Cytopathology Specimens and Molecular Testing for Solid Tumors in the Era of Personalized Medicine
Dara L. Aisner, MD, PhD
Department of Pathology, University of Colorado
The recently expanded availability of targeted therapies for solid tumors has greatly increased the number of tumor types treated with targeted agents and how new therapeutic agents are evaluated for efficacy. Increasingly, the evaluation and use of targeted therapy is associated with testing for specific molecular changes to predict responsiveness to the agent, so called 'personalized medicine.1 As the number of targeted therapies and indications for use grows, so too will the demand for testing. The trend towards minimally invasive procedures drives the desire to perform testing on smaller samples. Thus the question often arises whether small samples such as those seen in cytopathology, often initially intended only for diagnosis, are a suitable source of material for molecular testing. Many specimens that are extremely small are actually suitable for testing.
Molecular testing in cytopathology specimens has grown with rapidly evolving technology, such as seen with testing for HPV in liquid based cervical cytology and fluorescence in situ hybridization (FISH) for chromosomal aberrations in voided urine specimens and bladder washing (for example, Urovysion® Abbott Molecular). These assays demonstrate the ease and feasibility of performing molecular assays on cytologic specimens, and portend the broader application of molecular testing on cytologic material.
Tumor mutational assays have become a key part of the targeted therapy of several tumor types, including non-small cell lung cancer (NSCLC), colorectal carcinoma and melanoma.2 It is common for testing to be requested on cytologic specimens in these settings, especially if the primary tumor is not resectable or available for testing. Metastatic lesions are frequently accessible by approaches amenable to fine needle aspiration (FNA). The suitability of molecular testing in cytology specimens has been well studied in NSCLC, in which most patients are not candidates for resection of the primary tumor. Numerous studies have demonstrated that cell block preparations of specimens from NSCLC are suitable for molecular testing.3,4 Similar studies of metastatic colorectal carcinoma have shown that imprint and smear specimens can be utilized for KRAS mutational studies.5,6 These methods are likely to be applicable across all organ systems.
Another modality of testing for targeted therapy is FISH, which is used to test alterations such as gene amplification and chromosomal rearrangements and can be applied to cytology specimens. Typically, FISH testing requires a minimum of between 50-100 tumor cells per section or slide, and thus can be readily implemented on cytology cell blocks and in many cases, imprints or smears.
Multiple cytopathology specimen types are potentially useful for testing. Most testing is currently geared towards cell block preparations, as the processing mimics seen in surgical pathology, but this also exposes the cytology material to formalin, resulting in reduced nucleic acid quality. However, other cytology specimen types are amenable to testing, although validation of individual specimen types is needed before clinical implementation in individual laboratories. Testing of specimens directly from stained smeared or imprinted slides is often feasible, and has been adopted as a clinical practice in some laboratories, although it requires that the selected slide be relinquished for testing. In fact, because these specimens are not subjected to formalin, the nucleic acid quality is typically higher than that seen in cell blocks.7
Multiple parameters are involved in the evaluation of specimens for suitability for testing. The first assessment typically is evaluation of specimen quantity. In the setting of an initial diagnostic workup, which often involves special stains, assessing the total amount of tissue remaining is an important first step. It is not uncommon for the workup to deplete a small specimen, rendering further testing difficult, but not necessarily impossible, as methods exist to recover tissue from prepared slides, such as utilizing negative control slides from immunohistochemistry, or utilizing a stained smear. Because most assay platforms for molecular testing involve target or signal amplification or analysis of a small number of cells, in actuality, a limited number of molecular tests can be performed on extremely small specimens, sometimes with as few as 10 cells per section!
For mutational testing, a parameter, which is perhaps more critical than total specimen size, is the tumor content of the specimen, expressed as 'tumor cellularity' indicating the percentage of all nucleated cells in a specimen represented by tumor cells. Different methodologies for detecting mutations have varying abilities to detect the aberration in the setting of abundant background wild-type copies. This is defined as 'analytic sensitivity' or 'technical sensitivity' and for a mutational test is expressed as the percentage of mutant alleles that must be in the test sample in order to be reliably detected by the assay. Methodologies that have a low analytic sensitivity, such as Sanger sequencing, require that a specimen have a high tumor cellularity to reliably indicate mutation status, while assays with a high analytic sensitivity can use specimens with lower tumor cellularity. Thus, ‘dilution’ of tumor cells in specimens by a large quantity of non-tumor cells is challenging, and an extremely small specimen with high tumor cellularity may be superior to an abundant specimen with low tumor cellularity.
