Section 1 -

Bone and Gynecologic Specimens Anthony Montag, M.D. Aliya N. Husain, M.D. Jerome B. Taxy, M.D.

Frozen Sections: a Brief History [1] 1891 William Welch produces frozen section for Halstead at Johns Hopkins. Unfortunately the operation was completed long before the slide was ready [2] 1895 Cullen, working in Welch's lab, reports a 15 minute frozen technique 1896 Plenge (Heidelberg) and Pick (Berlin) report frozen section techniques. Pick credit Cullen, Plenge claims credit 1897 Kanthack and Pigg (Cambridge) publish their method for frozen section, which includes boiling the tissue in water for 5 minutes prior to freezing 1905 Wilson (Mayo clinic) develops frozen technique using methylene blue and publishes in JAMA. Generally credited with developing the frozen section. 1905-1925 Frozen section primarily done in private hospitals 1921 American Society of Clinical Pathology formed as professional society to elevate the status of Pathologist that provided services for live patients (rather than autopsy and research). 1927 Bloodgood (surgeon, Hopkins) states that surgery should not be performed in hospitals that do not have frozen section readily available. Indications for Frozen Section: To make an intraoperative decision concerning further surgery (excision, staging) Margins To assess adequacy of specimen when an open biopsy is being preformed To plan work-up and special studies on the specimen Quality Assurance Indicators in Frozen Section: Frozen section turn around time. [3] 90% of frozen section slides should be ready to read within 15 minutes of receipt 90% of frozen section diagnoses should be reported within 20 minutes of receipt Frozen section discrepancy The discrepancy rate between frozen and permanent diagnoses is approximately 2% [4] Most are sampling issues The acceptable rate for diagnostic error is 1% Deferral The more experienced the laboratory, the lower the deferral rate. Mayo defers 2% [4] A private hospital based study deferred 4.6% [5] Case 1. Chondroblastoma The differential diagnosis of an epiphyseal lesion is chondroblastoma, giant cell tumor, and clear cell chondrosarcoma. Before viewing the frozen section slide, this differential can be ranked by additional features such as age and radiologic features. Chondroblastoma is more centrally located than giant cell tumor, which is typically eccentric and frequently expands the contour of the bone. Clear cell chondrosarcoma radiographically resembles chondroblastoma, with central location, although it may extend from epiphysis to the articular cartilage, and may be heavily mineralized. Both clear cell chondrosarcoma and giant cell tumor occur in an older age group than chondroblastoma, albeit overlapping. Histologically chondroblastoma presents a sheet of polygonal cells with disturbingly irregular and dark nuclei on close inspection. Touch preps highlight the oval nucleus, occasionally with a longitudinal cleft. True pleomorphism is absent, and mitotic figures are rare. Scattered osteoclast-like giant cells may mimic giant cell tumor, however the stromal cell of giant cell tumor is less polygonal and has finer chromatin. Giant cells are usually more diffusely distributed in giant cell tumors. Chondroblastomas have a stromal background varying from a fine chicken wire pattern of calcified cartilage matrix, to hyaline cartilage, frequently calcified, to pink zones of "osteoid-like" cartilage matrix. Although reactive bone formation may be seen at the edge of the lesion or associated with fracture, scattered well-formed spicules of lamellar bone, as frequently seen in clear cell chondrosarcoma, are never seen in chondroblastoma. Features of epiphyseal lesions Chondroblastoma Giant Cell Tumor Clear Cell Chondrosarcoma Radiology Epiphyseal, central, usually not expanding the contour of the bone Epiphyseal, often with metaphyseal involvement, expansile, lytic Epiphyseal, central, not expansile, but may extend from epiphysis to articular cartilage Age 10 to 20 20 to 40 20 to 40 Cells Polygonal Oval to plump spindled stromal cells Abundant clear cytoplasm Nucleus Smudgy, notched, atypical looking, but uniform Oval and similar to giant cell nuclei Uniform, large, clear blown out chromatin, large nucleoli Matrix Cartilage, chicken wire calcified matrix, pink osteoid-like matrix Reactive bone at edges or secondary to fracture. No matrix Loose cartilage matrix with calcification. Well formed spicules of lamellar bone Giant cells Scattered