Abstract
Purpose:
Supportive oncodermatology has been shown to improve several aspects of care for patients with cancer, but research showing improved diagnostic accuracy as a benefit of supportive oncodermatology is largely lacking. We thus aimed to evaluate different dermatologist groups’ diagnostic accuracy for heterogenous cutaneous toxicities, using cutaneous immune-related adverse events (cirAEs) from immune checkpoint inhibitors (ICIs) as a test model.
Methods:
Billing/requisition codes were used to identify patients who initiated programmed death-1/ligand-1 (PD-1/PD-L1) ICIs between 2010–2019 at Dana-Farber Cancer Institute/Brigham and Women’s Hospital/Massachusetts General Hospital and underwent a subsequent skin biopsy. For each biopsied cirAE, pre-biopsy clinical diagnoses and post-biopsy clinico-pathologic diagnoses were retrospectively obtained from the medical record. Each biopsy-ordering dermatology provider was categorized as a general dermatologist or supportive oncodermatologist on the basis of providing clinical care within a cancer-center or attending on a hospital/clinic service dedicated to anti-cancer drug-related skin toxicities.
Results:
Of 4,183 patients who initiated anti-PD-1/PD-L1 therapy between 2010–2019, 101 (2.4%) patients collectively had 104 biopsied cirAEs. In more than one-third of all reviewed biopsied cirAEs (n=39, 37.5%), histopathology results frequently led to revision of the pre-biopsy clinical diagnosis. The rate of initial cirAE misclassification amongst supportive oncodermatologists was significantly lower than that amongst general dermatologists (18/66, 27.3% vs. 21/38, 55.3%; Fisher’s-exact-test p=0.006).
Conclusion:
Experienced supportive oncodermatologists may benefit patient care through increased diagnostic accuracy for skin toxicities from ICIs. Collectively, these results underscore that both skin biopsy from any dermatology provider and oncodermatology referral (where available) are valuable resources that should be integrated into supportive cancer care.
Keywords: supportive oncodermatology, immune checkpoint inhibitors, immune-related adverse events, oncodermatology, skin toxicity, diagnostic accuracy
Introduction
As recently summarized in Supportive Care in Cancer, supportive oncodermatology is a collaborative subspecialty focused on the dermatologic care of patients with cancer. Supportive oncodermatology has vast benefits for patients, including increased adherence to skin toxicity management guidelines and reduced anti-cancer treatment discontinuation.1–4
The involvement of supportive oncodermatology may improve the accuracy of delivered diagnoses for cutaneous toxicities, with one study reporting diagnostic revision amongst 59% of 204 dermatology consultations provided to hematology/bone-marrow-transplant hospital services.1,5 Further research supporting improved diagnostic accuracy as a benefit of supportive oncodermatology is lacking. Additionally, studies to date have qualified the involvement of any dermatologist in the care of patients with cancer as “supportive oncodermatology;” in contrast, comparisons of diagnostic accuracy across dermatologists’ experience in oncodermatology have not been reported.
Aim
To investigate the diagnostic accuracy of supportive oncodermatologists versus general dermatologists, we considered the cutaneous immune-related adverse events (cirAEs) caused by immune checkpoint inhibitors (ICI). CirAEs occur frequently in at least one-third of ICI-treated patients, and their substantial clinical heterogeneity makes them a suitable model for studying diagnostic accuracy.6–8 In addition, skin biopsy and histopathology have been shown to support accurate diagnosis for selected skin toxicities, and thus are important to guide specific cirAE treatment by specific cirAE classification.9 However, cirAEs are only infrequently biopsied (5% in one study)6, and the overall rate of post-biopsy diagnostic revision also remains unknown.
The involvement of any dermatologist in the care of ICI-treated patients with cirAEs improves patient outcomes, including overall-survival, progression-free-survival, and odds of cirAE treatment.8 To address the aforementioned gaps surrounding diagnostic accuracy, we investigated post-biopsy cirAE diagnostic change amongst specialized oncodermatologists compared to less specialized (general) dermatologists.
