Section 1 -

Origins, Contemporary Applications and Future of Diagnostic Electron Microscopy Applied to Soft Tissue Tumours Brian Eyden Department of Histopathology Christie Hospital Manchester, UK

1 – Origins The origins of the discipline of diagnostic electron microscopy of tumours lie in the development of instrumentation and technique, on the one hand, and of an intellectual basis for the discipline, on the other. Technical Developments Ruska in Germany was the first to construct an electron microscope in 1933 (Shampo and Kyle, 1997). While in practical terms this first instrument only imaged an aperture – nothing very interesting for pathologists – it nevertheless contained the potential to image structures at the nanometre scale, which would include intracellular structures. The Second World War delayed any significant development of the electron microscope (although see Agar, 1996) but 1945 onwards saw rapid commercialisation and increased availability of electron microscopes. The earliest specimens to be examined were particles and viruses supported on electron-transparent films, and another decade was required for the development of consistent thin-sectioning techniques that permitted the analysis of intracellular structure. The 1960s saw the introduction of high-quality epoxy resins (Luft, 1961), aldehyde fixation as a preliminary fixative to the metal-fixatives in use before that time (Sabatini et al 1963; Rosai and Rodriguez, 1968), while the 1970s saw the important technical step of using dewaxed tissue for electron microscopy (Johannessen, 1977). the Intellectual Basis for Tumour Diagnosis by Electron Microscopy The intellectual basis for the diagnosis of tumours by electron microscopy has several components: (1) – a knowledge of the ultrastructural characteristics of named tumours (for example, that leiomyosarcomas typically have smooth-muscle filaments and lamina); 2 – ultrastructural interpretational expertise, permitting the recognition of named cell structures or organelles [desmosomes, melanosomes and so forth]); 3 – a knowledge of the significance of these structures as markers of cellular differentiation (for example, that thick-and-thin filament arrays and Z-disks are characteristic of rhabdomyoblastic differentiation). There is an additional optional point, namely, that many investigators find it useful or interesting to know the functional significance of the cell structures identified. This can influence diagnostic interpretation. In short, it is necessary to know how to interpret tumour ultrastructure, and to know which structures are significant as indicators of particular lines of differentiation. Development of the Literature on Tumour Ultrastructure One of the first papers on tumour ultrastructure was Gessler's in 1949 (Notes on the electron microscopy of tissue sections and segregations; submicroscopic spherical bodies in human carcinoma), whereas one of the earliest papers on the ultrastructure of a soft tissue tumour – desmoplastic fibroma – was by Rabhan and Rosai in the Journal of Bone and Joint Surgery of America, 1968. Rosai was also, incidentally, among the earliest pathologists to recognise that ordinary buffered histological formalin gave acceptable structural preservation for diagnostic electron microscopy studies (Rosai and Rodriguez, 1968). The later 1960s, the 1970s and to a decreasing extent the 1980s witnessed an explosion of papers on human tumour ultrastructure, mainly by pathologists. Many seminal papers appeared in mainstream pathology journals: American Journal of Surgical Pathology, Human Pathology, American Journal of Clinical Pathology, Histopathology, Journal of Pathology, Journal of Clinical Pathology, Virchows Archiv, as well as Cancer, currently not an essentially pathological journal. These older publications are still useful today, even though, inevitably, more recently recognised tumour entities, such as PEComa and GANT, are not represented. At this time also, two specialised ultrastructural journals were established – the Journal of Submicroscopic Cytology (1969)(later renamed the Journal of Submicroscopic Cytology and Pathology), and Ultrastructural Pathology (1980). Further, a number of textbooks and review series became available on the ultrastructure of human tissues and human tumours: Bloom and Fawcett – A Textbook of Histology (the 1962 edition was one of the first to include good-quality electron micrographic illustrations); Johannessen – Electron Microscopy in Human Medicine (in 11 volumes, volume 4 of which is devoted to soft tissues); Trump and Jones – Diagnostic Electron Microscopy (in 4 volumes), and Russo and Sommers – Tumor Diagnosis by Electron Microscopy (in 3 