A. Gross handling and selection of sections. The approach to specimen handling is largely one of common sense. Small biopsy specimens should be submitted for sectioning in their entirety. If there is any doubt about whether they
contain bone, they should be fixed, briefly decalcified, and rinsed. Most bone biopsies performed with needles are sufficiently thin for adequate fixation and decalcification, whereas curettings may sometimes need to be sliced into thinner fragments. The amount of curettings to submit for sectioning depends on their volume and the uniformity of the curettings. When it is feasible, all curettings should be submitted. If the lesion curetted is a hyaline cartilage tumor, as much of the histology as possible should be reviewed to identify atypical chondrocytes as well as any subtle interface with normal surrounding bone.

Other large specimens such as total joint replacements and bone resections also need to be sliced into thinner fragments. Although this may be accomplished with large band saws or other power-type saws, motorized saws are dangerous and somewhat time-consuming to maintain properly, especially in a laboratory that receives a limited number of bone specimens. Vibrating or oscillating saws, which are usually available in autopsy suites, should be avoided if possible, because they do not section uniformly and their oscillating movement creates tension and compression artifacts that often make bone sections impossible to interpret properly. A very easy approach is to hold bone specimens steadily in a tabletop vise or clamp and to cut them with a hacksaw in which two finetooth blades are separated by 2- to 3-mm-thick washers. Such an apparatus is easily and cheaply made, although there are commercial instruments available for the same purpose. It must be emphasized that double-bladed instruments should be scrupulously cleaned between every specimen to avoid tissue cross-contamination between different cases.

The handling and disposition of larger resection specimens depends on the reason for the procedure. For malignant tumors in which patients have not received neoadjuvant chemotherapy (after biopsy but prior to resection), grading, staging, and adequacy of resection are the major clinical issues. Amputations from these patients should include sections from the soft tissue and vascular margins as well as those from the tumor itself. Tumor sections should be taken in such a way as to document the pertinent tumor histology, whether the tumor involves the medulla and/or cortex, and how far the tumor extends into soft tissues. The specimen should be cut in such a way as to disclose the greatest extent of tumor; review of the imaging studies can guide the selection process. Careful attention should be paid to taking sections from any areas that are grossly disparate from the appearance of the majority of the tumor. Radical resections for malignant tumors that are not amputations need the same sectioning methods, but any area of the resection constituting a margin must be sectioned and appropriately designated. This includes the bone resection margin, overlying soft tissue dissection margins, and the margins of any skin and soft tissue encompassing a prior biopsy site.

Specimens resected from patients who have received neoadjuvant chemotherapy (currently used in osteosarcoma and in Ewing sarcoma/primitive neuroectodermal tumor [EWS/PNET]) need more extensive sampling to estimate the extent of treatment-associated necrosis. This means that one or more thin slabs should be cut through the entire extent of the bone and tumor, and that the entire slab or slabs should be fixed, decalcified, mapped, and examined not only for tumor stage, but also for the extent of necrosis. The slabs should be photographed so as to produce a section map; if a specimen x-ray machine is available, specimen radiographs can be used both as section maps and as controls for adequate specimen decalcification (e-Fig. 48.3). Additional sections may be taken if there are areas not in the slab selected that appear as though they might be viable; the pathologist’s task in this enterprise is to find viable tumor if any is present. It is worthwhile to remember that to extrapolate the degree of necrosis in a single slab into necrosis of the tumor as a whole makes
the assumption that what is present in that particular slab is representative of the entire lesion.