Methods
Billing and pathology requisition codes were used to retrospectively identify patients at Mass General-Brigham/Dana-Farber Cancer Institute who received anti-programmed death-1/ligand-1 (PD-1/PD-L1) therapy between 2010–2019, and subsequently developed a cirAE that was biopsied by any dermatologist. During screening, cirAEs were defined as skin reactions consistent with established morphologic categories and attributed to ICI by the biopsy-ordering clinician.6,8,10 All potential cases were eligible for inclusion if they were consistent with cirAE classifications reported from large-scale retrospective U.S. epidemiologic claims data.11 Furthermore, all cases were secondarily reviewed by one study oncodermatologist (N.R.L.;S.T.C.), who retrospectively confirmed that skin reactions were attributed to ICIs and consistent with morphologies or disease phenotypes previously reported as cirAEs. For skin reactions that occurred secondary to ICI/non-ICI combination-therapies, only cases attributed specifically to the ICI by both the biopsy-ordering clinician and reviewing study oncodermatologist were included. For included cirAEs, pre-biopsy clinical diagnoses and post-biopsy clinico-pathologic diagnoses were retrospectively abstracted.
Biopsy-ordering dermatologists were classified as oncodermatologists if, at the time of biopsy, they: (1) provided dermatologic care for patients experiencing skin toxicities while embedded within a cancer-center, or (2) served as a consultant on an inpatient hospital-service dedicated to managing patients with cancer experiencing drug-related skin toxicities. Remaining providers were classified as general dermatologists.
Results
Of 4,183 patients who initiated anti-PD-1/PD-L1 therapy between 2010–2019, 101 (2.4%) patients collectively had 104 biopsied cirAEs (Supplement). Patients mostly received anti-PD-1 (n=89, 88.1%) therapy, often received ICI monotherapy (n=83, 82.2%), and were most frequently treated for melanoma (n=23, 22.8%) or lung cancer (n=23, 22.8%). The cases biopsied by oncodermatologists (n=66, 63.5%) and those seen by general dermatologists (n=38, 36.5%) did not differ significantly, with the exception that oncodermatologists saw a greater proportion of patients with non-skin cancers, such as lung/genitourinary malignancies (p=0.001, Supplement). The distributions of post-biopsy cirAE diagnoses did not differ across provider groups (p=0.13, Supplement). Mean cirAE-onset-to-biopsy time between provider groups did not differ significantly (oncodermatologists: 60.3 days; generalists: 54.0 days, p=0.64, Supplement).
In over one-third of all biopsied cirAEs (n=39, 37.5%), histopathology results changed the determined post-biopsy cirAE classification from the pre-biopsy clinical diagnosis. The rate of initial misclassification amongst oncodermatologists was significantly lower than that amongst general dermatologists (18/66, 27.3% vs. 21/38, 55.3%; Fisher’s-test p=0.006). Initial diagnostic accuracy was similar across all oncodermatologists (72.7%), cancer center-affiliated oncodermatologists (72.9%), and inpatient dermatology service-affiliated oncodermatologists (75.0%) (Supplement). Misclassification rates per final diagnosis varied widely: while no cases of ICI-induced vitiligo were misclassified prior to biopsy, nearly half of ICI-induced lichenoid (10/25, 40.0%) and eczematous (10/24, 41.7%) toxicities were initially misclassified (Table 1). Complete diagnostic patterns for common and uncommon cirAEs are provided (Table 1).