volumes). More recently, two large volumes on tumour ultrastructure and a major treatise on cell and matrix ultrastructure have become available: Dickersin – Diagnostic Electron Microscopy. A Text/atlas (1994 and 2000); Erlandson – Diagnostic Transmission Electron Microscopy of Tumors (1994); and Ghadially – Ultrastructural Pathology of the Cell and Matrix (1997). Books on functional ultrastructure include: Alberts' Molecular Biology of the Cell (2002), Pollard and Earnshaw's Cell Biology (2002), and Pavelka and Roth's Functional Ultrastructure (2005). Most recently, we have begun to see work in CDROM format (Dardick and Robb, 2004). All of these works – the early mainstream pathology journals of the 1970s and 1980s, the specialist ultrastructural journals, the monographs on normal human tissue ultrastructure, the monographs and edited series on human tumours and on functional ultrastructure – have contributed to a reservoir of data on human tumour ultrastructure which we use today in our contemporary diagnostic practice. One further paper and person meriting special mention here is Jan Vincents Johannessen, not only because of his vision in establishing the journal, Ultrastructural Pathology, but also because of his paper emphasising the value of the wax block as a source of material for ultrastructural diagnosis (Johannessen, 1977). He made the point that although such material was not optimal for electron microscopy to the extent that organelles might be damaged, enough characteristic cell structure can often be retained to make interpretation possible and provide diagnostically useful information. This is relevant today, at least for this author, who is increasingly asked to diagnose on the basis of tumour tissue from the wax block. Partly, this is because a younger generation of pathologists is establishing itself in the workplace without much awareness of the value of electron microscopy in tumour diagnosis, and, partly because of workplace procedures which make it impracticable to save a piece of each lesion in formalin for eventual electron microscopy. 2 – Contemporary Utility and Applications Continuing Utility – Rarity of Soft Tissue Tumours – Increasing Diagnostic Confidence – Academic Value Soft tissue tumours continue to be examined routinely by electron microscopy for diagnostic purposes, and papers continue to be published on tumour ultrastructure. Do we still need this activity, and does it have any value? It is clear that some soft tissue tumour pathologists do not need much of an ultrastructural input into their practice. Very largely, for such pathologists, the diagnostic and prognostic requirements for patient management are met by clinical data, histological sections, immunohistochemistry, cytogenetics and molecular techniques such as FISH and PCR. On the other hand, many pathologists continue to feel a benefit from electron microscopy. It needs to be remembered that with specific regard to soft tissue tumours, these are rare. Sarcomas (many if which will be soft tissue tumours) occur with an incidence of only about 30/million population/year (Fletcher et al, 2002) compared with about 5000 total malignancies/million. This low incidence can make soft tissue tumours more difficult to diagnose because unless a pathologist is specialising in soft tissue tumours or working in a particularly large institution, it may not be easy to gain sufficient exposure to these tumours to generate maximum diagnostic confidence. Electron microscopy can therefore increase the level of diagnostic confidence, and some pathologists continue to need this diagnostic support. Diagnostic precision is of paramount importance, of course, because the differential may include tumours of disparate differentiation which might need different treatment protocols: sarcomas can mimic malignant melanoma, for example (Landas & Urdaneta, 1991; Zelger et al, 1997). Moreover, some pathologists enjoy the academic process of furthering their understanding of tumour cell biology and of the mechanisms of disease – an academic approach that is sadly in decline – even though the data accumulated may have no immediate significance for patient treatment. In fact, if at all possible, and with only a few exceptions, we at the Christie Hospital in Manchester endeavour to examine by electron microscopy each and every soft tissue and related tumour that comes our way. Litigation – Continuing Education In the face of an increasingly educated and ligitious population, pathologists also value the confirmation of even firmly held diagnoses in the case of uncommon tumours. In more general terms, pathologists, I believe, benefit from an ultrastructural input to their work because exposure to