A Pathologist Dilemma, H&E, Immuno, or Molecular?

Until the 80’s we made diagnosis reading H&E slides, we learned to recognize patterns, and the studies of different entities were done based on the evolution of a uniform group of cases characterized by specific symptoms, gross and microscopic features. Excellent studies included 50 o 100 cases with extensive follow-up. This is how we learned about serous carcinoma of the endometrium, invasive vs non-invasive implants of serous borderline tumors, o separating a dedifferentiated endometrial cancer from a grade 2 endometrioid.

In the 90’s we started using immunohistochemistry and it was a great advance. At the beginning we had cytokeratin, and vimentin. It was difficult to publish a case of an unusual carcinoma if the keratin was negative. NSE was a very interesting marker, we should have learned from it. When this marker became available, it was considered to be specific for neuroendocrine tumors, but after many studies, the specificity was lost, and the joke was that NSE became Non-Specific Enolase. After many years working with tumors and immunos we have seen many similar problems. A very common one is how to interpret p53 when it is heterogenous. Often, to render a diagnosis we ignore some markers, or we find some rare reports supporting our interpretation and saying” some cases can be positive or negative”

Now the problem is bigger, because molecular analysis finds mutations that, again, help in many cases, but are very difficult to interpret in other cases.

The way I see this is very simple.

Molecular tests are telling us what the cells are thinking,

Immunohistochemistry tells us what the cells are going to do.

H&E slides confirm what the cells have done.

If we were going to be judged by what we are thinking, or what we would like to do, many of us would have a different life.

How can we solve this dilemma? We need to think about the patient who is going to be treated, and this must be based of solid studies. We cannot render a diagnosis of serous carcinoma of the endometrium when the tumor looks like a low-grade endometrioid but has some areas diffusely positive for p53. It is possible that this case might behave more aggressively but there are not enough studies to justify aggressive chemotherapy in these patients. These studies are not difficult to do, we have archival material, and we already know or can find out the follow-up of the cases. With molecular studies the problem is worst, we are creating new entities with two or three cases. The urgency to publish something is more important than doing good pathology studies. If we do not do studies with many cases and an acceptable follow-up, then the evidence-based studies are going to be replaced with eminence based decisions like “I have seen pilomatrixoma cases without ghost cells”  or ‘cannot be high-grade serous carcinoma because the p53 is wild type”.

Nobody can stop progress, but we should remember that based on our diagnosis patients will be treated and clinicians receive only some words, not the dilemma that we had to render those few words. They must have a solid base.