Table 1:
Rates and patterns of cirAE misclassification following biopsy amongst oncodermatologists and general dermatologists
| Post-biopsy clinico-pathologic diagnosis | Total (%)a | Presented to OD (%) | Misclassified by OD (%) | Presented to GD (%) | Misclassified by GD (%) | Total misclassified per diagnosis (%)b | Pre-biopsy misclassifications | Clinical imagec |
|---|---|---|---|---|---|---|---|---|
| Common cirAE classifications | ||||||||
| Lichenoid eruption | 25 (24.0) | 14/25 (56.0) | 4/14 (28.6) | 11/25 (44.0) | 6/11 (54.5) | 10/25 (40.0%) | Autoimmune bullous (2), eczematous (2), erythema multiforme (1), morbilliform (1), psoriasiform (1), rash NOS (1), SCLE (1), eruptive keratoacanthoma (1) | 1A |
| Eczematous eruption | 24 (23.1) | 19/24 (79.2) | 8/19 (42.1) | 5/24 (20.8) | 2/5 (40.0) | 10/24 (41.7%) | Lichenoid (4), pityriasis rosea (1), psoriasiform (2), urticaria (1), fixed drug eruption (1), acute cellulitis (1) | 1B |
| Dermal hypersensitivity (morbilliform) reaction | 14 (13.5) | 6/14 (42.9) | 2/6 (33.3) | 8/14 (57.1) | 4/6 (66.7) | 6/14 (42.9%) | Eczematous (3), lichenoid (2), miliaria rubra (1) | 1C |
| Autoimmune bullous eruption | 12 (11.5) | 8/12 (66.7) | 1/8 (12.5) | 4/12 (33.3) | 1/4 (25.0) | 2/12 (16.7%) | Eczematous (1), granulomatous (1) | 1D |
| Psoriasiform eruption | 9 (8.7) | 7/9 (77.8) | 2/7 (28.6) | 2/9 (22.2) | 1/2 (50.0) | 3/9 (33.3%) | Lichenoid (2), rash NOS (1) | 1E |
| Uncommon cirAE classifications | ||||||||
| Vitiligo | 4 (3.8) | –d | 0/4 (0.0%) | N/A | N/A | |||
| Mucositis, with or without CTD | 2 (1.9) | –d | 1/2 (50.0%) | Rash NOS (1) | N/A | |||
| Panniculitis | 2 (1.9) | –d | 0/2 (0.0%) | N/A | N/A | |||
| SJS/TEN-like eruption | 2 (1.9) | –d | 2/2 (100.0%) | Autoimmune bullous (1), erythema multiforme (1) | N/A | |||
| Sweet syndrome | 2 (1.9) | –d | 1/2 (50.0%) | Erythema multiforme (1) | N/A | |||
| Vasculitis | 2 (1.9) | –d | 1/2 (50.0%) | Lichenoid (1) | N/A | |||
| Eosinophilic fasciitis | 1 (1.0) | –d | 0/1 (0.0%) | N/A | N/A | |||
| Erythema multiforme | 1 (1.0) | –d | 1/1 (100.0%) | Autoimmune bullous (1) | N/A | |||
| Follicular eruption | 1 (1.0) | –d | 1/1 (100.0%) | Lichenoid (1) | N/A | |||
| Granulomatous eruption | 1 (1.0) | –d | 0/1 (0.0%) | N/A | N/A | |||
| Phototoxic drug eruption | 1 (1.0) | –d | 1/1 (100.0%) | Lichenoid (1) | N/A | |||
| SCLE | 1 (1.0) | –d | 0/1 (0.0%) | N/A | N/A | |||
Percentages are out of the total number of biopsied cirAEs (N=104).
Comparisons of misclassification rates stratified by provider type are not presented for diagnosis classes with fewer than 5 biopsied cirAEs.
See Supplemental Figure 1 for representative clinical images of common cirAE classifications.
Abbreviations: CTD, connective tissue disease; GD, general dermatologist; N/A, not applicable; NOS, not otherwise specified; SCLE, subacute cutaneous lupus erythematosus; SJS, Stevens-Johnson syndrome; OD, supportive oncodermatologist; TEN, toxic epidermal necrosis.
