EXPERT’S CORNER: Evolution of uncertainty in surgical pathology. Part 1: Recognition & Acceptance

Paul C. Stromberg DVM, PhD, DACVP
Professor Emeritus, Ohio State University

It is human nature to desire certainty in our affairs. We want to see a world of binary choices; of issues that are black and white without shades of grey. If we can see the data clearly and unambiguously we think this leads to confidence in our judgment. This is especially true for health care professionals because we feel the responsibility to make an accurate diagnosis. Anything that simplifies decision making and judgment supports our confidence and seemingly certainty that we are correct. But certainty in medicine is an illusion and the sooner we abandon the pursuit of this, the sooner we will be more effective diagnosticians.

In his insightful book “How Doctors Think” (Houghton Mifflin Company, 2007) Jerome Groopman states. “…every doctor is fallible. No doctor is right all the time. Every physician, even the most brilliant, makes a misdiagnosis or chooses the wrong therapy.” This is intuitively true for veterinarians as well as physicians and that of course includes veterinary pathologists. However, the majority of the errors are not caused by lack of medical facts but how we think about what we see. This is especially problematic for pathologists (and radiologists) whose principle clinical skill is visual pattern recognition, a highly subjective, experience-based ability to observe patterns of pathologic processes and interpret them. It is not an objective evaluation of quantitative data generated by a machine and reported in metric units with a reference range for normal. A hematocrit of 32 with a reference range of 40-50 can be confidently interpreted to be slightly anemic. But the visual pattern of cells in a lymph node is either malignant lymphoma or it’s not. It’s never “slightly lymphoma.” You can’t hedge. So dealing with uncertainty in pathologic evaluation challenges our confidence which in turn can affect our judgment. We have a unique skill possessed by no one else in the biomedical community but the mastering of it is a long arduous journey that never ends (“Life is short, the art is long”). The best example of pattern recognition is the approach used in dermatopathology recently reviewed in Veterinary Pathology (Vol 60. No. 6, November 2023). But the idea of recognizing patterns of pathologic processes is useful in other organs and tissues and at the gross, subgross and ultrastructural levels as well. Although it’s not as well developed as in dermatopathology, the process is essentially the same. It’s the observation and recognition of one or several abnormal anatomic changes seen in the tissue that when assembled constitutes a definition of a disease pattern that we call a morphologic diagnosis (i.e. a type of inflammation, a neoplasm, a degenerative change) and which we ultimately connect to a clinical disease entity. The rationale for the histopathology cases and essays on certification examinations is to test how well the candidate can recognize the various histologic changes and assemble them into a morphologic diagnosis (es) then relate it to a clinical disease. There is great subjectivity in our judgment of patterns which can lead to uncertainty in the process and it requires constant reinforcement. Generally I am interpreting biopsies 5-6 days a week. I have noticed when I am away from my microscope for a couple of weeks I am not as sharp the first day when I resume reading cases. It takes me a bit longer to work through the cases. My uncertainty is greater. I have to be more deliberate in my judgments. But within a day, I return to my normal state. Much of this is the uncertainty in what we see that accompanies lack of daily reinforcement.

We all began our exploration of pathology in residency training or earlier which was fraught with uncertainty as we began to learn the various visual patterns of different disease processes. It was (and still is) a bewildering collection of visual images gleaned from formal lectures reinforced by case material from the autopsy and surgical biopsy services supplemented by seminars, textbooks, research papers and study sets. In the main we learned by spending many hours over a double-headed microscope with an expert pathologist who guided us through the patterns; a very effective but highly inefficient method of learning. The learning was complicated by two foundational principles in anatomic pathology which make the process more difficult, further fueling our uncertainty. One, lesions have a range of expression.

