Data from Molecular Subtype Not Immune Response Drives Outcomes in Endometrial Carcinoma

Abstract

Listen

Translate article

<div>AbstractPurpose:<p>Tumors with high mutation load are thought to engender stronger immune responses, which in turn promote prolonged patient survival. To investigate this, we assessed tumor-infiltrating lymphocytes (TILs) and immunosuppressive factors across the 4 molecular subtypes of endometrial cancer, which have characteristic mutation rates ranging from low to ultra-high.</p>Experimental Design:<p>A total of 460 endometrial cancers were stratified by ProMisE (Proactive Molecular Risk Classifier in Endometrial cancer) into 4 molecular subtypes: mismatch repair-deficient (MMRd), <i>POLE</i> mutant (POLE), p53 abnormal (p53abn), and p53 wild-type (p53wt). Immune markers (CD3, CD8, CD79a, CD138, PD-1, PD-L1, FoxP3, IDO-1) were quantified by multiplex IHC and tested for associations with ProMisE subtype, survival, and other clinicopathologic parameters.</p>Results:<p>Two major TIL patterns were observed. TIL<sup>high</sup> tumors harbored dense T- and B-lineage infiltrates and multiple immunosuppressive features and were common in molecular subtypes associated with high mutation load (MMRd and POLE); however, equally strong responses were seen in significant numbers of p53abn and p53wt tumors, which have characteristically low mutation loads. TIL<sup>low</sup> tumors were generally devoid of immunologic features and were more prevalent in p53abn and p53wt endometrial cancers, yet were also seen in MMRd and POLE subtypes. In multivariable models involving ProMisE subtype, T-cell markers, and TIL clusters, only ProMisE showed independent prognostic significance.</p>Conclusions:<p>Immune response correlates with endometrial cancer molecular subtype but does not carry independent prognostic significance. Profound variation in immune response is seen across and within endometrial cancer molecular subtypes, suggesting that assessment of immune response rather than molecular subtype may better predict response to immunotherapy.</p><p><i>See related commentary by Mullen and Mutch, p. 2366</i></p></div>