electron microscopy also provides continuing education in that his or her interpretations are continually challenged: in short, electron microscopy makes good pathologists better pathologists. Limitations of Immunohistochemistry In soft tissue pathology as in other tumour pathology areas, some of the diagnostic problems arise from uncertainties from immunostaining results due to limitations of immunohisto- chemistry. The most important are: non-specificity of antibodies; limited immunoreactivity in some classes of tumour (for example, fibroblastic tumours); distinction between weak-positive and negative staining; and differing results for the same antibody in different laboratories (Fisher, 2006). Specific Applications As mentioned, many diagnostic difficulties arise from immunostaining uncertainties, but, equally, most cases referred for ultrastructural analysis will have a differential diagnosis which has usually been narrowed down by immunohistochemistry. Here, some recently encountered cases are illustrated to show the continuing value of electron microscopy in achieving diagnostic precision. Cases which by Light Microscopy Do not Pose Diagnostic Problems But which Reveal New Information on Electron Microscopy. Case 1: Malignant myxoid tumour presenting as a gluteal mass in a 54 year old man Tumour resembled a conventional extraskeletal myxoid chondrosarcoma. By electron microscopy the tumour contained abundant mitochondria and neuroendocrine granules. Subsequent immunostaining confirmed neuroendocrine differentiation and a diagnosis of neuroendocrine extraskeletal myxoid chondrosarcoma was made (Harris et al, 2000), a diagnosis which would not have been rendered without the scientific curiosity of the pathologist. Cases Revealing an Unexpected Phenotype Compared with Light Microscopy Case 2: Mass described clinically as "sarcoma of the left groin" in a 65 year old man Histologically, this was a pleomorphic sarcoma with solid and myxoid areas. It was myogenic on the basis of immunostaining for both striated and smooth muscle markers. The diagnosis of rhabdomyosarcoma was favoured but leiomyosarcoma was part of the differential. Ultrastructrually, the cells were packed with rough endoplasmic reticulum, no striated muscle cell differentiation was detected, while rare cells possessed some smooth-muscle myofilaments. Cell processes and prominent lysosomes were also seen. In purely descriptive terms, this tumour could be regarded as a myogenic fibrohistiocytic sarcoma. Distinguishing Between Several Diagnoses Within a Differential Case 3: Malignant round-cell tumour of the pelvis/broad ligament area of a 33 year old woman Tumour consisted of cords and trabeculae of monomorphic, round, oval or polygonal cells. The initial differential was: sex-cord stromal tumour, epithelioid leiomyosarcoma and intra-abdominal (non-desmoplastic) small-round-cell tumour. Immunostaining and the input of a specialist opinion narrowed the differential to sex-cord stromal tumour v. epithelioid leiomyosarcoma. a-smooth-muscle actin (SMA), desmin, inhibin and calretinin were positive. Ultrastructural analysis revealed glandular organisation (lumina with microvillous processes, junctional complexes and basal lamina). No myofilaments were detected. Desmin and SMA suggested leiomyosarcoma, while sex-cord stromal tumour was suggested by inhibin positivity (the case was complicated by inhibin being negative in one laboratory and positive in another) and calretinin. However, it seemed to us that the ultrastructure was inconsistent with smooth-muscle differentiation, and more acceptable for sex-cord stromal tumour. Sertoli cell differentiation would be a possibility given that Sertoli cells sit on a basal lamina and Sertoli cell tumours can express a glandular epithelial organisation. Here, the desmin and cytokeratin staining could be interpreted as aberrant immunostaining. Identifying Myofibroblastic and Smooth-muscle Differentiation in Tumours Case 4: Spindle cell malignancy on the pinna of the ear of a 77 year old man Malignant melanoma had been suspected clinically, and this diagnosis continued to be considered following histopathological examination and the demonstration of S100 protein in a recurrence. It was HMB45 negative, and although it was no longer suspected of being malignant melanoma, an alternative diagnosis was not forthcoming. Ultrastructural examination revealed the three main features of myofibroblastic differentiation – abundant rough endoplasmic reticulum, peripheral myofilaments and fibronexus junctions – which justified a diagnosis of myofibroblastic sarcoma. Case 5: Spindle cell sarcoma of the neck in a 45 year old man, strongly positive for desmin The