Background: Acral melanoma (AM) is less responsive to immunotherapy than nonacral cutaneous melanoma. Variable responses are seen during immunotherapy, including pseudoprogression, hyperprogressive disease (HPD) and heterogeneous responses. There are currently no studies on the response patterns of patients with AM treated with immunotherapy and the impact on the outcome.
Objectives: To evaluate the response patterns and prognosis of patients with AM treated with anti-programmed death (PD)-1 antibodies.
Methods: Patients with advanced AM treated prospectively in five clinical trials of anti-PD-1 monotherapy at Peking University Cancer Hospital were included. Responses of individual metastases and heterogeneous responses were evaluated during immunotherapy. Cox proportional hazards regression analysis was conducted to identify the possible predictive factors and generate a nomogram to predict the risk of 1-year and 2-year mortality.
Results: The overall response rate was 18·0%, the disease control rate was 36·1%, median progression-free survival was 3·5 months [95% confidence interval (CI) 1·7-5·3] and median overall survival was 17·5 months (95% CI 15·1-19·9) for anti-PD-1 monotherapy. Overall, 9·8% of patients met the criteria of HPD, and displayed a dramatically worse outcome than patients without HPD. In total, 369 metastatic lesions were assessed, with the highest response rate in lymph nodes (20·4%) and the lowest in the liver (5·6%). Homogeneous response, heterogeneous response and heterogeneous or homogeneous progression had different prognoses from the best to the worst. A predictive model was constructed and achieved good accuracy with a C-index of 0·73 (95% CI 0·63-0·84) in the training set and 0·74 (95% CI 0·61-0·86) in the validation set.
Conclusions: HPD during immunotherapy serves as an essential biomarker of poor prognosis in advanced AM. Metastases in different sites respond distinctively to immunotherapy. Clinically heterogeneous responses to immunotherapy affect the outcome of patients. A predictive model was built to distinguish the prognosis of acral melanoma under immunotherapy.
© The Author(s) 2022. Published by Oxford University Press on behalf of British Association of Dermatologists.