Endometrial Cancer Arising in Adenomyosis (EC-AIA): A Systematic Review

Simple Summary A systematic review of the literature was performed to assess the clinicopathological characteristics and survival outcomes of endometrial cancer arising in adenomyosis (EC-AIA). From our analysis, EC-AIA is a rare disease that mainly affects menopausal women and shows symptoms similar to endometrial cancer, but has a challenging preoperative diagnosis. The higher prevalence of the non-endometrioid histotype, advanced FIGO stages, and p53-signature might be behind the worse prognosis of EC-AIA compared to endometrial cancer. Abstract Endometrial cancer arising in adenomyosis (EC-AIA) is a rare uterine disease characterized by the malignant transformation of the ectopic endometrium within the adenomyotic foci. Clinicopathological and survival data are mostly limited to case reports and a few cohort studies. We aimed to assess the clinicopathological features and survival outcomes of women with EC-AIA through a systematic review of the literature. Six electronic databases were searched, from 2002 to 2022, for all peer-reviewed studies that reported EC-AIA cases. Thirty-seven EC-AIA patients from 27 case reports and four case series were included in our study. In our analysis, EC-AIA appeared as a rare disease that mainly occurs in menopausal women, shares symptoms with endometrial cancer, and is challenging to diagnose preoperatively. Differently from EC, it shows a higher prevalence of the non-endometrioid histotype, advanced FIGO stages, and p53-signature, which might be responsible for its worse prognosis. Future studies are necessary, to confirm our findings and further investigate this rare condition.


Introduction
Endometrial cancer (EC) is the most common gynecological cancer in developed countries [1][2][3]. In 22.6% of cases, EC coexists with adenomyosis, a benign gynecologic condition, defined as the migration of glands and stroma from the basal layer of the endometrium to the myometrium [4]. In less than 1% of cases, EC has origins in the malignant transformation of ectopic endometrium within the adenomyotic foci, causing a rare disease known as endometrial carcinoma arising in adenomyosis (EC-AIA) [5,6].
In 1959, Colman and Rosenthal established the criteria for the diagnosis of this disease [5], adapting Sampson's criteria for diagnosis of ovarian cancer arising in endometriosis to adenomyosis. Thus, according to Colman and Rosenthal's criteria, EC-AIA was defined by the presence of the following histopathological characteristics: absence of carcinoma in the endometrium or elsewhere in the pelvis; demonstration of carcinoma arising from the epithelium of adenomyosis and not invading from other sites; presence of endometrial stromal cells surrounding the epithelial glands, to support the diagnosis of adenomyosis [5].
Given its rarity, data about the clinicopathological characteristics and survival of EC-AIA are poor and mostly limited to case reports and a few cohort studies [5,7,8].
The aim of this study was to systematically review the literature, to assess the clinicopathological features and survival outcomes of women with EC-AIA.

Study Protocol
This study was performed following an a priori defined protocol. All review stages, including search strategy, study selection, risk of bias assessment, data extraction, and data analysis, were performed independently by 2 authors (M.M., A.R.). In case of disagreement, consensus was achieved through discussion among all authors. Reporting of the whole study followed the PRISMA statement and checklist [9].

Search Strategy
Eligible studies were collected by searching MEDLINE, Web of Sciences, Scopus, ClinicalTrial.gov, Cochrane Library, and Google Scholar from January 2002 to October 2022. Several combinations of the following words were used: endometr*; malignancy; tumour; tumor; neoplas*; cancer; carcinoma; endometrial cancer arising in adenomyosis; EC-AIA; ECAIA. Reference lists of all relevant studies were searched to check for possible eligible studies missed.

Search Selection
All peer-reviewed studies assessing women with EC-AIA were included. We a priori considered the following exclusion criteria: studies with no extractable data, studies with overlapping study populations, review articles, studies reported in languages other than English.

Risk of Bias Assessment
The risk of bias within studies was evaluated using the methodological index for non-randomized studies (MINORS) [10]. In detail, seven domains related to risk of bias were assessed, when applicable, as follows: (1) study aim (had the study a clearly stated aim?); (2) patient inclusion (were all eligible patients included during the study period?); (3) data collection (was data collection performed following an a priori defined protocol?); (4) study endpoints (were the study endpoints appropriate to the study aim?); (5) unbiased study endpoints (were Colman and Rosenthal's criteria for diagnosis of EC-AIA clearly reported? In particular, Colman and Rosenthal's criteria were the following: absence of carcinoma in the endometrium or elsewhere in the pelvis; demonstration of carcinoma arising from the epithelium of adenomyosis and not invading from other sites; presence of endometrial stromal cells surrounding the epithelial glands to support the diagnosis of adenomyosis [6]); (6) follow-up (was follow-up at least 24 months?; such a follow-up is considered enough long for women with EC [11]); (7) loss to follow-up less than 5% (were patients lost to follow-up less than 5% of the total sample?). All seven domains were applicable for case series, while only four domains (i.e., domains #1, #4, #5, and #6) were applicable for case reports. Authors judged each included study for each domain as at "low risk", "unclear risk", or "high risk" of bias if data were "reported and adequate", "not reported", or "reported but inadequate", respectively.

