Adequacy Assessment of Endoscopic Ultrasound-Guided, Fine-Needle Aspirations of Pancreatic Masses for Theranostic Studies: Optimization of Current Practices Is Warranted.
Recently, several novel clinically active chemotherapy regimens for adenocarcinoma of the pancreas have been described. (11,12) This represents a significant paradigm shift from the preceding 20 years, when gemcitabine was the only standard of care for this disease. The ability to choose among several active regimens raises for the first time the possibility of "personalizing" chemotherapy treatment options. However, very little is known about our ability to procure adequate tissue for theranostic studies on which to base treatment decisions. For patients with unresectable disease and for those who receive neoadjuvant therapy prior to resection, the initial diagnostic cytology material may be the only specimen that has viable tumor cells for diagnosis and for any potential ancillary theranostic tests that may be requested.
Endoscopic ultrasound-guided (EUS) fine-needle aspiration (FNA) is the most commonly used technique to obtain a diagnosis in patients with suspected pancreatic malignancies. Procuring sufficient material for diagnosis has been the priority of EUS-FNA, and adequacy assessments currently address the amount of material that is available for a diagnosis. Examination of formalin-fixed, paraffin-embedded cell block materials created from the EUS-FNA traditionally contributes minimally to the cytopathologic diagnosis. However, collection of materials for the creation of a cell block is typically performed for potential diagnostic immunohistochemistry that may be used in difficult cases with uncommon tumors, such as neuroendocrine or acinar cell tumors.
In order to meet the burgeoning field of personalized care in pancreatic diseases, (13) there will likely need to be a shift in the paradigm on intent of EUS-FNA. There will need to be a dual focus on obtaining adequate cells for diagnosis while optimizing the harvesting of enough material (cells/DNA/ protein) to perform advanced molecular analysis of critical pathways. Our study aims to correlate cellularity of cell block material with several clinicopathologic variables in an effort to understand factors that will allow us to optimize the collection of adequate material not only for diagnosis, but for pretreatment or theranostic testing.
MATERIALS AND METHODS
Cytology cases of pancreatic mass lesions that had cell blocks prepared were identified during a 12-month period (October 2010-October 2011). Only cases with a positive diagnosis of epithelial malignancy or neoplasms were included for examination. All cases with negative and indeterminate diagnoses, as well as lymphoid and mesenchymal neoplasms, were excluded. Based on availability of stained smears and hematoxylin-eosin cell block slides for analysis, a total of 169 cases were evaluated for smear and cell block cellularity. The clinical and pathologic variables recorded were campus location (our academic center has cytology centers at 2 hospitals, University of Pittsburgh Medical Center [UPMC] Presbyterian and UPMC Shadyside), EUS-FNA operator, presence or absence of on-site evaluation, number of passes, pathologist involved, size of mass, and cell block cellularity.
Based on the results of the retrospective analysis, a 3-month prospective analysis was performed using a new FNA algorithm. This encompassed 5 FNA passes without adequacy evaluation in all cases. In some cases, ProCore (Cook Medical, Bloomington, Indiana) ultrasound biopsy needles were used. Droplets of aspiration fluid from the first 2 passes were used to make a total of 4 smears (slides), and the remainder of the aspirate (pushing stylet back in to express material, including tissue fragments) was submitted for cell block preparations. Then, the material from 3 additional passes was placed directly in formalin for the cell block.
Cytologic materials included smear preparations stained with modified Romanowski stain (or Diff-Quik stain when rapid on-site evaluation was performed for adequacy assessment) and Papanicolaou stain. Cell blocks were prepared using standard protocol, detailed as follows: The specimen, procured in 10% neutral buffered formalin, was centrifuged for 10 minutes at 2000 rpm. The supernatant was separated with care to retain pellet material. Next, 1 to 3 drops of Histogel (Thermo Scientific, West Palm Beach, Florida) was added to the specimen and it was allowed to stand in the refrigerator for 10 minutes; the pellet was then pulled onto biopsy wrap paper (Cancer Diagnostics, Inc., Morrisville, North Carolina), which was then folded and placed in a tissue cassette. The block was then submitted to histology for routine processing and hematoxylin-eosin staining.
A total of 7 EUS-FNA operators (endoscopists) performed the initial EUS-FNA procedures at both sites. All procedures were performed using 25-, 22-, or 19-gauge needles. There were a total of 11 pathologists involved in the original diagnostic interpretation at both sites. The smears and cell block materials from all cases were evaluated for overall cellularity by a single cytopathologist (S.N.). Cellularity was scored on a scale of 1 to 4 in an effort to quantitate the amount of lesional cells in the specimen: score of 1 for fewer than 50 lesional cells (Figure 1, A), score of 2 for 50 to 100 lesional cells (Figure 1, B), score of 3 for 100 to 200 lesional cells (Figure 1, C), and score of 4 for more than 200 lesional cells (Figure 1, D). Cell blocks with scores 3 or 4 (>100 cells) were deemed adequate for ancillary studies. This study was approved by the institutional quality assurance committee.
