Dedifferentiated endometrial carcinoma arising from serous carcinoma: Diagnostic challenges and recommendations

Highlights

• •Dedifferentiated endometrial carcinoma arising from serous carcinoma (null-type P53).
• •Differential diagnosis including carcinosarcoma, FIGO grade 3 endometrioid carcinoma.
• •Also need to rule out SMARCA4-deficient uterine sarcoma and other sarcomas.

1. Introduction

Undifferentiated endometrial carcinoma (UEC) are uncommon aggressive uterine tumors. Most patients are more than 50 years of age, and some patients may have Lynch syndrome. Histologically, UC exhibits a solid growth pattern and consist of monomorphic, discohesive, small to medium-sized round cells, with absent or minimal neuroendocrine differentiation. If UEC is associated with a component of differentiated carcinoma, the term “dedifferentiated endometrial carcinoma (DDEC)” applies (Mirkovic, 2022, Bennett and Oliva, 2021, Zhang and Zheng, 2022). Initially, the differentiated component of DDEC was strictly defined as a low grade (FIGO grade 1 or 2) endometrioid carcinoma. It is now recognized that DDEC can also arise in the background of high-grade endometrial carcinoma. Here we report a rare case of DDEC where the differentiated component is an endometrial serous carcinoma with null-type P53. We'll review the literate on dedifferentiated endometrial carcinoma rising from serous carcinoma.

2. Case report

The patient is a 70-year-old white woman presented to her annual well woman visit with a chief complaint of seeing blood on wipe after urination. The symptom was seen about three times each day and it started three weeks prior to her presentation. On exam, a tiny polypoid lesion was identified at cervical os with blood coming off the lesion. Pelvic ultrasound showed thickened endometrial lining. Given these findings, a hysteroscopy with dilatation and curettage (D&C) was done and endometrial tissue biopsied. The biopsy result was “high-grade endometrial carcinoma” with a wide differential diagnosis. MMR proteins were retained. A subsequent CT scan confirmed a hypoattenuating mass within endometrial cavity without evidence of parametrial tumor invasion or distant metastasis within abdomen or pelvis.

The patient received robot assisted total hysterectomy, bilateral salpingo-oophorectomy, sentinel lymph node sampling and omental biopsy. Grossly, the lesion was gray-white, soft, and fleshy. It filled and expanded the entire endometrial cavity. The lesion measured 5.2 × 4.9 × 3.8 cm in size and infiltrated less than half the thickness of the myometrium. Ovaries, fallopian tubes, and omentum were unremarkable.

Histology of the resection specimen showed a tumor containing both differentiated (Fig. 1A) component and undifferentiated component (Fig. 1C) with abrupt transition (Fig. 1B). The differentiated component showed high grade tumor cells arranged in papillary or glandular structures. The undifferentiated component was characterized by solid sheets of poorly cohesive and monomorphic cells. There was no glandular or papillary architecture in the undifferentiated component. A 1.8 mm microscopic focus of sarcoma was also identified (Fig. 1D). The differentiated areas were positive for AE1/AE3, EMA, and ER. All these markers, as well as E-cadherin, were negative in the undifferentiated component (Fig. 2A, 2B, and 2C). P53 was null-type (complete loss of staining) in all components including the differentiated and undifferentiated carcinoma components (Fig. 2D). PAX-8 was positive in both the differentiated and undifferentiated carcinoma components while it was negative in the sarcoma component (Fig. 2E). In the sarcoma component, focal positivity of CD10 was seen (Fig. 2F). INI-1(SMARCB1) and BRG1 (SMARCA4) were retained; synaptophysin was negative. The tumor was predominantly exophytic, with superficial myoinvasion (Fig. 1E). Lymphovascular invasion was identified (Fig. 1F); however, sentinel lymph node sampling was negative for metastases. The tumor also showed areas with bizarre nuclei which demonstrated patchy AE1/AE3 and decreased E-cadherin expression. These ares are considered degenerative and/or poorly fixed. Overall, a diagnosis of DDEC with a minor sarcoma component was established. The carcinoma component was consistent with serous carcinoma, given the aberrant P53 expression and histologic features. The final FIGO stage was IA.

Fig. 1.

A. WD component (H&E, x100); B. Abrupt transition from WD to UD component (H&E, x40); C. UD component (H&E, x100); D. Sarcomatous component (H&E, x100); E. Pushing invasion of the tumor into the myometrium (H&E, x40); F. Lymphovascular invasion (H&E, x100) (WD: well differentiated UD: undifferentiated H & E: Hematoxylin and Eosin).

Fig. 2.