Data Extraction and Analysis
Data were extracted from the included studies without modifications. For each study, data extracted were the study country, study design, period of enrollment, patients' characteristics, EC-AIA histological features, EC-AIA histotype, EC-AIA International Federation of Gynecology and Obstetrics (FIGO) stage, death, recurrence, and expression of immunohistochemical markers.
Kaplan-Meier survival analyses for the risk of recurrence or death were performed and reported graphically via Kaplan-Meier curves; Statistical Package for Social Science (SPSS) 18.0 package (SPSS Inc., Chicago, IL, USA) was used as software for data analysis.

Study Selection
A total of 431 studies were identified through electronic searches. Forty-one articles remained after duplicate removal and abstract screening, and they were assessed for eligibility ( Figure 1). Finally, 31 articles were included in this systematic review .
of endometrial stromal cells surrounding the epithelial glands to support the diagnosis of adenomyosis [6]); (6) follow-up (was follow-up at least 24 months?; such a follow-up is considered enough long for women with EC [11]); (7) loss to follow-up less than 5% (were patients lost to follow-up less than 5% of the total sample?). All seven domains were applicable for case series, while only four domains (i.e., domains #1, #4, #5, and #6) were applicable for case reports.
Authors judged each included study for each domain as at "low risk", "unclear risk", or "high risk" of bias if data were "reported and adequate", "not reported", or "reported but inadequate", respectively.

Data Extraction and Analysis
Data were extracted from the included studies without modifications. For each study, data extracted were the study country, study design, period of enrollment, patients' characteristics, EC-AIA histological features, EC-AIA histotype, EC-AIA International Federation of Gynecology and Obstetrics (FIGO) stage, death, recurrence, and expression of immunohistochemical markers.
Kaplan-Meier survival analyses for the risk of recurrence or death were performed and reported graphically via Kaplan-Meier curves; Statistical Package for Social Science (SPSS) 18.0 package (SPSS Inc., Chicago, IL, USA) was used as software for data analysis.

Study Selection
A total of 431 studies were identified through electronic searches. Forty-one articles remained after duplicate removal and abstract screening, and they were assessed for eligibility ( Figure 1). Finally, 31 articles were included in this systematic review .
The mean follow-up time ±SD was 15.4 ± 22 months; fourteen (37.8%) patients were lost to follow-up. Of women with follow-up available, eight (34.8%) reported recurrence: three patients had a primary diagnosis of clear cell EC-AIA, while five had a primary diagnosis of endometroid EC-AIA. Death was reported for one (4.3%) patient, who had undergone total hysterectomy and bilateral salpingo-oophorectomy without adjuvant treatment for a stage IC, clear-cell EC-AIA. The patient developed an inguinal lymph node metastasis 3 months after surgery and died from metastatic disease 60 months after surgery.
We graphically reported disease-free survival using a Kaplan-Meier curve (Figure 2), while we were unable to perform Kaplan-Meier analysis for death, as the event "death" was observed in only one EC-AIA woman, who also showed the longest follow-up time within the whole patient cohort.

Assessment of Risk of Bias among Studies
For the "study aim", "data collection", and "loss of follow up" domains (when applicable), all studies were considered at low risk of bias.

Assessment of Risk of Bias among Studies
For the "study aim", "data collection", and "loss of follow up" domains (wh plicable), all studies were considered at low risk of bias.
Results of the risk of bias assessment are graphically shown in Figure 3.

