Keywords
Pancreatobiliary (ERCP/PTCD) - Tissue diagnosis - Pancreas - Endoscopic ultrasonography
- Fine-needle aspiration/biopsy - GI Pathology
10.1055/a-2240-0678
Pancreatic cystic lesions (PCL) are common with a prevalence ranging from 24% to 57%
in population studies [1]
[2]. Due to the widespread use of cross-sectional imaging, these lesions are often identified
incidentally [3]. PCLs represent a broad spectrum of pathology which includes inflammatory, non-neoplastic,
and neoplastic cysts. In one operative cohort of 851 patients, the most common diagnoses
included intraductal papillary mucinous neoplasms (IPMNs) (38%), mucinous cystic neoplasms
(23%), serous cystic neoplasms (16%), and cystic neuroendocrine neoplasms (7%) [4].
Accurately classifying a PCL helps inform surveillance and management, including the
need for surgical resection. While clinical context and cross-sectional imaging findings
can be suggestive, endoscopic ultrasound (EUS) provides additional diagnostic and
prognostic information. In many cases, endosonographic characteristics alone are insufficient
to differentiate between mucinous (MN) and non-mucinous lesions (n-MN) or to exclude
the presence of malignancy, thus prompting EUS-guided sampling of PCLs [5]
[6]
[7]. There is significant interest in optimizing tissue acquisition and accurate histologic
diagnosis, as up to 22% of patients inappropriately undergo major pancreatic resections
for benign disease [8].
In the current issue of Endoscopy International Open, Castro et al. report the retrospective outcomes of a cohort of 145 patients with a
PCL who underwent EUS-guided tissue acquisition (EUS-TA) using a 20-gauge EUS-guided
fine needle biopsy (FNB) needle to aid in the differentiation of MN and n-MN lesions
as well as evaluate for the presence of malignancy [9]. Ultimately, 81 patients were diagnosed with MNs (67 benign, 14 malignant) and 64
patients with n-MNs (53 benign, 11 malignant) with the final diagnosis largely defined
by histologic findings from the FNB itself (81, 55.9%) and surgery (58, 40%). The
calculated sensitivity was 92.6% and 92% for the diagnosis of MN and malignancy from
FNB, respectively. In the cohort of patients who underwent surgery, sensitivity was
88% for differentiating malignant from non-malignant lesions. An overall adverse event
rate of 2.7% was noted with three cases of bleeding and one case of pancreatitis.
The concept of sampling the wall of PCLs for histology has been well described including
an early report in 2005 which used a relatively-novel “trucut” needle to enhance histologic
analysis [10]. Several EUS-compatible “cutting” needles are now available, but most devices maintain
similar design principles to allow for core tissue acquisition. The authors utilized
a 20-gauge needle iterated with lateral pores on the needle shaft aimed to optimize
the ability to obtain cyst fluid and histology. While needle design certainly can
impact how a needle behaves during a procedure (e.g. traversing tissue or use in angulated
scope positions), it is unlikely that small tip variations translate to significant
improvements in clinical outcome. The majority of studies show equivalent tissue adequacy
regardless of needle-tip design [11]
[12]. Furthermore, data from solid pancreatic neoplasms suggest that FNB is not necessarily
superior to fine needle aspiration (FNA) in regard to cost-effectiveness, tissue acquisition,
and diagnosis [13], and the incremental benefit it may provide in PCLs is similarly not well established
[7].
Current guidelines suggest the use of FNA as the gold standard for PCL fluid acquisition
and cytologic analysis [7]. It is notable that only 40% of the patients in the current study had cyst fluid
available for biochemical analysis, which is lower than one might expect with the
use of a 20-gauge needle. FNA with a 22-gauge needle remains the standard of care
in our practice and is preferentially performed in PCLs. Certain technical considerations
are worth highlighting which can maximize the yield of cyst fluid aspiration including
the use of a stylet to clear the needle tip before aspirating and adapting a pneumatic
insufflator to apply negative pressure to collect viscous fluid. Numerous studies
have been published on FNA with one meta-analysis suggesting a specificity of 88%,
sensitivity of 63%, and area under the curve of 0.89 for differentiating MNs from
n-MNs [14]. Sensitivity for the presence of malignancy is generally lower, ranging from 27%
to 55% [15]
[16]. A hybrid technique of obtaining biopsies of the cyst wall using microscopic forceps
(Moray Micro Forceps; Steris, Mentor, Ohio, United States) after needle access is
obtained seems to improve the diagnostic performance of FNA but is more technically
challenging, associated with a higher risk of adverse events, and it has not seen
widespread adoption [17].
The diagnostic performance of FNB with histology in the current study appears quite
robust when compared with FNA data, with seemingly adequate tissue in all samples
and a sensitivity for malignancy of 92%. However, this must be evaluated in the context
of the study which may have been confounded by selection bias of the retrospective
cohort. For the majority of patients, it is also important to note that the reference
gold standard was a diagnosis obtained by the sampling methodology being investigated,
and the duration of clinical follow up (12 months) is likely insufficient to clinically
exclude occult malignancy. External validity is also impaired by a population that
likely differs from the PCLs seen by endoscopists in everyday practice, including
52 patients (36%) presenting with symptomatic cysts and a surgical intervention rate
of 40%. Previous work has been more equivocal than what is reported here, with studies
showing a specimen adequacy ranging from 46% to 87% [18]
[19]. Even in cases in which FNB might be theoretically advantageous (e.g. presence of
solid components), tissue adequacy was less than 50% using a 22-gauge biopsy needle
[18]. There is inadequate comparative data to truly assess diagnostic performance of
the technique, and therefore, it is our practice to reserve FNB for select cases.
Generally, FNA can be performed and the same needle may be used to sample a thickened
cyst wall or solid components after partial or complete decompression of the lesion,
a technique which has been described as “targeted cyst wall puncture” [20].
While the techniques discussed in this article all have high specificity if malignancy
is detected, they are limited by relatively poor sensitivity, requiring clinicians
to interpret negative pathology in the context of each individual patient. This is
driven by inherent limitations of sampling techniques as well as the diversity and
pathophysiology of PCLs which, for example, may be associated with field defects and
the development of carcinoma outside the reference lesion [21]. No sampling approach is a panacea and none can replace a thorough clinical history,
risk factor evaluation, and review of characteristic imaging features, which must
be performed in parallel with EUS-TA.
The recently available data on EUS-FNB are certainly intriguing and at the very least
highlight the technical feasibility, applicability, and safety of its use in the evaluation
of PCLs; however, there are no high-quality data suggesting that a particular combination
of needle type, size, or technique is superior to another. There are likely multiple
ways to obtain an adequate specimen to supplement clinical, radiographic, and endosonographic
evaluation of PCLs. The specific technique should be driven by lesion morphology,
endoscopist preference and comfort with specific devices, and locally available pathology
expertise. Continued investigation with prospective, randomized trials and ongoing
innovation in this area are essential to continue to refine the ability of clinicians
to detect pre-neoplastic and neoplastic lesions and provide patients with appropriate,
timely care.
