The diagnostic performance of endoscopic ultrasound–guided fine-needle aspiration is strongly dependent on the availability of an onsite cytopathologist. The diagnosis of some rare tumors may require ancillary testing for which a histologic core biopsy is required. There is increasing interest in evaluating core tissue for molecular markers that may serve as prognostic predictors and targets for focused chemotherapy in patients with cancer. If core tissue can be procured efficiently and reliably at endoscopic ultrasound, this will eliminate the need for an onsite cytopathologist, accurately diagnose tumors that are occasionally missed by fine-needle aspiration cytology, and enable the assessment for molecular markers.
Key points
- •
Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) is an accurate technique for establishing tissue diagnosis in patients with tumors or lesions in or adjacent to the gastrointestinal tract.
- •
The diagnostic performance of EUS-FNA is strongly dependent on the availability of an onsite cytopathologist.
- •
The diagnosis of some rare tumors may require ancillary testing for which a histologic core biopsy is required.
- •
Within oncology, there has been increasing interest in evaluating core tissue for molecular markers that may serve as prognostic predictors and targets for focused chemotherapy in patients with cancer.
The impact of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) on the practice of pancreatic pathology is considered a disruptive innovation. EUS-FNA rightly fits the description of a disruptive technology: “technically straight forward, consisting of off-the-shelf components put together in a way that is often simpler than prior approaches.” EUS-FNA consists of three technologies, each of which is disruptive in its own way: fiberoptic endoscopy was disruptive to rigid endoscopy, ultrasound to radiograph, and cytology to histology. A recent study analyzed the pattern of pancreatic pathology examinations during a 20-year period in a tertiary referral center with relationship to implementation of a EUS-FNA program. The sensitivity and specificity for cancer diagnosis improved from 55% and 78% to 88% and 96%, respectively, after implementation of the EUS-FNA program. Unsatisfactory (7% vs 1%), atypical (16% vs 4%), and suspicious (16% vs 3%) diagnosis were significantly reduced. After implementation of the EUS-FNA program, whereas the average percentage of annual cases managed by cytology alone increased from 19% to 51%, the percentage managed by histology alone declined from 56% to 23%. Consequently, non–EUS-guided FNA cytology decreased from 36% to 1% and needle biopsies from 29% to 9%. Because EUS-FNA brought about a much-needed improvement in diagnostic accuracy, it resulted in the displacement of histologic diagnosis in management of pancreatic diseases. This is evident from a recent study conducted by these authors that examined the 5-year trend (2006–2010) in tissue acquisition in pancreatic diseases in the United States using the Medicare database: The use of EUS-FNA increased by 69.3%, surgical biopsy declined by 41.7%, and the use of percutaneous biopsy has remained stable.
Although EUS-FNA can be regarded as a disruptive innovation, it is important for endosonographers to recognize the imperfections of EUS and make constant refinements and incremental improvements to existing techniques, technologies, and practice patterns. This self-evaluation is important because it enables one to see “what is next.” This might occasionally result in technological development that is offered in excess of what the consumer (endosonographers) demands or needs. Herein, the consumer weighs-in to differentiate between a sustaining “valued-based” technology and just “another” technology. Although biased investigators and industry-driven studies may make this distinction difficult, fortunately in medicine only those technologies that impact patient care in a meaningful manner eventually withstand the test of time.
Limitations of EUS-guided FNA cytology
There are three limitations to EUS-guided FNA that if overcome can propel the technology even further: (1) the availability of onsite cytopathology support, (2) the occasional dependence on histology to establish a diagnosis that is otherwise not possible with FNA cytology, and (3) the ability to reliably assess the tissue sample for molecular markers so that patients can be risk stratified and treated with tailored chemotherapeutic agents.
On-site Cytopathology Support
All studies have shown that the presence of an onsite cytopathologist improves the diagnostic yield, decreases the number of inadequate or unsatisfactory samples, and limits the number of FNA passes required to establish a diagnosis. In addition, two recent meta-analyses on EUS-FNA of pancreatic masses reached the same conclusions: the presence of an onsite cytopathologist was associated with a diagnostic sensitivity of 88% to 95% compared with 80% or less in the absence of a cytopathologist.
For institutions limited by resources, this challenge can be mitigated by adapting a combination of measures. (1) If the endosonographers have basic training in cytopathology to assess for onsite diagnostic adequacy, appropriate samples can be collected and sent for off-site assessment by a cytopathologist. In a recent study of 138 patients who underwent EUS-FNA of solid pancreatic mass lesions, two endosonographers underwent proctoring in cytopathology and found that by assessing for onsite diagnostic adequacy themselves, the rates of diagnostic accuracy improved by 22% and inconclusive diagnosis declined by 18%. (2) If adequate FNA passes can be performed to provide an adequate cell block, a reliable diagnosis can be achieved in most patients. In a prospective study of 91 patients with pancreatic masses who underwent EUS-FNA using a 22-G or 25-G needle, two passes were made for cell block analysis. The specimens were immediately fixed in formalin and processed later by the cell block method. By adapting this technique, adequate tissue was procured in 88 (97%) of 91 patients. Of these 88 patients, the accuracy of the cell block method for diagnosing carcinoma was 99% and for neuroendocrine tumor was 100%. The caveat with this method is that, in patients with large tumors, the aspirate may contain nonviable necrotic material and a cell block may still be nondiagnostic. (3) The need for an onsite cytopathologist can be obviated if reliable core tissue can be procured for histologic assessment. Currently, none of the specially designed biopsy needles or a 19-G needle can guarantee reliable histologic core tissue procurement or demonstrate a diagnostic accuracy of greater than 95%.
