March 2006
ANATOMIC PATHOLOGY
Multifocal and multicentric breast cancer: Does each focus matter?
Coombs NJ, Boyages J.
The authors conducted a study to see if multiple tumors in the breast are associated with increased nodal involvement when compared with similar staged unifocal disease. Their study compared two methods of tumor size assessment to predict tumor behavior in the relationship between size and axillary node involvement for patients with multifocal and multicentric breast cancer. The authors examined the histologic reports of every patient with multifocal breast cancer treated in New South Wales between April 1995 and September 1995. Tumors were assessed using the largest tumor focus diameter and the aggregate diameters of all tumor foci. The dimensions were compared with unifocal tumors and against nodal positivity. The authors found that 94 (11.1 percent) of 848 women had multifocal breast cancer, and of these, 49 women (52.1 percent) had axillary node involvement, compared with 37.5 percent with unifocal breast cancer (P=0.007). The use of aggregate dimension reclassified significant numbers of multifocal tumors at a more advanced stage. Using this method to stage cancers, rather than largest tumor size, removed the excess node positivity when compared with unifocal, stage-matched breast carcinomas. The authors concluded that the tendency of breast tumors to metastasize is a reflection of the total tumor load. Failure to measure the additional tumor burden provided by multiple small foci may understage a woman's disease. This may deny patients the opportunity of adjuvant therapies if doctors ignore the contribution of the smaller foci to the incidence of node positivity and survival.
J Clin Oncol. 2005 Oct 20;23(30):7497-502
Spindle Cell (Sarcomatoid)
Carcinoma of the Breast: A Clinicopathologic and Immunohistochemical Analysis
of 29 Cases
Carter MR, Hornick JL, Lester S, Fletcher CD.
Spindle cell (sarcomatoid) carcinoma of the breast is a rare variant of breast cancer that has been classified under the broad rubric of metaplastic carcinoma. Because the term "metaplastic carcinoma" comprises a heterogeneous group of tumors, it has been difficult to reliably predict biologic potential or to determine optimal therapy. To better characterize the spindle cell subset of metaplastic breast carcinomas, we reviewed 29 cases. All patients were adult females ranging from 40 to 96 years of age (median, 68 years). Tumor size ranged from 1.5 to 15 cm (median, 4 cm). Treatment was by excision and/or mastectomy with axillary node evaluation in most cases, often combined with postoperative radiation and/or chemotherapy. All cases were clinically of breast origin, showed >/=80% spindled/sarcomatoid morphology, and demonstrated keratin positivity and/or close association with ductal carcinoma in situ. Immunohistochemical studies showed evidence suggesting myoepithelial differentiation as exhibited by immunoreactivity for smooth muscle actin, cytokeratin 14, and p63 in a subset of cases (39%). Twenty-seven cases exhibited pure spindled or sarcomatoid morphology of variable appearance and nuclear grade, whereas 2 contained high-grade invasive ductal carcinoma comprising </=20% of the tumor mass. Two cases exhibited heterologous elements (1 rhabdomyosarcoma and 1 with both chondrosarcoma and osteosarcoma) and 4 were associated with ductal carcinoma in situ. Follow-up data were available on 24 of 29 patients (range, 1-120 months; median, 20 months). Of 20 cases in which axillary nodes were biopsied, definitive nodal metastases were identified in only 1 (5%), and this was in a case with a significant component of invasive ductal carcinoma. Three patients developed local recurrences. Extranodal metastases occurred in 11 of 24 patients (46%), most commonly to the lungs. Ten of 24 patients (42%) died of disease at a median interval of 11.5 months (range, 1-46 months) and 3 patients were alive with metastatic disease. Eight patients were alive with no evidence of recurrent or metastatic disease (median, 29.5 months). Based on this series, spindle cell/sarcomatoid carcinoma of the breast is a highly aggressive neoplasm with a high rate of extranodal metastases. Purely spindled/sarcomatoid tumors have a significantly lower rate of nodal metastases than conventional ductal and lobular breast carcinomas.
Am J Surg Pathol. 2006 Mar;30(3):300-309.
The use of routine special
stains for upper gastrointestinal biopsies
Wright CL, Kelly JK.