to numerous Usually diffuse and numerous Scattered Case 2 Aneurysmal Bone Cyst Case 3 Telangiectatic Osteosarcoma The intraoperative differentiation of cellular cystic lesions of bone is difficult, and is accompanied by the risk of misdirecting the surgeon to do a primary curetting of a malignant lesion, contaminating the surgical site with tumor and losing the opportunity to perform limb-sparing surgery. Both aneurysmal bone cyst and telangiectatic osteosarcoma may present as an expansile, purely lytic, cystic mass, frequently with fluid-fluid levels on CT or MRI examination, indicating layers of fresh and old blood in a cystic space. Grossly, the biopsy material may appear similar: hemorrhagic, fleshy, with cystic spaces and little bone matrix. Microscopically, both ABC and telangiectatic osteosarcoma have cyst-like spaces without an endothelial lining. The septae are generally more cellular in telangiectatic osteosarcoma, and are frequently broader. Osteoclast-like giant cells are present in both. Frank pleomorphism and frequent and atypical mitotic figures are seen in most cases of telangiectatic osteosarcoma, while the stromal cells of ABC are uniform and display occasional mitoses. Osteoid production is difficult to find in most cases of telangiectatic osteosarcoma, but when present is typically a fine meshwork of woven bone. Osteoid may be present in ABC, but tends appear as broader seams of woven bone with histologically benign osteocytes and osteoblasts. In the absence of fracture, the presence of necrosis in a cystic lesion of bone is an ominous sign is highly suggestive of telangiectatic osteosarcoma. Features differentiating aneurysmal bone cyst from telangiectatic osteosarcoma Aneurysmal Bone Cyst Telangiectatic osteosarcoma Radiology Expansile, lytic, cystic fluid-fluid levels Expansile, lytic, cystic fluid-fluid levels Peak Age 10 to 20 10 to 20 Low power Cystic spaces, webs, blood Cystic spaces, webs, blood Cells Short spindled stromal cells, fibroblasts Spindled to polygonal stromal cells, pleomorphic giant cells Nucleus Sometimes reactive but uniform Pleomorphic, size variability, Mitoses Occasional, but normal Atypical Giant cells Osteoclast-like, frequently at edges of cyst wall Osteoclast-like and pleomorphic Matrix Reactive woven and lamellar bone may be present Rare fine network of osteoid Necrosis Hemorrhage, but not tumor necrosis Tumor necrosis is highly suggestive of Telangiectatic osteosarcoma Case 4 Chronic (Granulomatous) Osteomyelitis, North American Blastomycosis Case 5 Langerhans Cell Histiocytosis (Histiocytosis X, Eosinophilic Granuloma) A radiograph showing a lytic intramedullary lesion with ill-defined borders and periosteal reaction raises the differential of a "marrow cell process", which includes chronic osteomyelitis, Langerhans cell histiocytosis, Ewing's sarcoma and metastasis. Although chronic osteomyelitis may occur at any age, hematogenously seeded osteomyelitis is most common in children and immunocompromised adults. Histiocytosis X and Ewing's are both most common before the age of 30. Metastases are more common in adults. Histologically Ewing's separates from the differential diagnosis on low power because of the mixed cell population, histiocytes, and inflammatory cells, while chronic osteomyelitis and Histiocytosis X may appear similar. The folded nucleus of the Langerhans cell and the bilobed nucleus of the eosinophils characterize Histiocytosis X. It is useful to note both the cytoplasmic staining and nuclear features of the segmented cells, as poorly stained frozen sections frequently do not show eosinophilic granules. Fragmented devitalized bone, displaying empty lacunae and a bluish dustiness to the edge, usually accompanied by neutrophils, is highly suggestive of osteomyelitis. In the absence of fracture eosinophilic granuloma does not usually display necrosis. Features differentiating Langerhans histiocytosis from chronic osteomyelitis Langerhans cell histiocytosis Chronic osteomyelitis Radiology Intramedullary ill defined lytic diaphyseal lesion. Frequently craniofacial bones Intramedullary ill defined lytic diaphyseal lesion Rarely effects craniofacial bones Peak Age <30 Hematogenously spread is more common in children Low power Mixed infiltrate of inflammatory and mononuclear cells Mixed infiltrate of inflammatory and mononuclear cells. Devitalized bone Cells Langerhans cells. Eosinophils, but sometimes neutrophils or plasma cells