Abstract

Cutaneous graft versus host disease (cGVHD) and erythema multiforme (EM), Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN) share similar clinical and histologic features. This overlap makes it difficult to distinguish these diseases in patients receiving hematopoietic transplants, leading to challenging treatment decisions. Prior studies have identified desmoplakin (Dp) I and II autoantibodies in patients with EM/SJS/TEN and loss of immunoreactivity to Dp I/II in patients with EM/SJS/TEN compared to cGVHD. We sought to further elucidate the utility of Dp I/II immunohistochemical stain in differentiating EM/SJS/TEN and cGVHD. 118 skin biopsies from patients with cGVHD (n=58; 25 Grade 2, 20 Grade 3, 13 Grade 4) and EM/SJS/TEN (n=60) were treated with mouse mAb IgG1 to Dp I/II and scored for the degree of Dp I/II staining. We found a statistically significant decrease in Dp I/II staining between EM/SJS/TEN and cGVHD (all grades; p<9.7e-05) as well as between EM/SJS/TEN and Grades 2+3 cGVHD (p<4.7e-06). We did not find a statistically significant difference in Dp I/II staining between EM/SJS/TEN and Grade 4 cGVHD (p<0.706). Receiver operating characteristic (ROC) curves demonstrated an area under the curve (AUC) of 0.692 (95% CI 0.603 to 0.779) for the ability of Dp I/II staining to differentiate between cGVHD (all grades) vs EM/SJS/TEN, and an AUC of 0.735 (95% CI 0.645 to 0.819) for differentiating Grades 2+3 cGVHD vs EM/SJS/TEN. Sensitivities and specificities at optimal thresholds showed sensitivities of 82.76% (cGVHD all grades vs. EM/SJS/TEN) and 88.89% (Grades 2+3 cGVHD vs. EM/SJS/TEN), and the same specificity for both these groups of 51.67%. Our data further validate that there is decreased Dp I/II expression in EM/SJS/TEN compared to cGVHD. In the appropriate clinical setting, Dp I/II immunostaining might be particularly helpful for differentiating EM/SJS/TEN from Grades 2+3 cGVHD.

Financial Disclosure:
No current or relevant financial relationships exist.