Since “nothing can be said to be certain, except death and taxes”1, pathologists are frequently presented with cases that are probably, but not certainly, a specific entity. However, typical pathology report structures, either traditional or synoptic, do not do a good job of conveying our level of confidence in the diagnosis.
Why does this matter?
This matters for several reasons. First, and most important, the level of certainty in the diagnosis has a direct bearing on patient treatment2. Some level of certainty is required before initiating expensive, invasive, and potentially harmful treatment. Conversely, expensive diagnostic tests may not be run if there is a high degree of certainty in the diagnosis.
Second, the level of certainty is important when using case materials for retrospective studies. Depending on the study, the initial diagnosis may be relied upon in the selection of case material. If this diagnosis is only tentative, material may be included that does not meet the study criteria, diluting out findings and confounding interpretation of data.
Methods for conveying certainty
While few studies have examined this in veterinary medicine, pathologists generally use three methods for conveying the level of certainty in their diagnosis. First, modifiers may be added to the morphologic diagnoses in the diagnosis section of the report. This may include “presumptive”, “likely”, or another word or phrase indicating doubt. Second, the comments section may be used to qualify the diagnosis; phrases such as “While this is the most likely diagnosis, another diagnosis cannot be completely ruled out” in the comments indicate some degree of doubt. Finally, the submitting clinician may be contacted directly by the pathologist to discuss the level of uncertainty.
However, each of these has limitations. The primary problem is determining what level of certainty is implied by any specific phrase. While this has not yet been examined in veterinary medicine, studies in human medicine3,4 have found that pathologists and clinicians do not always agree on the level of certainty conveyed by a specific word or phrase. Further, there is disagreement between pathologists and clinicians about causes of miscommunication; clinicians see “inconclusive word choice in final diagnosis” as a major cause of miscommunication more than 70% of the time, while pathologists see “misinterpretation of the final diagnosis” as the most common major cause of miscommunication.
This becomes particularly important when moving to synoptic reporting formats, as these attempt to standardize reporting items. As synoptic reports decrease a pathologist’s “personal voice” in reporting, having a method to convey certainty becomes more important for both clinicians and retrospective reports.
Standardization
Most important in this effort is determining a standard method and lexicon for conveying certainty. Pathologists and clinicians must be able to agree on the pathologist’s confidence in a diagnosis, or else the pathology report loses much of its value. A number of techniques for conveying this are possible, and a working group of pathologists and clinicians should determine the method that imposes the least workload on pathologists while conveying the most information to clinicians. This should also be in a format that is readily parsable for future retrospective studies.
A modest proposal
As a closing thought, using a modifier at the start or end of the morphologic diagnosis when a diagnosis is not 95-100% certain seems like the least amount of work. One of the works in human hospitals proposed the scheme “diagnostic of”, “compatible with”, and “suspicious for” in decreasing levels of certainty. This would lead to the following scheme:
| Morphologic Diagnosis | Level of Certainty |
|---|---|
| Mammary carcinoma | 95-100% |
| Diagnostic of mammary carcinoma | 75-95% |
| Compatible with mammary carcinoma | 50-75% |
| Suspicious for mammary carcinoma | Less than 50% |
References
1. Sparks, Jared (1856). “The Writings of Benjamin Franklin, Vol. X (1789-1790)“. Macmillan. p. 410. Retrieved 11 November 2022.
2. Thompson JF, Scolyer RA. Cooperation between surgical oncologists and pathologists: a key element of multidisciplinary care for patients with cancer. Pathology. 2004 Oct;36(5):496-503. doi: 10.1080/00313020412331283897. PMID: 15370122.
3. Amin, A., DeLellis, R.A. & Fava, J.L. Modifying phrases in surgical pathology reports: introduction of Standardized Scheme of Reporting Certainty in Pathology Reports (SSRC-Path). Virchows Arch 479, 1021–1029 (2021). https://doi.org/10.1007/s00428-021-03155-w
4. Blake A. Gibson, Elizabeth McKinnon, Rex C. Bentley, Jeffrey Mohlman, Benjamin L. Witt, Eric J. Yang, Daniel Geisler, Marie DeFrances; Communicating Certainty in Pathology Reports: Interpretation Differences Among Staff Pathologists, Clinicians, and Residents in a Multicenter Study. Arch Pathol Lab Med 1 July 2022; 146 (7): 886–893. https://doi.org/10.5858/arpa.2020-0761-OA
I agree with this concept. Communication of these levels of certainty to both pathologists and clinicians is critical. Also. Is there some distinction between “Morphologic Diagnosis” and “Diagnosis”?
On the topic of certainty – does the use of the term “interpretation” in place of “diagnosis” imply a different level of certainty?
Good questions – I suppose that depends on the reader. I don’t see a difference between morphologic diagnosis and diagnosis, although I suppose I think of morphologic diagnoses as being a little more formalized (less so for neoplasms where there aren’t the descriptors included that we do for inflammation).
I don’t read a difference between “interpretation” and “diagnosis”, but I think many practitioners might – pathologists understand (all too well!) that our diagnoses are just our best interpretations of the case material we have at hand rather than something that is set in stone.