Main Findings and Interpretation
This study showed that EC-AIA is a rare disease that most commonly affects menopausal women and shows symptoms similar to EC, with abnormal uterine bleeding as the most frequent symptom. The preoperative diagnosis can be misleading, while postoperative histological examination shows a lower prevalence of the endometrioid histotype and early FIGO stages than EC; conversely, an abnormal expression of p53 and the need for adjuvant treatment are more common. These findings might explain the worse prognosis of EC-AIA compared to EC. EC-AIA arises by transformation of the endometrium within adenomyotic foci. Similarly to EC, it can present with different histotypes, such as endometrioid, serous, clear cell, and primary uterine müllerian mucinous borderline tumor [5].
The causes of the neoplastic degeneration of adenomyosis are still unknown. However, adenomyosis and EC share common genetic mutations in the molecular pathways regulating cellular proliferation [4]. Among these, EC and adenomyosis share low expression of mRNA of the Phosphatase and Tensin Homolog (PTEN), mutations in phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR), and Catenin Beta 1 (CTNNB1) signaling pathways [43], and loss of heterozygosity in the DNA mismatch repair genes [8].
While EC with coexistent adenomyosis has been investigated and its clinicopathological and survival outcomes assessed [4,[44][45][46], only three cohort studies [5,7,8] have been reported on EC-AIA in the last sixty years. In detail, Matsuo et al. and Matchida et al. [7,8] compared cases of EC-AIA reported in the literature to women with EC [7] or EC with coexistent adenomyosis [8] treated in their centers. In both these studies, the authors reported an increased risk of deep myometrial invasion and serous or clear cell histology [7,8] in the EC-AIA group. However, although the presence of EC-AIA was associated with decreased disease-free survival, it was not found to be an independent risk factor for overall survival [7,8]. In a large retrospective observational cohort study by Chao et al. [5], of 2080 patients who underwent surgical treatment for EC, 28 endometroid EC-AIA were identified. When compared to endometroid EC and endometroid EC with coexistent adenomyosis, endometroid EC-AIA showed more favorable histological prognostic factors, such as grade 1, smaller maximum diameter, and less commonly deficient expression of mismatch repair proteins. Moreover, no recurrence or death occurred in these women [5]. Thus, the data from these three studies appeared to be in conflict, with a possible impact from the quality of individual studies in the exploratory analyses by Matsuo et al. and Matchida et al. [7,8] [5].
In our study, we systematically reviewed the literature to assess the clinicopathological features and survival outcomes of women with EC-AIA. We found that this rare disease appeared to be similar to EC regarding the age at diagnosis, menopausal status, and symptoms, while differing from it in prevalence of postoperative histological characteristics and prognosis. Moreover, the disease showed a challenging preoperative diagnosis. In fact, the most frequent preoperative diagnoses in women with EC-AIA were uterine sarcoma, myoma/adenomyoma, and atypical myoma, with an EC-AIA diagnosis never being preoperatively suspected. This might be due, on the one hand, to the rarity of the lesion, which has led to a lack of specific features at ultrasound or magnetic resonance imaging, and, on the other hand, to a possible normal endometrium with negative endometrial biopsies at hysteroscopy, as the disease arises in adenomyotic foci. In addition, the symptoms and phenotype also do not seem to help in the preoperative diagnosis, as they appeared to be similar to those of EC and EC with adenomyosis [44]. Indeed, EC-AIA occurred in postmenopausal women and presented with abnormal uterine bleeding, pelvic pain, and vaginal discharge as the main clinical manifestations. This challenging preoperative diagnosis might have contributed to the lower prevalence of early FIGO stages that we found in women with EC-AIA compared to EC. However, the EC FIGO staging system might need to be adapted for this rare disease: since EC-AIA orginates from the adenomyotic foci into the myometrium, the distinction between FIGO stages IA and IB might be meaningless.
Regarding histology, the most common histotype was well-differentiated endometroid. The hyperestrogenic environment due to adenomyosis might explain the higher frequency of the endometroid histotype among EC-AIA patients [4]. However, the prevalence of such histotype in EC-AIA patients appeared lower than that in women with EC. Indeed, we found that 42% of EC-AIA patients in our study showed non-endometroid histotypes. In particular, similarly to cancer arising from ovarian endometriosis, the clear cell histotype was the most frequent among non-endometroid histotypes. Moreover, the p53-signature also showed a higher prevalence (50%) in women with EC-AIA than that with EC. Since this signature has been related to non-endometrioid histotypes [47], it might explain the higher prevalence of non-endometrioid histotypes in EC-AIA women. In fact, 73% of p53-mutated ECs showed a non-endometrioid histotype, while 42.5% of clear cell ECs showed a p53 mutation [48]. Furthermore, the p53-signature might explain the worse survival outcomes of EC-AIA compared to EC. In detail, in our study, of women with available follow-up data, 34.8% showed recurrence and 4.3% died from metastatic disease.
However, studies investigating the The Cancer Genome ATLAS (TCGA) signature in women with EC-AIA might clarify the independent prognostic value of the molecular signature and the origin of the cancer from the malignant transformation of ectopic endometrium within the adenomyotic foci.

Strengths and Limitations
To the best of our knowledge, this is the first systematic review to assess the clinicopathological features and survival outcomes of women with EC-AIA. In particular, differently from previous reviews and case-control studies [4,6,7], we systematically reviewed the literature and evaluated the risk of bias within studies. Moreover, within this evaluation, our study may be the first study to assess adherence to Colman and Rosenthal's criteria for diagnosis of EC-AIA. Histologic criteria to identify EC-AIA and separate it from EC with coexistent adenomyosis are crucial and have been debated over the last sixty-years [4]. Colman and Rosenthal established criteria for diagnosis, adding to Sampson's criteria the presence of endometrial stromal cells into myometrium [49]. In addition, Kumar and Anderson argued for the necessity of demonstrating transitional changes from benign to malignant endometrium within adenomyosis, to allow a stricter diagnosis of EC-AIA [49]. However, since the transition from benign to malignant endometrium is an inconsistent finding [4], we exclusively considered Colman and Rosenthal's criteria within the risk of bias assessment.
However, our results might have been limited by the small sample size, the study design (mostly case reports), and the overall low quality of the included studies. Nevertheless, our study can provide updated data about a rare disease and direct future studies in the field.

Conclusions
EC-AIA is a rare disease that mainly occurs in menopausal women, shares symptoms with EC, and is challenging to diagnose preoperatively. Differently from EC, it shows a higher prevalence of the non-endometrioid histotype, advanced FIGO stages, and p53signature, which might explain its worse prognosis.
Future studies are necessary to confirm our findings and further investigate this rare condition.