Helicobacter pylori and intestinal metaplasia (IM) are readily seen in hematoxylin and eosin-stained slides of gastric and/or esophageal biopsies, yet many pathology laboratories perform routine special stains on all of these biopsies. Authors wished to determine if special stains are necessary for every single gastric and/or esophageal biopsy. We prospectively studied 613 gastric and/or esophageal biopsies from 494 consecutive patients. The slides were stained with hematoxylin and eosin, toluidine blue (TB) for H. pylori, and Alcian blue (AB) for IM. The hematoxylin and eosin slide was classed as positive or negative for H. pylori and IM. Then it was determined if the case needed a TB or AB stain. A total of 436 cases (71.1%) were identified as H. pylori-negative and not needing a TB stain, and none was TB+. A total of 126 (20.6%) of hematoxylin and eosin slides were inconclusive for H. pylori and were regarded as needing a TB stain. Twenty of these (15.9%) were TB+. Fifty-one biopsies (8.3%) were regarded as H. pylori+ on hematoxylin and eosin; the TB stain was also positive in 49. IM was present in 113 (18.4%) hematoxylin and eosin biopsies. Hematoxylin and eosin slides were IM-negative in 498 cases (81.2%). The AB stain revealed rare goblet cells in 3 of 498 cases (0.6%). Only one of those biopsies was esophageal, and that had one goblet cell that was missed on hematoxylin and eosin. Only 2 (0.3%) were regarded as needing an AB stain. Authors conclude that routine special stains for all gastric and/or esophageal biopsies are not required, and hematoxylin and eosin assessment combined with selective ordering of these stains will identify virtually all cases of H. pylori gastritis and intestinal metaplasia.
Am J Surg Pathol. 2006 Mar;30(3):357-61.
Immunohistochemical detection of
WT1 protein in a variety of cancer cells
Nakatsuka SI, Oji Y, Horiuchi T et al
WT1 was first identified as a tumor suppressor involved in the development of Wilms' tumor. Recently, oncogenic properties of WT1 have been demonstrated in various hematological malignancies and solid tumors. Because WT1 has been identified as a molecular target for cancer immunotherapy, immunohistochemical detection of WT1 in tumor cells has become an essential part of routine practice. In the present study, the expression of WT1 was examined in 494 cases of human cancers, including tumors of the gastrointestinal and pancreatobiliary system, urinary tract, male and female genital organs, breast, lung, brain, skin, soft tissues and bone by immunohistochemistry using polyclonal (C-19) and monoclonal (6F-H2) antibodies against WT1 protein. Staining for C-19 and 6F-H2 was found in 35-100 and 5-88% of the cases of each kind of tumor, respectively. WT1-positive tumors included tumor of the stomach, prostate, and biliary and urinary systems, and malignant melanomas. A majority of the positive cases showed diffuse or granular staining in the cytoplasm, whereas ovarian tumors and desmoplastic small round cell tumors frequently showed nuclear staining. Glioblastomas, some of soft tissue sarcomas, osteosarcomas, and malignant melanomas of the skin showed extremely strong cytoplasmic staining as compared with other tumors. Western blot analysis showed that WT1 protein was predominantly expressed in the cytoplasm of the tumor cells in two cases of lung adenocarcinoma, supporting the intracytoplasmic staining for WT1 using immunohistochemistry. Immunohistochemical detection with routinely processed histologic sections could provide meaningful information on the expression of WT1 in cancer cells.
Mod Pathol. 2006 Mar 17; [Epub ahead of print]
In stent restenosis: bane of the stent era
[REVIEW]
DEFINITION AND CLASSIFICATION OF IN STENT RESTENOSIS
ISR can be defined clinically or angiographically. Clinically it is defined as the presentation of recurrent angina or objective evidence of myocardial ischaemia, whereas angiographic ISR is the presence of >50% diameter stenosis in the stented segment. Traditionally, ISR has been classified based on the length of the lesion, as focal (<10 mm) or diffuse (>10 mm).
Angiographic classification appears to be more complete and classifies the ISR lesions into six groups (I–IV) according to the pattern and extent of restenosis with relation to the affected vessel. According to this classification, focal pattern I can be further subdivided into IA–ID and diffuse pattern can be divided into II–IV.
The long term outcome of stent implantation is affected by a process called in stent restenosis (ISR). Multiple contributory factors have been identified, but clear understanding of the overall underlying mechanism remains an enigma. ISR progresses through several different phases and involves numerous cellular and molecular constituents. Platelets and macrophages play a central role via vascular smooth muscle cell migration and proliferation in the intima to produce neointimal hyperplasia, which is pathognomic of ISR. Increased extracellular matrix formation appears to form the bulk of the neointimal hyperplasia tissue. Emerging evidence of the role of inflammatory cytokines and suppressors of cytokine signalling make this an exciting and novel field of antirestenosis research. Activation of Akt pathway triggered by mechanical stretch may also be a contributory factor to ISR formation. Prevention of ISR appears to be a multipronged attack as no therapeutic “magic bullet” exists to block all the processes in one go.
