Main Article Content
psoriasis, machine learning, clinical utility, biomarkers, transcriptomics, biologics
Objective: This study (MATCH) was designed to assess the clinical utility of a machine-learning based tool (Mind.Px) that predicts patient response to the most common biologic drug classes used in the management of psoriasis patients.
Methods: The MATCH study was designed to assess Mind.Px utility in physician decision making and patient outcomes. Psoriasis patients who were biologic naïve or those who were approaching a change due to non-response were enrolled into the study (N=112). At baseline, a dermal biomarker patch was applied to lesional skin and Mind.Px test results were provided to physicians prior to biologic selection. The choice of biologic for each patient was recorded and in the case of physician non-concordance with Mind.Px test results, a questionnaire completed to determine the reason for non-concordance. Patients were evaluated at weeks 4, 8, 12, and 16 after baseline using PASI, PGA, and BSA. Statistical analysis between groups was performed using Fisher’s exact test.
Results: Physician prescribing behavior was measured with and without the inclusion of Mind.Px test results in the decision-making process. This data set was compared to a previously obtained data set in which dermal biomarker patches were applied at baseline, but Mind.Px results were not provided to physicians at any point during treatment (N=180). Statistical analysis of concordance between the Mind.Px-informed and Mind.Px-uninformed groups within the MATCH study (84.4% vs 53.8%, respectively) showed that when given access to Mind.Px results, physician behavior was significantly altered (p = 0.0022). Furthermore, analysis of those patients who have completed the study showed improved clinical outcomes in those patients whose physicians were provided Mind.Px test results. Specifically, this cohort reached PASI75 sooner than those who were not provided test results (p = 0.004).
Conclusion: These results provide an interim measurement of the clinical utility of Mind.Px by demonstrating that physicians will utilize this test in psoriasis biologic decision making and by doing so, this leads to improved patient outcomes. These improved patient outcomes can also potentially translate into tremendous cost savings for healthcare systems. Mind.Px can effectively minimize the trial-and-error approach to psoriasis treatment, and provide physicians, patients, and payers with a powerful tool for improving psoriasis patient management.
2. Menter A, Strober BE, Kaplan DH, et al. Joint AAD‐NPF guidelines of care for the management and treatment of psoriasis with biologics. J Am Acad Dermatol. 2019;80:1029‐1072.
3. Bagel J, Wang Y, Montgomery P III, et al. A machine learning-based test for predicting response to psoriasis biologics. SKIN The Journal of Cutaneous Medicine. 2021;5:621–638.
4. Strober B, Fox J, Jekowsky E, et al. Evidence threshold for a precision medicine test that predicts optimal response to a biologic agent in patients with psoriasis: A consensus panel. J Drugs Dermatol. 2022;21:630-636.
5. Strober B, Pariser D, Deren-Lewis A, et al. A survey of community dermatologists reveals the unnecessary impact of trial-and-error behavior on the psoriasis biologic treatment paradigm. Dermatol Ther (Heidelb). 2021;11:1851-1860.