Discussion
In this multi-institutional study of all available dermatologist-biopsied cirAEs, we present cirAE initial diagnosis patterns and compare diagnostic revision rates across provider experience with supportive oncodermatology. Given the clinical heterogeneity of cirAEs and their association with improved overall survival and progression-free survival in individuals with melanoma and non-melanoma malignancies, the early diagnosis and appropriately targeted management of these toxicities is important.12–17 As hypothesized, oncodermatologists demonstrated greater pre-biopsy diagnostic accuracy in classifying cirAEs, corroborating that supportive oncodermatology specifically benefits patient care through frequent, accurate diagnostic revision.5
Though oncodermatologists misclassified 27.3% of biopsied cases, these include a sub-population of atypical, challenging-to-diagnose cirAEs (i.e., panniculitis, vasculitis, connective tissue disease). Furthermore, general dermatologists correctly classified nearly one-half of biopsied cases, despite rarely evaluating ICI-treated patients. Synthesizing this, we emphasize that the involvement of any dermatologist – oncodermatologist or generalist – can provide a well-established dermatological diagnosis to guide appropriate cirAE management, and improve the odds of skin toxicity-directed treatment and patients’ cancer outcomes.8
We also underline that diagnostic revision occurred in over one-third of all cases, and histopathological information is therefore important in accurately diagnosing cirAEs. This diagnostic accuracy importantly affects toxicity-directed treatment: receiving even low doses of systemic steroids for irAE management has been associated with worsened overall survival in ICI-treated individuals, and targeted steroid-sparing agents for cirAEs vary by specific diagnosis (i.e., rituximab/omalizumab for bullous pemphigoid; apremilast/acitretin for psoriasis).6,18,19 The early determination of cirAE classification can thus facilitate the delivery of appropriately tailored cirAE-directed treatment.6 Collectively, these results underscore that both skin biopsy from any dermatologist and oncodermatology referral (where available) are valuable resources that should be integrated into supportive cancer care.1,7,20
Limitations include a small sample, retrospective design, and inability to report cirAE biopsy rates per provider group. Our sample may not accurately represent the real-world cirAE population, given possibly skewing towards severe/atypical cirAEs that more frequently prompt specialist referral/biopsy. While the distributions of post-biopsy cirAE diagnoses did not significantly differ between provider groups, we are unable to similarly compare distributions of cirAE severity, due to limited retrospective data from which to extrapolate severity. As supportive oncodermatology expands, future investigations should assess the diagnostic accuracy of oncodermatologists for a broader set of skin toxicities to generalize these findings.
Supplementary Material
Supplemental Table 1: Sample characteristics
All p-values represent comparisons of the two provider groups. All p-values from comparing categorical patient variables were calculated from Fisher’s exact tests; p-values from comparing continuous variables were assessed by Wilcoxon-ranked-sum tests; p-values less than 0.05 was considered statistically significant.
a Other cancers included: head or neck cancers (4), non-Hodgkin lymphoma (3) breast cancer (3), neuroendocrine cancer (2), ovarian cancer (2), thyroid cancer (1), cervical cancer (1), prostate cancer (1).
b One patient had two separate cirAEs simultaneously, which were each biopsied. One patient had three separate cirAEs simultaneously, which were each biopsied. The total number of biopsied cirAEs is 104.
c A total of 18 cirAEs were clinically attributed to ICI therapy while the patient was also receiving another therapy: anti-PD-1 + carboplatin/pemetrexed (3), anti-PD-1 + eribulin (2), anti-PD-1 + ICOS antagonist (1), anti-PD-1 + anti-TGF-beta (1), anti-PD-1 + dabrafenib/trametinib (1), anti-PD-1 + lirilumab (1), anti-PD-1 + bevacizumab (1), anti-PD-1 + ribociclib (1), anti-PD-L1 + emactuzumab (2), anti-PD-L1 + vemurafenib/cobimetinib (2), anti-PD-L1 + anti-IDO (1); anti-PD-L1 + bevacizumab (1); anti-PD-L1 + prexasertib (1).
