplastic, (2) epithelial versus mesenchymal neoplasm, or (3) benign versus malignant neoplasm. In the case that a lesion cannot be classified as neoplastic or nonneoplastic, epithelial or mesenchymal, or benign or malignant, one could diagnose this lesion as “unclassified lesion of uncertain malignant potential.” However, in such a situation, one should obtain consultation on the case from an expert in that particular field to make certain that a lesion is truly unclassifiable and/or of uncertain malignant potential before rendering such a diagnosis.

After determining the nature of a tumor as either epithelial or mesenchymal, and either benign or malignant, the tumor may be classified according to its cellular origin, direction of differentiation, and gross or microscopic growth patterns. As nuclear features are very useful in determining malignant potential, cytoplasmic characteristics give more information with regards to cellular origin and direction of differentiation. Tumor classification is, in most cases, possible on H&E stained histologic tissue sections. Sometimes, histochemistry, immunohistochemistry, electron microscopy, and molecular pathology are required for further classification and for demonstration of prognostic and predictive markers of tumors.

PROGNOSTIC FACTORS OF TUMOR

Currently the most important and well validated prognostic factors of cancer are tumor stage and histologic grade (category I prognostic factors). These factors are directly used for determining the patient's treatment and evaluation of prognosis. Therefore, pathologic diagnosis should include findings for evaluating tumor stage and grade, which will play a crucial role in predicting patient outcome and selection of modalities of further treatment. Other prognostic factors are classified into category II or III and require additional study before for they are ready for clinical usage (Table 4).5

T4-6TABLE 4:

Prognostic Factors

TNM Classification of Tumor (Table 5)6

In principle, depth of invasion in tumors of hollow viscous, such as the gastrointestinal tract and urinary bladder, and size of tumor in solid organs, such as breast and kidney, are used to determine the primary tumor (T) stage. Regional lymph node (N) stage is classified as N0, N1, N2, or N3 according to the number of lymph node metastasis and/or site of lymph node metastasis. Metastasis in other than regional lymph nodes (nonregional lymph node metastasis), for example, mediastinal lymph node metastasis of testicular cancer or retroperitoneal lymph node metastasis of lung cancer, is considered as distant metastasis (M) rather than nodal (N) stage classification. When a distinct tumor nodule without adjacent lymphoid tissue or complete nodal effacement is identified in the soft tissue adjacent to the main tumor, such a nodule should be considered as a regional lymph node metastasis, provided the nodule is round, has discrete margins, and has a capsule-like structure. However, if the nodule has irregular margins with infiltration to adjacent tissue, it is considered to be invasion and direct extension of the primary tumor and is thus classified according to T stage. In addition, when tumor directly invades into an adjacent lymph node it is still considered to be a nodal metastasis. Distant metastasis (M) stage is classified as M0 or M1 according to the absence or presence, respectively, of distant organ metastasis and/or site of metastasis. Overall disease stage is classified into stage I, II, III, or IV using the combination of T, N, and M stages. Usually, stage I and II are considered as localized disease whereas stage III and IV are considered as advanced disease. Stage is expressed by Roman numerals: I, II, III, and IV.

T5-6TABLE 5:

Tumor Stage

Tumor Grade

Grading systems for tumors fall into 3 general categories: histologic grading systems, nuclear grading systems, and combined histologic and nuclear grading systems.1,2,7–10 Histologic grade is based on degree of differentiation of tumor, originally classified as 4 grades by Broders' grading system, which is the prototypical histologic grading system. When more than 75% of a tumor is consisted of well-differentiated areas resembling its normal (tissue of origin) counterpart, the tumor is classified as grade 1. When well differentiated areas comprise 50% to 75%, 25% to 50%, or <25%, the tumor is classified as grade 2, grade 3, and grade 4, respectively.7 Using the 4-tiered grading system, there were problems of interobserver and intraobserver reproducibility in classification of grade 2 and grade 3 tumors. Therefore, Broders' grading was modified into a 3-tiered system, in which grade 1 was used for well differentiated tumors, grade 2 (grade 2 and 3 of the original Broders' grading) for moderately differentiated tumors, and grade 3 (grade 4 of the original Broders' grading) for poorly differentiated tumors. Recently, 2-tiered systems (ie, low grade and high grade) have been used increasingly to decrease interobserver discrepancy and to increase clinicopathologic relevance.1,2

