It is still controversial whether keratoacanthoma is to be considered as a well differentiated variant of squamous cell-carcinoma or a separate entity. As opposed to malignant potential of squamous cell-carcinoma, keratoacanthoma is characterized by a spontaneous regression. However, in some cases, otherwise typical keratoacanthoma can behave aggressively showing the signs of perineural and perivascular invasion and metastases in regional lymph nodes. The most important feature that separates these two closely related entities is a tendency of keratoacanthoma to regress. Causes and detailed mechanism of this regression are still not completely elucidated. Within the past few years, it has become evident that the molecular events regulating cell survival and apoptosis are important contributors to the overall kinetics of benign and malignant cell growth. Immunological mechanisms have been implicated in a phenomenon of spontaneous tumor regression. Recent studies suggested that the tumor regression is dependent mainly on the immune response mediated by cytotoxic T lymphocytes (CD8+), together with helper T cells (CD4+). Cytotoxic T lymphocytes can kill tumor cells and mediate tumor regression in vivo through two distinct molecular mechanisms: Fas/Fas ligand and granzyme B/perforin mediated pathways. Tumor cells are capable of developing different escape mechanisms in order to overcome their sensitivity to apoptotic signals. However, granzyme B, contained in cytolytic granules released upon target cell recognition, can also cause tumor cell death and consequently tumor regression by direct damage to non-nuclear structures through a caspase-independent pathway. Therefore, we propose a key role of plasticity in the granzyme B mediated cell death pathway in the killing of changed tumor cells, resulting in keratoacanthoma regression through apoptosis or direct damage of tumor cells. On the other hand, insufficient activation of cytotoxic T lymphocytes and decreased release or activity of granzyme B could be responsible for squamous cell-carcinoma progression and occasional aggressive behavior in keratoacanthomas. As a first step in confirming or refuting our hypothesis, we suggest a thorough immunohistochemical study of the presence of granzyme B and its activity in keratoacanthoma and squamous cell-carcinoma samples. To our knowledge, no such study has been performed so far.