A total of 543 pancreas FNA specimens were submitted to the cytology laboratory during the time period of this study, 453 of which were from UPMC Presbyterian and 90 from UPMC Shadyside. Of these, 270 were cystic lesions and 273 were mass lesions. Of the mass lesions, 216 lesions were evaluated at UPMC Presbyterian and 57 at UPMC Shadyside. Of the 273 pancreas FNA specimens from mass lesions, 231 were positive for malignancy, and 42 cases (15%) received an indeterminate diagnosis. After exclusion because of indeterminate diagnosis, lack of slide availability, presence of lymphoma diagnosis, and rare metastatic tumors, a total of 169 cases were available for evaluation (Figure 2). Adenocarcinoma was the predominant diagnosis (in 88% of cases), and other diagnoses such as neuroendocrine tumor, solid pseudopapillary neoplasm, and acinar cell carcinoma comprised a smaller proportion (12%).
Most procedures were performed using 25-gauge needles (n = 140, 82%), with fewer cases using 22-gauge (n = 23, 13%) and 19-gauge (n = 6, 4%) needles. Clinicopathologic variables evaluated in correlation with cell block cellularity included lesional characteristics such as mass size and diagnosis (adenocarcinoma versus other) and technical characteristics such as number of passes, FNA needle gauge, presence or absence of on-site adequacy evaluation, smear cellularity, and EUS-FNA operator/endoscopist.
Overall, a relatively small proportion, 12.4% (21 of 169), of the positive cases had a cell block cellularity score that was considered adequate for ancillary theranostic studies (score of 3 or 4; see Figure 3, A). Cell block cellularity score was not associated with on-site evaluation (Figure 3, B), the gauge of the needle (Figure 3, C), or the number of passes (mean number of passes was 4.6 and 5.0 at each hospital, with a mean of 4.7 passes for adequacy scores of 0, 1, and 2 and mean of 4.3 passes for adequacy scores of 3 and 4; Figure 3, D). Despite the fact that adenocarcinoma was the most common diagnosis, fewer adequate cell blocks had a diagnosis of adenocarcinoma (66.7% in adequate category versus 90.5% in inadequate category, with P = .006; Figure 3, E). Similar to cell blocks, the majority (88%) of smears had inadequate cellularity scores of 1 or 2. Although low in number, cases that had adequate smear cellularity of 3 or 4 tended to also have adequate cell block cellularity scores of 3 or 4; P < .001 (Figure 3, F). Best versus worst cellularity as well as mean cell block cellularity showed correlation with endoscopists, P = .04 (Figure 4).
Thirty-six patients during this time period had subsequent pancreatic resections. Ten of the patients with resections received neoadjuvant therapy and as such were excluded from analysis of histomorphologic findings. The remaining 26 cases included 10 well-differentiated neuroendocrine tumors and 16 adenocarcinomas, undifferentiated carcinomas, and neuroendocrine carcinomas. Tumors that yielded high cell block cellularity were slightly larger compared with low-cell block cellularity tumors (5.4 versus 3.4 cm), but this did not quite reach statistical significance (P = .09; Figure 5, A).
Nineteen tumors (73%) had increased fibrosis, as judged qualitatively as more collagen then cells, and only 3 (16%) of the tumors with increased fibrosis yielded high (adequate) cell block cellularity (scores of 3 or 4; Figure 5, B). Conversely, only 7 tumors (27%) had low fibrosis, as judged qualitatively as more cells then collagen, and 3 of the tumors with low fibrosis (43%) yielded high (adequate) cell block cellularity, although this was not statistically significant. Because neuroendocrine tumors are typically cellular neoplasms, when they were analyzed separately, a size difference was detected (tumors that yielded high-cellularity cell blocks were larger, P = .02; Figure 5, A). In addition, all of the neuroendocrine tumors (n = 3) that yielded high-cellularity 926 Arch Pathol Lab Med--Vol 138, July 2014 cell blocks had low fibrosis, but this did not reach statistical significance, likely because of small sample size (not shown). No differences were detected for adenocarcinomas only.