Journal Of Clinical Pathology Volume 59(3), March 2006, pp 232-239
CYTOPATHOLOGY
Body Cavity Fluid Cytology
[Editorial]
The pleural, peritoneal, and pericardial serous cavities, usually referred to as body cavities, are lined by a single layer of mesothelial cells and, in the healthy person, contain very little fluid. Serous effusions occur when fluid, either as a transudate or exudate, accumulates in the cavities. Transudates have a low protein content and specific gravity and are mostly caused by cirrhosis or congestive cardiac failure. Exudates, in contrast, have a high protein content and specific gravity. They have a broader etiology and include inflammatory, neoplastic, connective tissue, and iatrogenic disorders. Serous effusions can occur at any age and may be the initial presentation of the disorder or a later manifestation. Most serous cavity-fluid specimens are effusions, but some are obtained by using a saline wash to exfoliate cells from the mesothelium.
Cytologic examination of the cellular features of fluids is a valuable adjunct to patient diagnosis and the staging and management of tumors, in particular, gynecologic tumors. The German-language literature contains the earliest references to the cytology of malignant cells in fluid specimens. Lücke and Klebs in 1867 may have been the first to make reference to their appearance in an ascitic fluid, but Quincke in 1882 is credited with the first detailed description of ovarian and pulmonary malignancies in fluid samples. Almost a hundred years later, Keettel and Elkins from Iowa conceived the idea of washing the peritoneal cavity with normal saline to examine the spread of ovarian cancer and in 1956 published their results.
Today, body-cavity fluid cytology is a routine diagnostic tool. Preparation of the specimen has evolved from unstained wet smears to protocols that generally include centrifugation and the generation of stained smears and a cell block. The smears may be alcohol-fixed direct smears, cytospins, or a liquid-based preparation, and they are usually stained with the Papanicolaou method. The cell block is processed as a tissue specimen and usually stained with hematoxylin and eosin. The fluid is also suitable for ancillary testing using histochemistry, immunohistochemistry, electron microscopy, flow cytometry, and molecular techniques. The cell block is the optimal specimen for immunohistochemical testing.
The 6 cases presented in this issue taken separately represent common diagnostic problems in effusion cytology and illustrate the importance of ancillary testing, as well as clinicopathologic correlation. Taken as a whole, they represent a comprehensive review of effusion cytology.
Drs. Bhatti and Tabbara begin our issue with a case presentation of malignant mesothelioma and address one of the fundamental diagnostic problems in effusion cytology: the differentiation of reactive mesothelial effusions, mesothelioma, and metastatic disease. They present a comprehensive discussion of the cytomorphologic features and histochemical and immunohistochemical stains that aid in the diagnosis.
The identification of metastatic disease in a body-cavity fluid may be the first indication of an underlying malignancy. Drs. Filie and Jones extend discussions from the first case to present the workup of malignant effusion from an unknown primary site. Although clinical statistics suggest a differential diagnosis based on gender and location of the malignancy (pleural, pericardial, or pelvic cavity), immunohistochemistry still represents the best ancillary testing available to confirm a primary site.
Lymphocytic effusions are fairly common, although their etiologies may range the gamut from reactive and infectious to lymphomas. Primary effusion lymphomas (PEL) are rare and typically associated with HIV infections. Dr. Caraway presents a case of PEL and focuses on its distinctive immunophenotypic and molecular features. The differential diagnosis of lymphomatous effusions rounds out her discussion.
Dr. Boerner uses a straightforward case as a platform for an extensive discussion of mimicry in effusions. He addresses 2 major areas of concern: false negatives, in which malignancies may resemble mesothelial cells or may blend into the background and be overlooked; and false positives, in which mesothelial cells may resemble adenocarcinoma or the “benign interlopers,” cells that are occasionally seen in effusions that cause overreaction.
Dr. Jain discusses the role of pelvic washings in the clinical management of gynecologic tumors. Normal peritoneal cytology, as well as endosalpingiosis, is reviewed and compared with the cytomorphology of the various gynecologic tumors that may spread to the peritoneal surface. The utility of immunohistochemical staining is also reviewed.
The role of molecular techniques in cytopathology is increasing and evolving almost exponentially. In our last case presentation, Dr. Demetrick explores the role of forensic DNA identification as a method for troubleshooting rare cases where concerns of contamination or misidentification may have profound clinical consequences. Molecular diagnostic techniques are highly sensitive and specific and are often readily adaptable to cytologic specimens. Their efficacy is proven and their once-exorbitant costs are becoming cost-effective.
Authors sincerely thank the contributors for their valuable and insightful contributions. Authors hope that this series of cases has provided not only a comprehensive review of effusion cytology but has also addressed many of the issues, challenges, and difficulties we all encounter in our own practice.
Pathology Case Reviews Volume 11(2), March/April 2006, pp 65-66
Mimicry and Pitfalls in Effusion Cytology
[Case Review]
Effusion cytology is one of the most common and yet challenging of nongynecologic cytology specimens. The difficulty in interpretation of the cytomorphology of the cells in effusion samples results from extensive mimicry. Benign conditions may mimic malignancy through introduction of unanticipated interlopers or through mesothelial cell alterations masquerading as malignant cells. Alternatively, malignant tumors may evade detection in effusions by subterfuge and adopting a deceptively benign appearance mimicking mesothelial cells or by stealth, allowing them to blend into their background. Despite their mastery of disguise, both the benign and malignant conditions leave behind morphologic clues to their true identity that may be unearthed through diligent examination and deductive reasoning.
