What are keratoacanthomas (KAs)?

Updated: Oct 07, 2019
  • Author: Guy J Petruzzelli, MD, PhD, MBA, FACS; Chief Editor: Gregory Gary Caputy, MD, PhD, FICS  more...
  • Print


Keratoacanthomas (KAs) are benign cutaneous neoplasms that occur in elderly patients who are frequently exposed to the sun. Often presenting as solitary lesions, they have a predilection for the face. The spontaneous involution and eventual disappearance of these lesions, within months, characterizes the self-healing nature of KAs.

Keratoacanthoma Keratoacanthoma

Histologically, the fully formed domelike lesion reveals a mass of proliferating squamous epithelial cells, which created the central core of keratin. These proliferating cells exhibit no cytologic abnormalities and have a normal nuclear-to-cytoplasmic ratio. The surrounding normal epidermis is displaced upward with the lesion, providing the characteristic buttress appearance of KAs. [4]

KAs are well-contained lesions, but case reports have shown perineural and vascular invasion. [5]

Clinically, the lesion rapidly increases in size within 6-8 weeks, with the eventual formation of a central keratin plug. During the immunologically mediated regression phase, the central keratin plug is extruded, and only a flat scar remains.

KA lesions can easily be mistaken for squamous cell carcinoma (SCC). Sudden onset and rapid growth are the distinguishing features of KAs. SCCs display a more indolent growth rate. Nodular melanoma can also mimic KAs. [6]

In order to distinguish between KA and SCC, further histologic studies can be used, including proliferating-cell nuclear antigen immunostaining. [7]  In KAs, cells that stain positive with proliferating-cell nuclear antigen immunostaining are distributed only in the outer edges of the tumor, corresponding to the proliferating squamous epithelial cells. In contrast, cells within an SCC that stain positive with proliferating-cell nuclear antigen immunostaining are more diffusely distributed.

A class of drugs that treats melanoma with BRAF mutations by inhibiting BRAF has been noted to induce the formation of KAs. [8]

In treating KAs, several problems are often encountered:

  • Some umbilicated papules or plaques look like KAs but are actually SCCs. A biopsy is needed to differentiate between the two, and if the biopsy cannot firmly establish the diagnosis, excision is the preferred treatment option.
  • Hypertrophic lupus can histologically imitate SCC and KA. Specimens should be evaluated by a trained dermatopathologist. If the patient has lupus and a KA or SCC is suspected, a second opinion should be obtained.

  • KA involutes in most cases. However, the involution can lead to tissue destruction and scarring that permanently impairs function. Thus, on the sensitive eyelid, KAs should be treated promptly. Mohs surgery with reconstruction is the optimal treatment for KA on the eyelid.

  • Multiple KAs should raise suspicion of Muir-Torre syndrome, a genodermatosis with a defect in the  MSH mismatch repair gene that is associated with sebaceous neoplasms and gastrointestinal cancers. 

  • KAs have been noted in multiple reports to arise in tattoos. [9, 10]  Pseudoepitheliomatous hyperplasia must be distinguished from a squamous neoplasm.

  • Primary cutaneous CD30+ anaplastic large cell lymphomas mimicking KAs have been reported. [11]

Although these lesions are self-healing, several reports have advised a more aggressive approach for treatment. Beham and colleagues advocated surgical excision because of the difficulties in distinguishing between KAs and SCCs upon gross examination. [12]

Gray and colleagues proposed a more conservative approach, using topical 5-fluorouracil (5-FU) therapy to accelerate the regression phase of the keratoacanthoma without the need for diagnostic biopsy. [13]  With 5-FU treatment, the KAs in a limited number of patients had a partial response within 3 weeks and complete resolution within 8 weeks. On the other hand, SCC has a poor response to 5-FU treatment.

Studies have noted that topical Imiquimod demonstrates some efficacy against KAs.

Did this answer your question?
Additional feedback? (Optional)
Thank you for your feedback!