As part of departmental quality assurance activity, this information was shared with both cytopathologists and Adequacy for Theranostic Testing--Navina et al clinical gastroenterologists at our institutions. A decision was made by the clinical gastroenterology team to standardize the approach of sampling mass lesions, as described in "Materials and Methods." The new approach included aspiration using 22- and 25-gauge needles and performing 5 passes total without suction syringe. Then, droplets of aspiration fluid from the first 2 passes were used to make 4 slides total, and the rest (pushing stylet back in to express material) was submitted for cell block preparations. Finally, 3 additional passes were performed solely for cell block material. As follow-up, cell block adequacy information from a subsequent time period (January-March 2012) was gathered to assess potential changes in cellularity. During this time, a total of 93 pancreas mass specimens were procured by EUS-FNA. Of these, 75 were positive for malignancy by cytology diagnosis, and 18 received an indeterminate diagnosis (suspicious/atypical category). Cell block cellularity distribution was similar to that in the initial time frame of the study (Figure 6): 9 of 75 positive cases (12%) showed score 3 cellularity, 37 of 75 (49%) had score 1 to 2 cellularity, and 29 of 75 (39%) had score 0 cellularity. The needle gauge distribution was 32 cases with 25-gauge and 42 cases with 22-gauge needles. Of note, ProCore ultrasound biopsy needles were used for 52 cases (22-gauge needles were used in 42 cases and 25-gauge in 10 cases); there was no difference in cell block adequacy when the ProCore needles were used (P > .99).
The workup of patients with suspected pancreatic malignancies has traditionally included EUS-FNA to obtain cells for a definitive diagnosis. One recent meta-analysis concluded that EUS-FNA is a highly accurate diagnostic test for solid lesions of the pancreas and should be included in algorithms for the management of patients with solid pancreatic neoplasms. (14) With the advent of personalized medicine and the desire to perform testing on smaller samples obtained by minimally invasive techniques, there is now a need for the cytopathologist not only to provide an accurate diagnosis, but to assess cytology specimens for adequacy for potential ancillary theranostic studies. (16) Because pancreatic cancer is a rare disease, personalized medicine has not yet been established as it has, for example, in colorectal cancer. However, biomarker profiling of pancreatic cancer tissue for therapeutic efficacy is currently offered commercially, new clinical trials will likely incorporate tumor profiling into their algorithms, (16) and assessment of key genes (such as CDKN2A/p16, TP53, and SMAD4) may provide a prognostic tool that can direct optimal therapeutic strategies for patients. (17)
Although numerous papers in the cytology literature address the assessment of adequacy for diagnosis, only recently have studies begun to address the need to procure enough tissue for ancillary theranostic/molecular studies. (15,18) With the knowledge that the primary focus of our clinical FNA practice is on procuring sufficient material for a definitive diagnosis, we set out to investigate whether a different approach would be needed to ensure that adequate tumor is available for ancillary testing. In our study, which took place in a high-volume hospital-based practice, 73% of all FNA specimens of solid pancreatic lesions received a diagnosis of a primary epithelial malignancy. Of all cases with a diagnosis of a solid primary epithelial malignancy, only 12.4% had adequate cell block cellularity for possible ancillary studies. In a similar study of routine endobronchial EUS-FNA procedures that were positive for malignancy, a higher rate (43%) of cell block adequacy for potential ancillary studies was reported, (18) but this rate is still suboptimal.
Our study indicates that a combination of technical variables (pathologist's involvement/on-site evaluation), operator skills/experience, and tumor type may all impact specimen and cell block cellularity, which in turn could impact adequacy for important ancillary, theranostic, and/or molecular tests. For example, we found no association with on-site evaluation. This may be related to the role of the onsite evaluation, which is for the pathologist to inform the endoscopist when enough cells are obtained for a diagnosis, hence ending the procedure rather than allowing for additional passes for additional material. We did, however, find an association between cell block adequacy and the endoscopist doing the procedure, an observation that may be related to operator experience (this was not further examined in the current study). Although adenocarcinoma was overall the most common diagnosis, it comprised a significantly lower proportion of adequate cases, suggesting that some lesional characteristics may affect aspirate cell block cellularity. Based on the cytology-histology correlation findings, it appears that tumor characteristics such as tumor size and presence or amount of fibrosis may potentially impact the cellular yield of FNA, but further study is needed.
Changes in practice, such as obtaining dedicated passes for ancillary studies or using currently available biopsy needles, may not be enough to improve the theranostic utility of EUS-FNA in pancreatic neoplasia. Two recent studies found similar performance for diagnostic adequacy when standard FNA needles were compared with ultra sound biopsy needles, and both studies noted that the ultrasound biopsy needles did not increase the yield of material in cell blocks for histologic analysis. (19,20) Other methods to improve tumor cell yield, such as novel cell block preparations, use of direct smears, (18) and incorporation of filter paper cards, (21) need to be further investigated. In addition, EUS-FNA needle designs that reliably provide histologic core samples would be greatly welcomed.
