To the editor:
Immune checkpoint inhibitor-induced bullous pemphigoid (ICI-BP) is a rare antibody-mediated skin toxicity, reportedly affecting up to 1.0% of ICI-treated patients.1 This dermatologic immune-related adverse event (dAE) has been associated with improved tumor responses, but frequently causes significant morbidity or necessitates permanent ICI discontinuation.1,2 Diverse treatments for ICI-BP include systemic corticosteroids and biologics, with observational evidence for rituximab, dupilumab, and omalizumab.3,4 However, the clinical impact of these immuno-biologics on the ICI-mediated anti-tumor response is unknown.
We queried the MassGeneral-Brigham/Dana-Farber Cancer Institute registries by ICD codes/BP-associated key-terms (Supplement) to retrospectively identify ICI-BP cases from January 2006-December 2020, confirmed by three dermatologists (J.T.;S.T.C.;N.R.L.). Records were reviewed from ICI initiation until end-of-study date to create three treatment groups: biologics and systemic corticosteroids (BSS), systemic corticosteroids without biologics (SS), and no systemic corticosteroid/non-biologic therapy (NSNB).
Overall survival (OS), time-to-progression, and progression-free-survival (PFS) were calculated from ICI-BP onset to December 31, 2020 or death (progression defined by RECIST, v1.1).5 Outcomes were adjusted in inverse probability-weighting regressions (IPWR), controlling for age, cancer stage, mucosal involvement, toxicity grade, performance-status, and pre-ICI chemotherapy (Supplement). To minimize guarantee-time-bias, weighted Cox-proportional-hazards controlled for the number of pre-toxicity ICI cycles. Weighted hazard-ratios were adjusted for multiple-hypothesis-testing (Tukey-Kramer method).
Thirty-five ICI-BP cases (median follow-up 11.7 months, Table 1) were sub-divided into BSS (n=11), SS (n=15), and NSNB (n=9) treatment groups. BSS-treated patients were more likely to be male (p=0.03) and have grade-3 toxicity (p=0.002). Biologics included rituximab (n=6), omalizumab (n=4), and dupilumab (n=2). Biologic therapy resulted in a temporally associated complete response of ICI-BP without flares following subsequent ICI cycles in 8/11 cases.
Table 1:
Clinical characteristics and toxicity-directed treatments amongst patients with immune checkpoint inhibitor-induced bullous pemphigoid
| Total (N=35) (%) | Biologics and Systemic Steroids (N=11) (%) | Systemic Steroids without Biologics (N=15) (%) | Non-Steroids, Non-Biologics only (N=9) (%) | p-value | |
|---|---|---|---|---|---|
| Follow-up time, months – median (range) * | 11.7 (1.6, 37.1) | 15.8 (3.5, 32.0) | 10.8 (7.2, 37.1) | 10.3 (1.6, 33.7) | – |
| Age, years – mean (SD) | 72.6 (8.8) | 74.1 (9.6) | 70.7 (9.3) | 74.0 (7.3) | 0.53 |
| Sex | 0.03 | ||||
| Female | 10 (28.6) | 0 (0.0) | 6 (40.0) | 4 (44.4) | |
| Male | 25 (71.4) | 11 (100.0) | 9 (60.0) | 5 (55.6) | |
| Race | 1.00 | ||||
| White | 34 (97.1) | 11 (100.0) | 14 (93.3) | 9 (100.0) | |
| Asian / Pacific Islander | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Cancer type | 0.83a | ||||
| Melanoma | 12 (34.2) | 5 (45.5) | 5 (33.3) | 2 (22.2) | |
| Lung | 8 (22.9) | 2 (18.2) | 3 (20.0) | 3 (33.3) | |
| Other | 15 (42.9) | 4 (36.4) | 7 (46.7) | 4 (44.4) | |
