Abstract
Since the Food and Drug Administration approved the anti-CCR4 monoclonal antibody, mogamulizumab, for the treatment of relapsed/refractory MF/SS in 2018, several studies have described an entity called mogamulizumab-associated rash (MAR). In the MAVORIC trial, 24% of patients in the mogamulizumab group developed a “drug eruption”. MAR is thought to be due to excessive immune reaction and has been suggested as a positive predictor of response. However, studies investigating MAR have been limited in their size due to the rarity of MF/SS. We undertook this study to leverage a large real-world database to look for association between emergence of new cutaneous reactions after mogamulizumab administration and overall survival.
Methods We performed this analysis using data from the TriNetXTM database; a platform for real-world de-identified patient-level data. The cohort of interest was generated using the Research database (155,725,981 patients globally) with a combination of ICD-10-CM codes C84.0 (Mycosis fungoides) or C84.1 (Sezary disease), ICD-0 codes 9700/3 (Mycosis fungoides) or 9701/3 Sezary syndrome. The resulting database of patients was downloaded, and the statistical analysis was carried out using Stata BE 18 (StataCorp LLC). Patients were filtered according to whether they received mogamulizumab using RxNorm codes (2054068, J9204, C9038). The corresponding encounter IDs were used to determine the date of first administration of mogamulizumab. The following ICD-10 codes corresponding to drug-related cutaneous adverse events were used to classify patients: D72.12, L27.0, L27.1, L27.9, L43.2. Patients with a new diagnosis corresponding to this list within 48 months of mogamulizumab administration were included in the MAR+ group while the rest formed the MAR- group. Survival was calculated from the date of first mogamulizumab administration to death due to any cause. Follow up duration was capped at 48 months from mogamulizumab administration.
Results A total of 24,547 patients in the Research database had a diagnosis of MF/SS. Of these, 342 had received mogamulizumab. The median age of this cohort was 67 years (range 22 to 87). Females comprised 45.6% (156) of the cohort. More than half of the patients, 58.7% (201) had SS while the rest had MF. Of the 343 patients, 57 (16.6%) were placed in the MAR+ group and 285 (83.4%) in the MAR- group. The MAR+ group had a higher proportion of patients with SS (71.9%) compared with the MAR- group (56.1%) and this difference was statistically significant (p-value 0.027). There was no statistically significant difference in the mean age or sex composition between the two groups. The average time from mogamulizumab administration to diagnosis of MAR was 7 months (range 11 days to 43 months). For the 342 patients analyzed, the median follow-up duration was 38 months (95% CI 44.3 to 56.3). There were 84 deaths (24.6%) during this period. Median overall survival (OS) for the whole cohort was not reached. The 4-year OS rate was 69.8% [95% CI 63.4 to 75.2]. There were 75 deaths in the MAR- group and 9 in the MAR+ group. The respective 4-year OS rates were 67.0% [95% CI 60.1 to 73.1] and 78.9 % [95% CI 61.9 to 88.9]. The data did not violate the proportional hazard assumption based on Schoenfeld residuals. Using the Cox proportional hazard model, the univariable hazard ratio for MAR as a predictor returned as 0.55 [p-value 0.048, 95% CI 0.25 to 0.99] favoring the MAR+ group and was statistically significant. Age and sex were not found to be independent predictors of survival on multivariable modelling. The HR for MAR as a predictor on the multivariable model returned as 0.54 (p –value 0.086, 95% CI 0.27 to 1.09). A diagnosis of SS was not found to be a predictive of survival on both univariable (HR 0.85, 95% CI 0.55 to 1.31) and multivariable modelling (HR 0.86, 95% CI 0.54 to 1.37).
Conclusion We used the TriNetX database to identify surrogate markers for MAR and predict survival in 342 patients with MF/SS. The MAR+ rate stood at 16.6%. The 4-year OS rate was longer in the MAR+ group (78.9% vs 67.0%) and MAR was found to be a positive predictor of survival on univariable regression analysis but not multivariable analysis. A diagnosis of SS was not found to have an influence on survival. A major limitation of this study is the lack of ability to precisely query for MAR using this database as well as the clinical challenge of distinguishing between MAR and disease relapse.