Please Note: Illustration(s) are not available due to copyright restrictions.
(1.) American Cancer Society. Cancer facts & figures. http://www.cancer.org/ Research/CancerFactsFigures/index. Accessed January 5, 2013.
(2.) Bilimoria KY, Bentrem DJ, Ko CY, Stewart AK, Winchester DP, Talamonti MS. National failure to operate on early stage pancreatic cancer. Ann Surg. 2007; 246(2):173-180.
(3.) Cameron JL, Crist DW, Sitzmann JV, et al. Factors influencing survival after pancreaticoduodenectomy for pancreatic cancer. Am J Surg. 1991; 161(1):120-124.
(4.) Glimelius B, Hoffman K, Sjoden PO, et al. Chemotherapy improves survival and quality of life in advanced pancreatic and biliary cancer. Ann Oncol. 1996; 7(6):593-600.
(5.) Moore MJ. Pancreatic cancer: what the oncologist can offer for palliation. Can J Gastroenterol. 2002; 16(2):121-124.
(6.) Katz MH, Pisters PW, Evans DB, et al. Borderline resectable pancreatic cancer: the importance of this emerging stage of disease. J Am Coll Surg. 2008; 206(5):833-846; discussion 846-848.
(7.) Stokes JB, Nolan NJ, Stelow EB, et al. Preoperative capecitabine and concurrent radiation for borderline resectable pancreatic cancer. Ann Surg Oncol. 2011; 18(3):619-627.
(8.) National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology: pancreatic adenocarcinoma. Version I.2011. Published 2011. http://www.nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed February 4, 2011.
(9.) Le Scodan R, Mornex F, Girard N, et al. Preoperative chemoradiation in potentially resectable pancreatic adenocarcinoma: feasibility, treatment effect evaluation and prognostic factors, analysis of the SFRO-FFCD 9704 trial and literature review. Ann Oncol. 2009; 20(8):1387-1396.
(10.) Varadhachary GR, Wolff RA, Crane CH, et al. Preoperative gemcitabine and cisplatin followed by gemcitabine-based chemoradiation for resectable adenocarcinoma of the pancreatic head. J Clin Oncol. 2008; 26(21):3487-3495.
(11.) Warsame R, Grothey A. Treatment options for advanced pancreatic cancer: a review. Expert Rev Anticancer Ther. 2012; 12(10):1327-1336.
(12.) Sarris EG, Syrigos KN, Saif MW. Novel agents and future prospects in the treatment of pancreatic adenocarcinoma. JOP. 2013; 14(4):395-400.
(13.) Crane CH, lacobuzio-Donahue CA. Keys to personalized care in pancreatic oncology. J Clin Oncol. 2012; 30(33):4049-4050.
(14.) Hewitt MJ, McPhail MJ, Possamai L, Dhar A, Vlavianos P, Monahan KJ. EUS-guided FNA for diagnosis of solid pancreatic neoplasms: a meta-analysis. Gastrointest Endosc. 2012; 75(2):319-331.
(15.) Aisner DL, Sams SB. The role of cytology specimens in molecular testing of solid tumors: techniques, limitations, and opportunities. Diagn Cytopathol. 2012; 40(6):511-524.
(16.) Hidalgo M, Von Hoff DD. Translational therapeutic opportunities in ductal adenocarcinoma of the pancreas. Clin Cancer Res. 2012; 18(16):4249-4256.
(17.) Oshima M, Okano K, Muraki S, et al. Immunohistochemically detected expression of 3 major genes (CDKN2A/p16, TP53, and SMAD4/DPC4) strongly predicts survival in patients with resectable pancreatic cancer. Ann Surg. 2013; 258(2):336-346.
(18.) Knoepp SM, Roh MH. Ancillary techniques on direct-smear aspirate slides: a significant evolution for cytopathology techniques. Cancer Cytopathol. 2013; 121(3):120-128.
(19.) Bang JY, Hebert-Magee S, Trevino J, Ramesh J, Varadarajulu S. Randomized trial comparing the 22-gauge aspiration and 22-gauge biopsy needles for EUS-guided sampling of solid pancreatic mass lesions. Gastrointest Endosc. 2012; 76(2):321-327.