Nuclear grading systems classify tumors based on tumor areas with the worst nuclear grade. Fuhrman nuclear grade3 for renal cell carcinoma and Black nuclear grade9 for breast carcinoma are typical nuclear grading systems and are widely used for clinical practice. Fuhrman nuclear grading is based on observation of nucleoli and uniformity of nuclei, and is a relatively simple and reproducible grading system. When nucleoli are easily identifiable on medium magnification (100×, or 10× objective), nuclear grade is either 3 or 4; if not, nuclear grade is either 1 or 2. Nuclear grade 3 and 4 are distinguished by nuclear uniformity. If the nuclear size is uniform, nuclear grade is 3. If there is nuclear pleomorphism, the grade is 4. Nuclear grade 1 and 2 are distinguished by nuclear detail under high magnification (400×, or 40× objective). When nucleoli or nuclear details are observed at high power, the nuclear grade is 2. When nucleoli are not observed or when nuclear chromatin is condensed, the nuclear grade is 1 (Table 6). In Black nuclear grading,9 5 nuclear characteristics including nuclear size, stainability, nuclear pleomorphism, nucleoli, and mitoses are considered.

T6-6TABLE 6:

Fuhrman Nuclear Grade

The third category of grading system is combined grading. The best example is The International Federation of Gynecology and Obstetrics grading for endometrial cancer,10 which combines histologic grade and nuclear grade. First, the endometrial cancer is divided into well (5% or less solid areas), moderately (6% to 50% solid) and poorly differentiated (more than 50% solid) based on gland formation and proportion of solid areas. After histologic grade is assigned, nuclear grading is evaluated. If the nuclear grade is high in a well or moderately differentiated tumor, the final grade is upgraded to the next tier, so that the end result is moderately or poorly differentiated, respectively. Tumors of each different organ may have their own individualized grading systems. Tumor grade is expressed by Arabic numbers: 1, 2, 3, and 4.

CONCLUSIONS

In pathology practice, a systematic approach after an algorithmic flow of thoughts facilitates an easier arrival at the correct diagnosis. The algorithmic steps are as follows: (1) Neoplastic or non-neoplastic? (2) Epithelial or mesenchymal? and (3) Benign or malignant? The next step is determination of the tissue of origin for malignant tumors using morphology and ancillary studies. Finally, prognostic factors for malignant tumors must be assessed. The most important prognostic factors are stage and grade. Continuous effort to increase knowledge of staging systems for each organ and to maintain consistency in tumor grading is highly encouraged to provide the highest possible quality in clinicopathologic correlation and the most relevant information with regards to patient prognosis.

REFERENCES 1. Cotran RS, Kumar V, Collins T Robbins Pathologic Basis of Disease. 19996th ed Philadelphia W.B. Saunders Co:260–327 2. Lieberman MW, Lebovitz RMIvan Damjanov, James Linder. Chapter 24, neoplasia Anderson's Pathology. 199610th ed St. Louis Mosby:513–547 3. Fuhrman SA, Lasky LC, Limas C. Prognostic significance of morphologic parameters in renal cell carcinoma Am J Surg Pathol.. 1982;6:655–663 4. Beckwith JB. Wilms' tumor and other renal tumors of childhood: a selective review from the National Wilms' Tumor Study Pathology Center Hum Pathol.. 1983;14:481–492 5. Hammond ME, Fitzgibbons PL, Compton CC, et al. College of American Pathologists Conference XXXV: Solid tumor prognostic factors—which, how and so what? Summary document and recommendations for implementation. Cancer Committee and Conference Participants Arch Pathol Lab Med.. 2000;124:958–965 6. Greene FL, Page DL, Fleming ID, et al. AJCC Cancer Staging Manual. 20026th ed New York Springer 7. Giacomarra V, Tirelli G, Papanikolla L, et al. Predictive factors of nodal metastases in oral cavity and oropharynx carcinomas Laryngoscope.. 1999;109:795–959 8. Jemal A, Siegel R, Ward E, et al. Cancer Statistics, 2009 CA Cancer J Clin.. 2009;59:225–249 9. Black MM, Speer FD. Nuclear grade in cancer tissues Surg Gynecol Obstet.. 1957;105:97–102 10. Bilgin T, Ozuysal S, Ozan H. A comparison of three histological grading systems in endometrial cancers Arch Gynecol Obstet.. 2005;172:23–25

留言 (0)

沒有登入
gif