d Other pre-biopsy clinical diagnoses included: rash NOS (4), Stevens-Johnson syndrome/erythema multiforme (3), panniculitis (2), SCLE (2), acute cellulitis (1), eosinophilic fasciitis (1), eruptive keratoacanthoma (1), fixed drug eruption (1), granulomatous (1), miliaria rubra (1), mucositis (1), pityriasis rosea (1), Sweet syndrome (1), urticarial reaction (1), vasculitis (1).
e Other post-biopsy clinicopathologic diagnoses included: Mucositis (2), Panniculitis (2), Stevens-Johnson syndrome/erythema multiforme (3), Sweet syndrome (2), vasculitis (2), eosinophilic fasciitis (1), follicular (1), granulomatous (1), SCLE (1), phototoxic drug eruption (1).
f P-value for time from cirAE onset to biopsy is calculated from a two-tailed t-test.
Abbreviations: cirAE, cutaneous immune-related adverse event; ICOS, inducible T cell costimulator; IDO, indoleamine 2,3-dioxygenase; IQR, interquartile range; NC, not calculated; NOS, not otherwise specified; PD-1, programmed death-1; PD-L1, programmed death-ligand 1; SCLE, subacute cutaneous lupus erythematosus; TGF, transforming growth factor.
Supplemental Table 2: Comparison of diagnostic accuracy between oncodermatologist sub-groups and general dermatologists
For all cirAE cases included in this cohort, the biopsy-ordering dermatology provider was categorized as a general dermatologist or supportive oncodermatologist. Each provider was classified as a supportive oncodermatologist if, at the time of documented clinical care, they: (1) provided dermatologic clinical care for patients experiencing skin toxicities while embedded within a cancer center, or (2) served as a consultant on a hospital service dedicated to managing cancer patients experiencing drug-related skin toxicities. The providers meeting criteria 1 were considered “cancer center-affiliated,” and providers meeting criteria 2 were considered “inpatient dermatology service-affiliated.” The providers meeting neither of these criteria were classified as general dermatologists. Two-by-two Fisher’s exact tests were used for pairwise comparisons as specified. The rates of initial correct cirAE classification were similar amongst the total OD group (72.7%), cancer center-affiliated OD group (72.9%), and hospital dermatology-affiliated OD group (75.0%). The cancer center-affiliated OD group has a statistically significantly increased frequency of initial correct cirAE classification compared to the GD group (72.9% vs 44.7%, p=0.010). The hospital dermatology-affiliated OD group also had greater frequency of initial correct cirAE classification compared to the GD group (75.0% vs 44.7%, but this did not achieve statistical significance (p=0.072), likely due to small sub-group size (N=16).
Abbreviations: GD = general dermatologist. OD = oncodermatologist. OD(CC) = oncodermatologist, cancer-center affiliated. OD(IDS) = oncodermatologist, inpatient dermatology service-affiliated.
Supplemental Figure 1: Clinical photographs of common cirAE classifications
1A: Lichenoid eruption. 1B: Eczematous eruption. 1C: Dermal hypersensitivity (morbilliform) reaction. 1D: Autoimmune bullous eruption. 1E: Psoriasiform eruption.
Acknowledgements:
Accepted for presentation at the Medical Dermatology Society Annual Meeting, originally scheduled for March 15, 2020 in Denver, Colorado, USA; presented at the American Academy of Dermatology VMX, held from April 23–25, 2021 online.
Funding:
Nicole R. LeBoeuf is supported by NIH/NCI grant U54-CA225088. Steven T. Chen is supported by a Medical Dermatology Career Development Award from the Dermatology Foundation.
Footnotes
Consent to participate: Not applicable; given retrospective nature, patients were not actively or involved in study beyond chart review. There is no identifiable information at the individual level in this study.
