The prognostic and predictive markers that might be tested in cytologic samples, such as ER, PR, HER2 status in breast cancer, ALK gene rearrangement in lung cancer, CD117 expression in gastrointestinal stromal tumor, Ki67 index in neuroendocrine and other tumors, and CD38 expression in B-cell SLL/CLL, have been addressed in Chaps. 4 and 5.

Detection of human papillomavirus (HPV) DNA in squamous carcinoma from the uterine cervix and oropharyngeal origins is not only helpful in the diagnosis but also of prognostic significance. Patients with HPV-positive oropharyngeal cancer tend to present with advanced-stage disease due to spread to the lymph nodes of the neck but paradoxically have a better prognosis than those with HPV-negative tumors. FNA materials of squamous carcinoma from head and neck sites, either cell block or direct smear, are often used for HPV testing using in situ hybridization, Hybrid Capture 2 or Third Wave (Cervista) Technologies.

Molecular revolution has led to an increasing effort to identify specific tumor mutations to enable adoption of genetically informed medicine for optimizing treatment and improving clinical outcome. In non-small cell lung cancer, detection of EGFR and KRAS mutation status on cytologic specimens is frequently requested by treating physicians to determine a patient’s eligibility for anti-EGFR therapy (such as gefitinib/Iressa or erlotinib/Tarceva). Two classes of EGFR mutations , short deletions in exon 19 and the L858R point mutation in exon 21, are the most frequent mutations. The presence of EGFR mutations is associated with response to tyrosine kinase inhibitors. Major biomarkers EGFR (15–20 % in non-small cell lung cancer), KRAS (15–30 %), ALK (2–7 %), and ROS-1 (1–2 %) are generally mutually exclusive. The tumors with KRAS mutations will not response to anti-EGFR therapy and likely have poor prognosis. NGS is a high-throughput sequencing technique that allows simultaneously identifying numerous DNA mutations that may have targeted therapies available.

To ensure a successful molecular testing, high quality of tumor sample is highly desirable. Cell block material or cells scraped from direct smears are suitable for the molecular studies with similar adequacy. Some studies show that smears are equivalent or slightly superior to cell block in mutational analysis. The efficacy of the molecular tests is affected to a great extent by the amount of tumor cells available for analysis. A total of approximately10 ng of DNA is usually sufficient for NGS-targeted molecular profiling, depending on the testing platforms. A tumor percentage of >20 % is required for both smear and cell block preparations at MD Anderson. The number of slides required depends on the cellularity. For smears (either Diff-Quik stained or Papanicolaou stained), an average of 2 slides (range, 1–3) is usually required. For cell block, if the corresponding H&E section contains >300 tumor cells, an average of 5 unstained slides (range, 3–10) should be sufficient for NGS testing. In cases with <20 % tumor fraction, microdissection may be needed to enrich the tumor cells. Usually, if a molecular testing is anticipated, a pathologist should request more materials during rapid onsite evaluation, cut additional unstained slides upfront to avoid the loss of tissue due to repeatedly trimming of the cell block, judiciously select immunomarkers using systematic tired diagnostic approach, and consider dual or multiplex staining or transfer technique to save tissue. The goal is to maximize the cell availability for potential molecular studies, especially for a putative lung cancer.

Gene expression profiling has been used to discover prognostic and drug response signatures for various tumor types and appears to be promising to facilitate personalized medicine because it allows for simultaneously measuring thousands of gene products from a single tumor sample. Gene signatures appear to provide more accurate prognostic and predictive information than any single gene measurement alone. In breast carcinoma, this technique has been used in identifying intrinsic subtypes, in developing prognostic signatures, such as the 70-gene signature (MammaPrint) , and in exploring the gene signatures that predict tumor response to neoadjuvant chemotherapy, endocrine therapy, and other targeted therapies. FNA samples collected in RNA later can yield adequate amounts of total RNA in experienced hands.