September 2011
Feature Story

William Check, PhD

Computers are so taken for granted today that it’s hard to remember what our lives were like before they became commonplace. In pathology, immunohistochemistry has much the same status. Surgical pathology was a thriving and valuable discipline long before the 1970s, but the rapid emergence of IHC late in that decade radically altered the field and took it to a much higher level. “Not a day goes by in my practice that I don’t marvel at what we can do using immunohistochemistry that we couldn’t do 30 years ago,” says Stuart J. Schnitt, MD, director of the Division of Anatomic Pathology at Beth Israel Deaconess Medical Center and professor of pathology, Harvard Medical School.

Dr. Schnitt began his pathology residency in 1979 at which time, in the department he was in, no immunohistochemistry was done other than immunofluorescence to evaluate renal and skin biopsies. Having lived through those early years has allowed him to “truly appreciate how far we have come,” he says, adding, “It’s amazing to think of the changes since then.” He likens it to the difference between Kodachrome slides and PowerPoint presentations: “In the Kodachrome days, I packed sleeves and sleeves with hundreds of slides to take with me for my USCAP courses. Now I can put 15 presentations on a small device that fits on my keychain.”

These days IHC is not only automated but also practically transparent, as invisible as the computer monitors on which pathologists submit reports.

This month at the CAP ’11 annual meeting, Dr. Schnitt and others led sessions in which they talked about using IHC to help solve diagnostic problems. Last month, they spoke with CAP TODAY.

Several sessions at CAP ’11 focused on breast cancer. Dr. Schnitt taught a course on evaluating breast core needle biopsies with Sandra J. Shin, MD, chief of breast pathology and associate professor of pathology and laboratory medicine at the New York Presbyterian Hospital-Weill Cornell Medical College. Dr. Shin cites several challenging aspects of core needle biopsies that can enhance the value of IHC as an adjunctive diagnostic tool.

“For core needle biopsy from the breast, the target is either something seen during mammographic screening—usually calcifications but also it can be a nonpalpable mass or tissue distortion—or a palpable lesion discovered by the patient or physician,” Dr. Shin says. “In all cases it comes to the same problem. The tissue sample may not show all the features of the targeted lesion.” Sometimes, she says, it’s enough to make a diagnosis, but other times it’s what she calls a jigsaw puzzle where you may have some, but not all, of the pieces. “So the pressure is there right from the get-go with breast core needle biopsy lesions.”

A second problem with core needle biopsy specimens is that they are inherently fragmented. “The architecture of the lesion is compromised, which is one of the two main features we use to make a diagnosis,” Dr. Shin says. Diagnosis is based on the cytology of lesional cells and how they relate to each other.

In addition, breast lesions are often heterogeneous. “Some lesions in breast pathology have a higher degree of subjectivity than others,” Dr. Shin says. “Distinguishing atypical ductal hyperplasia from low-grade ductal carcinoma in situ [DCIS] can be problematic even in excisional specimens.” Adding heterogeneity to the issues inherent to core needle biopsy specimens makes rendering a diagnosis even more difficult. To ramp up the pressure one more step, Dr. Shin notes, “What you call the lesion will have immediate clinical management consequences.”

All of these difficulties make IHC especially useful in the context of breast core needle biopsy. “Immunohistochemistry helps to even the playing field,” Dr. Shin says. “After having so many things taken away from us in terms of making a diagnosis, IHC gives us back some advantage.”

IHC is called for in several specific situations, Dr. Schnitt says. “Based on my consultation experience,” he says, “there are certain recurring problems we see where IHC can make diagnoses that used to be very subjective much more objective.” These situations occur not only in core needle biopsies but also in excision biopsy specimens.

One is to help distinguish invasive carcinoma from benign proliferations with a similar morphological appearance. “It can be invasive carcinoma,” Dr. Shin says, “but it can also be in situ carcinoma or a completely benign lesion.” In this diagnostic scenario, myoepithelial markers are most useful. Invasive carcinomas are not invested in a myoepithelium, whereas benign proliferative lesions and most preinvasive lesions will have an intact myoepithelium. Dr. Shin uses a panel of markers: p63, calponin, smooth muscle myosin, CD10, and smooth muscle actin (SMA). SMA also stains myofibroblasts in adjacent stroma, which can give the appearance of positive staining around invasive carcinoma mimicking the presence of myoepithelium.

Papillary lesions as a class also call for IHC. “They can be difficult to diagnose by morphology alone because of their heterogeneous and complex features,” says Dr. Shin, who calls the papillary lesion “nobody’s favorite breast lesion.”

When using myoepithelial IHC stains for papillary lesions, Dr. Shin says, “We need to show the presence of myoepithelium not only around the periphery but also in the fibrovascular stalks of the papillary lesion.” Papillary carcinomas lack myoepithelial staining in fibrovascular stalks, while benign papillary lesions (papillomas) retain this lining. “You also need to know that some preinvasive papillary lesions can lack a myoepithelium,” Dr. Shin says. This can be seen in examples of solid papillary carcinoma and intracystic variant of papillary carcinoma. “There is some question about whether these lesions are indolent invasive or preinvasive carcinomas,” she says. Currently, Dr. Shin says, patients with these lesions are being treated similar to those with DCIS.