Consent for publication: Not applicable; given retrospective nature, patients were not actively or involved in study beyond chart review. There is no identifiable information at the individual level in this study.
Supplemental Materials: This manuscript is supported by Supplemental Tables 1–2 and Supplemental Figure 1.
Conflict of Interest Disclosures: Nicole R. LeBoeuf is a consultant and has received honoraria from Bayer, Seattle Genetics, Sanofi, Silverback and Synox Therapeutics outside the submitted work. Steven T. Chen receives Honoria for serving on a social media advisory board for Pfizer outside the submitted work.
Ethics approval: This study was approved by the Mass General Brigham IRB (Protocol No. 2017P000501).
Availability of data and material:
Available upon written reasonable request.
References
- 1.Long V, Choi ECE, Tan CL. Supportive oncodermatology—a narrative review of its utility and the way forward. Support Care Cancer. Published online 2021. doi: 10.1007/s00520-021-06124-w [DOI] [PubMed] [Google Scholar]
- 2.Barrios DM, Phillips GS, Freites-Martinez A, et al. Outpatient dermatology consultations for oncology patients with acute dermatologic adverse events impact anticancer therapy interruption: a retrospective study. J Eur Acad Dermatology Venereol. Published online 2020. doi: 10.1111/jdv.16159 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Balagula Y, Rosen ST, Lacouture ME. The emergence of supportive oncodermatology: The study of dermatologic adverse events to cancer therapies. J Am Acad Dermatol. Published online 2011. doi: 10.1016/j.jaad.2010.06.051 [DOI] [PubMed] [Google Scholar]
- 4.Yu Z, Dee EC, Bach DQ, Mostaghimi A, Leboeuf NR. Evaluation of a Comprehensive Skin Toxicity Program for Patients Treated with Epidermal Growth Factor Receptor Inhibitors at a Cancer Treatment Center. JAMA Dermatology. Published online 2020. doi: 10.1001/jamadermatol.2020.1795 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tracey EH, Forrestel A, Rosenbach M, Micheletti RG. Inpatient dermatology consultation in patients with hematologic malignancies. J Am Acad Dermatol. Published online 2016. doi: 10.1016/j.jaad.2016.05.014 [DOI] [PubMed] [Google Scholar]
- 6.Phillips GS, Wu J, Hellmann MD, et al. Treatment Outcomes of Immune-Related Cutaneous Adverse Events. J Clin Oncol. Published online June 2019:JCO.18.02141. doi: 10.1200/JCO.18.02141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Curry JL, Tetzlaff MT, Nagarajan P, et al. Diverse types of dermatologic toxicities from immune checkpoint blockade therapy. J Cutan Pathol. 2017;44(2):158–176. doi: 10.1111/cup.12858 [DOI] [PubMed] [Google Scholar]
- 8.Thompson LL, Li EB, Krasnow NA, et al. Effect of dermatological consultation on survival in patients with checkpoint inhibitor-associated cutaneous toxicity. Br J Dermatol. Published online 2021. doi: 10.1111/bjd.20074 [DOI] [PubMed] [Google Scholar]
- 9.Sellheyer K, Bergfeld WF. A retrospective biopsy study of the clinical diagnostic accuracy of common skin diseases by different specialties compared with dermatology. J Am Acad Dermatol. Published online 2005. doi: 10.1016/j.jaad.2004.11.072 [DOI] [PubMed] [Google Scholar]
- 10.Sibaud V Dermatologic reactions to immune checkpoint inhibitors. Am J Clin Dermatol. 2018;19(3):345–361. doi: 10.1007/s40257-017-0336-3 [DOI] [PubMed] [Google Scholar]
