Infertility is a global medical problem that affects individuals, families, and society. With the development of technology, assisted reproductive technology (ART) has solved the fertility problems of most infertile patients, but 10–15 % of patients still experience recurrent implantation failure (RIF) (Clifford et al., 1999; Coughlan et al., 2014; Cruz and Bellver, 2014; Quenby et al., 1999; Yin et al., 2017; Zhang et al., 2022). Meanwhile, with the development of society, the reproductive needs of older women are increasing. These people often requires ART to solve reproductive problems, and frozen embryo transfer(FET)has become the main means of treatment. However, embryo implantation is a complex process, and its success depends largely on embryo quality, endometrial receptivity, and synchronization of the two. The core of the precise synchronization between the endometrial environment and the embryo is a finely regulated immune balance, which is crucial for successful embryo attachment and subsequent pregnancy (Brazdova et al., 2016, Ebrahimi et al., 2024).

The immune system of the endometrium plays a key role in embryo implantation and pregnancy. Endometrial immune cells play an important role in establishing a receptive environment that supports embryo attachment and growth (Blazheva et al., 2024; Liang et al., 2015; Mor and Cardenas, 2010; Russell et al., 2011). Abnormalities of these immune cells levels or functions can lead to inflammatory responses that affect the receptivity of the endometrium, thereby disrupting embryo implantation and causing RIF (Liang et al., 2015; Saito, 2024). Studies have shown that changes in T cell, B cell, and natural killer (NK) cell populations are associated with embryo implantation failure and recurrent pregnancy loss (RPL)(Ebrahimi et al., 2024; Zhao et al., 2024). Women with RIF have altered ratios of CD56hiCD16-NK cells, T cells, and macrophages, which may lead to immune imbalance at the maternal-fetal interface (Huang et al., 2021, Quenby et al., 1999). CD56hiCD16-NK cells and macrophages are key players in regulating local immune responses, and their dysregulation may lead to pathological conditions, such as inflammation or insufficient endometrial receptivity (Gellersen and Brosens, 2014, Mor et al., 2017). In patients with RIF, studies have shown elevated levels of pro-inflammatory cytokines and altered ratios of immune cells, such as neutrophils and macrophages, may create a hostile intrauterine environment and lead to implantation failure (Clifford et al., 1999, Ebrahimi et al., 2024, Jia et al., 2022).

Although age is known to be an important factor affecting immune function and embryo implantation, it is not an important factor in FET success in young patients (Seifer et al., 2011). Previous studies on the role of endometrial immune cells in reproduction rarely exclude the influencing factor of age and have often included age to build prediction models (McLernon et al., 2016). However, age is closely related to levels of multiple immune cells. Suppose age was included in immune microenvironment studies. In that case, it can easily prevent researchers from discovering the immune cell subpopulations that truly affect outcomes and excessively affect the performance of immune microenvironment models due to age.

Endometrial immune microenvironment is composed by multiple immune cells and their interactions (Lédée et al., 2016). However, there is still no clear conclusion regarding which immune cells are the best choice for evaluating endometrial conditions. Most existing studies have focused on T cells and their subsets, CD56hiCD16-NK cells, macrophages, etc., whereas there are few studies on the relatively small number of immune cell groups in the endometrium. Although the immune cells that have been studied can reflect the situation of the intrauterine environment to a certain extent, they are not sufficient to fully reflect the complexity of the endometrial immune environment. To gain a deeper understanding of endometrial receptivity and the complexity of the immune environment, in this study, we expanded the scope of immune cells, added plasma cells, γδT, and other relatively rare endometrial immune parameters. We subdivided the various subgroups of NK, including CD56dimCD16+NK, CD56hiCD16-NK and CD56hiCD16+ NK, to gain a more comprehensive understanding of the endometrial immune environment. Meanwhile, we used machine learning to establish an immune model that predicts FET outcomes independent of age.