predominate. Neutrophils, lymphocytes and plasma cells. Sometimes prominent histiocytic infiltrate or granulomas Nucleus Grooved Langerhans nuclei, bilobed eosinophils Reniform macrophages Mitoses Not prominent Not prominent Giant cells May have giant cells composed of Langerhans cells, with similar nuclei Osteoclasts, or foreign body giant cells if chronic (fungal) or Langhans if TB Matrix No matrix unless fracture or reactive border Devitalized bone and remodeling may be present Necrosis With fracture Frequent Case 6: Metastatic Renal Cell Carcinoma In adults, metastatic lesions by far outnumber primary bone tumors and as a rule, any lytic lesion in a patient over 40 years old is a metastasis till proven otherwise. That being said, it is vital during the intraoperative consultation to differentiate a metastasis from a known or occult primary from a primary bone tumor. The usual therapeutic approach to a metastatic lesion that threatens to fracture is to rod, plate, or otherwise stabilize the bone, without attempt to keep the surgical field free of tumor. Primary lesions are candidates for limb salvage, and rendering an incorrect intraoperative diagnosis of a metastatic lesion may lead to subsequent amputation. One study found a local recurrence rate of 83% for erroneously curetted osteosarcoma cases that were subsequently treated by limb salvage, as opposed to 8% for cases that were not curetted. [6] Radiographically, a lytic or mixed lytic and blastic lesion may be either a primary or a metastatic lesion. With the exception of primary vascular lesions and fibrous dysplasia, most primary bone lesions are unifocal. However unifocal metastases are common, with nearly half of all osseous renal cell metastases presenting as solitary lesions. [7] Since metastases frequently present as fractures, secondary repair changes often complicate the histologic picture. The first specimen submitted for frozen is frequently periosteum with fracture callus or procallus, and may tempt a diagnosis of benign or malignant primary bone lesion. The majority of carcinomas are easily differentiated from a primary bone tumor, however 10% of renal cell carcinoma have areas of sarcomatoid differentiation, mimicking a high grade primary sarcoma of bone. Primary bone lesions may also mimic epithelial tumors, particularly in the case of epithelioid hemangioendothelioma or epithelioid angiosarcoma. Features of primary vs metastatic bone lesions Primary Metastasis Age All ages Most common over 40 Radiograph Lytic or blastic Lytic (lung, kidney) or blastic (prostate, breast) Purely lytic lesions may be silent on bone scan Cortical based lesion more commonly renal cell Multifocality Rare, usually in same limb Vascular tumors More common History of primary elsewhere Doesn't rule out a second primary in bone Approximately 30% have skeletal met as first sign [8] Site Appendicular skeleton most common, but anywhere Trunk, proximal appendicular, marrow bearing areas Presentation Pain, fracture Pain, fracture Case 7: Osteoblastic Osteosarcoma Case 8: Fracture Callus Complicating Intraosseous Cyst Many benign and malignant lesions present acutely as fractures. The differentiation of malignant bone formation from reactive bone on frozen section is particularly difficult since fracture callus can present with many of the elements of osteosarcoma: sheets of large osteoblasts, mitotic activity, woven bone, immature cartilage matrix, and cellular fibroblastic proliferation. Recent fractures of less than a week are unlikely to display osteoid matrix, however small subclinical fractures may have occurred before presentation. Maturation stages of fracture Days after fracture Feature < 3 days Hemorrhage, edema, tissue necrosis 3 to 7 days Reactive myofibroblasts, tissue culture 7 to 10 days Increasing cellularity, early wisps of osteoid >10 days Osteoid and cartilage matrix 2 to 3 weeks Osteoblast rimming, broad seams of osteoid Features of osteosarcoma and fracture on Frozen Section Osteosarcoma Fracture Radiology Evidence of osteoid formation, Codman's triangle Early; underlying lesion is prominent Late; callus has symmetrical expansion of contour Low power Sheet-like and disorganized Borders infiltrate and entrap pre-existing spicules of lamellar bone Zonal distribution of elements Gradual transitions between fibroblastic, osteoblastic and cartilaginous areas Borders tend to be sharper, with no infiltration and trapping of lamellar bone, but with remodeling of