Artificial intelligence (AI) image recognition models have been relatively successful in diagnosing cutaneous manifestations in individuals with light skin tone. However, when these models are test...

Objectives: Bullous pemphigoid, an autoimmune dermatological disease, may be associated with the use of a relatively new anti-cancer drug class, PD-1 inhibitors, which includes pembrolizumab and nivolumab. This paper analyzes the signals between PD-1 inhibitors and bullous pemphigoid based upon the reported real world data. Methods: A pharmacovigilance analysis was performed on the publicly available Adverse Event Reporting System database of Food and Drug Administration. Disproportionality ratios were used to examine a signal between PD-1 inhibitors and bullous pemphigoid. A heat map was generated to depict the signal between PD -1 inhibitor use and skin toxicity adverse events. Results: The analysis indicated that there is a significant signal (PRR = 13.82 [95% CI: 9.99-19.11], Chi-squared with Yates’ correction = 420.48) between pembrolizumab use and bullous pemphigoid and that there is a significant signal (PRR = 13.19 [95% CI: 10.57-16.46], Chi-squared with Yates’ correction = 869.71) between nivolumab use and bullous pemphigoid. The signals remained statistically significant after stratifying for sex and age for both pembrolizumab and nivolumab. The signal is supported by 35 case reports in which there was evidence of PD-1 inhibitor use and a pemphigoid adverse event. Conclusion: When prescribing PD-1 inhibitors, physicians should monitor closely for symptoms of bullous pemphigoid.

Background/Objectives: Several treatment options with differing mechanisms exist for plaque psoriasis. The objective of this analysis was to compare the time to onset of action among the available systemic therapies for plaque psoriasis. Methods: Randomized controlled trials that investigated two or more therapeutics for the management of moderate to severe plaque psoriasis were included. A weighted average time for 50% of patients to reach PASI75 and PAI90 with each of the therapeutics was calculated. A network meta-analysis was performed to determine which therapeutics were significantly faster in time to meaningful clinical response than others. Results: IL-17 inhibitors had the shortest time to achieve PASI75 and PASI90 followed by risankizumab in the weighted mean analysis. In the meta-analysis, the fastest time to PASI75 was seen with bimekizumab, brodalumab and ixekizumab. No significant (p < 0.05) difference was seen in the time to meaningful clinical response between these drugs; however, bimekizumab was significantly faster in time to PASI75 among the remaining therapeutics. In the meta-analysis for PASI90, the fastest time was seen with ixekizumab, bimekizumab, risankizumab, secukinumab and guselkumab with no significant differences in between these therapeutics. However, bimekizumab was significantly faster than the remaining therapeutics for PASI90. Conclusions: IL-17 and IL-23 inhibitors may be considered as requiring the shortest time for meaningful clinical response in plaque psoriasis. In addition to the time to onset, the safety profile of each drug needs to be considered when deciding on a therapeutic to initiate.

Background: The performance of deep-learning image recognition models is below par when applied to images with Fitzpatrick classification skin types 4 and 5. Objective: The objective of this research was to assess whether image recognition models perform differently when differentiating between dermatological diseases in individuals with darker skin color (Fitzpatrick skin types 4 and 5) than when differentiating between the same dermatological diseases in Caucasians (Fitzpatrick skin types 1, 2, and 3) when both models are trained on the same number of images. Methods: Two image recognition models were trained, validated, and tested. The goal of each model was to differentiate between melanoma and basal cell carcinoma. Open-source images of melanoma and basal cell carcinoma were acquired from the Hellenic Dermatological Atlas, the Dermatology Atlas, the Interactive Dermatology Atlas, and DermNet NZ. Results: The image recognition models trained and validated on images with light skin color had higher sensitivity, specificity, positive predictive value, negative predictive value, and F1 score than the image recognition models trained and validated on images of skin of color for differentiation between melanoma and basal cell carcinoma. Conclusions: A higher number of images of dermatological diseases in individuals with darker skin color than images of dermatological diseases in individuals with light skin color would need to be gathered for artificial intelligence models to perform equally well.