CA: Cancer Antigen; CEA: Carcinoembryonic Antigen; CT: Computed Tomography; MRI: Magnetic Resonance Imaging; FIGO: International Federation of Gynecology and Obstetrics; MPA: Medroxyprogesterone Acetate; PET-CT: Positron Emission Tomography-Computed Tomography; DNA: Deoxyribonucleic Acid; IVF-ET: In Vitro Fertilization and Embryo Transfer; D&C: Dilatation and Curettage; H&E: Hematoxylin and Eosin; Pt: Patient; EA: Endometrioid Adenocarcinoma; G1-3: Grade 1-3; AEH: Atypical Endometrial Hyperplasia; LN: Lymph Node; HP: Hydroxyprogesterone; MA: Megesterol Acetate; DG: Dihydrogesterone; LNG: Levonorgestrel; EE: Ethinyl Estradiol; TMX: Tamoxifen; dLA: depot Leuprolide Acetate; rt: right; lt: left; IUD: Intrauterine Device; DOD: Died of Disease; NED: No Evidence of Disease; NA: Not Available; C/S: Cesarean Section

Preservation of the uterus by treatment with high-dose progestin in young patients with endometrial adenocarcinoma is an important therapy for fertility preservation. This therapy is indicated for patients with grade 1 endometrioid histology and very early stage cancer. Good hormonal response rates between 57% and 75% have been achieved, but the relapse rates after remission are also high, varying between 47% in follow-up of 7 to 22 months and 89% after follow-up of over 30 months [1-3].

Relapse within the uterus can be treated relatively easily by radical treatment. However, relapse occurring outside the uterus has less chance for a complete cure and is likely to reduce survival. Progression to high-grade disease during therapy or follow-up has been reported, although this is rare. Here, we report a case that relapsed as advanced stage carcinosarcoma during follow-up after apparent complete remission in a patient with grade 1 endometrioid stage Ia endometrial adenocarcinoma. We also review previous cases to emphasize the importance of caution in management of patients after fertility-sparing treatment.

In October 2007, a 30-year-old 0-0-0-0 Japanese woman presented to Nippon Medical School Chiba Hokusoh Hospital with suspected endometrial carcinoma diagnosed at a local in vitro fertilization clinic. Her body height was 154 cm, body weight was 57.4 kg, and blood pressure was 117/70 mmHg. She had no anemia and normal biochemistry laboratory findings. Serum CEA was 0.5 ng/ml, CA19- 9 was 18.5 U/ml, and CA125 was 29.2 U/ml. Endometrial thickness was 11.5 mm on transvaginal ultrasound. Diagnostic imaging with CT and MRI did not indicate myometrial invasion or metastasis to regional/distant lymph nodes or remote organs. Endometrial biopsy by multidirectional curettage showed grade 1 endometrioid adenocarcinoma (Figure 1a).

Figure 1: a: Histological diagnosis of endometrioid adenocarcinoma of grade 1 was obtained through multidirectional endometrial curettage. This procedure demonstrated confluent glands with a cribriform bridging arrangement in most areas (H&E, low power magnification). b: Microscopic findings for the endometrial curetted specimen after 16 weeks of progestin treatment showed small atrophic glands scattered in the background of predecidualstroma (H&E, low power magnification). This finding indicates complete remission. c: T2-weighted MRI (sagittal section) showed no abnormal findings and a thin endometrium. d: T2-weighted MRI (sagittal section) showed no endometrial abnormality, but a 3 cm abnormal mass (white arrowheads) was observed between the cervical wall and rectum, with a moderate amount of ascites. e: T2-weighted MRI (coronal section) indicated a suspected metastasis at the surface of the right hepatic lobe (white arrow). f: In laparotomy, a cul-de-sac tumor was found to have invaded the rectal wall. The cervical wall of the uterus and both ovaries looked normal. g: Microscopic observation (H&E, low power magnification) revealed foci of carcinomatous elements with strong epithelial connections, and sarcomatous elements with atypical cells losing epithelial arrangement and osseous formation. In other areas, various levels of differentiations of endometrioid, squamous, and clear cell carcinomas were observed with immature cartilaginous formation. h,i: Low power magnification of the ovarian cortex showed multiple (independent) foci of carcinoma. This finding suggests that the ovarian malignancy was more metastatic than the primary origin.

