In our study, we firstly explored the differences in clinical characteristics between three types of rare EC (UCCC, USC, and UMC) and type I EC (UEC). We found that compared to patients with UEC, patients with high-risk pathological types of EC (UCCC, USC, and UMC) were older and had a higher incidence of menopause status, which was consistent with previous research results [14, 15]. This phenomenon could be partly explained by the UEC being caused by higher estrogen exposure. Furthermore, we found that age was an independent risk factor for patients’ prognosis in the entire EC population included or some subgroup analyses. Previous studies have suggested a correlation between the occurrence of EC and high BMI [16]. However, we did not find significant differences between different pathological types in BMI. BMI was not an independent risk factor for patient prognosis in the entire EC population included, nor was it in subgroup analysis. The impact of BMI on carcinogenesis and patient prognosis in EC needs further exploration in the future.

We also explored whether different pathological types could serve as independent prognostic factors for EC. We concluded that the pathological subtypes of USC and UCCC were unfavorable prognosis factors for OS and PFS, while the UMC subtype was not. Compared to UEC, further subgroup analyses revealed that UCCC and USC were unfavorable prognosis factors for OS only in the early (stage I–II) and advanced stages (stage III–IV), respectively. On the contrary, UCCC or USC were no longer considered unfavorable prognosis factors in the early (stage I–II) and advanced (stage III–IV) stages for PFS as no significant differences were achieved in the corresponding subgroup analysis. In the future, it is necessary to collect more patients and further explore the impact of different pathological types on patients’ prognoses through more nuanced groups.

Although USC solely accounts for 10% of EC, it leads to nearly 40% of EC-related deaths [17]. Similar to previous studies, we also found that the prognosis of USC was far worse than that of UEC. In our study, the fractions of USC with stage III–IV (34.9%), myometrial infiltration (46.2%), cervix involvement (21.7%), and lymph node metastasis (29.2%) were the highest among all pathological subtypes, which could partly explain its negative effects on unfavorable prognosis. USC was an independent unfavorable prognosis factor for OS when patients were diagnosed at stage III-IV, indicating that once USC had pelvic and peritoneal metastasis, its biological behavior was closer to that of ovarian high-grade serous carcinoma, namely metastatic dissemination. Previous studies also have revealed that USC could share a similar biological behavior with advanced ovarian serous cancer, with high genomic mutation rates of HRD signaling pathway and disordered cell-cycle regulation [18,19,20]. All these findings may deepen the pathogenesis of USC, and contribute to finding suitable therapeutic treatments.

UCCC is another rare pathological subtype with high malignancy risk, accounting for approximately 2 to 5% of all EC cases [21, 22]. Previous studies revealed that patients with UCCC were usually diagnosed at an advanced stage, and could be susceptible to chemoresistance [12, 23]. Here, we found that among UCCC patients, 20.5% were in stage III/IV, a proportion significantly higher than that observed in patients with UEC (7.3%). Patients with UCCC had significantly poorer OS and PFS than those with UEC. Further subgroup analyses revealed that only patients with UCCC in stage I/II achieved unfavorable OS, while those with USC or UMC did not. Actually, we found that 47.8% of included patients with UCCC in stage I/II did not undergo postoperative adjuvant radiotherapy or chemotherapy, which was inconsistent with current NCCN guidelines [24]. Combined with its remarkably negative impact on the prognosis in stage I/II, we speculated that the poorer prognosis of early-stage UCCC could be due to the low proportion of postoperative adjuvant therapy, similar to some previous studies [25, 26]. Based on the above findings, our study supports the application of postoperative adjuvant treatment (chemotherapy, radiotherapy, or chemoradiotherapy) in early-stage UCCC patients. In the future, more UCCC samples should be included for further analysis.

UMC, as an extremely rare pathological type, accounting for approximately 3–8% of EC cases, has drawn attention in recent years. In 2014, the World Health Organization (WHO) defined UMC as a mixed EC composed of two or more pathological types, with at least one type II EC accounting for 10% [27]. Currently, whether the coexistence of type II EC components will affect the prognosis of patients remains elusive. A large-scale clinical study containing 934 patients compared the prognostic differences between UMC and pure USC, and no significant differences were found regarding OS and PFS [28]. The conclusions drawn from other studies with small sample sizes also varied greatly. Boruta et al. found that when the proportion of USC in UMC components was greater than 50%, patients had poorer PFS and OS [29]. Nevertheless, Nikolaos Thomakos et al. found that there was no difference in the prognosis between UMC and other type II EC, regardless of the proportions of other type II EC components in UMC [30]. In our study, we compared the prognostic differences between UMC and UEC, and we found there was no significant difference in prognosis between UMC and UEC, which may be due to the components of involved pathological types. Here, the major components of UMC were endometrioid carcinoma and other type II EC (82.35%), and the presence of endometrioid adenocarcinoma may improve the prognosis of patients to some extent. However, for those UMC patients completely composed of type II EC, it is still uncertain whether it will lead to a worse clinical prognosis due to the sample size of this study. Moreover, different molecular typing can also have a certain impact on the prognosis of patients. In the future, it is necessary to further expand the sample size and improve molecular typing for better analysis.

Although our study has concluded some novel findings, it still has its inherent limitations. Firstly, as few type II EC patients with different pathological types were included, we were unable to identify specific factors affecting patient prognosis for each pathological subtype. Secondly, the proportion of included patients receiving surgery was very high (UEC: 396/396; UCCC: 32/34; USC: 102/106; UMC: 34/34), so surgery could not be used as a prognostic factor for subsequent analysis. Thirdly, the included patients rarely received postoperative adjuvant radiotherapy alone, so radiotherapy alone could not be further analyzed in subgroup analysis. Last but not least, uterine carcinosarcoma is one of the main types of type II EC. However, we found that only limited EC patients with uterine carcinosarcoma met the inclusion criteria, and we did not include them in our study as a subgroup for subsequent analysis. We should include more eligible uterine carcinosarcomas by performing a multicenter retrospective analysis in the future.