Pyoderma gangrenosum (PG) is an ulcerating neutrophilic dermatosis that is often associated with the underlying systemic disease. For example, PG is often a common presenting symptom in patients with hematologic malignancies, most commonly myelodysplastic syndrome (MDS). Here, we present the case of a patient who developed PG and a lichenoid drug eruption after the initiation of ipilimumab and nivolumab immune checkpoint inhibitor (ICPI) therapy. Lichenoid drug eruptions are well known to be associated with ICPI therapy, particularly nivolumab. However, only one case of PG has been reported in association with ipilimumab and no cases have ever been reported with nivolumab. Awareness that PG can be associated with ICPI therapy in patients with MDS can allow physicians to be better prepared to help in early recognition and early treatment to prevent the spread of disease.
Pyoderma gangrenosum (PG) is a neutrophilic dermatosis that is commonly associated with inflammatory bowel disorders, polyarthritis, and hematologic disorders. Here, we present the case of a patient who developed PG and a lichenoid drug eruption after the addition of ipilimumab and nivolumab therapy to his treatment regimen of azacitidine. While PG is a well-known event associated with hematologic malignancies, it has only been reported once in relation to ipilimumab and has never been reported with nivolumab therapy. Lichenoid drug eruption has, however, been well reported with immune checkpoint inhibitors (ICPIs), particularly nivolumab. Due to severity of PG and different approaches to treatment, it is important for physicians to be aware of this potential cutaneous adverse event (CAE) in the setting of lichenoid dermatitis in patients being treated with ICPI therapy for myelodysplastic syndrome (MDS).
A 75-year-old Caucasian male with no previous dermatologic history presented with a rash that started 2 weeks after ipilimumab and nivolumab initiation for MDS. He was diagnosed with MDS 3 years prior and had completed 17 cycles of azacitidine, which was stopped due to progression of disease. He was started on a Phase II clinical trial of triple therapy on azacitidine (75 mg/m2 intravenous [IV] daily for 5 days every 4 weeks), nivolumab (3 mg/kg IV every 2 weeks), and ipilimumab (3 mg/kg IV every 3 weeks) for MDS.
The rash consisted of dusky, indurated, erythematous papules with central pseudovesiculation distributed on the extremities including dorsal hands and feet but sparing palms and soles [Figure 1]. A 4-mm punch biopsy of the left forearm showed lichenoid dermatitis. This rash was treated with triamcinolone 0.1% cream and a 20-mg prednisone course with 14-day taper without much improvement. The patient was switched to clobetasol ointment with improvement of rash by 14 weeks and complete resolution 30 weeks after rash development.
Sixteen weeks after treatment was started, there were resistant-indurated purpuric plaques on the bilateral lower legs, with focal 2 cm × 1.7 cm ulceration and suppurative drainage of the right lower leg lesion. A wound culture was done which showed Pseudomonas aeruginosa, Serratia marcescens, and Enterococcus faecalis. At first, this was thought to be stasis dermatitis with a venous ulcer on the right lateral leg, but over time, as both lesions enlarged, they became more concerning for PG [Figure 2]. Ipilimumab was permanently discontinued by the treatment team due to concern that ipilimumab was the cause of the PG. After negative tissue culture for bacteria, fungus, and atypical mycobacteria, he was treated with 1 cc of 10 mg/cc intralesional triamcinolone and he coincidentally was on a 5-day course of 60-mg PO prednisone due to concern for autoimmune nephritis while on treatment.
The suspected PG worsened, and a wound culture was done which showed few P. aeruginosa and moderate E. faecalis. He was treated with ciprofloxacin, amoxicillin/clavulanic acid, and isovuconazole. A 4-mm punch biopsy of the right posterior leg showed diffuse dermal edema with fibrosis, increased small blood vessels, lymphocytes, histiocytes, and numerous neutrophils. These findings are nonspecific and most consistent with granulation tissue. Clinicopathologic correlation suggested a diagnosis of PG, which is a diagnosis of exclusion.
Over the next 4 months, the PG slowly improved. The rash was complicated by multiple incidents of cellulitis (treated with intermittent 14-day courses of ciprofloxacin) and venous stasis (treated with compression garment). To date, the patient has received 6 intralesional 10 mg/cc triamcinolone injections. He additionally treats the rash with clobetasol 0.05% cream under occlusion. The left leg purpuric plaque was treated with three intralesional triamcinolone injections and resolved without ulceration. The ulcer on the right leg is still ongoing but is improving in size and depth.
PG is a neutrophilic dermatosis that in 50%–70% of patients is associated with the underlying disease. PG has been reported as the initial symptom in many hematologic malignancies including MDS and is often mistaken to have an infectious etiology. PG has been reported once before in a patient on ipilimumab for metastatic melanoma but has not been reported acutely after the initiation of ICPI therapy for MDS. There were no reports found relating PG to nivolumab or a combination of ipilimumab and nivolumab.
Nivolumab is a programmed death-1 (PD-1) checkpoint inhibitor and ipilimumab is a cytotoxic T-lymphocyte-associated antigen 4 checkpoint inhibitor. They are both used in advanced melanoma and are being tested for efficacy in many other malignancies. Both nivolumab and ipilimumab are monoclonal antibodies that augment the body’s natural antitumor response while simultaneously increasing the risk for immune-related adverse events. CAEs seen in ipilimumab treatment include eczema, morbilliform rash, pruritus, and vitiligo, among others. CAEs seen in nivolumab include lichenoid reactions (as seen in our patient with the initial rash), eczema, and vitiligo. As both nivolumab and ipilimumab enhance the activity of the immune system, it would reason that either could cause PG, which is hypothesized to be an abnormal immune response. Another possible mechanism would be that ICPI could unmask PG in MDS patients. However, there has been one report in the literature of PG after four ipilimumab treatments for metastatic melanoma, which does support that PG can occur in patients on ICPI that do not have the underlying MDS. Other possible drug etiologies include colony-stimulating factors, and more recently, multikinase inhibitors.[8, 9] This is important to consider when a patient is exposed to one of these drugs; however, our patient never received.
Neutrophilic dermatoses have been shown to be a poor prognostic indicator in patients with MDS, which correlates to our patient who developed PG after change in treatment secondary to disease progression. Patients with neutrophilic dermatosis had a 1.8-fold increase in mortality in the setting of MDS. In addition, patients with MDS and deletions in 5q were found to have an increased risk of PG. The deletion of 5q has been associated with NF-κB signaling leading to upregulation of genes involved in pro-inflammatory responses. This supports the idea that increased immune response could put patients at higher risk for developing PG. Whether drug-induced PG is also a poor prognostic indicator is unknown due to limited report of cases. In addition, having a CAE on a PD-1 inhibitor has been linked to longer progression-free intervals, which could correlate to better treatment response. Our patient had PG and lichenoid drug eruption at the same time, indicating that there was off-target immune upregulation in the skin. It is unknown whether PG onset was enhanced or caused by the ICPI; however, it is important to note that PG can occur while on ICPI and can occur simultaneously with other more common CAEs, such as a lichenoid drug eruption. As PG is a diagnosis of exclusion, considering PG in a differential diagnosis in this patient population will allow for earlier recognition and treatment, avoid unnecessary treatments for possible infections, and could have implications for the prognosis of patients.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
The authors declared no funding related to this study.
Conflicts of interest
The authors disclosed no conflicts of interest.
For reprints contact:firstname.lastname@example.org