High-dose progestin fertility-sparing treatment was started for a presumptive clinical stage of FIGO Ia, according to our protocol (Table 1) and with informed consent. A daily dose of oral Medroxyprogesterone Acetate (MPA) 600 mg was given for 16 weeks until endometrial histology showed atrophic glands among predecidual stroma (Figure 1b), and CT and MRI showed normal findings (Figure 1c). A daily dose of MPA 200 mg was subsequently given for another 12 weeks, after which the patient returned to the local clinic for in vitro fertilization and embryo transfer.

Table 1: Treatment regimen of high-dose progestin therapy for uterine preservation at the Nippon Medical School Chiba Hokusoh Hospital.

In January 2009, she returned to our hospital for infusion therapy for emesis; however, her pregnancy was terminated at gestational age 14 weeks due to intrauterine fetal death. Two months later, she came to our hospital complaining of dull but intolerable pain around the perianal area. Rectovaginal examination revealed a palpable hard mass of 20×23 mm² between the posterior vaginal fornix and rectum. An ultrasound examination showed echo-free space that led to a suspicion of ascites. A biopsy indicated atypical glands and recurrence was suspected. A systemic check-up including PET-CT, MRI, and fiberscopy over the whole gastrointestinal tract was performed. MPA was started again based on a pre-surgical diagnosis of a recurrent tumor invading the rectum and a metastatic tumor near the surface of the right hepatic lobe (Figure 1d,e).

In August 2009, the patient was treated with modified radical hysterectomy with low anterior bowel resection, bilateral salpingooophorectomy, pelvic and para-aortic lymph node dissection, subcolic omentectomy, and full thickness diaphragm resection around the metastatic tumor (Figure 1f). Grossly, the resected specimen showed tumor invasion of less than half the thickness of the uterine wall, a cul-de-sac tumor with direct invasion of the rectal mucosa and vaginal wall, and metastasis to the bilateral ovaries and diaphragm (Figure 1g). Histologically, the glandular component of the tumor showed various levels of differentiation of endometrioid, squamous, and clear cells, while the stromal component included osseous and cartilaginous formation (Figure 1h,i).

The final diagnosis was intraperitoneally disseminated carcinosarcoma, heterologous type (suspected to be recurrence of endometrial carcinoma). After two years with completion of six courses of a triweekly paclitaxel-carboplatin regimen, another recurrence on the right hepatic lobe was suspected on a CT scan and was removed by partial resection of the liver in April 2012. In histology, welldifferentiated endometrioid adenocarcinoma was detected again. One year later, a new lesion appeared in the liver near the great vessels and was treated by proton irradiation. At the end of 2013, one month after completion of irradiation, the size of the new lesion had decreased. The patient still requires careful follow-up.

Recurrence of endometrial adenocarcinoma after fertility-sparing treatment has been widely reported. The site of recurrence is usually limited to the endometrium and histology typically shows the same well-differentiated endometrioid type. In contrast, endometrioid carcinoma recurring as carcinosarcoma (as in our case) is rare, with only four reported cases, three of which occurred in older patients who were diagnosed with high-grade carcinoma at an advanced stage [4]. Only one case has been described in a younger patient after fertilitysparing treatment, with Fujiwara et al. [5] describing a patient who was initially presumed to have early stage endometrial adenocarcinoma and was finally diagnosed with heterologous carcinosarcoma of the uterus following hysterectomy after 16 weeks of unsuccessful conservative therapy. Our case differs from this case because recurrence occurred as multiple intra-abdominal disseminations after apparent complete remission.

Recurrence restricted to the uterus and with a well-differentiated endometrioid histology is more likely to respond to additional hormonal therapy [6]. However, recurrence outside the uterus may result in serious life-threatening complications. We performed a MEDLINE search for cases of extrauterine metastasis or recurrence after temporal cure for early stage low-grade endometrioid type endometrial adenocarcinoma or endometrial hyperplasia. Kempson and Pokorny used uterusconserving therapy with hydroxyprogesterone caproate and found one ovarian recurrence in a patient under 40 years of age among 22 cases treated between 1930 and 1967 [7]. Only four cases were treated hormonally and none of these cases had ovarian recurrence; therefore, they were excluded from our list of cases. Cases in which extrauterine metastasis occurred during conservative therapy, such as those reported by Imai et al. and Ota et al., were also excluded because such cases can be treated by standard surgery. Thus, 13 cases of extrauterine recurrence after temporary remission are listed in Table 2 [3,8-14]. Six of these cases had ovarian lesions as the extrauterine site. These cases were mostly thought to be independent primary or concurrent ovarian endometrioid adenocarcinoma based on histological findings showing that both lesions were very early stage and were rarely confirmed by a heterozygous pattern at the human androgen receptor gene locus. It is of note that Huang et al. found similar microscopic patterns, immunohistochemical profiles, and DNA ploidy in a small ovarian lesion of 2×2 mm² compared to a primary endometrial lesion [9,11,14-16].