IHC: A. AE1/AE3 is positive in the WD component and negative in the UD component; B. ER positive in the WD component and negative in the UC component​; C. E cadherin positive in WD component and negative in the adjacent UC component; D. P53 null in both WD and UD component; E. PAX8 positive in the WD and UD components; F. CD10 positive in the microscopic sarcoma component​; (WD: well differentiated UC: undifferentiated).

The patient was managed with six cycles of carboplatin and paclitaxel chemotherapy and brachytherapy. On one-year follow up, there were no signs of recurrence, and the CEA level was within normal limit.

3. Discussion

Undifferentiated endometrial carcinoma is an aggressive pathologic subtype of endometrial cancer. It was defined as a malignant epithelial neoplasm characterized by a total absence of glandular differentiation and a patternless solid growth of tumor cells, with absent or minimal neuroendocrine differentiation (Altrabulsi et al., 2005). DDEC is an uncommon variant of endometrial carcinoma. Silva et al in 2006 first recognized this entity. They reviewed 25 cases of endometrial carcinoma in which a low-grade endometrioid adenocarcinoma was combined with an undifferentiated carcinoma and designated them as DDEC. They postulated that the undifferentiated component represents a form of tumor progression from a preexisting low-grade endometrioid adenocarcinoma (Silva et al., 2006). Before the 2020 WHO classification of female genital tract tumors, DDEC was strictly defined as composed of undifferentiated carcinoma and a low-grade endometrioid carcinoma (FIGO grade 1 or 2) components.

It is now recognized that DDEC can also arise in the background of high-grade endometrial carcinoma. A case of DDEC rising from clear cell carcinoma was reported by Lee SE et al in 2017 (Lee et al., 2017). A case series published by Al-Hussaini M et al included two cases of undifferentiated endometrial carcinoma with serous carcinoma, which by today’s guidelines, should qualify a diagnosis of DDEC (Al-Hussaini et al., 2018). Another case series by Busca A in 2020 included one case of DDEC arising from serous carcinoma (Busca et al., 2020). Based on these observations, several entities that were not considered to be close mimicker of DDEC have now entered the differential diagnosis.

Similar to DDEC, uterus carcinosarcomas show a low-power biphasic appearance and may exhibit myxoid stroma. Both DDEC and carcinosarcoma have high mitotic activity. Unlike DDEC, where the dedifferentiated component shows sheets of discohesive, monotonous epithelioid tumor cells, carcinosarcoma is characterized by a malignant mesenchymal component that is usually diffusely atypical and pleomorphic (Mirkovic, 2022). Spindle cell morphology, as well as heterologous differentiation, are in keeping with a true mesenchymal component. DDEC may contain areas of spindled cells, but this should be a focal finding. DDEC may also have areas of prominent myxoid matrix deposition, which by itself is not diagnostic of sarcomatous transformation. The vast majority of carcinosarcomas exhibit p53 mutation/abnormal immunohistochemistry expression. On the other hand, undifferentiated carcinoma, exhibits frequent microsatellite instability and MMR immunohistochemistry loss (Hoang et al., 2014 Sep, Jenkins et al., 2020 Jun).

It is important to note that undifferentiated carcinoma can rarely be seen as a component in a carcinosarcoma, most frequently in those that are MMR-deficient (41% of MMR-deficient carcinosarcomas in a recent study) (Segura et al., 2020). Also, sarcomatous transformation in UEC/DDEC was described recently in 3 cases, one of which exhibited true heterologous differentiation (Vroobel and Attygalle, 2020). The ambiguity between these two diseases has led to diagnostic challenge in our case as both dedifferentiated component and sarcomatous component were confidently identified. In our case, there were several features that would otherwise support a diagnosis of carcinosarcoma. The differentiated component is serous carcinoma, from which carcinosarcoma often rise. Immunophenotypically, our case is characterized by intact MMR protein and null-type P53 which again is more commonly seen in carcinosarcoma compared to DDEC (Huvila et al., 2021). In our case, the dedifferentiated area is about 9 mm in size and sarcomatous component is <2 mm. Based on the size and the histology feature, we favor a diagnosis of DEC rising from serous carcinoma with a minor focus of sarcomatous transformation.

