Insights From: John G. Gosney, MD, FRCPath, Royal Liverpool and Broadgreen University Hospital NHS Trust
John R. Gosney, MD, FRCPath: As pathologists, of course, we have no control over what happens until we receive the specimen. That is entirely dependent on the clinician who takes the biopsy. Hence, the importance of good communication.
However, we do have control of what we do with that tissue once we’ve got it. And 1 of the problems—which we find generally throughout Europe, and indeed in the United States as well—is that a lot of the initial diagnostic work-up of lung cancer specimens is in peripheral hospitals, district general hospitals, where nonexpert pathologists are required to make an initial diagnosis. And because they lack the expertise and the confidence, they often do unnecessary and extensive immunochemistry to classify the tumor, to exclude the possibility, even if it’s minimal that the tumor is metastatic and so on. And this uses tissue, so that when those specimens are referred centrally for the genomic profiling—which these days is the crucial information—there is insufficient for analysis.
So we as pathologists need to be aware of that problem. Don’t waste tissue. This is a precious resource. Whatever we receive from the clinician—whether it’s good or not so good—we need to treasure it and do the best with it, and we should not waste it doing unnecessary tests that compromise the profiling that counts.
The use of cytology specimens in Europe is widespread. And I think it’s important to realize first of all about the difference between what we tend to consider histology and cytology, and whether it’s in fact more imagined than real. When you think about it, a histology specimen is a piece of tissue where they structure, so that in the case of a non–small cell lung cancer, for example, you have the cells of the tumor arranged in 3 dimensions in a stroma. They are in context.
A cytology specimen essentially consists of those cells pulled out of their context and dispersed; that’s the essential difference. But I don’t think there is a fundamental difference. And providing what you’re looking for is integral to those individual cells. It really shouldn’t matter whether they are in their context, in the stroma, or whether they dispersed. And indeed, cytology specimens have been widely used for the last decade to look for single-driver genomic abnormalities and EGFR mutations with no trouble at all.
And especially with a good aspirate, the failure rates, because of inadequate numbers of cells or inadequate DNA, is no greater than that for histology specimens. There has been a rather a reluctance to use cytology specimens for PD-L1 [programmed death-ligand 1] expression, and this has arisen from the fact that cytology was not used in the clinical trials of the immune modulators that have now come to market, nor is it used by the diagnostics companies in developing their immunochemical assays. And that’s led to an understandable reluctance to promote cytology as a median.
However, if you look at the literature, you will see that there is no good evidence from any study, and there are now 10, maybe 15, over the last 2 years alone, to show that cytology is in any way inferior to histology as a medium for looking for PD-L1 expression. There are worries about fixation. There is evidence now to show that whether you fix in alcohol or formalin really doesn’t matter. And I think the only problem in reality is that they are tricky to interpret. You need a good pathologist with a lot of experience to assess PD-L1 expression in a cytology specimen. But notwithstanding that, cytology is an excellent substrate for anything that you might want to do with it when you’re profiling on small cell lung cancer.