differential diagnosis was inflammatory myofibroblastic tumour v. leiomyosarcoma. Desmin can be positive in both of these entities, and electron microscopy is recognised as being able to make a distinction between myofibroblastic and smooth-muscle differentiation on the basis of surface features: fibronexuses indicate myofibroblastic differentiation, while lamina, plaques and caveolae indicate a smooth-muscle phenotype. No fibronexuses were seen, while cell surface plaques and lamina were present. No myofilaments were detected, and the tumour was therefore judged to be a poorly differentiated inflammatory leiomyosarcoma. Case 6: Mass in the right thigh of an 11 year old boy Clinically, this tumour was thought to be a lipoma, despite the patient's age, but histopathologically it resembled and was diagnosed as solitary fibrous tumour (SFT). It had delicate smooth-muscle actin and H-caldesmon immunostaining, while desmin was negative. Ultrastructurally, the cells possessed poorly developed myofilaments and a well defined lamina. Lamina is not a feature of SFT, and this, in combination with the smooth muscle actin and H-caldesmon, suggests a kind of smooth-muscle tumour, possibly a myofibroblastoma. Despite the name, myofibroblastomas have been regarded by some authors as showing a primitive level of smooth muscle differentiation (Eyden et al, 1996, 1999; Eyden and Chorneyko, 2001). 3 – The Future Continuity and Contraction A few years ago in Histopathology, Gary Mierau and I gave views on the current status of tumour diagnosis by electron microscopy (Mierau, 1999; Eyden, 1999). My paper emphasised the modest objective of continuity, while Dr Mierau's was an affirmation of a resurgence of electron microscopy. When one thinks of the future of diagnostic electron microscopy applied to tumours, however, contraction of the discipline in the face of other emerging technologies inevitably comes to mind: indeed, this has happened over the years since the emergence of immunohistochemistry in the early 1980s (Fisher, 2006). Much in the future will depend on funding and economic activity, the will of individuals to keep the technique alive, and geography. Funding – Rationalisation – Large Centres Funding issues frequently lead to contraction of overall resources into smaller numbers of fairly well-funded laboratories. This is perhaps starting to happen in the United Kingdom, as representatives of the UK Association of Clinical Electron Microscopists are liaising with Government Department of Health officials to discuss the support of a small number of large diagnostic electron microscopy centres, which will not only work on tumours but also non-neoplastic fields, such as renal and neuromuscular disease. This is a parochial example, but generally, diagnostic electron microscopy is likely to continue in big hospitals or specialist cancer centres where the workload can justify the technique and the salaries of staff. the Ultrastructural Enthusiast Often, diagnostic electron microscopy (and probably many another technique) is pursued by the forceful personality of an individual who loves electron microscopy, and recognises its value. In these circumstances, the technique may disappear with the retirement of the individual and leave a vacuum of diagnostic expertise and intellectual activity. Geography As just mentioned, a sizeable workload is usually needed to justify a diagnostic electron microscope centre. Juan Rosai, at the 2005 4th Asia-Pacific IAP meeting in Beijing , gave a lecture illustrating how pathology has developed in terms of time (from 14th century to the present), geography and economic activity (Italy England Germany America). Rosai hinted at China being the next country in this sequence. China certainly has plenty of huge population centres, which will produce correspondingly large numbers of tumours: among these will be diagnostically problematical ones, and here diagnostic electron microscopy may have a role. 4 – Conclusions Diagnostic electron microscopy of tumours (including soft tissue tumours) arose from purely scientific (academic) enquiry, and the technique became practically useful throughout the 1970s and 1980s in particular. Despite a contraction due to the emergence of competing techniques, it continues to be useful for solving real diagnostic difficulties, as well as confirming fairly secure diagnoses for the purpose of providing maximum diagnostic confidence for the pathologist and a protection against litigation. Outside the narrow confines of providing a more precise diagnosis, electron microscopy arguably helps pathologists become better pathologists, provides material for furthering our understanding of tumour cell biology for academically orientated pathologists, and material for teaching. The future of the technique can be secure provided a combination of factors coincide: these include the continued funding of large or specialist centres (even if these are few), and the emergence and support of motivated individuals who recognise the value of the technique and are willing to help others appreciate its value. 