Symposium: Expert Insights into Pancreaticobiliary Endotherapy
Chicago, Illinois, May 2014*
Dr. Sahai’s Presentation Topic:
Cytology vs histology?
Variables Impacting Cytology Yield
Maximizing Diagnostic Yield and Getting a Core
Q&A with Dr. Sahai
*Sponsored by Boston Scientific Corporation
ANAND V. SAHAI, MD: So when I say pancreatic diagnostics, basically, what I want to talk about is tissue acquisition. What's great about these sort of forums, for me, is that it's a little less formal, and this is basically my opinion-- a lot of it-- about where tissue acquisition should be going. A certain amount of data, but with my biased interpretation of the data. So we'll start with just the whole idea of cytology versus histology. There's a lot of interest these days about getting histology instead of cytology for EUS-FNA. In my opinion-- and what we do, basically-- cytology is all you need for about 95% of what you're going to be doing in basic EUS, which is mostly solid tumors, metastatic lymph nodes, epithelial cancers, pancreatic tumors. We just make two slides. When I travel, I see people making tons of slides-- all different ways to prepare slides. I think two slides is enough. I think that the whole idea of getting histology is a little bit moot in the sense that you can also get a cell block, which is where you put your whole sample into alcohol. It's spun down into a pellet, then it's fixed in paraffin, then it's cut like histology slides. So again, a cytologist that doesn't like-- sorry, a pathology that doesn't like doing cytology-- you can convert them basically into histology and get basically what I call sort of a pseudo histology. And I think as a general rule here, and if you're looking at all the data on the EUS-FNA, this year at DDW there are literally, I think, 30 papers looking at different ways to do EUS-FNA. I think in your practice, you should be getting at least overall 90% sensitivity for cancer-- pancreatic-- a little higher maybe for lymph nodes. But I think if you're getting about 90%, you're doing basically what the best people-- or the most experienced people-- in the world are doing. I based that also-- if you look at this pool analysis on the EUS-FNA accuracy-- you look at the studies-- there are some people who are claiming they're getting like 96% sensitivity, this and that. Honestly, I find that a little bit hard to believe, unless they're doing only lesions that are extremely likely to be cancer. You know that sensitivity can be affected by the severity of the disease, so if you're only doing huge 4 cm pancreatic masses, well, then your sensitivity will probably be quite high. But if you're doing a good mix of hard lesions, small lesions, regular pancreatic masses, overall, when you look at the combined data, you should be getting about 90% as an overall yield. OK, hold on. What have I-- escape, there we go. These Windows computers. OK, so cytology or histology, basically, if you look at them-- if you compare them both-- the medium is different-- one is alcohol, one is formalin. They both get fixed in paraffin. You can do immunostaining on both of them. I do believe, in most cases, you get enough material to do tumor markers and a genetic analysis if you need to. There's a lot of interest in that these days. Honestly, in real life-- certainly in Canada with our, I guess, third-world health care system-- all this fancy, genetic markers-- when it comes down to pancreatic cancer and lung cancer-- certainly pancreatic cancer, anyway-- I don't think it's going to make much of a difference. Bottom line is, the five-year survival is still less than 5%, and whether-- we're improving that a little bit, but all this genetic markers so far isn't making a big difference. I do believe you can get a mitotic index. You can still get an on-site read. The only difference really with a cell block is that it's sort of mushed up cytology. You don't get the true tissue structure, whereas if you do get a core biopsy, you can get tissue structure. And that's probably useful only in lymphoma. Certainly in our experience, that's the one time you really need tissue structure.