To complicate matters, “Atypical ductal hyperplasia or in situ carcinoma can arise in an otherwise benign papillary lesion.” By IHC, positive myoepithelial staining is seen in the benign area with attenuated or absent staining in areas of atypia or in situ carcinoma. “The area of atypia or in situ carcinoma may not even be represented in the limited sample of a core needle biopsy,” Dr. Shin says. When reporting such lesions, she advises, “Definitely say what’s most worrisome without going overboard.”

Some stains other than those that mark myoepithelium can be used to evaluate papillary lesions. “Neuroendocrine markers such as synaptophysin, chromogranin, and CD56 can highlight neuroendocrine differentiation,” Dr. Shin says. “They are most commonly positive in solid papillary carcinomas.” Unlike papillomas, estrogen receptor (ER) is typically diffusely and strongly positive in cases of in situ papillary carcinoma.

A third place where IHC comes into the picture is in helping to assess intraductal proliferative lessions. High-molecular-weight cytokeratins, such as cytokeratin 14 and cytokeratin 5/6, can be helpful in distinguishing usual (benign) ductal hyperplasia from DCIS, says Melinda Lerwill, MD, assistant professor of pathology at Harvard Medical School and assistant pathologist at Massachusetts General Hospital. She led a separate session on breast cancer. “If you are having trouble distinguishing usual ductal hyperplasia from DCIS, these markers provide additional information. Usual ductal hyperplasia is positive,” she says, “while low- or intermediate-grade DCIS is almost always negative.” Some high-grade examples of DCIS can express these markers, but, Dr. Lerwill says, “Usually we don’t have trouble distinguishing usual ductal hyperplasia from high-grade DCIS.”

Determining whether an in situ lesion is lobular carcinoma in situ or DCIS is a fourth situation in which IHC proves its worth. In most cases, Dr. Lerwill says, the histologic features are quite different between the two. “However, we all encounter cases where it is hard to tell whether a proliferation is ductal or lobular.” Making the correct call has clinical management implications, Dr. Lerwill points out. DCIS requires clear margins. “And if it is treated with lumpectomy, it needs radiation,” she says. In contrast, the management for lobular carcinoma in situ is monitoring. “In ambiguous cases a stain for E-cadherin can be quite useful,” Dr. Lerwill says. DCIS retains expression of E-cadherin, whereas LCIS loses expression of this cell adhesion molecule. IHC for E-cadherin can also help clarify whether an invasive cancer is lobular or ductal.

Dr. Shin describes the fifth diagnostic conundrum that IHC can help resolve—spindle cell lesions. Spindle cell lesions represent the gamut of benign, atypical, and malignant lesions, the cells of which share a common morphologic shape. One possibility in the differential diagnosis is spindle cell metaplastic carcinoma. Dr. Shin orders a panel of cytokeratins that can react with this carcinoma—CKAE1/3, CK7, CAM5.2, p63, and 34 E12. “Immunoreactivity for cytokeratin stains can be highly variable in these tumors,” she says. Some pathologists might order one or two cytokeratin stains to rule it out. That is not enough, Dr. Shin cautions, and adds another caveat: “A negative immunopanel on a core needle biopsy does not exclude this diagnosis.”

Dr. Schnitt identifies the last area in which IHC can be contributory: to try to identify whether a metastatic lesion might originate in the breast. Choice of immunostains for this purpose depends on the differential diagnosis. “We usually try to narrow it down to breast versus something else,” Dr. Schnitt says. “In general, the markers that are helpful in defining breast origin are ER, gross cystic disease fluid protein-15 [GCDFP-15], mammaglobin, and CKs 7 and 20. If we are considering breast versus lung, we would add TTF-1,” he says. If he and colleagues consider breast versus ovary, they would include WT-1, CA125, and PAX-8.

Dr. Lerwill offers a note of caution in this situation: GCDFP-15 and mammaglobin can sometimes be expressed by non-breast tumors. “If you get a positive result in an adenocarcinoma of unknown primary, that certainly suggests breast origin,” she says. “However, neither marker is completely specific or sensitive. It is another piece of information to be integrated with the histology and clinical history to help identify the primary site of origin.”

More generally, Dr. Shin says, “Morphology comes first. Immunohistochemistry is there mainly as an adjunctive tool. If we use IHC first, it can lead us down the wrong path.” Moreover, it is important to know the limitations of individual IHC stains in a given diagnostic dilemma. The workup can become complicated, Dr. Shin notes, adding, “Yet somehow we manage.”