It has been said that if you have not undercalled and overcalled effusion samples, you have simply not seen enough. As much as we might like to ignore this statement, it is a truism. At times we seem unwilling to reveal our discomfort with effusions. Even though we may not openly discuss our dilemma, it refuses to remain hidden and manifests itself in our literature, where there is a virtual cornucopia of articles on ancillary techniques that have been and will continue to be proposed to provide the magic bullet when perplexed by effusion cytomorphology. Our undercall/overcall mantra implies that experience might ease the pain of an effusion sample. Inexperience can contribute to the problem, but it is not the sole source. Effusion samples are immensely common, and after a short time most pathologists have seen a large number of effusion cases. However, we must acknowledge that effusion cytology still challenges and, much to our chagrin, humbles even the most experienced of pathologists. Why, then, is it so difficult? The answer is the morphologic mimicry that occurs in effusion samples and the diagnostic pitfalls it generates.
Mimicry is common in all pathologic samples (histologic as well as cytologic) and is a well-recognized source of diagnostic errors. Yet it is in effusion cytology where mimicry reaches a zenith. But why is mimicry so commonplace in effusion cytology? The major reason for the mimicry comes from the physics of living in a fluid environment, something akin to the process of convergent evolution. In convergent evolution, the environment drives morphologic features towards similar ends that are most suited for that environment. Thus, the ichthyosaur, an ancient marine reptile from the Jurassic period, and the modern-day dolphin share a large number of morphologic similarities suited for their marine life, despite the fact that each animal is very different. A similar event occurs when cells are forced to live their aquatic life following desquamation into an effusion. Thus, mesothelial cells and any “foreign” cells begin to resemble one another due to the forces of physics in their fluid world. But, just as the ichthyosaur is readily recognized by its sagittally oriented caudal fin, which differs from the coronally positioned fin seen on the dolphin, differences in morphology persist that can be exploited to allow separation of the different pathologic entities despite their superficial similarities engendered by their environment. The key then is to learn what features may separate the true bill from the mimics.
CASE REPORT
A 64-year-old G2P2 woman presented for investigation of worsening stress and urge incontinence. As part of the investigation, an abdominal/pelvic ultrasound was performed, revealing ascites. A serum CA-125 was found to be elevated, at 44 U/mL (normal <34 U/mL). Upon referral for investigation of the ascites, the patient acknowledged a recent unintentional weight loss, a diminished appetite, and some abdominal pain with a feeling that “something was moving” within her abdomen when lying prone. The patient's past medical history included type II diabetes controlled by oral hypoglycemic agents and hypercholesterolemia but no other significant illnesses. Her gynecologic history was unremarkable, with normal menses, 2 pregnancies with vaginal deliveries, and menopause 9 years prior to presentation. An abdominal/pelvic CT scan demonstrated a small amount of particulate ascites with a 1-cm anterior-wall uterine mass, mesenteric stranding, and mild peritoneal enhancement but no other anatomic abnormalities. CA-125 on repeating testing had increased to 58 U/mL, but it was elected to follow the patient for 2 months. The follow-up CT scan was unchanged, with persistence of a small volume of ascites and the CA-125 that had increased to 83 U/mL. A diagnostic, ultrasound-guided paracentesis recovered 40 mL of clear yellow fluid.
Cytologic evaluation of the ascitic fluid showed a uniform and monomorphic population of dispersed single cells, with an appearance suggesting reactive mesothelial cells with scattered histiocytes and lymphocytes in the background). Cell clusters were infrequent and, when present, typically consisted of no more than 2 or 3 cells and appeared unidimensional. The nuclei were centrally or occasionally slightly eccentrically placed, with round contours and smooth nuclear membranes. The chromatin pattern was fine, with a single, small nucleolus. Mitoses were identified at a rate of approximately 1 mitoses per 700 cells, with scattered cells undergoing apoptosis. Careful examination uncovered numerous cells with cytoplasmic vacuoles. These vacuoles demonstrated internal structure, with microglobules that were typically cyanophilic but occasionally eosinophilic. In some cells, the vacuoles were sufficiently large to displace and then indent the nucleus, imparting a signet ring appearance to the cells. Although low-power examination had suggested a single monomorphic population, high-power examination demonstrated 2 cell populations that greatly resembled one another. The first type of cell was a reactive mesothelial cell, which differed subtly from an abnormal cell population characterized by an increased nuclear/cytoplasmic ratio with irregular nuclear contours but relatively bland chromatin. On the basis of the cytomorphology, a diagnosis of adenocarcinoma was made, with the suggestion of a possible lobular carcinoma of breast as the primary. Subsequent staining confirmed the presence of intracytoplasmic mucin on PAS/D and mucicarmine stains. Immunoperoxidase studies showed expression of pankeratins and estrogen receptor by the tumor cells, with no staining for progesterone receptor or calretinin. Follow-up clinical and imaging examinations demonstrated left axillary lymphadenopathy with diffuse bony metastases. Other than inversion of the left nipple, neither clinical examination or imaging studies were able to identify a breast lesion and no other potential primaries could be found. Further biopsies were not performed and the patient was treated with aromatase inhibitor, Letrozole. Eight months following the initial cytologic diagnosis, a bone marrow biopsy revealed metastatic carcinoma, compatible with a breast primary.