For centers that do not have access to onsite cytopathology support, a combination of all three measures must be adopted to achieve a diagnostic accuracy of greater than 95%. The endosonographer after (self) assessing for diagnostic adequacy must perform two to four dedicated passes for cell block or obtain a core biopsy.
Pitfalls in EUS-FNA Cytology
Although encountered rarely, false-positive FNA has been reported in the literature, particularly when evaluating pancreatic diseases. Chronic pancreatitis and autoimmune pancreatitis are the most common reasons for a false-positive diagnosis of malignancy. Some of the cytologic features that may mimic malignancy in chronic pancreatitis are occasional atypical cells, which include enlarged cells, enlarged nuclei with degenerative vacuoles, single cells, and occasional mitosis. Likewise, aspirations from autoimmune pancreatitis often show marked stromal reaction with embedded small clusters of atypical epithelial cells. It is also important to recognize that cytologic features of primary pancreatic carcinomas are similar to many other adenocarcinomas, which can metastasize to the pancreas. Immunohistochemical stains can reliably suggest the possible primary site of tumor. Therefore, it is important to perform dedicated passes to form a cell block to aid in performing immunohistochemical stains or other ancillary studies when the diagnosis is unclear.
A false-negative diagnosis may occur because of technical difficulties, sampling error, or interpretive errors. For a cytopathologist, offering diagnosis on hypocellular samples is one of the more common causes for delivering a false diagnosis. It is also possible that the marked desmoplasia of pancreatic adenocarcinoma might result in an inadequate specimen, which is commonly encountered in the setting of chronic pancreatitis. Diagnosing well-differentiated adenocarcinoma can be challenging because they tend to lack the typical hyperchromasia, display minimal architectural disorder of epithelial fragments, and have only modestly increased nuclear-cytoplasmic ratios. The differential diagnosis includes reactive epithelial changes, which is often associated with ductal stenting or other instrumentation. Under these circumstances, an increased number of passes may be necessary and the specimen needs to be fixed in alcohol for better delineation of nuclear morphology. Several biomarkers are increasingly available to distinguish reactive ductal epithelium from neoplastic cells. Occasionally, even a repeat EUS-FNA procedure is unrevealing and the patient may require an open surgical biopsy for histologic diagnosis.
Whenever possible, one or more dedicated passes must be obtained for cell block creation so that, should a need arise, ancillary testing can be performed. Alternatively, a core biopsy can be performed to facilitate the same outcome.
Assessment for Molecular Markers
Enormous strides have been made in the treatment of breast and lung cancers where delivery of chemotherapeutic agents is guided by expression of molecular markers in the tumor. These molecular markers help prognosticate the tumor and guide treatment algorithm. Despite poor clinical outcomes, treatment of pancreatic cancer has not shown much progress and available data on molecular-based treatment are primitive at best. There is growing evidence that immune cells in pancreatic ductal adenocarcinoma produce immunosuppressive signals that allow tumors to evade the immune response. Furthermore, the stromal fibroblasts provide a protective environment that not only supports and promotes pancreatic adenocarcinoma tumor growth and progression, but also likely suppresses development and/or access of antitumor immune responses. Strategies to deplete the desmoplastic stroma before immune therapy is instituted could possibly promote robust response against tumor cells. Likewise, recent studies have shown that the gene ANG2 is overexpressed in core biopsies of pancreatic neuroendocrine tumors and could potentially serve as a molecular marker or therapeutic target. It is increasingly clear that the evaluation of fibrous stroma for molecular markers may become an integral part of cancer therapy.
In a recent study, specimens obtained from EUS-FNA with a 22-G needle were processed by the standard cytologic approach and compared with another cohort that were formalin fixed to preserve microcores of tissue before histologic processing. EUS-FNA histology preserved the tumor tissue architecture with neoplastic glands embedded in the stroma in 67.8% of diagnostic cases compared with 27.5% with standard cytology. Furthermore, microcore samples were suitable for molecular analysis including the immunohistochemical detection of intranuclear Hes1 in malignant cells and the laser-capture microdissection-mediated measurement of Gli-1 mRNA in tumor stromal myofibroblasts.
Although data are limited, it is becoming apparent that histologic core tissue is preferred over a cytologic aspirate for assessment of molecular markers. It is also evident that the technique of specimen preservation/preparation is as important as the procedure itself. In future, we anticipate that advanced endoscopy fellows will be trained not only to procure tissue but also to assess for diagnostic adequacy and have the requisite knowledge in specimen processing, because these have important implications for patient management.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