B. Decalcification. The main difference between processing of bone specimens and of softer tissues is the requirement for an extra step of decalcification. Removal of calcium insures that bone collagen is no harder than the paraffin in which it is embedded, and that microtomy of bone tissues will approximate that for other types of specimens. Decalcification may be performed in a number of ways. In acid decalcification, hydrogen ions are in effect substituted for calcium ions. Electrolysis in effect accomplishes the same end, but is performed in an electrolyte solution with a weak electrical current. Ionic exchange is the slowest method but is the most gentle on tissue and results in the fewest artifacts. In practice, most histology laboratories rely on weak acid decalcification because it is the quickest and there is pressure from eager clinicians for rapid turnaround times in diagnosis. With use of acid decalcification methods, a few caveats must be kept in mind. First, the tissue must always be fixed adequately prior to decalcification to prevent artifacts that interfere with adequate staining or that can degrade the tissue after sections are prepared. This means that the tissue must be adequately thin (no more than 3- to 4-mm thick) prior to fixation, and that the tissue has remained in formalin or some other suitable fixative for an interval adequate to coagulate the proteins for routine staining. In addition, if immunohistochemistry needs to be performed, adequate fixation helps to insure that the decalcification process will less alter tissue antigens. Second, when decalcification is performed with acid solutions, specimens must be rinsed in running water to ensure that the residual pH of the tissue is sufficiently neutral for hematoxylin staining. Failure to neutralize the acid not only results in understaining with hematoxylin, but also will cause stained sections to lose their hematoxylin staining in an accelerated manner. If time is insufficient for adequate specimen rinsing, the specimen should be neutralized in a dilute basic solution such as sodium bicarbonate. Third, if sections are left in dilute acid for a much longer period than necessary for calcium removal, tissue hydrolysis will remove the nucleic acids that cause nuclear hematoxylin staining and nuclei will appear acidophilic. This so-called overdecalcification artifact is generally not reversible. Overdecalcification may not interfere with many diagnostic interpretations, but it is important not to mistake this artifact for tissue necrosis, particularly in postchemotherapy specimen interpretation.

Adequate decalcification will vary by the tissue being decalcified. For example, woven bone, even though it tends to have higher calcium concentrations than lamellar bone, will often section adequately with incomplete decalcification because the former has less organization and less cutting resistance. The decision regarding whether the tissue is ready for embedding is often subjective and revolves around whether the tissue is pliable, trims easily, or can be penetrated with a needle. Complete decalcification is best judged either by testing the supernatant fluid with a colorimetric indicator or by comparing specimen radiographs prior to and after decalcification. These tests are seldom practical in a very busy general surgical pathology practice.

C. Approach to the interpretation of bone specimens. Patient complaints related to the musculoskeletal system constitute nearly one-third of physician office visits in the United States, so orthopedic problems are extremely common. While surgical pathologists are often asked to rule out bone tumors as the etiology of a clinical problem, it is useful to keep in mind that fractures alone are about 3000 to 4000 times more common than all primary bone tumors combined, and that metastatic tumors to bone are at least 20 times more common than primary bone tumors. The accurate diagnosis of bone diseases requires the correlation of patient demographics along with the clinical history and imaging studies to put the problem in its correct context prior to any histologic examination of the
tissue. Symptoms and signs are fairly similar in orthopedic diseases; these consist of pain, loss of function, deformities, and (in the case of tumors) sometimes a mass or a sense of fullness. Pain is the most common symptom, and although it may vary considerably, pain severe enough to wake a patient from sleep is the type suspicious for neoplastic diseases.

D. Importance of radiologic findings. The surgical pathology of orthopedic diseases most often consists of defining the nature of bone lesions that are space occupying on imaging studies, advising the clinician if an infection may be present, and histologically documenting miscellaneous bone diseases that are not diagnosable by imaging studies alone. Because surgical pathologists usually render biopsy diagnoses with the assumption that a biopsy is representative of the pathologic process, it is natural to assume the same parameters in bone biopsies. This is a potentially dangerous assumption, because most orthopedic diseases are invisible without imaging studies. This means that to assure that a biopsy is representative of a process, the smaller the biopsy specimen, the greater the need to review the imaging studies defining that process. Because bones are deep seated, imaging studies are required to grasp the extent and behavior of bone lesions. For a surgical pathologist, correlating imaging studies with histologic findings depends on knowledge of normal bone and joint anatomy, normal anatomy as represented in radiographic images or other imaging studies, and the rudimentary alterations in imaging studies produced by pathologic processes (not only what a process does to normal bone, but also how normal bone alters the process) (Adv Anat Pathol. 2005;12:155). The majority of the radiographic image produced by long bones is due to beam attenuation by cortical bone in the shafts and by cancellous bone in the ends (e-Fig. 48.4). The attenuation produced by flat bones is primarily due to cortical bone, and that of irregular bones depends on the proportion of bone elements in any given part of the bone.

