This is a histological classification of tumours and tumour-like lesions of the exocrine pancreas which also includes those tu- mours showing a mixture of exocrine and endocrine elements. The classification is based principally on standard microscopic observations, but whenever indicated it incorporates diagnosti- cally valuable immunohistological findings. In addition, the most important immunohistological findings which are helpful in cat- egorizing pancreatic tumours are summarised in Table 1. The major guideline of this classification scheme is the group- ing of the pancreatic exocrine tumours according to their biologi- cal behaviour. Thus, the neoplasms are broadly divided into benign (adenoma) and malignant tumours (carcinoma). How- ever, in recent years we have learned that this division is not a sharp but rather a gradual transition. We therefore added a third group which we call "tumours of uncertain malignant potential" representing a borderline category analogous to that recognized for some ovarian tumours. This group includes mucinous cystic tumour, intraductal papillary-mucinous tumour and solid- pseudopapillary tumour. These neoplasms are defined by the grade of dysplasia and/or potential to become malignant. Mucinous cystic tumours of uncertain malignant potential, for instance, exhibit moderate epithelial dysplasia, but do not show severe dysplasia/carcinoma in situ changes, nor carcinomatous invasion of the cyst wall or the adjacent pancreatic tissue. Solid- pseudopapillary tumour has a benign looking histological ap- pearance, but metastases may occur. Biologically, all these neoplasms are primarily slow-growing lesions and have an excel- lent prognosis when adequately treated by complete resection.

This is a classification of tumours and tumour-like lesions of the liver. It is based primarily on the microscopic characteristics of the tumours, and is therefore concerned with morphologically identi- fiable cell types and histological patterns. The haematoxylin- and eosin-stained section remains the mainstay of morphological diag- nosis, but special histochemical stains are often helpful and have been referred to in the explanatory notes. Readers interested in specific special stains mentioned in the text should consult Labora- tory Methods in Histotechnology of the Armed Forces Institute of *Pathology, Washington, D. c., USA. 1 The results of immunohisto- chemical methods for identifying various tumour "markers" have also been noted whenever indicated. The present classification incorporates all the previously classi- fied tumours, but also includes several new lesions, viz. biliary papil- lomatosis, the fibrolamellar variant of hepatocellular carcinoma and epithelioid haemangioendothelioma. Several subtypes of hepato- blastoma are mentioned. A serous type of bile duct cystadenoma is described. The section on tumour-like lesions has been expanded to include focal fatty change and inflammatory pseudotumour. The section on adenomatous hyperplasia, including macroregenerative nodules, has been amplified. The number of photomicrographs has been increased from the original 56 to 150. Unlike the first edition, the photomicrographs in the second edition are mostly black and white. All are new and were taken of representative cases on file at the Armed Forces Institute of Pathology.

M. K. Gospodarowicz, P. Hermanek, and D. E. Henson Attention to innovations in cancer treatment has tended to eclipse the importance of prognostic assessment. However, the recognition that prognostic factors often have a greater impact on outcome than available therapies and the proliferation of biochemical, molecular, and genetic markers have resulted in renewed interest in this field. The outcome in patients with cancer is determined by a combination of numerous factors. Presently, the most widely recognized are the extent of disease, histologic type of tumor, and treatment. It has been known for some time that additional factors also influence outcome. These include histologic grade, lymphatic or vascular invasion, mitotic index, performance status, symptoms, and most recently genetic and biochemical markers. It is the aim of this volume to compile those prognostic factors that have emerged as important determinants of outcome for tumors at various sites. This compilation represents the first phase of a more extensive process to integrate all prognostic factors in cancer to further enhance the prediction of outcome following treatment. Certain issues surround ing the assessment and reporting of prognostic factors are also considered. Importance of Prognostic Factors Prognostic factors in cancer often have an immense influence on outcome, while treatment often has a much weaker effect. For example, the influence of the presence of lymph node involvement on survival of patients with metastatic breast cancer is much greater than the effect of adjuvant treatment with tamoxifen in the same group of patients 5]."

The fourth edition of the TNM Classification was published in 1987,1 and a revision in 1992.2 It was the result of efforts by all national TNM Committees towards a worldwide uniform classification. The classifica tion criteria are identical with the fourth edition of the Manual for 3 Staging of Cancer of the American Joint Committee on Cancer (AJCC). Although the classification has found wide acceptance, some workers have pointed out that individual definitions and rules for staging are not sufficiently detailed. This can lead to inconsistent application of the clas sification. the antithesis of standardization. This source of differences in interpretation applies not only to the classification of individual organs but also to the general rules of the system, especially to the definitions of the requirements for the pathological classification (pT, pN). These are specified only for carcinoma of the breast; for other sites, reference must be made back to the general rules. which can lead to variable interpreta tions. The TNM Project Committee of the UICC has addressed this prob lem and collected and considered the criticisms and suggestions from the national TNM Committees as well as from cancer registries, oncolo gical associations and individual users. The result was the decision to complement the 4th edition of the TN M Classification 1.2. 3 with the publi cation of a TNM Supplement containing recommendations for the uni form use of TNM."