A low tumor cellularity in a specimen can often be overcome with enrichment methodologies, most commonly macro- or micro-dissection. This is a common practice, but can be more challenging to apply to cytologic specimens compared to surgical specimens, as cytology specimens can have diffusely spread tumor which may be more difficult to identify at the low magnification used for enrichment. Macro-dissection indicates that a region of interest is identified by the pathologist, and scraped from the slide without the assistance of microscopy. Microdissection is typically performed under a dissecting or standard light microscope, and is the preferred approach for cytology specimens when enrichment is needed. Some specimen types are more problematic than others for molecular analysis. Malignant effusions can be among the most difficult specimen type to evaluate by molecular methodologies because they frequently contain abundant reactive non-tumor cells that morphologically mimic tumor cells at low magnification.
In addition to predictive/therapy defining tests, molecular testing also has application in helping to define an initial diagnosis. One example is the evaluation of thyroid aspirates for molecular alterations such as BRAF mutation or defined translocations commonly seen in thyroid carcinomas. BRAF mutational analysis and reverse transcriptase PCR testing for RET/PTC rearrangements have documented utility in the setting of indeterminate thyroid cytology, and the implementation of this testing is growing.8 There are other examples of molecular tests that can aid in diagnosis, including Urovysion® (Abbott Molecular), sarcoma translocation assays, and a growing list of complex multi-analyte tests, which look for specific profiles, or 'signatures' to help in classification of lesions.
As the demand for testing and the number of targets for evaluation grows, communication between the pathologist and molecular testing laboratory is essential. Laboratories that do not accept small specimens have sound reasons for doing so; however if testing on a small specimen is the only option for patient care, a laboratory willing to accept the specimen can often be identified. Recognition of the limitations of such testing is essential, as is communication of these limitations to the requesting clinician. Similarly, as the cytopathologist is often the only link between the ordering clinician and the molecular laboratory, communication of testing priorities, especially when multiple tests are necessary is another essential role.
This underscores another important issue in the testing of cytologic specimens: a negative result must always be interpreted in the context of the specimen submitted. A negative result in a specimen that is borderline for testing adequacy should be interpreted with caution. As the demand for testing small specimens increases, the burden of determining whether a specimen should be referred for molecular testing will be increasingly placed on the pathologist. Although molecular laboratories which perform this testing will also evaluate the suitability of the specimen, familiarity with the criteria for suitability will assist in determining which specimens are sufficient for referral.
1. Hamburg MA, Collins FS. The path to personalized medicine. The New England Journal of Medicine 2010;363(4):301-4.
2. Igbokwe A, Lopez-Terrada DH. Molecular testing of solid tumors. Archives of Pathology & Laboratory Medicine 2011;135(1):67-82.
3. da Cunha Santos G, Saieg MA, Geddie W, Leighl N. EGFR gene status in cytological samples of nonsmall cell lung carcinoma: Controversies and opportunities. Cancer Cytopathology 2011;119(2):80-91.
4. Malapelle U, Bellevicine C, Zeppa P, Palombini L, Troncone G. Cytology-based gene mutation tests to predict response to anti-epidermal growth factor receptor therapy: A review. Diagnostic Cytopathology 2010.
5. Malapelle U, Bellevicine C, Russo A, Salatiello M, Palombini L, Troncone G. KRAS testing on colo-rectal carcinoma cytological imprints. Diagnostic Cytopathology 2010.
6. Troncone G, Malapelle U, Cozzolino I, Palombini L. KRAS mutation analysis on cytological specimens of metastatic colo-rectal cancer. Diagnostic Cytopathology 2010;38(12):869-73.
7. Killian JK, Walker RL, Suuriniemi M, Jones L, Scurci S, Singh P, Cornelison R, Harmon S, Boisvert N, Zhu J and others. Archival fine-needle aspiration cytopathology (FNAC) samples: untapped resource for clinical molecular profiling. The Journal of Molecular Diagnostics : JMD 2010;12(6):739-45.
8. Nikiforov YE. Molecular diagnostics of thyroid tumors. Archives of Pathology & Laboratory Medicine 2011;135(5):569-77.