DISCUSSION
When presented in this manner, the case may not seem like much of a diagnostic challenge as the diagnostic features have been carefully illustrated in the figures provided. However, what cannot be demonstrated is the difficulty in finding these features on routine review of a case. This is a case where a malignant tumor's mimicry of mesothelial cells may falsely lead to the interpretation of a benign effusion. But this is actually only one form of mimicry that occurs in effusion cytology. Excluding contamination of the sample in the laboratory, mimicry in effusion samples and associated errors can be grouped into 5 categories.
1. Benign interlopers mimicking malignant cells (overcall, false-positive errors)
2. Mesothelial cells mimicking malignant cells (overcall, false-positive errors)
3. Malignant cells mimicking mesothelial cells (undercall, false-negative errors)
4. Stealth malignancies mimicking background cells (undercall, false-negative errors)
5. One type of malignancy mimicking another (classification errors)
Benign Interlopers
The benign interlopers cause problems by mimicking the “foreign” cell population of classic cytology teaching. In other words, they are present when they “are not supposed to be there” or because the cells look “weird” and thus foreign. Classic cytology teaching emphasizes the identification of the foreign cell population to aid in the diagnosis of malignancy in a fluid. The cells that “are not supposed to be there” are cells readily recognized as foreign to effusions but can become a source of overinterpretation if faulty logic is followed. The faulty logic assumes that the mere presence of foreign cells, and in particular epithelial cells, in a fluid sample proves malignancy. This logic is incorrect, as the benign interlopers prove. It is only when the foreign cells are malignant that there is proof of a malignant effusion. The benign interlopers can be grouped as:
1. Procedural (needle) pickups
2. Müllerian epithelium
3. Shedding of cells that are rarely seen in effusions
The inadvertent pickup of tissue by the centesis needle at the time of collection of effusions samples is usually readily recognized. The tissues that may be inadvertently collected include portions of epidermis or skin adnexal structures during penetration through the skin, as well as tissue fragments of skeletal muscle or internal organs such as liver or intestinal epithelium. Careful examination of the tissue fragment architecture will reveal cohesive, well-ordered epithelial structures where the nuclear features of malignancy will be lacking. Fortunately, procedural pickups are uncommon and not a significant cause of concern in most effusion cases.
The identification of Müllerian epithelium, in the form of endosalpingiosis and endometriosis, is an infrequent event in spontaneous effusions but presents more of a diagnostic challenge than procedural pickups. Endometriosis is a well-established source of pleural, peritoneal, and rarely pericardial effusions. In many cases, the cytologic examination of effusions with endometriosis will be nondiagnostic, but occasionally, as with endosalpingiosis, the pathologist is confronted by glandular epithelium, often somewhat atypical, in a fluid sample. Remembering that the mere occurrence of epithelium in a fluid sample does not prove malignancy is a long step towards avoidance of a significant error. Only when the cells are cytologically malignant is there carcinoma in the fluid. In fact, the lack of overt malignant features by the Müllerian epithelium is frequently appreciated, and the presence of this foreign but mildly atypical population will many times trigger a diagnosis of “suspicious for carcinoma” rather than a definitive diagnosis of malignancy. Both endosalpingiosis and endometriosis are more frequently seen in peritoneal washings and are discussed at length in another article.
The prototype of the “weird” cells in an effusion is megakaryocytes that have found their way into an effusion. These cells are easy to locate during screening because of their large size and “bizarre” morphology. This situation is most likely to arise where there is extramedullary hematopiesis affecting a pleural or peritoneal effusion. It has also been described in leukemic involvement of the serous cavity and very rarely as a result of a hemorrhagic diathesis in which the megakaryocytes resident within the pulmonary microvasculature are washed into a hemorrhagic pleural effusion. The large size and bizarre appearance of megakaryocytes may suggest a poorly differentiated malignant tumor, but the low nuclear/cytoplasmic ratio, pale and granular cytoplasm, multilobed nuclei, indistinct nucleoli, and occasional cytoplasmic budding belie their true kindred. Here, familiarity with the morphology of normal megakaryocytes and bone marrow elements is the greatest chance for correctly recognizing the situation.
