Patient Detection of Suspicious Skin Lesions
Malignant melanoma (MM) incidence has been rising for the past six decades, with 87,100 cases diagnosed in the United States in 2017.
[1]Over half of melanomas are first detected by patients or their families,
[2] with diagnosis of early-stage, thin MM dramatically improving 10-year survival rates.
[3] These two facts taken together underscore the importance of patient education to encourage early MM detection.
Since its development in 1985, the ABCD rule (A=asymmetry, B=border irregularity, C=color variegation, D=diameter larger than 6 mm) has been taught as a simple pattern recognition tool to distinguish normal melanocytic nevi from melanomas. This simple rule has been a boon to early MM detection but may fail to pick up small or amelanotic MM, even when adding the "E" for "evolving lesion."
[4] In contrast, a newer screening tool instructs patients to compare their cutaneous lesions in search of any outliers—the so-called "ugly duckling" (UD) sign.
[5]The UD rule stems from the notion of the "signature nevus"—that each person forms a characteristic pattern of melanocytic nevi based on genetic determinants such as skin phototype. Hence, by encouraging patients to look for cutaneous lesions that break their typical nevus pattern, the UD sign may show superior sensitivity and/or specificity for MM detection when compared with the more static ABCD rule.
Which Detection Rule Is Best?
To test this hypothesis, Ilyas and colleagues
[6] randomly assigned 101 adult volunteers into two groups: those to be taught the ABCD rule (n=51) versus those to be taught the UD sign (n=50). Volunteers were recruited from an outpatient multidisciplinary clinic (Mayo Clinic, Arizona), given a tutorial on their respective screening rule, and then asked to categorize nine skin lesions as MM or not MM. For the UD sign, images included a background of Photoshop-generated nevi.
Both volunteer groups were comparable with respect to age, history of melanoma (8.9%), education level (85% college level or higher), prior knowledge of melanoma recognition techniques, and prior dermatologic care.
The study findings included the following:
Both rules showed high sensitivity: 99% (ABCD group) vs 100% (UD group).
The UD sign group demonstrated better specificity than the ABCD group (88.3% vs 57.4%).
The UD sign group also demonstrated superior accuracy of MM recognition (90.9% vs 66.7%).
These differences in group specificity and accuracy persisted even when adjusted for volunteer age, education, and dermatologic history.
Discussion
In 2014 (the most recent year with available data), 76,665 new primary cutaneous MMs were diagnosed in the United States, and 9324 people died from MM, making this by far the most lethal form of skin cancer.
[7]Physicians and the media have done a good job of educating the public about the dangers of MM, primarily through teaching the ABCDE criteria, which emphasize evaluation of each individual lesion for asymmetry, border irregularity, color variegation, large diameter, and evolution.
[8] In contrast, the UD rule relies more on intuitive pattern recognition, with patients using their own background nevi as controls to detect any outlying lesions of concern.
As Ilyas and colleagues have demonstrated, the UD sign yielded more accurate MM detection among laypeople than the ABCD rule—albeit in a simulated scenario. It would have been even more compelling if volunteers (or their partners) trained to look for ugly duckling lesions proved better at detecting melanomas in a real-life cohort study.
Nevertheless, teaching patients to add the UD sign to their screening criteria for MM seems like a no-brainer. As Daniel Jensen and Elewski
[9] recently suggested, it may be time to add yet another letter to the ABCDE rule, adding "F" for "funny looking."
Dermatology
It is difficult to understand how hemorrhage in the stratum corneum is not preceded by hemorrhage in the papillary dermis. In order for the blood/hemosiderin to get there, it must have exited the vessel prior to entering the epidermis. I accept that clinically observed black dots in warts reside in the stratum corneum, but some hemorrhage in the papillary dermis must have come first. The a priori notion that microthrombi needed to be seen may not be true. The authors don't comment on hemorrhage in the papillary dermis, which, to my mind, may (must) underlie the findings in the stratum corneum. Otherwise – abracadabra!