ABSTRACT: Cutaneous paraneoplastic syndromes are a group of dermatoses that demonstrate a range of morphological and pathological findings. These syndromes may precede, be concurrent with, or follow the diagnosis of an underlying malignancy. Treatment of the malignancy is often associated with improvement in or resolution of the mucosal and cutaneous disease; however, this is not the case with paraneoplastic pemphigus (PNP). PNP is a rare syndrome that was first described in 1990, and it occurs almost exclusively in patients with lymphocytic neoplasms. Pulmonary manifestations occur in 30% to 40% of cases, and it is the only form of pemphigus that attacks epithelium other than squamous epithelium in an antibody-mediated fashion. The mortality rate for PNP associated with malignancy is greater than 90%. Treatment guidelines are not available, but case series point to the use of rituximab(Drug information on rituximab) (Rituxan) as well as corticosteroids and various other immunomodulating agents. Here we present a diagnostic and treatment dilemma in a 39-year-old active-duty male who developed PNP in the setting of treatment with rituximab, cyclophosphamide(Drug information on cyclophosphamide), doxorubicin(Drug information on doxorubicin), vincristine, and prednisone(Drug information on prednisone) (R-CHOP) for grade 3 follicular lymphoma. This case report is followed by a review of the diagnosis and treatment of other cutaneous paraneoplastic syndromes that are associated with hematologic malignancies.
Paraneoplastic pemphigus (PNP) is an autoimmune-mediated, blistering, erosive, and painful mucocutaneous disease that is diagnosed almost exclusively in patients with lymphocytic neoplasms. The most commonly associated neoplasms are, in decreasing order of frequency, non-Hodgkin lymphoma, chronic lymphocytic leukemia, Castleman disease, thymoma, retroperitoneal sarcoma, and Waldenström macroglobulinemia. When PNP is associated with an underlying malignancy, it is extremely resistant to treatment and carries a 90% mortality rate within a few months following diagnosis, independent of the course of malignancy. PNP is characterized by the presence of IgG autoantibodies that react to desmosomal and hemidesmosomal plakin proteins as well as to desmosomal transmembrane proteins called desmogleins, and these reactions result in the loss of cell adhesion of epithelial cells.
The original diagnostic criteria for PNP were devised in 1990 based on initial case reports; these criteria included:
1. Mucosal ulcerations and blisters, and polymorphous skin lesions in the context of an underlying neoplasm.
2. Histological findings of vacuolar interface change, keratinocyte necrosis, and intraepidermal acantholysis.
3. Epidermal intercellular deposition of IgG and C3, with or without linear deposition at the basement membrane zone, as demonstrated by direct immunofluorescence (DIF).
4. Serum autoantibodies that bind to the cell surface of skin mucosa in a pattern typical for pemphigus—but that also bind to simple, columnar, and transitional epithelia, as demonstrated by indirect immunofluorescence (IIF) in murine esophageal and bladder tissue.
5. Serum autoantibodies that immunoprecipitate a complex of high–molecular weight proteins from keratinocytes, with relative molecular weights of 250 kDa (desmoplakin 1), 230 kDa (bullous pemphigoid antigen [BP Ag 1]), 210 kDa (desmoplakin 2 and envoplakin), 190 kDa (periplakin), and 170 kDa (possible novel epidermal adhesion molecule antibody).
With additional experience, it became obvious that not all persons with PNP would meet all these criteria. Subsequently, minimal criteria for the diagnosis of PNP were proposed (Table 1).
PNP is generally resistant to therapy, especially with regard to stomatitis and lung involvement. Although there are no available treatment guidelines, and although PNP response may not follow the course of tumor resolution, initial treatment is aimed at the underlying malignancy in an attempt to slow the autoantibody production. Numerous immunosuppressive agents have been tried, with varying results. Systemic corticosteroids are used the most frequently, often in combination with other immunomodulatory modalities, such as rituximab, azathioprine(Drug information on azathioprine), cyclosporine, methotrexate(Drug information on methotrexate), mycophenolate mofetil, cyclophosphamide, and IVIg.
A 39-year-old active-duty male with grade 3 follicular lymphoma was transferred from Germany for therapy. He was treated with and responded to 2 cycles of R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone), with a 30% decline in the size of his largest mass (Figure 1). Between cycles 1 and 2, the patient developed what appeared to be grade 2 mucositis, dry eyes, and a viral rash (Figure 2). After multiple admissions, evaluations, and biopsies by ophthalmology, dermatology, and oral maxillary facial surgery for continued ocular discomfort, new-onset photophobia, and grossly worsening mucosal ulcerations, the patient was diagnosed with mucosal PNP and lichen planus of the skin. The diagnosis of PNP was not obvious given the nonspecific purple, polygonal, pruritic skin papules (Figure 3), which can also be seen in lichen planus, and a shave biopsy that revealed lichenoid interface dermatitis (Figure 4). Also, the first oral biopsies only demonstrated inflammation.
It was not until progression of his mucosal manifestations that a larger left buccal surgical biopsy was performed and the specimen sent for direct immunofluorescence (DIF), which strongly favored PNP (Figure 5). On receipt of the DIF results, the patient was started on 1mg/kg of oral prednisone, and a specimen was sent for indirect immunofluorescence (IIF). The staining pattern was diagnostic for PNP: prominent IIF staining was observed when the patient’s serum was cultured on mouse bladder and mouse tongue. IIF staining on monkey esophagus was also observed at a dilution of 1:80.
Within three days of initiation of systemic corticosteroids, the patient experienced a notable improvement in his oral and dermal lesions. However, one week later, he was admitted with shortness of breath and a new oxygen requirement. Bronchoscopy demonstrated bronchiolitis obliterans with no underlying infections. Oral cyclosporine was started.