Mesothelial Cells Mimicking Malignancy
Mesothelial cells themselves are one of our greatest foes in achieving the correct interpretation of effusion cytology due to the variety of alterations that mesothelial cells may develop that mimic malignancy>. The etiologies that may generate mesothelial cell atypias Among the causes of benign mesothelial cell abnormalities, hemorrhagic pericarditis is notorious for generating wildly atypical or, if you will, malignant-appearing benign mesothelial hyperplasia. Therapy effects have also been associated with mesothelial cell changes that may mimic malignancy. The effects of chemotherapy on mesothelial cells have not received much attention, although chemotherapy has been reported to cause malignant cells to undergo cytoplasmic and nuclear enlargement, chromatin alterations with hyperchromasia, giant cells and abnormal mitoses, particularly noted with Taxol therapy. Radiation treatments have been associated with bizarre giant multinucleated cells, cytoplasmic and nuclear enlargement and vacuolation, and clumped or smudged chromatin with pyknotic nuclei. It is important to note that, while these features have been seen in effusion samples from patients treated with radiation, it is difficult to know what is actually causing the changes. Similar morphologic alterations can occur in reactive mesothelial cells in patients not undergoing treatment, and patients may be receiving both chemotherapy and radiation therapy concurrently, making it impossible to trace the source of the changes. In fact Wojno et al concluded that distinctive cytologic changes were not evident in the pleural effusions of patients treated with irradiation.
In some ways, the etiology of the changes in mesothelial cell morphology that allows them to mimic malignancy is a moot point. What is significant is the means by which to detect the mimics from the real McCoy. The separation of reactive mesothelial cells from malignancy is the topic of another article in this journal; suffice it to say that the clues to the correct diagnosis will be found by searching for the characteristic features identifying the cells as mesothelial cells and rigid requirement for the strict cytologic features of malignancy before concluding the cells are malignant.
Malignancies That Mimic Benign Effusions
Fortunately, most malignant tumors involving effusions are readily identified, but some malignancies are notorious for their ability to mimic benign effusions. The malignancies that may most mimic reactive mesothelial cells include those that have reduced cell adhesion and present with a dispersed population of single cells with few and small clusters such as malignant melanoma and, as illustrated in this case, some carcinomas of breast. Other malignancies that have very strong mesothelial cytologic features and thus mimic benign mesothelial cells include low-grade serous carcinoma and mesothelioma. Another mechanism by which malignancies elude detection is by adopting a stealth approach and mimicking the background cells. Thus, the cancer hides itself among the benign background cells and is so unobtrusive as to be easily overlooked. Among the deceivers within this group are lobular carcinoma of breast, as well as other signet-ring cell adenocarcinomas and small-cell anaplastic carcinoma. Low-grade non-Hodgkin lymphomas may adopt an approach similar to this in that they may mimic a benign lymphocytic effusion.
Unmasking the Disguise
Although it is important to be armed with the knowledge of tumors that may mimic a benign effusion and how this mimicry is achieved, it is more important to know the morphologic clues to the presence of the subtle malignancy. In some cases, a specific feature characterizing a tumor may be evident, such as melanin pigment in melanoma. In fact, this is a rare occurrence, and the tumors with specific defining features usually lack those features in the cases that pose diagnostic problems. Additionally, features are not always as specific or diagnostic as one would wish. It is widely appreciated that psammoma bodies are not restricted to tumors of serous derivation and may be seen in a variety of primaries and furthermore may be seen in benign conditions so that their discovery in an effusion is not proof of a malignant etiology. In fact, the chance of malignancy being discovered when psammoma bodies are evident in an effusion is dependent on the site of the effusion and, more importantly, on the cell population in the cytology. Thus, reliance on specific cytomorphology features of a given primary will not typically be of great assistance in identifying the covert malignancy.
Perhaps the most useful features for exposing the subtle case are what I consider the “unnerving” findings in an effusion sample. Some of these features have been suggested previously to aid in identifying malignant effusions. It must be remembered that none of these findings in and of themselves is proof of a malignancy, and all can be imitated by benign processes. However, these findings act as warning flags to the potential for a subtle malignancy. The features that I consider unnerving include:
1. Cell clusters with 3-dimensional (3D) cell groupings
2. Cells with vacuoles containing something
3. Loss of the “spectrum of changes” (the mono-/dimorphic population)
4. Necrosis/apoptosis
5. Mitoses
As the number of warning flags increases, the potential for a subtle malignancy increases. It should be noted that our case illustrated 4 of the 5 unnerving features, lacking only the cell clusters.
