Anti-PD-1-Induced Pneumonitis Is Associated with Persistent Imaging Abnormalities in Melanoma Patients.

Pneumonitis may complicate anti-programmed death-1 (PD-1) therapy, although symptoms usually resolve with steroids. The long-term effects on respiratory function, however, are not well defined. We screened melanoma patients treated with anti-PD-1, with and without ipilimumab (anti-CTLA-4), and identified 31 patients with pneumonitis. Median time to radiographic findings was 4.8 months. Twenty-three patients (74%) presented with respiratory symptoms, whereas 8 (26%) were asymptomatic, and 11 (35%) were hospitalized. With 22.1 months median follow-up, 27 patients (87%) had resolution of symptoms, whereas 4 had persistent cough, dyspnea, and/or wheezing. By contrast, the rate of radiographic resolution was lower: Only 11 (35%) had complete radiographic resolution, whereas 14 (45%) had improvement of pneumonitis with persistent scarring or opacities, and 6 (19%) had persistent or worsened ground-glass opacities and/or nodular densities. Persistence (vs. resolution) of radiographic findings was associated with older age and initial need for steroids but not with need for hospitalization, timing of onset, or treatment regimen (combination vs. monotherapy). Among patients with serial pulmonary function tests, lung function improved with time. Although symptoms of anti-PD-1-induced pneumonitis resolved quickly, scarring or inflammation frequently persisted on computerized tomography. Therefore, further study of subclinical pulmonary effects of anti-PD-1 is needed.

Clinical characterization of colitis arising from anti-PD-1 based therapy.

Colitis is a frequent, clinically-significant immune-related adverse event caused by anti-programmed death-1 (PD-1). The clinical features, timing, and management of colitis with anti-PD-1-based regimens are not well-characterized. Patients with advanced melanoma that received either anti-PD-1 monotherapy ("monotherapy") or combined with ipilimumab ("combination therapy") were screened from 8 academic medical centers, to identify those with clinically-relevant colitis (colitis requiring systemic steroids). Of 1261 patients who received anti-PD-1-based therapy, 109 experienced colitis. The incidence was 3.2% (30/937) and 24.4% (79/324) in the monotherapy and combination therapy cohorts, respectively. Patients with colitis from combination therapy had significantly earlier symptom onset (7.2 weeks vs 25.4 weeks, p < 0.0001), received higher steroid doses (median prednisone equivalent 1.5 mg/kg vs 1.0 mg/kg, p = 0.0015) and experienced longer steroid tapers (median 6.0 vs 4.0 weeks, p = 0.0065) compared to monotherapy. Infliximab use and steroid-dose escalation occurred more frequently in the combination therapy cohort compared to monotherapy. Nearly all patients had resolution of their symptoms although one patient died from complications. Anti-PD-1 associated colitis has a variable clinical presentation, and is more frequent and severe when associated with combination therapy. This variability in checkpoint-inhibitor associated colitis suggests that further optimization of treatment algorithms is needed.