At CAP ’11, another course in which the contribution of IHC was discussed was one on difficult melanocytic lesions, taught by Maxwell A. Fung, MD, associate professor of clinical dermatology and pathology and director of the dermatopathology service at the University of California, Davis.

Clinical data such as age and anatomic site can provide clues to the nevus/melanoma distinction. Melanoma is rare in children and increases with age, but it’s not uncommon in those in their 20s and 30s. So, Dr. Fung says, “the stakes are high” for melanoma, which frequently metastasizes if not detected and removed at an early stage. Nevi on the palm or sole can exhibit features that overlap to some extent with melanoma, making differential diagnosis difficult.

One marker that helps in problematic cases is HMB-45, which is fairly specific for melanocytes and virtually 100 percent specific for melanocytes in the epidermis. “For HMB-45, the pattern of staining usually differs between nevi and melanoma,” Dr. Fung says. The invasive component of a melanoma is often diffusely immunopositive for HMB-45 and may have strong staining in the deepest portion of the invasive component. In nevi, in contrast, HMB-45 generally stains only the most superficial layers and shows progressive loss of staining with descent into the deeper dermis.

Another commonly employed marker is Ki-67, a proliferation marker that is positive in dividing cells and so measures the growth fraction of cells. “Growth fraction is generally higher in melanomas and lower in nevi,” Dr. Fung says. However, Ki-67 growth fraction overlaps in nevi and melanomas, so there is no specific cutoff value. “Some melanomas have low staining and some nevi have high staining,” Dr. Fung says. “That limits its utility and means you have to use Ki-67 as one component of an overall composite assessment of the case.”

For melanocytic lesions, too, progress is being made with newer markers. One that is useful for staging melanoma is phosphohistone H3 (PHH3), which indicates mitosis and thus dividing cells. “In the latest 2010 AJCC [American Joint Committee on Cancer] staging criteria for melanoma there is an increased emphasis on mitotic rate,” Dr. Fung says. “Traditionally this is assessed on H&E, but that can be imprecise and time-consuming. So people have been looking for an IHC stain for mitosis. At this time, PHH3 is the best studied in melanoma.” Dr. Fung calls PHH3 a “cutting edge” marker for mitotic rate in melanoma. “Its sensitivity for mitosis is about double that of H&E,” he says. “In the future, immunohistochemical markers such as PHH3 may become a standard method by which dermal mitotic activity is measured in melanoma.”

A four-color FISH test is available that has three probes for sites on chromosome six and one on chromosome 11. One can also order array comparative genomic hybridization (aCGH), which Dr. Fung calls “probably the most complete method of genetic analysis available.” In aCGH, the entire genome of a tumor is scanned for gain or loss of loci and changes in chromosome copy number. “Melanomas generally show loss of one or more loci, while nevi generally don’t,” he says.

Allen M. Gown, MD, medical director and chief pathologist at PhenoPath Laboratories, Seattle, was not a CAP ’11 faculty member but he is a member of the CAP’s Immunohistochemistry Committee. In 2007, Dr. Gown was a co-author of an article in the “Best Practices in Immunohistochemistry” series that discussed hepatocellular carcinoma versus metastatic neoplasms (Kaker S, et al. Arch Pathol Lab Med. 2007;131:1648–1654). “The liver is a very common site for metastasis,” Dr. Gown notes. So a tumor in the liver could be primary HCC or a metastasis from another organ. Because these non-hepatic carcinomas are very common, tumors in the liver more often end up being metastases than primary HCC. Limited tissue is available with core and fine-needle biopsies, so choosing the best immunostains for IHC becomes crucial.

In the 2007 article, two markers were identified as being of superior value in making the distinction between primary HCC and metastasis: Hep Par 1, a marker of HCC, and MOC-31, an antibody against a surface glycoprotein found on tumors metastatic to the liver. Other proteins, alpha-fetoprotein and CD10, for example, were found not to be as useful.

“Some newer markers have superseded those we described in the paper,” Dr. Gown says. One doubly useful marker is the oncofetal protein glypican-3, which is markedly elevated in HCC. A neoplasm in the liver presents not only the problem of HCC versus metastasis, but also the question of malignant versus benign liver tumor. Glypican-3 can help with this second question as well, Dr. Gown says. Glypican-3 is not generally expressed in normal liver or adenoma, so its presence signals the probability of a malignant liver tumor.

A second marker of HCC, arginase-1, which is superior to Hep Par 1, has also arisen since publication of the 2007 article. “Arginase-1 is even more sensitive and specific than Hep Par 1,” Dr. Gown says. “If we were to rewrite that paper, we would include it as well.” He has introduced these two newer markers into his practice.

IHC has become an integral part of the pathology laboratory, Dr. Gown notes. Does it have the “sex appeal of molecular studies”? he asks. No. “But it’s a more mature technology and accessible to the vast majority of pathology laboratories.” And, just as with computers, it’s hard to remember having to do without it. Says Dr. Schnitt, “It allows us to refine our diagnoses in ways we were not able to do before.”