5 – References Agar AW. The story of European commercial electron microscopes. Adv Imaging Electron Phys 96, 415-584, 1996 Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular Biology of the cell, 4th ed, Garland Science, New York, 2002 Bloom W and Fawcett DW – A textbook of Histology, WB Saunders, Philadelphia , 1962 Dardick I, Robb I. Primer of Diagnostic Electron Microscopy for Pathologists-in-Training [CD-ROM]. Society for Ultrastructural Pathology/Pathology Images Inc, 2004 Dickersin GR. Diagnostic electron microscopy. A text/atlas. Igaku-Shoin, New York , 1988 and 2000 Erlandson RA. Diagnostic Transmission Electron Microscopy of Tumors. Raven Press, New York , 1994 Eyden B. Electron microscopy in tumour diagnosis: continuing to complement other diagnostic techniques (part of an "Expert Opinion" feature: Electron microscopy for tumour diagnosis: is it redundant? Histopathology 35: 99-108 1999) Histopathology 35, 102-108, 1999 Eyden BP, Chorneyko KA. Intranodal myofibroblastoma: study of a case suggesting smooth-muscle differentiation. J Submicrosc Cytol Pathol 33, 157-163, 2001 Eyden BP, Harris M, Greywoode GIN, Banerjee SS. Intranodal myofibroblastoma: report of a case. Ultrastruct Pathol 20, 79-88, 1996 Eyden BP, Shanks JH, Ioachim E, Ali HH, Christensen L, Howat AJ. Myofibroblastoma of breast: evidence favoring smooth-muscle rather than myofibroblastic differentiation. Ultrastruct Pathol 23, 249-257, 1999 Fisher C. The comparative roles of electron microscopy and immunohistochemistry in the diagnosis of soft tissue tumours. Histopathology 48, 32-41, 2006. Fletcher CDM, Rydholm A, Singer S, Sundaram M, Coindre JM. Soft tissue tumours: epidemiology, clinical features, histopathological typing and grading. Pathology and genetics of tumours of soft tissue and bone. Fletcher CDM, Unni KK and Mertens F (eds), IARC Press, Lyon , 2002 Gessler AE, McKarty KS et al . Notes on the electron microscopy of tissue sections and segregations; submicroscopic spherical bodies in human carcinoma. Exp Med Surg 7, 237-268, 1949 Ghadially FN. Ultrastructural Pathology of the Cell and Matrix, 4th edition. Butterworth-Heinemann, Boston , 1997 Harris M, Coyne J, Tariq M, Eyden BP, Atkinson M, Freemont AJ, Varley J, Attwooll C, Telford N. Extraskeletal myxoid chondrosarcoma with neuroendocrine differentiation. A pathologic, cytogenetic, and molecular study of a case with a novel translocation t(9;17)(q22;q11.2). Am J Surg Pathol 24, 1020-1026, 2000 Johannessen JV. Use of paraffin material for electron microscopy. Pathol Annu 12, 189-224, 1977 Johannessen JV. Electron Microscopy in Human Medicine. Volume 4 Soft tissues, bones and joints, McGraw-Hill , New York, 1981 Landas SK, Urdaneta LF. Pleomorphic melanoma mimicking sarcoma. Southern Med J 84, 770-772, 1991 Luft JH. Improvements in epoxy resin embedding methods. J Biophys Biochem Cytol 9, 409-414, 1961 Mierau G. Electron microscopy in tumour diagnosis: not redundant – resurgent! (part of an "Expert Opinion" feature: Electron microscopy for tumour diagnosis: is it redundant? Histopathology 35, 99-108 1999) Histopathology 35, 99-101, 1999 Pavelka M, Roth J. Functional Ultrastructure. At Atlas of Tissue Biology and Pathology. Springer-Verlag, Wien, 2005 Pollard TD, Earnshaw WC. Cell Biology. Saunders, Philadelphia , 2002 Rabhan WN, Rosai J. Desmoplastic fibroma. Report of ten cases and review of the literature. J Bone and Joint Surg Am 50, 487-502, 1968 Rosai J, Rodriguez HA. Application of electron microscopy to the differential diagnosis of tumors. Am J Clin Pathol 50, 555-562, 1968 Russo J, Sommers SC. Tumor Diagnosis by Electron Microscopy (in 3 volumes), Macmillan , New York, and Field & Wood, New York, 1988-1990 Sabatini DD, Bensch KG, Barrnet RJ. Cytochemistry and electron microscopy: The preservation of cellular ultrastructure and enzymatic activity by aldehyde fixation. J Cell Biol 17, 19-58, 1963 Shampo MA, Kyle RA. Ernst Ruska – inventor of the electron microscope. Mayo Clin Proc 72, 148, 1997 Trump BF, Jones RT. Diagnostic Electron Microscopy (in 4 volumes). Wiley, New York, 1978-1983 (volume 3 contains a chapter on soft tissue tumors by Taxy JB and Battifora H) Zelger BG, Steiner H, Wambacher B, Zelger B. Malignant melanomas simulating various types of soft tissue tumors. Am Soc Dermatol Surg 23, 1047-1054, 1997