The Molecular Edge
The demise of the microscope has been repeatedly predicted with each new and supposedly more quantitative technologic improvement, such as flow cytometry and immunohistochemistry. All have turned out to be useful adjunct techniques, but have not done much to replace basic morphologic examination.
Are molecular techniques going to replace the microscope in anatomic pathology? I believe this is highly unlikely and see the two areas as being more complementary than antagonistic.
First, microscopic examination of H&E tissue sections is actually quite useful for clinical diagnosis in the real world, and molecular testing is not needed for many common lesions such as basal cell or squamous cell carcinoma. In these cases, an H&E-stained slide, a microscope and a pathologist are all that is needed for a diagnosis.
Also, determining tumor margins and other spatial parameters, including depth of invasion in melanoma, are primarily done by morphologic examination and are important in determining patient prognosis and treatment. Therefore, while molecular testing gets a lot of press, many anatomic pathology labs send out less than 10% of their workload for molecular analysis. This percentage will no doubt increase over time, but histology will remain a relatively cheap, efficient method for tissue diagnosis of common lesions for the foreseeable future.
Second, without morphologic examination of tissue analyzed for molecular testing, the conclusions derived from molecular assays may be confusing or incorrect, since you do not know what you are analyzing-viable tumor, necrosis, normal tissue, stroma-or even whether the correct diagnosis is attached to the analyzed specimen. This is particularly true with solid tumor cases, where it is not unusual to receive a block of tissue without any tumor, with the wrong tumor diagnosis or of poor quality.
However, there are more complex cases where molecular testing is absolutely necessary for diagnosis or therapeutic prediction and histology alone cannot provide the answer. These include leukemic subclassifications, translocations in soft tissue tumors, mismatch repair mutation analysis in colon cancer, gene analysis profiles for lymphoma and breast cancer, and companion diagnostics such as FISH for HER2/neu amplification and KRAS mutation analysis. In these cases, the two techniques work together, as the molecular test is only relevant based on the morphologic classification of the tumor.
Molecular tests are not context-independent and are therefore not usually definitive alone for tumor classification, prognosis or predicting response to treatment. Apparently identical translocations, for instance,can be found in histologically different tumors, such as the EWS-ATF1 and EWS-CREB gene fusions found in both angiomatoid malignant fibrous histiocytoma and clear cell sarcoma. Hematopathology requires molecular testing to subclassify leukemia and myelodysplastic syndromes, but lymphomas can usually be diagnosed by histology and immunohistochemistry alone.
Genomic analysis signatures will not always predict prognosis without histology, even within a single prognostic group. One example is the basal-like breast carcinoma prognostic group considered to predict poor prognosis; however, this group contains adenoid cystic carcinoma, a subtype histologically distinct from other members of this group, with generally good prognosis.1
More Evidence for Interdependency
Perhaps the best argument for interdependency of molecular biology and microscopy, however, is that the two techniques give quite different data. Histologic exam is broad, going beyond tumor diagnosis to achieve an overall picture of disease pathophysiology in a specific tumor at a specific time-where is the tumor border? Is it pushing or infiltrative? How heterogeneous and anaplastic is the tumor? How much necrosis or apoptosis is present? Is there any inflammation and what type?
Molecular techniques are narrow but much more detailed and quantitative, able to identify what specific mutations are present and to what degree they are expressed.
If new molecular tests are developed that can be directly appreciated in the context of the tumor, more questions may be answered than by using either histology or molecular biology alone. The microscopic anatomy of the tumor gives vital spatial context for understanding the impact of molecular data. Which cell types carry the mutation or overexpression of protein? How do these relate to other proteins in the same or adjacent cells, and what signaling pathways may therefore be active? What subcellular compartment is the target protein expressed in, and how has that changed from where it is normally expressed?
Dr. Coberly is director, Department of Pathology, Amgen, San Francisco. Views are those of the author and do not reflect those of Amgen.
1. Weigelt B, et al. J Pathol 2010; 220:263-280.