Space-occupying lesions within bone usually cause bone destruction, bone production, or some combination of the two. Destructive lesions are not seen in a single radiographic view until at least 40% of the bone in the path of the x-ray beam is destroyed. This means that almost an entire thickness of cortex must be destroyed to see the lesion if an intact and a destroyed cortex are superimposed in one view, or that at least 40% of cancellous bone must be destroyed in a bone end. It is partly for this reason that orthogonal views of bones are taken (e.g., posteroanterior and lateral views) so that destructive lesions may be isolated in routine radiographs. Radiodense lesions superimpose on the extant bone, causing more attenuation and easier visibility on a routine radiograph. In contrast, a lesion that is less dense than bone may fill the entire medullary cavity of a bone, but if it does not destroy the cortex it will be invisible regardless of the view because the dominant attenuator of the x-ray beam is the cortex, not the medullary fat or marrow. It is for these reasons that other imaging studies such as computed tomography (CT) scans and magnetic resonance imaging (MRI) are performed. These studies yield information that is complementary to that derived from routine radiographs. While they may be more sensitive in yielding information, a particular type of imaging modality should be used in concert with routine radiographs to answer a particular clinical question not answered by the radiographs.

A. Congenital and developmental diseases. Very few of these disorders come to the attention of surgical pathologists, since most are diagnosed on the basis of their clinical and imaging appearances. Some of them, such as the histiocytoses and storage disorders, may be confirmed histologically, or may be seen incidentally, such as when there is a hip replacement for avascular necrosis associated with Gaucher disease. Many others, such as most of the sclerosing dysplasias exclusive of diseases with specific histologies (e.g., osteopetrosis), demonstrate bone of increased density but are not specific or separable from one another histologically without demonstrating the changes seen in the radiographs (e-Fig. 48.5).

B. Traumatic disorders. Fractures are numerically the most frequent bone and joint disorders. They do not usually come to the attention of surgical pathologists because most treatments are closed or do not produce tissue for diagnosis. In contrast, open fractures requiring debridement and acute fractures of the femoral neck undergoing joint replacement are sometimes received in pathology laboratories. Acute fractures usually demonstrate some degree of accompanying hemorrhage, reactive changes such as dilated sinusoids in still viable nearby marrow, and fragmented bone trabeculae. Subacute fractures will also demonstrate devitalization of the bone at the fracture site (empty osteocyte lacunae and necrosis of marrow), although histologic evidence of healing is usually not evident for 7 to 10 days. Fractures that do not heal and fractures that are thought to be pathologic are sometimes sampled to rule out the presence of tumor or infection. It is very important to know that there is a history of trauma when reviewing tissue, or there is some danger of misinterpreting microcallus or reactive changes as matrix production by a tumor. Even if the history is not available, imaging studies will reveal if a tumor is present, and it is worthwhile to remember that primary bone tumors that produce bone matrix are rarely the sites of fracture. There are histologic parameters to separate bone and cartilage
formation by tumor from that of trauma, although it takes some experience to recognize them. Bone or osteoid production by tumor matrix is often lacelike and becomes sheetlike as more bone is produced. Bone produced as a repair phenomenon may be focally lacelike, but more often it rapidly acquires a microtrabecular architecture and then becomes trabecular as it matures. In reactive bone there is almost always a zonation of maturity that is dependent on both the area in the lesion sampled and the time from trauma. While bone and cartilage are common findings in both osteosarcoma and fracture callus, cartilage tends to disappear as callus matures but it persists in osteosarcoma (e-Fig. 48.6). In addition, the progression from bone to cartilage and back to bone is orderly in reactive processes but is totally random in bone tumors.

D. Paget disease. This disease has some histologic features in common with high-turnover metabolic bone diseases, but it is not a metabolic disease because it does not diffusely affect the entire skeleton and has no known associated metabolic defect (metabolic bone diseases are covered in Chap. 49). Paget disease is characterized by an imbalance or uncoupling of osteoclastic and osteoblastic activities, with osteoclastic bone resorption predominating early in the disease and osteoblastic activity persisting late in the disease. These histologic manifestations are correlated radiographically with characteristic radiolucency early in the disease, radiodensity in the late stages, and a mixed pattern for most of the interval between (Skeletal Radiol. 1995;24:173). Because the bone microarchitecture is altered, there is loss of the normal bone contour radiographically, and there is gradual loss of the normal cortical appearance and an increasingly coarse appearance to the bone trabeculations. A biopsy from an early radiolucent lesion demonstrates large bizarre osteoclasts producing large and irregular resorption pits (Howship’s lacunae) on trabecular surfaces. These are often accompanied by paratrabecular fibrosis and dilated marrow sinusoids. As the resorption pits become filled in by osteoblast activity, irregularly shaped cement lines (sometimes likened to grout lines in a mosaic) mark the demarcation between the old and new bone. The bone on either side of these cement lines demonstrates either lamellar bone, in which the layers are discontinuous on either side, or lamellar bone on one side and woven bone on the other side. As the disease progresses and osteoclast activity slows, the bone becomes thicker and more interconnected than normal, but its arrangement and increased irregular cement lines make it more prone to deformities and fractures (e-Fig. 48.10).