Not every case of pemphigus or toxoplasmosis looks alike. Not every histiocytoma or mast cell tumor looks the same. Cirrhosis, pneumonia, enteritis all present with varying patterns. I think of this as a bell curve with the common classic “textbook” lesion as the mean in the middle and the variations from that on the shoulders of the curve. As our experience grew we learned to sift through the variable nature of pathologic entities and build confidence in our abilities to recognize not only the classic patterns but the variations around them. But with broadening experience we became aware of a second principle of anatomic pathology; that the variable lesion patterns of different pathologic entities may overlapmaking it difficult to confidently differentiate them. Histiocytomas and reactive histiocytosis look very similar. Some spindle cell tumors can look like immature granulation tissue. It is the ability to operate confidently in the ranges of variation and lesion pattern overlap that marks the expert or experienced pathologist.

We worked hard to master this unique skill in preparation for the certification examination and our eventual career. Success brought a new professional status and the confidence it engenders. We had arrived. We’re Diplomates. We’re experts! But at that point we entered “The Danger Zone,” a space in which we worked with supreme confidence in our ability but impunity with respect to error because we had not yet experienced a major mistake. None of us wants to admit error (and we still don’t want to confront it in ourselves!). Time in the Danger Zone is highly variable among individuals but eventually we all have “The Crisis” which ends the illusion of experience and certainty. For me it was the misdiagnosis of a nasal fibrosarcoma in a dog that was actually granulation tissue. I was depressed for a week. Thought I was incompetent; did not deserve to be a member of the ACVP. How could I have done this? The feeling was enhanced by a well-meaning mentor who wanted to reinforce the lesson by adding to my guilt and depression. I eventually got over this and accepted the knowledge that error was possible if I was not careful. I had emerged from the “Danger Zone.” Three months later, another illusion about experience was burst when my mentor made the exact same error of diagnosis in the nasal cavity of a rat. Experts are fallible. Experience is no guarantor against mistakes as Dr. Groopman tells us. (“Experience delusive”).

Several times since, I subjected myself to an informal audit to get an estimate of my potential for error. Both times my error rate was estimated to be in the 1% range; not bad, huh? But considering I was reading about 10,000 cases per year that amounted to about 100 erroneous diagnoses per year or about 8-9/month. Admittedly some of the errors were moot. Whether it’s a sebaceous adenoma or sebaceous hyperplasia does not impact the clinical course. Nevertheless, 8-9 errors/month seemed to be too many. But I wanted to better manage it somehow and that led to the issue of uncertainty in interpreting pathologic processes, how it can impact judgment and lead to diagnostic errors.

Everyone must learn to practice with uncertainty if you are going to maintain your effectiveness as a pathologist. For you younger folks, uncertainty does not go away as you gain experience. You just get used to it. Uncertainty is with us all the time but it can be managed. Remember it’s not ignorance of medical facts but how you think about what you see that is the problem. The majority of your cases will be straight forward and present little uncertainty. But it’s that 10-15% of cases which initially seem confusing that make the difference in your performance. A rock solid principle of science and medicine
is to value or weight objective data more than subjective data. But this is a problem for the surgical pathologist who may only have a small portion of tissue with an image pattern that may overlap with the image pattern of other pathological entities. You often get little objective information in a biopsy submission such as history, physical findings, a description of the gross lesions and what the clinician thinks; information that could reinforce your judgment and reduce your uncertainty. Histiocytomas can be highly variable and by themselves can generate considerable uncertainty. But if you are told it’s a solitary 1cm circular, red, raised ulcerated “Button lesion” on the ear of a 6-month old boxer, that supports your diagnosis and decreases your uncertainty. This is quite different from autopsy pathology where you have most of the facts about the case and the luxury of seeing the entire patient and the gross lesions.

In Part Two of this essay I will discuss how to manage your uncertainty and the steps you can take to increase your confidence, improve your judgment and potentiate the correct diagnosis. Part of the solution is to educate veterinary students, and clinicians about the importance of their role in the surgical biopsy process. Hint: It’s not just collecting a piece of tissue, putting it in a jar and waiting for the report.