devitalized bone in fracture zone Osteoid Confluent, fine net or mesh-like osteoid, not organizing architecturally Tendency to form broader spicules of woven bone with linearity When mature, palisades of Roman arches Rimming Osteoblast rimming usually not seen in conventional osteosarcoma Osteoblast rimming prominent Atypia Osteoblasts have nuclear size and chromatin heterogeneity When woven bone trabeculae are present, the entrapped osteoblasts are atypical Osteoblasts are reactive with abundant cytoplasm, often with prominent Golgi apparatus, but not heterogenous or overtly atypical. Osteocytes within broader trabeculae of osteoid become blander Mitoses Frequent and atypical Frequent Stroma Stroma between trabeculae is cellular and has fine osteoid Stroma between trabeculae is vascular and represents a granulation tissue response Extraosseous extension Periosteal elevation, usually not trapping muscle or fat Elements injured by the fracture, such as muscle and fat, may be engulfed by the callus Case 9: Low Grade Central Osteosarcoma, Fibrous Dysplasia-like Low-grade central osteosarcoma may present with several histologies resembling benign bone lesions, including chondromyxoid fibroma and fibrous dysplasia. Low-grade central osteosarcoma is usually favored clinically over fibrous dysplasia by the aggressiveness of the radiographic appearance. Cortical erosion, involvement of the epiphysis, irregular borders and coarse trabeculation all favor low-grade central osteosarcoma. A clinical history of pain is also suggestive, as most cases of fibrous dysplasia present with deformity without chronic pain. Histologically the lesions can appear very similar, but the recognition of long spicules of mixed lamellar and woven bone favors OS, even in the absence of overt anaplasia. If suspected, it is best to request additional diagnostic material for permanents and defer. Features of low cellularity osteoid producing lesions Low grade central osteosarcoma Fibrous dysplasia Age 20 to 40 peak age Usually < 30 Location Long bone lower extremity metaphysis Any bone, metaphyseal Symptom Pain Swelling deformity Radiograph Lytic medullary lesion, ground glass, may destroy cortex but usually no periosteal reaction. If coarsely trabeculated, think low grade central OS Lytic medullary lesion, lucent or ground glass, but more uniform, usually better circumscribed. Does not involve epiphyseal plate May be multifocal Stroma Low cellularity, may be fibromatosis-like. May coalesce around spicules Low cellularity Osteoid Mixed woven and lamellar bone. Characteristic long parallel streamers, as in parosteal OS Woven bone only, short trabeculae are more curled Osteoblast rimming Usually can find easily Only focal Atypia May be subtle to impossible to see on FS Absent Mitoses Rare Rare Case 10: Enchondroma Differentiating enchondroma from low-grade chondrosarcoma is a difficult differential on frozen section, and should be approached with full knowledge of clinical history, symptoms, radiology, site, and surgical intraoperative impression. Several caveats should be heeded: Asymptomatic lesions are usually benign Lesions of the hand and feet, although they may be cellular and atypical enough to warrant a diagnosis of Grade II chondrosarcoma in other sites, are benign until proven otherwise Small lesions and lesions in patients less than 30 y/o are almost always benign Myxoid stroma is more likely malignant Necrotic cartilage in the absence of a fracture is probably malignant Infiltrative growth pattern is diagnostic of malignancy Features of enchondroma and low grade chondrosarcoma Enchondroma Low Grade Chondrosarcoma Age Teens to elderly Most over 50 Symptoms Usually asymptomatic, incidentally discovered. If painful, usually associated with fracture Progressive pain, night pain, mass Distribution Hands and feet, femur, humerus, tibia Pelvis, femur, humerus, ribs, scapula Location Diaphysis, metaphysis, uncommon epiphyseal Same Size Usually 6 cm or less Larger Low power Islands of cartilage with ossification to lamellar bone at rim, no trapped lamellar bone Infiltrative pattern, trapped native lamellar bone, infiltration of Haversion canal, Lobules separated by fibrous bands Cellularity Low cellularity, Fewer than 25 cells per 400X hpf, likely benign Variable, but if greater than 100 cells per hpf (400x) then likely malignant Mitoses < 1/50 hpf More than 2/50 hpf Radiology Sharply defined radiolucency with cloud-like calcification. Without erosion of cortex and without periosteal extension Less well defined, may expand contour of bone, scallop the cortex, and show periosteal reaction Matrix Hyaline, not myxoid May be myxoid Cytology Small nuclei Atypia, prominent nucleoli Case 11: Granulosa Cell Tumor Case 12: Undifferentiated Ovarian Carcinoma Granulosa cell tumors are typically unilateral and although they may occur in all ages, are more common in adults. The tumors may be large, and occasionally present as an acute abdomen due to ovarian torsion, rupture, or hemoperitoneum. 