E. Infectious disorders. Infections of bone arise either by direct introduction of organisms into the bone due to open trauma or overlying infections of soft tissue, or by secondary hematogenous spread. Most hematogenous osteomyelitis occurs in the first two decades of life. Its usual site in the bone is in the metaphysis adjoining the growth plate of a long bone because the microcirculation is stagnant in this area. Osteomyelitis due to open trauma can occur at any age; osteomyelitis associated with overlying infections is most often associated with peripheral vascular disease and so is seen later in life. Most infections of the bone are bacterial, but infections with fungi and lower virulence organisms may occur in immunocompromised hosts.

The vast majority of hematogenous osteomyelitis is due to coagulase-positive Staphylococcus aureus, but many other organisms may infect bones. Histologically, microorganisms are seldom seen in bone biopsies of patients with osteomyelitis because the sheer number of organisms required for the sensitivity of high-power or oil-immersion microscopy to detect bacteria is very high. Because of this, bacterial cultures should always be taken when infections of bone are suspected clinically—preferably prior to the institution of antibiotic therapy. A single bacterial culture is on the order of 10 million times more sensitive than histology—even when special stains for organisms are added to the regimen. Infections in bone are often accompanied by necrosis of at least some of the affected bone; the primary reason for this is that edema accompanies
inflammation, and edema in the closed confines of the cortex compromises the medullary nutrient arteries and sinusoids due to resulting increased pressure. The innermost cortical circulation may be similarly compromised by increased intramedullary pressure. If the pressurized exudate finds its way into empty Haversian and Volkmann canals, it may push its way through these intracortical spaces and eventually dissect the periosteum, and its perforating arteries, from the cortex. If the cortex is deprived of its dual circulation, then it in turn becomes necrotic; this necrotic bone is called sequestrum. The combination of necrotic bone sequestrum, marrow fibrosis and/or fat necrosis, and mixed inflammatory infiltrates (usually including neutrophils and plasma cells) provides good histologic corroboration of osteomyelitis, but the demonstration of organisms is the gold standard for the diagnosis of infections (e-Fig. 48.11).

F. Iatrogenic disorders. Treatment-related disorders are seldom a major problem in the pathologic diagnosis of orthopedic disease, provided that an accurate clinical history is communicated to the surgical pathologist. For example, the diagnosis of osteosarcoma would be very unusual in a patient of the sixth decade without prior radiation of the site, or without some other underlying premalignant bone lesion. Administration of various therapeutic regimens may lead to secondary alterations in bones; perhaps the most notable of these is the amyloidosis of bones, tendon sheaths, and ligaments that develops from β2-microglobulin accumulation in long-term hemodialysis patients. Substances that have been given parenterally but that are not metabolized may also be deposited in bones or joints; without prior knowledge of therapeutic treatment, it may be difficult to make an accurate diagnosis (e-Fig. 48.12).

G. Neoplastic and tumor-like lesions. Primary tumors of bone are quite rare, accounting for only 0.2% of all malignancies, or an incidence of 1 per 100,000 individuals per year (Fletcher, CDM, Unni K, Mertens K, eds. Tumors of Soft Tissues and Bone. Lyon, France: IARC Press; 2002). There is a bimodal age distribution, with one peak in adolescence and a smaller one in patients older than 60 years. Among other characteristics, each bone tumor has its own age predilection, which is very useful from a differential diagnostic standpoint. Primary benign bone tumors are probably less common than primary malignant tumors if the very common nonossifying fibroma, osteochondroma, and enchondromas of the hands are excluded. In addition to benign and malignant bone neoplasms, there are a number of nonneoplastic lesions that can present in a manner similar to neoplastic conditions (Table 48.3); all of these lesions are discussed below, and their main features are presented in Tables 48.4 and 48.5. Pathologic stage is among the findings that are recommended to be reported for bone tumors (Hum Pathol. 2004;35:1173) and the American Joint Committee on Cancer (AJCC) Tumor, Node, Metastasis (TNM) staging scheme (Table 48.6) and/or the simpler Musculoskeletal Tumor Society scheme (Table 48.7) can be used for this purpose.