Table 2: Cases with extrauterine recurrence after temporary remission of endometrioid adenocarcinoma or atypical endometrial hyperplasia.

In our case, we were unable to identify the exact date of recurrence; therefore, the tumor might have already recurred at the time of IVFET treatment. It is generally thought that a high serum progesterone level during pregnancy suppresses the growth of well-differentiated endometrioid adenocarcinoma, but Mitsushita et al. reported that recurrence can occur before pregnancy [17] and speculated that carcinoma may remain during pregnancy and delivery after conservative therapy. If this is so, implantation of carcinoma cells could occur via the needle from the cul-de-sac tumor to the bilateral ovaries or from the ovary to the cul-de-sac tumor during ovum assembly in IVF-ET.

Egarter et al. described a retrograde intraperitoneal spread of endometrial cancer cells through tubal reflux during hysteroscopy [18]. Rose et al. reported complete response with conservative treatment with repeated D&C, hysteroscopic resection and oral progestin in a patient with endometrial cancer who developed pelvic peritoneal implants involving the sigmoid colon at 24 months after diagnosis [19]. However, a literature search over two decades by Revel provided no definite conclusion concerning the prognostic impact of diagnostic hysteroscopy for endometrial cancer [20].

Another possible explanation in our patient may be that dissemination might not have occurred as a recurrence, but secondary to ovarian carcinosarcoma that happened to grow as a double cancer, since intra-abdominal dissemination occurs more frequently in ovarian cancer than the endometrial origin. In support of this idea, there are more reports of endometrial cancer coexisting with ovarian cancer in younger patients (11%-25%) than in older patients and 92% of these cases have an endometrioid type histology [21,22]. Moreover, ovarian lesions may not be detected until a microscopic study is performed after surgery because the lesion is too small. In fact, in our case, the ovarian lesion was unclear at the time of laparotomy and microscopic observation of the lesion showed multiple foci in the cortical area. Thus, from a histological point of view, ovarian lesions are more likely to be metastatic, rather than primary. In this situation, one of the foci may have been the primary lesion and others could have been secondary lesions through implantation by repeated punctures for ovum sampling.

It is not unreasonable to assume that carcinosarcoma arose as a recurrence from grade 1 endometrioid adenocarcinoma and that the epithelial components of the carcinosarcoma showed various histology (including clear, squamous cells from G1 to G3 endometrioid histology) because carcinosarcoma can be thought of as a metaplastic carcinoma. Furthermore, since the metastatic liver lesion in the latest recurrence had the same histology as the primary endometrial cancer (i.e., well-differentiated endometrioid adenocarcinoma with squamous differentiation), a clone with worse histology could have been eliminated by chemotherapy; whereas a clone with chemo-resistant well-differentiated histology could have survived and recurred, as seen in most cases.

In conclusion, we cannot say with certainty that our case of disseminated carcinosarcoma is a recurrence from endometrial cancer or from a new malignant ovarian neoplasm that, by chance, occurred independently. However, it is clear that carcinosarcoma was present and extrauterine recurrence occurred after fertility-sparing treatment. Early stage, well-differentiated endometrioid adenocarcinoma of the uterine corpus is generally considered to have a good prognosis, but counseling of patients in life decisions is important, with provision of sufficient information about the possibility of extrauterine progression or disseminated recurrence.

We are grateful to the late Professor Yoshiharu Ohaki (Surgical Pathologist, Nippon Medical School; and former Director of the Department of Pathology, Nippon Medical School Chiba Hokusoh Hospital) for assistance with pathology and Mariko Ikeda (Department of Obstetrics and Gynecology, Nippon Medical School) for management of clinical information.