| NMSC | 5 (14.7) | 0 (0.0) | 3 (20.0) | 2 (22.2) | |
| Renal | 3 (8.6) | 2 (18.2) | 0 (0.0) | 1 (11.1) | |
| Bladder/Urothelial | 2 (5.7) | 0 (0.0) | 1 (6.7) | 1 (11.1) | |
| Esophageal | 1 (2.9) | 1 (9.1) | 0 (0.0) | 0 (0.0) | |
| Prostate | 1 (2.9) | 1 (9.1) | 0 (0.0) | 0 (0.0) | |
| Colon | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Breast | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Salivary gland | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Cancer stage | 0.99 | ||||
| Stage II | 1 (2.9) | 1 (9.1) | 0 (0.0) | 0 (0.0) | |
| Stage III | 5 (14.3) | 1 (9.1) | 3 (20.0) | 1 (11.1) | |
| Stage IV | 29 (82.9) | 9 (81.8) | 12 (80.0) | 8 (88.9) | |
| ECOG performance status | 0.16 | ||||
| 0 | 14 (40.0) | 7 (63.6) | 5 (33.3) | 2 (22.2) | |
| 1+ | 21 (60.0) | 4 (36.4) | 10 (66.7) | 7 (77.8) | |
| Inciting ICI agent(s) | 0.34 | ||||
| Nivolumab | 13 (37.1) | 5 (45.5) | 3 (20.0) | 5 (55.6) | |
| Nivolumab + Ipilimumab | 2 (5.7) | 1 (9.1) | 1 (6.7) | 0 | |
| Pembrolizumab | 20 (57.1) | 5 (45.5) | 11 (73.3) | 4 (44.4) | |
| Pre-BP ICI cycles received – mean (SD) | 13.7 (9.5) | 11.7 (10.1) | 12.0 (5.5) | 19.1 (12.6) | 0.31 |
| Concurrent chemotherapy | |||||
| No | 33 (94.3) | 11 (100.0) | 14 (93.3) | 8 (88.9) | |
| Yes | 2 (5.7) | 0 (0.0) | 1 (6.7)b | 1 (11.1)c | |
| Mucosal lesions | 0.13 | ||||
| No | 23 (65.7) | 5 (45.5) | 10 (66.7) | 8 (88.9) | |
| Yes | 12 (34.3) | 6 (54.5) | 5 (33.3) | 1 (11.1) | |
| Vesicles or bullae | 0.99 | ||||
| No | 6 (17.1) | 2 (18.2) | 3 (20.0) | 1 (11.1) | |
| Yes | 29 (82.9) | 9 (81.8) | 12 (80.0) | 8 (88.9) | |
| Maximum CTCAE grade | 0.002 | ||||
| 1 | 7 (20.0) | 2 (18.2) | 1 (6.7) | 4 (44.4) | |
| 2 | 22 (62.9) | 4 (36.4) | 14 (93.3) | 4 (44.4) | |
| 3 | 6 (17.1) | 5 (45.5) | 0 (0.0) | 1 (11.1) | |
| Topical therapy | – | ||||
| Topical steroids | 35 (100.0) | 11 (100.0) | 15 (100.0) | 9 (100.0) | |
| Topical tacrolimus | 2 (5.7) | 1 (9.1) | 0 (0.0) | 1 (11.1) | |
| Oral therapy | – | ||||
| Systemic steroids (prednisone) | 25 (71.4) | 11 (100.0) | 15 (100.0) | 0 (0.0) | |
| Methotrexate | 6 (17.1) | 2 (18.2) | 4 (26.7) | 0 (0.0) | |
| Mycophenolate mofetil | 2 (5.7) | 0 (0.0) | 2 (13.3) | 0 (0.0) | |
| Dapsone | 2 (5.7) | 0 (0.0) | 2 (13.3) | 0 (0.0) | |
| Tetracycline, no niacinamide | 6 (17.1) | 2 (18.2) | 2 (13.3) | 2 (22.2) | |
| Tetracycline, with niacinamide | 11 (31.4) | 4 (36.4) | 4 (26.7) | 2 (22.2) | |
| Hydroxychloroquine | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Acitretin | 1 (2.9) | 0 (0.0) | 1 (6.7) | 0 (0.0) | |
| Biologics | – | ||||
| Omalizumabd | 4 (11.4) | 4 (36.4) | 0 (0.0) | 0 (0.0) | |
| Rituximabd | 6 (17.1) | 6 (17.1) | 0 (0.0) | 0 (0.0) | |
| Dupilumab | 2 (5.7) | 2 (18.2) | 0 (0.0) | 0 (0.0) | |
| Other | – | ||||
| IVIG | 1 (2.9) | 1 (9.1) | 0 (0.0) | 0 (0.0) | |
| Best skin response to treatment | 0.12 | ||||
| Complete response | 14 (41.2) | 8 (72.7) | 5 (33.3) | 4 (44.4) | |
| Partial response | 19 (55.9) | 3 (27.3) | 10 (66.7) | 4 (44.4) | |
| No response | 1 (2.9) | 0 (0.0) | 0 (0.0) | 1 (11.1) |
P-values for mean age and treatment cycles received prior to bullous pemphigoid are calculated from Kruskal-Wallis tests. All other p-values for categorical data are calculated from Fisher’s exact tests.