(20.) Witt BL, Adler DG, Hilden K, Layfield LJ. A comparative needle study: EUS-FNA procedures using the HD ProCore and EchoTip 22-gauge needle types [published online ahead of print March 20, 2013]. Diagn Cytopathol. doi:10. 1002/dc.22971
(21.) da Cunha Santos G, Saieg MA, Geddie WR, Kamel-Reid S. Cytological preparations for molecular pathology: letter to the editor regarding "Ancillary techniques on direct-smear aspirate slides: a significant evolution for cytopathology techniques." Cancer Cytopathol. 2013; 121(5):275.
Sarah Navina, MD; Kevin McGrath, MD; Jennifer Chennat, MD; Vijay Singh, MD; Timothy Pal, MD; Herb Zeh, MD; Alyssa M Krasinskas, MD
Accepted for publication August 12, 2013.
From the Departments of Pathology (Drs Navina, Pal, and Krasinskas) and Surgery (Dr Zeh) and the Division of Gastroenterology, Department of Medicine (Drs McGrath, Chennat, and Singh), University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania. Dr Krasinskas is now with the Department of Pathology, Emory University, Atlanta, Georgia.
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Alyssa M. Krasinskas, MD, Department of Pathology, Emory University, 1364 Clifton Rd NE, Room H-180, Atlanta, GA 30322 (e-mail: email@example.com).
Caption: Figure 1. Cell block cellularity scores. Cell block cellularity was graded on a scale of 1 to 4 in an effort to quantitate the amount of lesional cells in the specimen. A, Score 1 represents cell blocks with fewer than 50 lesional cells. B, Score 2 for 50 to 100 lesional cells. C, Score 3 for 100 to 200 lesional cells. D, Score 4 for more than 200 lesional cells (hematoxylin-eosin, original magnifications X40).
Caption: Figure 2. Selection of cases for this study. A total of 169 cases were evaluated for cell block cellularity and only 21 cases (12.4%) were deemed adequate for ancillary studies. Abbreviation: FNA, fine-needle
Caption: Figure 3. Cell block cellularity. Scores 0, 1, and 2 are deemed inadequate, and scores 3 and 4 are deemed adequate for ancillary testing. A, All epithelial neoplastic diagnoses. Only 21 (12.4%) of the 169 positive cases had an adequate cell block cellularity score of 3 or 4. B, On-site evaluation. Although more on-site evaluations were performed at hospital 2 (H2; University of Pittsburgh Medical Center [UPMC] Shadyside) than hospital 1 (H1; UPMC Presbyterian), there was no association with cell block cellularity and on-site evaluation. C, Needle gauge. Cell block cellularity was not associated with needle gauge. Abbreviation: G, gauge. D, Number of passes. There was no association between cell block cellularity and mean number of passes per case. E, Diagnosis of adenocarcinoma compared with other epithelial neoplasms. Most cases (90.5%) with a diagnosis of adenocarcinoma had inadequate cell block cellularity (only 9.5% were adequate). Of the other epithelial neoplasms (neuroendocrine, solid pseudopapillary tumor, acinar cell carcinoma), there were proportionally fewer cases that had inadequate cell block cellularity (66.6% inadequate and 33.3% adequate); P =.006 for adenocarcinoma versus other neoplasms. F, Smear cellularity. Adequate smear cellularity correlated positively with adequate cell block cellularity. That is, cases with inadequate smear cellularity also had inadequate cell block cellularity, and conversely, cases with cellular smears also had cellular cell blocks; P < .001.
Caption: Figure 4. Correlation of cell block cellularity with individual endoscopists. The best versus worst cell block cellularity, as well as mean cell block cellularity, showed a correlation with endoscopists (P = .04 for endoscopist 1 versus endoscopist 3, respectively). Cell block cellularity scores of 0, 1, and 2 are deemed inadequate, and scores of 3 and 4 are deemed adequate for ancillary testing. Note: One endoscopist only had 2 cases, so this endoscopist was omitted from this figure.
Caption: Figure 5. Surgically resected tumors and correlation with cell block cellularity score. A, Tumors that yielded adequate cell block cellularity (scores 3 and 4) were larger than those that yielded low cell block cellularity; this difference reached statistical significance for neuroendocrine tumors (NET) (*P = .34; **P = .02). B, Tumors that had increased fibrosis tended to yield lower cell block cellularity, but this did not reach statistical significance, even when neuroendocrine tumors were analyzed separately (possibly because of small sample size).
Caption: Figure 6. Cell block cellularity based on the initial retrospective review of 169 cases (Before) compared with cell block cellularity obtained once a standardized approach was implemented on 75 cases (After). There was no difference in cell block cellularity after a standardized approach was implemented. Cell block cellularity scores of 0, 1, and 2 are deemed inadequate, and scores of 3 and 4 are deemed adequate for ancillary testing.