The appearance of a sample on low-power examination that should generate concern is the finding of cell clusters composed of greater than 3 cells arranged in 3D groupings and imparting a “clumpy” look to the slide. This is often the first feature that may be appreciated, pointing to a malignant process, and is one of the classic findings of the “foreign cell population.” Although mesothelial hyperplasia may generate cell clusters of 2 or 3 cells, clusters containing more than 3 cells are uncommon in a simple reactive effusion. Related to the number of cells in the cluster is the architecture of the cluster. As the number of cells in a cluster increases, the cluster takes on a 3D architecture, and this architectural arrangement is distinctly abnormal. A profound reactive mesothelial hyperplasia may rarely have 3D cells clusters, but this arrangement, so uncommon in reactive cells, is typical of neoplastic cells. If the cells show strong mesothelial cytologic features but persist in the abnormal 3D clusters, the possibility of a mesothelioma should be entertained.
Cells with holes in their cytoplasm are common in effusion samples and become an unnerving finding only under specific circumstances. Holes in the cytoplasm of cells reflect cytoplasmic vacuolation. Cytoplasmic vacuoles result from accumulation of glycogen, lipid, water (hydropic), mucin, or phagocytotic debris in the case of histiocytes. Glycogen and lipid are typically smaller vacuoles and not a source of diagnostic concern in most cases. Thus, it is the hydropic, mucin, and phagocytotic vacuoles presenting as macrovesicles seen on routine light microscopy that become vacuoles of interest. Phagocytotic vacuoles typically are not a problem as the cell containing the vacuoles can be appreciated as a histiocyte due to its multiplicity of vacuoles, abundance of cytoplasm, and characteristic rounded to kidney-shaped nucleus. Hydropic vacuolation is a very common occurrence in cells within an effusion. It can be seen in benign as well as malignant cells and in tumors of various lineages and thus is of lower diagnostic utility. However, the presence of mucin vacuoles is highly significant and explains why cells with holes may be so concerning. Therefore the separation of hydropic vacuoles from mucin vacuoles is paramount.
The main morphologic finding that separates mucin from hydropic vacuoles is perhaps too obvious to state. Mucin vacuoles appear to contain something (mucin) and hydropic vacuoles do not. Hydropic vacuoles do in fact contain water, but by microscopy they appear empty. Obviously, the presence of mucin will be proven by a mucin stain, but frequently this will not be necessary as the fact that there is mucin in the vacuole may be revealed in some Papanicolaou stains, in which the mucin will have an eosinophilic tinge. Even if the tinctorial characteristics of the mucin cannot be appreciated, the mucin within the vacuole is structured, and this structure is apparent on light microscopy. When well developed, the mucin may be so structured as to have a concentrically laminated appearance and has been aptly named as targetoid mucin. If less well structured, the vacuoles may appear to contain only microglobules within the mucin.
There are other clues to the presence of mucin in a vacuole. Both hydropic and mucin vacuoles displace the nucleus of the cells, and thus nuclear displacement is of little diagnostic utility. However, the hydropic vacuole is compliant and will wrap itself around the nucleus. This results in an “overlapping” appearance by light microscopy so that when moving in and out of the focal plane, the edge of the vacuole will overlap the edge of the nucleus. Mucin vacuoles are far less compliant because they contain mucin, which is more structured than water. Thus, mucin vacuoles indent or crush the nucleus. When taken to the extreme, the nucleus may be crushed against the cell membrane and impart a signet-ring appearance to the cell.
Perhaps it would seem to be a simple matter to diagnose signet-ring carcinomas when involving an effusion. In fact, this is often not the case, as well illustrated by lobular carcinoma of breast in which signet-ring forms may be numerous but very small in overall cell size. During the natural screening process where the “big and ugly cells” come under intense scrutiny, the stealthy small signet-ring cells may go unnoticed. Therefore, review of every effusion sample necessitates a careful examination for vacuolated cells, particularly of smaller size and hiding among the background cells, as these holey cells are not particularly saintly. Signet ring cells may be small and hide, but even when large, they may pose a diagnostic problem. Unfortunately, the presence of the mucin artificially lowers the nuclear/cytoplasmic ratio by increasing the relative abundance of cytoplasm, once again aiding these cells in their deceptive demeanor.
There are some final warnings about cells with cytoplasmic holes that must be stated. Although very useful as an unnerving feature, cells with vacuoles that contain something will be absent in a number of malignant tumors. Clearly, poorly differentiated adenocarcinomas that have lost mucin production or that are not mucinous in the first place will lack mucin vacuoles. Furthermore, not a single mucin vacuole will be found, no matter how hard one searches, when confronted with small-cell carcinoma or squamous carcinoma in an effusion. Therefore, the absence of cytoplasmic vacuolation is not synonymous with benignancy. Nor does the finding of a hydropic vacuole rather than a mucinous vacuole act as proof of benignancy. Serous carcinoma is the prototypic tumor in which hydropic vacuoles are common and almost a defining feature. Nevertheless, holey cells must be sought out and interrogated.