90% of cases present as stage-one lesions; standard management is to do unilateral oophorectomy and inspect peritoneal surfaces for implants. Particularly for women in the reproductive years, the misdiagnosis of granulosa cell tumor as an undifferentiated carcinoma can lead to inappropriate surgery. Granulosa cells can be recognized by the scant cytoplasm, relatively uniform nuclear size and shape, and lack of well-organized epithelial structures. Cordlike patterns or Call Exner bodies may, particularly in poorly executed frozen sections, lead to consideration of a poorly differentiated carcinoma. Touch preps can be particularly helpful in evaluating a granulosa cell tumor, as nuclear grooves and other cytologic features are much more readily seen in a touch preparation than on frozen section. Undifferentiated carcinoma of the ovary includes tumors with little or no differentiation towards the traditional histologic subtypes. Patients tend to be older, with average age in the mid-fifties. Approximately 50% of cases have bilateral ovarian involvement, and the majority present at an advanced stage. Microscopically, undifferentiated carcinoma may present sheets of cells with scant cytoplasm and deceptively uniform hyperchromatic nuclei, superficially resembling a granulosa cell tumor. Abortive gland formation may be mistaken for sex cords or Cal Exner bodies. Pleomorphism, increased mitotic activity atypical mitoses, and tumor necrosis not associated with torsion all suggest undifferentiated carcinoma as opposed to granulosa cell tumor. Cytologic examination is particularly useful to compare nuclear detail. The clinical history will generally separate these two diagnoses based on age, bilaterality and stage at presentation: a granulosis cell tumor presenting at an advanced stage in an older woman should be held suspect as a potential undifferentiated carcinoma. Features of Granulosa cell tumor vs undifferentiated carcinoma Granulosa cell tumor Undifferentiated carcinoma Age Peak 45-55 Peak 55-60 Stage 90% stage I 90% stage III or greater Bilaterality 5% 50% Gross Predominantly solid Predominantly solid Low power Solid sheets, cords, usually no necrosis if no torsion Solid sheets, poorly formed papillae or glands, necrosis High power Usually uniform nuclei, smaller, scant cytoplasm, grooves better seen on touch preps Pleomorphic, large nuclei scant cytoplasm Mitoses Occasional, rarely atypical Frequent and atypical Case 13: Luteoma of Pregnancy Ovarian lesions presenting in pregnancy reflect the tumors characteristic of the age group and the special functional tumor-like lesions that arise during pregnancy. The incidence of adnexal masses during pregnancy is variable, dependent of the use of imaging and the inclusion in studies of functional lesions that involute without being biopsied. Retrospective studies have found adnexal masses in 1/632 pregnancies [9] and 1/1000 pregnancies. [10] The majority of benign lesions are mature cystic teratomas (27% to50%), cystadenoma (20% to 34%) and functional lesions (13% to 18%). Of malignant or borderline tumors, germ cell tumors (30-45%), borderline tumor (30% to 35%) cystadenocarcinoma (5% to 10%) and sex cord stromal tumors (10-20%) predominate. Torsion with and without rupture is a common presentation of adnexal tumors during pregnancy. The differential of a solid eosinophilic lesion presenting in pregnancy includes luteoma of pregnancy, corpus luteum of pregnancy, stromal hyperthecosis, stromal hyperplasia, steroid (lipid) cell tumor, Leydig cell tumor, luteinized granulosa cell tumor, oxyphilic variant of clear cell carcinoma and metastatic carcinoma with marked stromal luteinization. Clear cell carcinoma is rare before the age of 40, and has cystic and papillary structures. Metastatic carcinoma typically has stromal reaction and epithelial islands in addition to the luteinized stoma cells. In terms of intraoperative