Follow-up times are calculated from time of ICI-BP initial presentation to last follow-up date within the study period or death, based on the number of patients still alive.
This p-value was calculated from a 3-by-3 Fisher’s exact test of lung cancer, melanoma, and other cancer types.
One patient was also receiving eribulin.
One patient was also receiving denosumab.
One patient in the BSS group received both omalizumab and rituximab.
Abbreviations: BP = bullous pemphigoid. CTCAE = Common Terminology Criteria for Adverse Events. ECOG = Eastern Cooperative Oncology Group. ICI = immune checkpoint inhibitor. IVIG = intravenous immunoglobulin. NMSC = non-melanoma skin cancer. SD = standard deviation.
Eleven deaths occurred in the study-period (3/11 BSS; 5/15 SS; 3/9 NSNB). In pairwise comparisons of treatments groups, OS in the BSS group was significantly longer compared to the NSNB group (HR 0.05, adjusted-95%-CI 0.01–1.00, p=0.05) (Table 2). No other comparisons achieved statistical significance. In binary analysis comparing all biologics-recipients (BSS) to all non-biologics-recipients (SS and NSNB), there were near-significant differences toward improved OS and PFS amongst biologics-recipients (pOS=0.06, pPFS=0.06).
Table 2:
Cancer-related outcomes across toxicity-directed treatment groups in two analyses
| Analysis 1: Pairwise Comparisons of Three Treatment Groups | |||
|---|---|---|---|
| OS Comparison | |||
| Treatment groups | Hazard ratio | Adjusted 95% confidence intervals (Tukey-Kramer) | p-value |
| BSS vs. SS | 0.07 | 0.01 to 1.3 | 0.09 |
| BSS vs. NSNB | 0.05 | 0.01 to 1.0 | 0.05 |
| TTP Comparison | |||
| Treatment groups | Hazard ratio | Adjusted 95% confidence intervals (Tukey-Kramer) | p-value |
| BSS vs. SS | 1.13 | 0.26 to 6.2 | 0.93 |
| BSS vs. NSNB | 2.27 | 0.26 to 19.3 | 0.65 |
| PFS Comparison | |||
| Treatment groups | Hazard ratio | Adjusted 95% confidence intervals (Tukey-Kramer) | p-value |
| BSS vs. SS | 0.24 | 0.03 to 1.8 | 0.22 |
| BSS vs. NSNB | 0.08 | 0.01 to 3.5 | 0.25 |
| Analysis 2: Pooled Binary Comparisons | |||
| OS Comparison | |||
| Treatment groups | Hazard ratio | 95% confidence interval | p-value |
| Biologics vs. No Biologics | 0.08 | 0.01 to 1.1 | 0.06 |
| TTP Comparison | |||
| Treatment groups | Hazard ratio | 95% confidence interval | p-value |
| Biologics vs. No Biologics | 1.53 | 0.39 to 6.0 | 0.54 |
| PFS Comparison | |||
| Treatment groups | Hazard ratio | 95% confidence interval | p-value |
| Biologics vs. No Biologics | 0.27 | 0.07 to 1.02 | 0.06 |
In the first analysis, overall survival, time to progression, and progression-free survival compared are compared between pairs amongst the three treatment groups of interest. P-values represent pairwise comparisons of hazard ratios between the indicated treatment groups, with Tukey-Kramer correction for multiple hypothesis testing. Statistical significance was set at p ≤ 0.05. In the second analysis, “Biologics” is identical to the “SSB” group presented in the first analysis. The “No Biologics” group is a pooled group of the “SS” and “NSNB” treatment group denoted in the first analysis.
