Keratoacanthoma: the tumour that may be a squamous cell carcinoma

No question in skin pathology has been reopened as often. The fourth edition of the WHO skin classification (2018) placed keratoacanthoma inside the squamous cell carcinoma section, and the UK reporting dataset recorded that position in terms: it notes the WHO terminology, which considers keratoacanthoma as synonymous with invasive squamous cell carcinoma of keratoacanthomatous type; it notes that there is still considerable national and international debate as to whether keratoacanthoma is truly a pathologically benign or malignant neoplasm; and it records that, to date, no one single criterion can make a reliable distinction between squamous cell carcinoma and keratoacanthoma, so each case must be approached using a constellation of clinical and pathological features, including a history of an initial period of rapid growth (RCPath G124). The fifth edition (2025) then moved it back out, making it a heading of its own beside the basal cell and squamous cell carcinomas, while conceding the same difficulty: reliable distinction between the two is not always possible, immunohistochemistry for p53, Ki-67 and p16 helps in many cases, and where it does not, the uncertainty can be reported as a squamoproliferative tumour of uncertain malignant potential.

The coding follows the confusion rather than resolving it. ICD-O-3.2 carries keratoacanthoma as a related term under 8071/3, squamous cell carcinoma, keratinizing, which is a malignant behaviour code. The UK dataset's own coding appendix does the opposite, coding keratoacanthoma as SNOMED M7286/0, a benign behaviour code, and keeping keratoacanthoma-like squamous cell carcinoma separately at M8071/3. A British pathologist reporting the same lesion under the two systems would file it as benign in one and malignant in the other.

The same dataset adds a safety rule that is the practical point for a reader. The term keratoacanthoma should perhaps be avoided in an immunosuppressed patient, and for large lesions, and in a subungual location under a nail, because all of those are considered to have greater potential for aggressive behaviour. In other words, the cases where calling it a keratoacanthoma would be most reassuring are the cases where it is least safe to.

This is why keratoacanthoma is a word here and not a page. The corpus gives a page to a lesion with a settled behaviour that registries register; this one has neither, and its position in the reference classification has moved in both directions within seven years. Anyone whose biopsy report carries the word should read the cutaneous squamous cell carcinoma page, which is where the management of the thing in front of them is written.