management, the separation of functional lesions from tumor is most important. Functional lesions such as luteoma and luteal cysts require no surgery beyond wedge excision, and will involute after pregnancy. Differential of ovarian tumors in pregnancy Luteoma Steroid cell tumor Granulosa cell tumor Age 20's to 30's 40's to 50's 20's to 70's Race 80% in African Americans No tendency No tendency Laterality Multifocal or bilateral Unilateral Unilateral Symptoms No symptoms or torsion Most are virilizing Hyperestrogenic Histology Uniformly pink cells, rarely lipid vacuoles Pink and clear cells, lipid Solid, cords, trabeculae, may have vacuoles Cytology Uniform nuclei, small nucleoli Variable nuclei, small nucleoli Nucleoli, but frequently grooves Case 14: Endometrial Carcinoma, Grade 3, Stage II B or Greater The diagnosis of endometrial carcinoma is typically made on biopsy or curetting prior to surgery. Intraoperative evaluation of hysterectomy specimens is done to document the extent of disease and tumor grade so that a decision regarding pelvic lymph node dissection may be made. Features indicating a high risk of metastasis are shown below. If no risk factors are present, there is a 2.8% risk of pelvic and <1% risk of paraaortic node involvement [11, 12, 13] Prognostically significant features for endometrial carcinoma Feature Relative risk increase of lymph node metastasis Serous or Clear cell histology 3X Grade 3 histology 3X Outer half myometrial involvement 5X Cervical involvement 4X Adnexal involvement 4X Missing significant prognostic factors during the intraoperative examination leads to the omission of pelvic node dissection, which usually leads to pelvic radiation therapy. Over diagnosis of bad prognostic factors leads to unnecessary pelvic node dissection and increased potential for morbidity (up to 15%). The serosa and adnexa of the hysterectomy specimen are inspected grossly for extrauterine involvement, which is rarely seen. The corpus is bivalved at 3 and 9 o'clock and the lining inspected. Bread-loafing cuts at 5 mm intervals are then made to identify areas of possible myometrial invasion. The deepest apparent focus is then frozen, ideally as a full thickness, or in a manner to preserve the overall depth and total thickness measurements. Any mucosal abnormality in the lower uterine segment and endocervical canal should also be frozen. Accuracy of intraoperative assessment of grade is between 80% and 96%, with under-grading the common problem, [14, 15] and usually due to sampling error. Gross assessment of myometrial invasion alone is difficult, particularly with high-grade tumors, and is overall probably no better than 70% accuracy. [16] Most discrepancies in depth involve superficial lesions. Case 15: Mucinous Cystadenocarcinoma Ovarian masses are rarely removed with a previous study confirming malignancy; a solid and cystic ultrasound appearance on ultrasound suggests malignancy but is correct only 25% of the time, and only a quarter of stage I ovarian tumors have elevated serum markers. Apparently localized lesions and other sites of involvement are found on staging in a quarter of cases. [14] Under-diagnosis leads to under-staging, while over-diagnosis of malignancy may result in inappropriate loss of fertility. Mucinous ovarian tumors present two clinical problems: are they benign, borderline or malignant; and are they primary or metastatic? The evaluation of mucinous tumors is the largest source of discrepancy in intraoperative examination of gynecologic specimens. [17, 18] Lesions with one or few smooth walled dominant cysts and no sold areas on gross examination prove to be benign more than 95% of the time, [19] and do not require frozen section examination. Multiloculated and solid and cystic mucinous lesions require frozen section, however gross inspection rarely indicates where to find the focal area of borderline or malignant histology. Although adequate sampling for permanent sections requires one section per centimeter of the largest diameter of the tumor, it is impractical to do more than one or two frozen sections. The predictive value of the frozen section diagnosis is 95% for benign, 99% for malignant, and 65% for borderline tumors. [20] Approximately a fourth of frozen section diagnoses of borderline tumor will have carcinoma in the final pathologic sections. For this reason a diagnosis of "at least borderline" is sometimes warranted. Metastases to the ovary from the intestinal