Abbreviations: NSNB = non-steroid, non-biologic treatment only. OS = Overall survival. PFS = progression-free survival. SS = systemic steroids, without biologics. BSS = biologics and systemic steroids. TTP = time to progression.
This multi-institutional retrospective cohort study demonstrates that biologics for ICI-BP is effective and associated with better OS compared to regimens that did not include systemic steroids or biologics, controlling for cancer and toxicity severity. Considering efficacy, these findings corroborate previously reported successful ICI-BP management with rituximab, dupilumab, and omalizumab and add, substantially, to available safety data.3,4 No outcomes were worsened amongst biologics-recipients suggesting that these biologics may preserve ICI efficacy. Patients that develop ICI-BP may have greater baseline anti-tumor responses; these data support that ICI treatment in this responsive population may be maximized by utilizing toxicity-directed biologics to prevent ICI discontinuation, uncoupling toxicity from anti-tumor efficacy.2
Limitations include our sample size that precluded per-biologic analysis, lack of a “biologics, without systemic corticosteroids” group, and possible residual confounding-by-indication, despite controlling for ICI-BP grade as a proxy for underlying anti-tumor efficacy in our IPWRs. All BSS-treated patients were male, thus sex was excluded from our IPWR. Future investigations should evaluate possible associations between receiving dAE-directed biologics and cancer outcomes in larger cohorts of ICI-treated patients.
Supplementary Material
Funding:
Steven T. Chen is supported by the Medical Dermatology Career Development Award from the Dermatology Foundation. Nicole R. LeBoeuf is supported by NIH/NCI grant U54-CA225088.
Footnotes
Conflicts of interest:
Arash Mostaghimi receives personal fees from Pfizer, hims, Digital Diagnostics, Concert, Lilly, AbbVie, and Bioniz; he is also an Associate Editor of JAMA Dermatology. Steven T. Chen received honoraria for serving on a social media advisory board for Pfizer and Novartis outside the submitted work. Nicole R. LeBoeuf is a consultant and has received honoraria from Bayer, Seattle Genetics, Sanofi, Silverback and Synox Therapeutics outside the submitted work. The remaining authors (Said, Talia, Wei, Semenov, Giobbie-Hurder) have no potential conflicts of interest involving this research or outside the submitted work.
Supplemental Methods support this manuscript; see “Supplement.”
See “Supplement” for Supplemental Methods.
REFERENCES
- 1.Siegel J, Totonchy M, Damsky W, et al. Bullous disorders associated with anti–PD-1 and anti–PD-L1 therapy: A retrospective analysis evaluating the clinical and histopathologic features, frequency, and impact on cancer therapy. J Am Acad Dermatol. Published online 2018. doi: 10.1016/j.jaad.2018.07.008 [DOI] [PubMed] [Google Scholar]
- 2.Nelson CA, Singer S, Chen T, et al. Bullous pemphigoid after anti-PD-1 therapy: a retrospective case-control study evaluating impact on tumor response and survival outcomes. J Am Acad Dermatol. Published online 2020. doi: 10.1016/j.jaad.2019.12.068 [DOI] [PubMed] [Google Scholar]
- 3.Phillips GS, Wu J, Hellmann MD, et al. Treatment outcomes of immune-related cutaneous adverse events. J Clin Oncol. Published online 2019. doi: 10.1200/JCO.18.02141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nelson CA, Singer S, Chen T, et al. Reply to: “Comment on Bullous pemphigoid after anti-PD-1 therapy: a retrospective case-control study evaluating impact on tumor response and survival outcomes.” J Am Acad Dermatol. Published online 2020. doi: 10.1016/j.jaad.2020.05.023 [DOI] [PubMed] [Google Scholar]
- 5.Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur J Cancer. Published online 2009. doi: 10.1016/j.ejca.2008.10.026 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.