The diagnosis of pigmented skin lesions is amongst the most challenging and potentially dangerous areas of skin cancer medicine and dermatopathology. In this article, we discuss the limitations of histopathological diagnosis, and point to ways that the clinician may be able to assist the pathologist to minimise adverse outcomes.

Reliability of pathological diagnosis

To a large extent, the pathological diagnosis of melanocytic lesions is subjective. Most lesions can be readily and reliably diagnosed by an experienced pathologist, but some lesions may possess morphological features that are not stereotypical. In such cases, different observers may render different diagnoses and, in a small number of cases, the diagnoses may range between benignity, on the one hand, and malignancy on the other. The extent of agreement between observers (or for a single observer making a number of observations) can be measured using a statistical tool called a kappa coefficient, where κ = 1 is perfect agreement, κ = −1 is perfect disagreement and κ = 0 is chance agreement. A number of studies have examined the reliability of pathologists in reporting pigmented skin lesions. Although almost all of these studies tend to be distorted because of case selection (with most papers studying ‘difficult cases’), the results are worth noting. In distinguishing melanoma from benign lesions, kappa coefficients around 0.5 (‘moderate agreement’) are reported. In one study, where cases were not selected, a kappa coefficient of 0.6 was achieved and, in more concrete terms, this meant that in 25% of cases, at least one of the four dermatopathologists in the study disagreed with the diagnosis of the remainder. Other studies have a narrower focus. In dividing naevi into dysplastic, banal or intermediate categories, κ = 0.34 (‘fair agreement’), and in grading dysplasia, κ = 0.2–0.4 (‘slight to fair agreement’). These disappointing results mean that, in practice, expert dermatopathologists agree on the degree of dysplasia in less than 50% of cases. Difficulties also arise in the diagnosis of ‘Spitzoid’ lesions. This rubric encompasses Spitz naevi, so-called ‘atypical Spitz naevi’ and melanomas with morphological features similar to those seen in Spitz naevi. In one study, world expert dermatopathologists were unable to agree on cases of Spitz naevi, except in one case. Unfortunately, this case ultimately proved to be a fatal melanoma. Other areas of poor diagnostic concordance relate to the diagnosis of melanoma in childhood, and in the diagnosis of cellular blue naevus.

The clinician’s role in assisting the pathologist

The problems which might arise from the inability of even expert pathologists to render reliable diagnoses in a subset of cases can be ameliorated by an insightful and prudent clinician. The surgeon must provide adequate clinical information and an adequate specimen. The importance of clinical information cannot be overstated. Professor Peter Soyer, from the University of Queensland, and colleagues, showed that incremental increases in submitted clinical information improved the concordance of diagnoses rendered by a group of expert dermatopathologists. Moreover, the pathologists reported a substantial increase in the level of confidence in the diagnosis. Unfortunately, it remains the case that, in Australia and New Zealand, useful clinical information is not provided in around one third of specimen submissions. In addition to the site and demographic details, the most pertinent information includes the clinical history, the size of the lesion and the clinical index of suspicion. Increasingly, more experienced clinicians provide dermatoscopic descriptions and photographs, and this additional information can be invaluable. Biopsy technique has a major impact in the reliability of dermatopathological diagnoses. As a general rule, any lesion suspected of being a melanoma, and where the major differential diagnosis is a naevus, should be excised. This is the case because the features which permit clinical diagnosis of melanoma have microscopic correlates which the pathologist uses to establish the diagnosis. In particular, the symmetry of the lesion, the circumscription of its borders and the regularity in which the melanocytes are disposed, all require substantial breadth in a specimen for assessment. Pathologists recognise that, on occasion, excisional biopsy is not feasible, but clinicians must be mindful that in parte biopsy specimens are associated with a very substantial increase in false negative and false positive diagnoses, and that interpretation of the pathology report should be tempered with this knowledge. In a recent study from Melbourne, shave biopsies were associated with a 2.5 times risk of misdiagnosis, compared with excisional biopsies, and punch biopsies, 17 times. False negative diagnoses with an adverse outcome were 20 times more likely with (partial) punch biopsies. It follows that the use of partial punch biopsies for diagnosis of melanocytic lesions is potentially dangerous, and this technique is inappropriate in most instances. Examples where their use might be appropriate are the biopsy of a papule to distinguish a dermal naevus from a small nodular basal cell carcinoma, and the biopsy of an irregular pigmented macule on the face, where the diagnosis is either a solar lentigo or melanoma. In both of these examples, the diagnosis can be made by identifying the cellular constituents of the lesion, that is, there is no requirement accurately to assess the architecture. The expert clinician cannot be a passive recipient of a pathology report. Often, reports will have diagnoses couched in terms like ‘suggestive of’, ‘features favouring’ or similar phrases, which reflect the pathologist’s lack of certainty in the diagnosis. In such circumstances, there is an obligation on the clinician either to clarify the diagnosis with the pathologist, or to ensure that an equivocating diagnosis will not adversely affect the patient. In some circumstances, this might require that an ambiguous lesion is completely excised, to ensure that it is extirpated. Where the pathological diagnosis is discordant with the clinical diagnosis, the clinician may wish to reconcile the diagnoses by asking the pathologist to review the slides, to examine additional tissue or to seek a further opinion. In Australia, pathology laboratories are obliged to retain pathology slides, tissue and reports for a minimum of seven years, and review of archival material is useful in select cases. This lesion was diagnosed as dysplastic naevus by four pathologists and melanoma by three pathologists. Benign and malignant melanocytic lesions may have features which overlap.

Conclusion

In summary, the diagnosis of naevi and melanomas presents considerable difficulties for the pathologist. In the majority of cases, the diagnosis is straightforward, but in a significant minority of cases, pathological diagnosis is unreliable. Part of the unreliability is an inevitable consequence of the subjective nature of histopathological diagnosis, where different morphological features are given different weights by different observers. However, at least in some circumstances, and using current dermatopathological techniques, morphology is not a perfect predictor of biology. Consequently, the clinician must assess every pathology report in concert with the other clinical information about the patient and about the lesion.

---

General Practice Pathology is a new fortnightly column each authored by an Australian expert pathologist on a topic of particular relevance and interest to practising GPs. The authors provide this editorial, free of charge as part of an educational initiative developed and coordinated by Sonic Pathology.