Photodistributed interface dermatitis on a checkpoint inhibitor
A 61-year-old man with metastatic melanoma has received nivolumab for 14 weeks. He develops symmetric, violaceous, scaly plaques over the dorsal hands, forearms, anterior chest, and V of the neck, together with Gottron-like papules over the metacarpophalangeal joints and periungual erythema. He has no proximal weakness; creatine kinase and aldolase are normal, and ANA is 1:80 speckled. A 4-mm punch biopsy from the forearm shows vacuolar interface change, scattered dyskeratotic keratinocytes, a sparse perivascular lymphocytic infiltrate with occasional eosinophils, and abundant alcian-blue-positive dermal mucin. Direct immunofluorescence from the same site shows granular C3 along the dermoepidermal junction and around eccrine ducts, with no IgG, IgM, or IgA. The local laboratory cannot perform myositis-specific antibody testing. He asks whether the eruption means his cancer treatment must stop.
1. Rank the differential diagnosis and identify the histologic findings that carry the most weight. Does the presence of eosinophils settle the question?
Read the worked answer
The differential is between a checkpoint-inhibitor lichenoid eruption (including a dermatomyositis-like variant), dermatomyositis, and subacute cutaneous lupus.
Findings that do the most work:
- Interface change with abundant dermal mucin, in a photodistributed eruption with Gottron-like papules, is the classic dermatomyositis pattern. Mucin is not a feature of a simple lichenoid drug reaction.
- The inflammation is sparse relative to the interface damage (a pauci-inflammatory interface dermatitis), which is another clue to dermatomyositis, particularly the anti-MDA5 phenotype.
Eosinophils are common in drug reactions, including checkpoint-inhibitor reactions, but they are neither sensitive nor specific; they also occur in dermatomyositis and lupus. They shift probability modestly toward drug and do not settle it. The temporal relationship to nivolumab is suggestive but not decisive, because checkpoint inhibitors can unmask or trigger a genuine autoimmune dermatomyositis-like syndrome.
The highest-yield next step is clinical and serologic, not another biopsy: examine for proximal weakness, nailfold capillary changes, and mechanics' hands; send a full myositis antibody panel; obtain pulmonary function tests; and consider MRI or EMG if symptoms emerge.
2. How do you interpret the direct immunofluorescence, and what DIF findings would have changed your interpretation?
Read the worked answer
The DIF is not diagnostic. Granular C3 at the dermoepidermal junction and around eccrine ducts is described in dermatomyositis and also in lupus erythematosus, and nonspecific junctional C3 can be seen in drug eruptions. The absence of IgG, IgM, and IgA makes a classic lupus band less likely, but the band is IgG-predominant and is best sought in lesional skin that is at least 2 to 4 weeks old, and in nonlesional sun-exposed skin when systemic lupus is suspected.
Findings that would have changed my interpretation:
- Strong granular IgG, with or without IgM, IgA, and C3, at the junction, especially if also present in nonlesional sun-exposed skin: favors lupus erythematosus.
- Linear or shaggy fibrinogen along the junction with cytoid bodies: favors lichen planus or a lichenoid drug reaction.
- A completely negative DIF: does not exclude dermatomyositis, which is often DIF-negative.
- Dermal vascular C3: described in dermatomyositis but nonspecific.
DIF is a supporting test, not a tiebreaker, and a nonspecific result should not override a convincing clinical and histologic picture.
3. Two weeks later he develops exertional dyspnea and a rising ferritin. How does that reshape the differential and your next steps, and what does it mean for continuing nivolumab?
Read the worked answer
New exertional dyspnea with a rising ferritin in this setting is a red flag for anti-MDA5 dermatomyositis with rapidly progressive interstitial lung disease. Anti-MDA5 disease is often clinically amyopathic, and the mucocutaneous findings (palmar violaceous papules, oral ulcers, a pauci-inflammatory interface dermatitis) may precede or accompany lung disease. Ferritin elevation correlates with disease activity and macrophage activation.
Next steps: urgent high-resolution CT of the chest, pulmonary function tests with diffusing capacity, and urgent pulmonology input; send the myositis-specific antibody panel including MDA5. If the syndrome is confirmed, treat aggressively with systemic corticosteroids plus a steroid-sparing agent such as a calcineurin inhibitor, mycophenolate, or rituximab according to local protocol, because untreated rapidly progressive lung disease carries high mortality.
Regarding nivolumab: this is now a systemic autoimmune syndrome with potential organ threat, so I would hold the checkpoint inhibitor and discuss the risk-benefit with oncology. Rechallenge would be a multidisciplinary decision, not a dermatology-only one.