Worsening of the oral lesions was noted seven weeks after prednisone was started, so a difficult multidisciplinary decision was made to start the patient on IVIg in the setting of ongoing CHOP therapy, oral cyclosporine, and a corticosteroid taper. Rituximab was not reinitiated because of the agent’s black-box warning regarding severe mucocutaneous reactions, limited data on its use in the setting of pulmonary involvement, and the fact that the patient began experiencing PNP-related symptoms while undergoing lymphoma treatment with rituximab. The addition of IVIg led to a complete resolution of the patient’s cutaneous lesions and a substantial improvement in his mucosal and ocular lesions, which allowed him to eat solid foods and to visualize objects clearly and without photophobia. One month later, the patient was admitted for fever and worsening shortness of breath; the repeat bronchoscopy demonstrated aspergillosis, and voriconazole(Drug information on voriconazole) (Vfend) was started. During the same hospital stay, a CT scan demonstrated stable lymphomatous masses after cycles 5 and 6 of chemotherapy, with an overall response of 39%. A repeat pulmonary evaluation one month later showed no signs of infection. However, the patient’s respiratory status declined over the next few months, and he experienced multiple admissions, increasing oxygen requirements, and the need for mechanical ventilation. He passed away 11 months after being diagnosed with follicular lymphoma and 9 months after recognition of the associated PNP.
PNP is a rare syndrome that was first described by Anhalt in 1990; approximately 150 cases have been reported to date. The clinical features vary greatly from patient to patient, but in all reported cases the patients developed severe treatment-resistant stomatitis. Unlike other types of pemphigus, PNP may involve not only stratified epithelium but also the columnar epithelium of the lungs. Pulmonary involvement, which is seen in approximately 30% of cases, manifests as obstructive lung disease, progresses to bronchiolitis obliterans, and almost always results in death. Thus, shortness of breath and hypoxia even in the setting of a clear chest X-ray are ominous signs and may be the first indication of pulmonary involvement.
Skin findings are polymorphous in nature and appear to change as the disease progresses. The skin lesions can appear as blisters that rupture and leave erosions, like those seen in bullous pemphigus, or they may appear as lichenoid eruptions. An important difference between PNP and pemphigus vulgaris is that PNP lesions are often seen on the palms and soles, while it would be very unusual to see lesions in these locations in pemphigus vulgaris.
The histopathology of PNP lesions is as varied as the clinical features of the disease. Biopsy specimens from areas of stomatitis are often misleading, since the mucosal disease is typically severe but yields nonspecific inflammatory changes, as seen in our patient. In such cases, it is helpful to biopsy perilesional epithelium, as this is likely to demonstrate suprabasilar acantholysis of the oral mucosa. Suprabasilar acantholysis is also seen in biopsy specimens of intact cutaneous blisters, while keratinocyte necrosis and lymphocytic infiltration into the epidermis can be seen in non-blistering lesions. The diverse pathological findings are likely secondary to the fact that PNP is presumed to be the result of an antitumor immune response involving a combination of both humoral and cell-mediated immunity. It is this spectrum of clinical findings and histopathology that leads to a broad differential diagnosis as well as a delay in diagnosis—and subsequently a delay in treatment, due to the need to wait for the more definitive immunofluorescence results.
Indirect immunofluorescence (IIF) assists in differentiating PNP from pemphigus vulgaris and pemphigus foliaceous. The pemphigus vulgaris antigen and the pemphigus foliaceous antigen are only expressed in stratified squamous epithelia. Patients with PNP have antibodies present in their serum that cross-react with desmoplakins, which are present in all epithelial structures and even in some nonepithelial structures. Therefore, the serum of a patient with PNP will react via IIF with both the stratified squamous epithelium of monkey esophagus and the transitional epithelium of mouse bladder. It is estimated that this serological test is 75% sensitive and 83% specific.
There are no treatment guidelines for PNP, but theoretically three treatment strategies are available. These options include: treatment of the underlying malignancy, treatment of the autoimmune process by reducing serum levels of autoantibodies, and suppression of the immune system. When PNP is associated with benign tumors such as Castleman tumors and thymoma, surgical resection is indicated and is associated with an improvement or complete resolution of symptoms, as well as a decline in the titer of circulating autoantibodies. However, resection is not possible in most cases that are driven by malignancy, and the course of the disease is progressive and independent of the underlying malignancy. In most cases, PNP progresses and leads to death from sepsis, gastrointestinal bleeding, multiorgan failure, or respiratory failure within weeks to two years of diagnosis.
While corticosteroids are the mainstay of therapy, additional immune-modulating agents are added for patients who do not respond to corticosteroids alone. Multiple agents, including gold, azathioprine, cyclosporine, methotrexate, mycophenolate mofetil, cyclophosphamide, intravenous immunoglobulins, and rituximab, have been utilized with varying results. No therapeutic regimen has shown consistent efficacy. One case series demonstrated some response to rituximab in five out of six cases, with the most improvement seen in skin lesions but only minimal response in oral lesions. Additional case reports show a 50% success rate with rituximab. However, it is important to keep in mind that rituximab also carries a black-box warning for severe, sometimes fatal, mucocutaneous reactions, including Stevens-Johnson syndrome, which can appear similar to the lesions seen in PNP. Rituximab has also been reported to cause lung injury. Our case and treatment dilemma are unique in that the patient began to experience oral, ocular, and dermal lesions, along with shortness of breath, while receiving rituximab.
The biggest medical-legal pitfall in the management of PNP is failure to make a prompt diagnosis. Our case highlights the need for continuous scrutiny of diagnoses, especially when treatments expected to be beneficial yield a suboptimal clinical response; the case also showcases the difficult decisions regarding treatment of PNP in a patient receiving rituximab.