This classification is based primarily on the microscopic characteris tics of tumours. It is therefore concerned with the identification of cell types and histological patterns as seen by conventional light micro scopy. In general, time-honoured terms have been retained. Syn onyms are listed only if they have been widely used or if they are con sidered to be important for understanding the disease process. In such cases, the preferred term is given first, followed by the synonym in parentheses. The individuality of a tumour manifests itself principally in its histological appearance and the extent of spread at the time of diag nosis. This volume is concerned only with the histological classific ation of tumours. Anatomical extent or staging is covered in the TNM Classification.! The histological classification of a tumour de pends on two main parameters, typing and grading, and a number of additional parameters which may apply to selected examples. Histological Typing This divides tumours of a given organ into different types according to their direction of differentiation. Although this may frequently in dicate the underlying histogenesis of the tumour, it may be difficult or impossible to identify the cell of origin. Note is taken of the structure and function of cell types, as well as the overall growth pattern of the tumour, with the aim of matching these features to those of a normal tissue found in the same organ.

Knowledge of tumours of the thyroid gland has advanced consider- ably in the 22 years that have elapsed since work was started on the first edition of Histological Typing of Thyroid Tumours. In the intro- duction to that volume it was recognized that the definitions and clas- sifications put forward would need revision in time, and the present text differs substantially from the first edition. As far as is possible, however, the framework of the classification proposed remains the same, as the original classification was widely accepted and proved useful in many studies. The link between the morphological type of thyroid tumour and its epidemiology, natural history, function, prognosis and response to therapy has been further strengthened since the first edition. In par- ticular, the decision taken to separate papillary and follicular carcino- mas and exclude a mixed papillary follicular type has been well justi- fied. One of the major changes has been the recognition that many tu- mours regarded 20 years ago as small cell carcinoma were really ma- lignant lymphoma, and this development has been incorporated into this edition, with increased importance given to primary malignant lymphoma of the thyroid. Much work has also been done on medul- lary carcinoma of the thyroid, its link with mUltiple endocrine neo- plasia syndromes, and its association in its inherited form with C-cell hyperplasia; this too is recognized by an expanded section on this tu- mour.

This volume deals with tumours occurring in the following sites: Nasal cavity and paranasal sinuses Nasopharynx Larynx, hypopharynx and trachea External ear Middle and inner ear The classification of tumours arising from each of the anatomical sites listed above is given separately. However, since the same tu- mour type may occur in more than one site, the definitions and il- lustrations have been grouped together. The classification is based on the histological characteristics of the tumours and is therefore concerned with morphologically identifiable cell types and histo- logical patterns as seen with conventional light microscopy. Al- though many of the histological terms and definitions have histo- genetic implications, this classification is not primarily based on histogenesis. The term "tumour" is used synonymously with neoplasm. The term "tumour-like" is applied to non-neoplastic lesions which clini- cally or morphologically resemble neoplasms; they are included in this classification because of their importance in differential diag- nosis. The terminology adopted for individual tumours is based on their general acceptance and world-wide usage. Synonyms are in- cluded only if they have been widely used in the literature or if they are considered helpful in understanding the lesions. In such cases the preferred terms are given first followed by the synonyms within parentheses.

This is a classification of tumours and tumour-like lesions of the gall- bladder and extrahepatic bile ducts, including the ampulla of Vater. Although most of the lesions are found in all three sites, variations in frequency of the histological types occur and will be noted. The incidence of carcinoma of the gallbladder varies in different parts of the world. Variation is also found in different ethnic groups within the same country. In the United States, for example, carcino- ma of the gallbladder is more common in American Indians than in Caucasians or in Blacks; the rate among female American Indians is 21 per 100000 compared with 1.4 per 100000 among Caucasian fe- males. In Latin American countries, the highest rates are found in Chile, Mexico and Bolivia. In other countries, such as Japan, the inci- dence rates are intermediate between those of American Indians and those of Caucasians. Despite certain features in common, carcinomas of the gallblad- der and carcinomas of the extrahepatic bile ducts show a number of differences. Gallbladder carcinomas are usually associated with cholelithiasis and have a strong female predominance. In contrast, extrahepatic bile duct carcinomas are seen less often, occur in both sexes with equal frequency, are usually not associated with choledo- cholithiasis, produce early biliary obstruction, and are better differen- tiated histologically as a group. Moreover, they are seen in patients with primary sclerosing cholangitis and ulcerative colitis.

This classification is based primarily on the microscopic characteris- tics of tumours. It is therefore concemed with the identification of cell types and histological pattems as seen by conventional light mi- croscopy. In general, time-honoured terms have been retained. Syn- onyms are listed only if they have been widely used, or if they are considered to be important for understanding the disease process. In such cases the preferred term is given first, followed by the synonym in parentheses. The individuality of the tumour manifests itself principally in its histological appearance and the extent of spread at the time of diag- nosis. This volume is concemed only with the histological classifica- tion of tumours. Anatomieal extent or staging is covered in the TNM 1 Classification. The histological classification of a tumour depends on two main parameters, typing and grading, and a number of additional parame- ters which may apply to selected examples.