- 11.Wongvibulsin S, Pahalyants V, Kalinich M, et al. Epidemiology and risk factors for the development of cutaneous toxicities in patients treated with immune checkpoint inhibitors: A United States population-level analysis. J Am Acad Dermatol. Published online 2021. doi: 10.1016/j.jaad.2021.03.094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fan Y, Xie W, Huang H, et al. Association of Immune Related Adverse Events With Efficacy of Immune Checkpoint Inhibitors and Overall Survival in Cancers: A Systemic Review and Meta-analysis. Front Oncol. Published online 2021. doi: 10.3389/fonc.2021.633032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tang K, Seo J, Tiu BC, et al. Association of Cutaneous Immune-Related Adverse Events With Increased Survival in Patients Treated With Anti-Programmed Cell Death 1 and Anti-Programmed Cell Death Ligand 1 Therapy. JAMA dermatology. 2022;158(2):189–193. doi: 10.1001/jamadermatol.2021.5476 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Thompson LL, Nadelmann ER, Blum AE, et al. Patterns and prognostic significance of cutaneous immune-related adverse events in non–small cell lung cancer. Eur J Cancer. Published online 2021. doi: 10.1016/j.ejca.2021.01.022 [DOI] [PubMed] [Google Scholar]
- 15.Thompson LL, Chang MS, Polyakov NJ, et al. Prognostic significance of cutaneous immune-related adverse events in patients with melanoma and other cancers on immune checkpoint inhibitors. J Am Acad Dermatol. Published online 2022. doi: 10.1016/j.jaad.2021.03.024 [DOI] [PubMed] [Google Scholar]
- 16.Thompson LL, Krasnow NA, Chang MS, et al. Patterns of Cutaneous and Noncutaneous Immune-Related Adverse Events among Patients with Advanced Cancer. JAMA Dermatology. Published online 2021. doi: 10.1001/jamadermatol.2021.0326 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Patel AB, Farooq S, Welborn M, et al. Cutaneous adverse events in 155 patients with metastatic melanoma consecutively treated with anti-CTLA4 and anti-PD1 combination immunotherapy: Incidence, management, and clinical benefit. Cancer. Published online 2022. doi: 10.1002/cncr.34004 [DOI] [PubMed] [Google Scholar]
- 18.Faje AT, Lawrence D, Flaherty K, et al. High-dose glucocorticoids for the treatment of ipilimumab-induced hypophysitis is associated with reduced survival in patients with melanoma. Cancer. Published online 2018. doi: 10.1002/cncr.31629 [DOI] [PubMed] [Google Scholar]
- 19.Lo JJ, Heberton MM, Pacha O, Huen AO, Patel AB. Biologic therapies for checkpoint inhibitor-induced cutaneous toxicities: a single-institution study of 17 consecutively treated patients. Support Care Cancer. Published online 2022. doi: 10.1007/s00520-021-06548-4 [DOI] [PubMed] [Google Scholar]
- 20.Chen ST, Molina GE, Lo JA, et al. Dermatology Consultation Reduces Interruption of Oncologic Management Among Hospitalized Patients with Immune-related Adverse Events: A Retrospective Cohort Study. J Am Acad Dermatol. Published online September 2019. doi: 10.1016/J.JAAD.2019.09.026 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Table 1: Sample characteristics
All p-values represent comparisons of the two provider groups. All p-values from comparing categorical patient variables were calculated from Fisher’s exact tests; p-values from comparing continuous variables were assessed by Wilcoxon-ranked-sum tests; p-values less than 0.05 was considered statistically significant.
a Other cancers included: head or neck cancers (4), non-Hodgkin lymphoma (3) breast cancer (3), neuroendocrine cancer (2), ovarian cancer (2), thyroid cancer (1), cervical cancer (1), prostate cancer (1).
b One patient had two separate cirAEs simultaneously, which were each biopsied. One patient had three separate cirAEs simultaneously, which were each biopsied. The total number of biopsied cirAEs is 104.
