IFN-γ signature enables selection of neoadjuvant treatment in patients with stage III melanoma
Irene L. M. Reijers, Disha Rao, Judith M. Versluis, Alexander M. Menzies, Petros Dimitriadis, Michel W.J.M. Wouters, Andrew J. Spillane, W. Martin C. Klop, Annegien Broeks, Linda J.W. Bosch, Marta López‐Yurda, Winan J. van Houdt, Robert V. Rawson, Lindsay G. Grijpink-Ongering, Maria Gonzalez, Sten Cornelissen, Jasper Bouwman, Joyce Sanders, Elsemieke I. Plasmeijer, Yannick S. Elshot, Richard A. Scolyer, Bart A. van de Wiel, Daniel S. Peeper, Alexander C. J. van Akkooi, Georgina V. Long, Christian U. Blank
Abstract
Neoadjuvant ipilimumab + nivolumab has demonstrated high pathologic response rates in stage III melanoma. Patients with low intra-tumoral interferon-γ (IFN-γ) signatures are less likely to benefit. We show that domatinostat (a class I histone deacetylase inhibitor) addition to anti-PD-1 + anti-CTLA-4 increased the IFN-γ response and reduced tumor growth in our murine melanoma model, rationalizing evaluation in patients. To stratify patients into IFN-γ high and low cohorts, we developed a baseline IFN-γ signature expression algorithm, which was prospectively tested in the DONIMI trial. Patients with stage III melanoma and high intra-tumoral IFN-γ scores were randomized to neoadjuvant nivolumab or nivolumab + domatinostat, while patients with low IFN-γ scores received nivolumab + domatinostat or ipilimumab + nivolumab + domatinostat. Domatinostat addition to neoadjuvant nivolumab ± ipilimumab did not delay surgery but induced unexpected severe skin toxicity, hampering domatinostat dose escalation. At studied dose levels, domatinostat addition did not increase treatment efficacy. The baseline IFN-γ score adequately differentiated patients who were likely to benefit from nivolumab alone versus patients who require other therapies.