A spectrum of cytologic changes characterizes reactive conditions in a variety of specimen types and from a variety of anatomic locations. This holds true as much for effusions as for other samples. The two patterns of neoplasia in effusions are the monomorphous population and dimorphous population. The polymorphous population, that is, the population with a spectrum of change, is far more likely to be reactive. The dimorphic population is the easy one to recognize as there is the population of benign cells and the “alien population” of the tumor cells. Thus, the sample with the dimorphic population is usually not the case that generates diagnostic dilemmas. It is the monomorphic population that is the pitfall. Much like the saying about the forest that cannot be seen through the trees, a monomorphic population that mimics mesothelial cells causes the observer to lose the reference point for “normal” and be more apt to let significant abnormalities pass unnoticed. Here, discipline is required to carefully examine the cells of a monomorphic population looking for the classic cytomorphologic features of malignancy, including increased nuclear/cytoplasmic ratio, hyperchromasia, irregular nuclear outlines with irregular nuclear membrane thickness, coarse chromatin with membranous margination, and nucleolar abnormalities. In benign conditions, the features of malignancy will not be identified, but in malignancy, the intense scrutiny of cytomorphology will unmask the deceivers and reveal their malignant face.
Necrosis and apoptosis are uncommon findings in effusion samples. Thus, these features are less commonly encountered and therefore less frequent as an unnerving finding. Furthermore, necrosis may have a benign etiology, such as the classic appearance of a rheumatoid nodule in a pleural fluid. However, necrosis and apoptosis are abnormal even if not malignant and necessitate careful examination for other features of malignancy.
Mitoses are not proof of malignancy. This must be remembered. Any benign effusion may have a mitosis or two, but mitoses are rare in benign effusions. I will tolerate only a few mitoses per slide in an effusion before the alarm is raised within me. Again, the mitoses are not proof of malignancy, but I will launch a very careful search for the other unnerving features and often will find sufficient cytomorphologic changes to prove the case is malignant. The tumors that are frequently exposed by mitoses are melanoma, small-cell carcinoma, and breast carcinoma.
Malignancies Mimicking Other Primaries
One final form of mimicry that needs to be addressed, at least in passing, is the ability of one primary to mimic the morphology of another primary when observed in effusion samples. Here again, the physics of their fluid world is shaping some of the morphologic features. Separation of primaries is the subject of another article in this journal but remains a significance source of mimicry, causing errors by misclassification of tumors, which may lead to inappropriate therapeutic choices.
CONCLUSIONS
Mimicry causes immense consternation in effusion cytology and sets both overcall and undercall pitfalls in our path. Fortunately, most of malignant cases openly declare themselves. In those cases where reactive changes mimic malignancy and malignant cells take on the guise of benignancy, careful review of the cytomorphology, with particular attention to the unnerving features, will help expose the majority of cases. In this review, I have addressed only the morphologic tools to help uncover the occult malignancy. This is because the morphology is our first line of defense, and if the case passes through this, we have already lost the battle.
Pathology Case Reviews Volume 11(2), March/April 2006, pp 85-91
BOTTOM LINE
A primer for the Nobel Prize
Martin B Van
Der Weyden
Editor, The Medical Journal of Australia
Nobel laureates are particularly prone to publishing their memoirs and sharing their wisdom. Examples include James Watson’s The double helix: a personal account of the discovery of the structure of DNA and Macfarlane Burnet’s Changing patterns: an atypical autobiography. The latest addition to this passing parade is Peter Doherty’s The beginner’s guide to winning the Nobel Prize. The title stimulated me, but it may deter other readers.
Doherty and Rolf Zinkernagel won the Nobel Prize in Physiology or Medicine in 1996.
The book has many themes: selected autobiographical details; forays into the history of the science of medicine; the story of the progress of immunology in the 20th century and the many Nobel Prizes gained along the way; the immunological research that won Doherty and Zinkernagel the Nobel Prize; the human aspects of the modern research enterprise and its culture; and the people and areas of research that are contenders for future Nobel Prizes. More accessible topics are Doherty’s views on science and religion, politics and the media, and the importance for a nation’s economy of supporting the research enterprise. All these topics are embellished with Doherty’s wisdom and wit.
The book’s encyclopaedic coverage is both its strength and its weakness. Its continuity is broken at times by poorly placed and unnecessary diversions. Despite this, I persisted, in pursuit of the advice promised in the book’s title. Ultimately, the answer was revealed, but you will have to read the book yourself for the revelation.
It is difficult to know to whom to recommend the book. It could be for anyone, from the curious citizen to the young scientist wanting to become street-smart for his or her journey to Stockholm and the Nobel Prize.
The beginner’s guide to winning the Nobel Prize. A life in science. Peter Doherty. Melbourne: Melbourne University Publishing, 2005 (304 pp, $34.95). ISBN 0 522 85120 7.