c A total of 18 cirAEs were clinically attributed to ICI therapy while the patient was also receiving another therapy: anti-PD-1 + carboplatin/pemetrexed (3), anti-PD-1 + eribulin (2), anti-PD-1 + ICOS antagonist (1), anti-PD-1 + anti-TGF-beta (1), anti-PD-1 + dabrafenib/trametinib (1), anti-PD-1 + lirilumab (1), anti-PD-1 + bevacizumab (1), anti-PD-1 + ribociclib (1), anti-PD-L1 + emactuzumab (2), anti-PD-L1 + vemurafenib/cobimetinib (2), anti-PD-L1 + anti-IDO (1); anti-PD-L1 + bevacizumab (1); anti-PD-L1 + prexasertib (1).
d Other pre-biopsy clinical diagnoses included: rash NOS (4), Stevens-Johnson syndrome/erythema multiforme (3), panniculitis (2), SCLE (2), acute cellulitis (1), eosinophilic fasciitis (1), eruptive keratoacanthoma (1), fixed drug eruption (1), granulomatous (1), miliaria rubra (1), mucositis (1), pityriasis rosea (1), Sweet syndrome (1), urticarial reaction (1), vasculitis (1).
e Other post-biopsy clinicopathologic diagnoses included: Mucositis (2), Panniculitis (2), Stevens-Johnson syndrome/erythema multiforme (3), Sweet syndrome (2), vasculitis (2), eosinophilic fasciitis (1), follicular (1), granulomatous (1), SCLE (1), phototoxic drug eruption (1).
f P-value for time from cirAE onset to biopsy is calculated from a two-tailed t-test.
Abbreviations: cirAE, cutaneous immune-related adverse event; ICOS, inducible T cell costimulator; IDO, indoleamine 2,3-dioxygenase; IQR, interquartile range; NC, not calculated; NOS, not otherwise specified; PD-1, programmed death-1; PD-L1, programmed death-ligand 1; SCLE, subacute cutaneous lupus erythematosus; TGF, transforming growth factor.
Supplemental Table 2: Comparison of diagnostic accuracy between oncodermatologist sub-groups and general dermatologists
For all cirAE cases included in this cohort, the biopsy-ordering dermatology provider was categorized as a general dermatologist or supportive oncodermatologist. Each provider was classified as a supportive oncodermatologist if, at the time of documented clinical care, they: (1) provided dermatologic clinical care for patients experiencing skin toxicities while embedded within a cancer center, or (2) served as a consultant on a hospital service dedicated to managing cancer patients experiencing drug-related skin toxicities. The providers meeting criteria 1 were considered “cancer center-affiliated,” and providers meeting criteria 2 were considered “inpatient dermatology service-affiliated.” The providers meeting neither of these criteria were classified as general dermatologists. Two-by-two Fisher’s exact tests were used for pairwise comparisons as specified. The rates of initial correct cirAE classification were similar amongst the total OD group (72.7%), cancer center-affiliated OD group (72.9%), and hospital dermatology-affiliated OD group (75.0%). The cancer center-affiliated OD group has a statistically significantly increased frequency of initial correct cirAE classification compared to the GD group (72.9% vs 44.7%, p=0.010). The hospital dermatology-affiliated OD group also had greater frequency of initial correct cirAE classification compared to the GD group (75.0% vs 44.7%, but this did not achieve statistical significance (p=0.072), likely due to small sub-group size (N=16).
Abbreviations: GD = general dermatologist. OD = oncodermatologist. OD(CC) = oncodermatologist, cancer-center affiliated. OD(IDS) = oncodermatologist, inpatient dermatology service-affiliated.
Supplemental Figure 1: Clinical photographs of common cirAE classifications
1A: Lichenoid eruption. 1B: Eczematous eruption. 1C: Dermal hypersensitivity (morbilliform) reaction. 1D: Autoimmune bullous eruption. 1E: Psoriasiform eruption.
Data Availability Statement
Available upon written reasonable request.