tract may mimic a primary ovarian mucinous carcinoma or borderline tumor. Useful clinical and histologic features have been described recently by Lee and Young [21]. Features of primary and metastatic mucinous tumors Primary Metastatic P value Laterality Unilateral 95% Bilateral 60% to 75% <.0001 Microscopic surface involvement Absent 79% <.0001 Nodular growth pattern Absent 42% <.0003 Infiltrative invasive pattern 16% 91% <.0001 Small glands/ tubules 12% 94% <.0001 Expansile invasive pattern 88% 18% <.0001 Complex papillae 60 8 .0004 Benign appearing areas 76 36 .008 Borderline with atypia 57 31 .035 Laterality, growth pattern, gland size and complex papillae are all highly predictive; areas of benign or borderline histology are not. The status of the other ovary and previous history of intestinal carcinoma should be established at the time of the intraoperative consultation. All patients with bilateral tumors or suspicious histology should have the gastrointestinal tract evaluated during the laparotomy. References: Wright JR, The development of the frozen section technique, the evolution of the surgical biopsy, and the origins of surgical pathology. Bull Hist Med, 1985. 59: p. 295-326. Carter D, Surgical Patholgoy at Johns Hopkins, in The history of American surgical pathology, Rosai J, Editor. 1997, American Registry of Pathology: Washington D.C. Novis DA, Z.R., Interstitutional comparison of frozen section turnaround time. A College of American Pathologists Q probe study of 32868 frozen sections in 700 hospitals. Arch Pathol Lab Med, 1997. 121(6): p. 559-67. Ferreiro JA, M.J., Bostwick DG, Accuracy of frozen section diagnosis in surgical pathology: review of a 1 year experience with 24,880 cases at Mayo clinic Rochester. Mayo Clin Proc, 1995. 70(12): p. 1222-3. Novis DA, G.G., Zarbo RJ; College of American Pathologists, Interinstitutional comparisons of frozen section consujltation in small hospitals: a College of American Pathologists Q-probes study of 18,532 frozen section consultation diagnoses in 233 small hospitals. Arch Pathol Lab Med, 1996. 120(12): p. 1087-93. Ayerza M, M.L., Aponte-Tinao L, Farfalli G. Misdiagnosis in osteoarcoma: effect of erroneous surgical procedure on recurrence and survival. in Connective Tissue Onocology Society. 2004. Montreal. Smith EM, K.E., Makley J, Resnick MI, Treatment of osseous metastases secondary to renal cell carcinoma. J Urol, 1992. 148: p. 784-787. Katagiri KH, t.M., Inagaki J, Sugiura H, Ito S, Iwata H, Determining the site of the primary cancer in patients with skeletal metastasis of unknown origin: a retrospective study. Cancer, 1999. 86(3): p. 533-7. Sherard GB, H.C., Willliams HJ, Semer DA, Hadi HA, Tait DL, Adnexal masses and pregnancy: a 12 year experience. Am J Obstet Gynecol, 2003. 189(2): p. 358-362. Hermans RHM, F.D., van der Putten HWHM, Einzmann T, Vos MC, Kieback DG, Adnexal masses in pregnancy. Onkologie, 2003. 26: p. 167-172. Mikuta JJ, Preoperative evaluation and staging of endometrial cancer. Cancer, 1995. 76: p. 2041-2043. Fanning J, T.Y., Piver MS, Intraoperative frozen section diagnosis of depth of myometrial invasion in endometrial cance. Gynecol Oncol, 1990. 37: p. 29-33. Creasman WT, M.C., Bundy BN, et al, Surgical pathologic spread patterns of endometrial cancer. A Gynecologic Oncology Group Study. Cancer, 1987. 60: p. 2035-2041. Acs G, Intraoperative consultation in gynecologic pathology. Sem Diag Pathol, 2002. 19(4): p. 237-254. Quinlivan JA, P.W., Nicklin JL, Accuracy of frozen section for the operative management of endometrial cnacer. Bjog, . 108(798): p. 798-803. Goff BA, R.L., Assessment of depth of myometrial invasion in endometrial adenocarcinoma. Gynecol Oncol, 1990. 38: p. 46-48. Boriboonhirunsarn D, S.A., Accuracy of frozen section in the diagnosis of malignant ovarian tumor. J Obstet Gynaecol Res, 2004. 30(5): p. 394-399. Twaalfhoven FC, P.A., Trimbos JB, et al, The accuracy of frozen section diagnosis of ovarian tumors. Gynecol Oncol, 1991. 41: p. 189-192. Lim FK, Y.C., Chong SM, Pre and intraoperative diagnosis of ovarian tumors: how accurate are we? Aust N Z J Obstet Gynaecol, 1997. 37: p. 223-227. Rose PG, R.R., Nelson BE, Hunter RE, Reale FR, Accuracy of frozen section (intraoperative consultation) diagnosis of ovarian tumors. Am J Obstet Gynecol, 1994. 171(3): p. 823-6. Lee KR, Young R., The distinction between primary and metastatic mucinous carcinomas of the ovary. Am J Surg Pathol, 2003. 27(3): p. 281-292.