Given the dicohesive and epithelioid appearance and high mitotic activity of the undifferentiated component, the possibility of two uterine sarcomas, SMARCA4-deficient uterine sarcoma and high-grade endometrial stromal sarcoma (HG ESS) should not be overlooked. SMARCA4-deficient uterine sarcoma, however, are more commonly seen in younger patients while DDEC predominantly affects peri- or postmenopausal women. Rhabdoid cell morphology and phyllodiform-like growth mimicking adenosarcoma would favor SMARCA4-deficient uterine sarcoma (Kolin et al., 2018 Sep, Kolin et al., 2020 Feb). Rhabdoid cells are not seen in our case. Among the HG ESSs, the ones with WYHAE-NUTM2 fusion are the most challenging to distinguish from undifferentiated component. Morphologically, the HG ESS can have a nested growth pattern, which is unusual in DDEC (Lee et al., 2012). Both sarcomas are completely negative for cytokeratins and EMA. In our case, it is easier to exclude both possibilities given the abrupt transition from differentiated component to dedifferentiated component, demonstrating a bi-phasic picture. Of note, some DDECs have aberrations of SWI/SNF complex subunits, such as loss of expression in ARID1A/B, SMARCA1 and SMARCB4 proteins, which may have an aggressive clinical course (Tessier-Cloutier et al., 2021, Santoro et al., 2021 May). Our case showed retained SMARCA1 (INI) and SMARCB4 (BRG1) stain, indicating intact SWE/SNF complex.

Another possibility to consider is a high grade/FIGO grade 3 endometrioid carcinoma (EEC).¹ In FIGO 3 EECs, it is not unusual to see the solid component that is non-cohesive, with high grade cell morphology and brisk mitotic activity, mimicking the dedifferentiated component of DDEC. Careful assessment of entire tissue and look for area of focal glandular formation is helpful to rule out a DDEC. In contrast, a diffuse lack of cellular cohesion and glandular formation would favor EEC. In the appropriate morphologic setting, immunohistochemistry stains can be helpful. In DDEC, BRG1 expression is lost in 30% of cases, while INI1 expression is lost in 5–10% cases (Karnezis et al., 2016, Ramalingam et al., 2017 Feb). Absent or focal CKs and EMA, negative PAX8, ER and E-cadherin, along with loss of INI1 or BRG1 would favor DDEC over FIGO grade 3 EEC.

Per ESGO/ESTRO/ESP guidelines for the management of patients with endometrial carcinoma, our patient had high risk disease due to her serous carcinoma with aberrant p53 expression and the recommended therapy for her would be external beam radiotherapy (EBRT) with concurrent and adjuvant chemotherapy (Concin et al., 2021). The patient decided to receive vaginal cuff brachytherapy due to the potential better side effect profile compared to EBRT. Our patient received carboplatin-paclitaxel; however, recent result from GOG 86P trial has shown that patient with aberrant p53 expression might benefit from adding bevacizumab to the regimen (Leslie et al., 2021).

In summary, it is important to know that DDEC can rise from high grade ECs such as serous carcinoma and clear cell carcinoma. Recognizing DDEC is important as it carries a poorer prognosis with recurrence or death from disease occurring in 55–95 % of cases. Given the existence of diverse components such as differentiated and sarcomatous components, a clear differentiation between DDEC and carcinosarcoma can be challenging. Management of DDEC and carcinosarcoma are similar. So far, there is not enough clinical data to determine whether there is a difference in the prognosis between the two groups of patients, although both DDEC and carcinosarcoma are considered high-grade endometrial carcinoma and should be managed aggressively. Other close mimicker to DDEC, such as FIGO grade 3 EEC or uterine sarcomas, should also be considered for differential diagnosis. IHC stains can help guide the differential diagnosis of challenging malignancies of the uterus; a panel of stains are listed in Table 1.

Table 1.

Immunohistochemical staining.

IHC stain	Usage in current case
AE1/AE3 and EMA	To confirm the tumor has epithelial origin (carcinoma)
ER	Positive in 50% serous carcinoma. ER is generally positive in endometrioid carcinoma, and negative in clear cell carcinoma or HPV-related cervical adenocarcinoma.
PAX8	Pan-Mullerian (endometrial, ovarian, and endocervical) marker.
P53	Abnormal P53 stain pattern supports serous carcinoma. Of note, aberrant P53 can also be seen in a subset of high grade endometrioid or clear cell carcinoma.
E-cadherin	Positive in carcinoma, negative in dedifferentiated carcinoma
CD10	Positivity supports the presence of sarcoma component. It is frequently positive in endometrial stromal sarcoma.
MLH1, PMS2, MSH2 and MSH6	To detect micro-satellite instability. MMR loss would support endometrioid carcinoma.
BRG1 (SMARCA4) and INI-1(SMARCB1)	BRG1 is lost in SMARCA4-deficient uterine sarcoma. Loss of BRG1/ INI-1 is seen in 20% of DDEC (Kolin et al., 2020 Feb).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.