Uterine corpus tumor with neuroectodermal differentiation and frequent ganglion-like cells in a postmenopausal woman.

•Uterine neuroectodermal tumors (NETs) are uncommon malignant neoplasm with poor prognosis.•Ganglion-like cells with fibrillary background as major component of uterine NETs are extremely rare.•We present a patient affected by uterine NET with frequent ganglion-like cells, resembling ganglioneuroblastoma.•This case report is important to define the pathogenesis and establish better treatments for neuroectodermal tumors.

Introduction

Uterine neuroectodermal tumors (NETs) are uterine neoplasms with a poor prognosis (Elizalde et al., 2016; Euscher et al., 2008; Novo et al., 2015; Prat et al., 2014). They are pathologically classified into 2 groups: 1) those resembling central nervous system (CNS) embryonal tumors (central-type NETs) (Euscher et al., 2008; McLendon et al., 2016; Prat et al., 2014), and 2) those resembling peripheral primitive neuroectodermal tumors/Ewing sarcomas (peripheral-type NETs) (Elizalde et al., 2016; Novo et al., 2015; Prat et al., 2014). Uterine NETs are also associated with endometrial adenocarcinomas, carcinosarcomas, and high-grade sarcomas (Prat et al., 2014). However, the pathogenesis of NETs remains unknown because of the rarity of this type of malignancy (Elizalde et al., 2016; Euscher et al., 2008; Novo et al., 2015; Prat et al., 2014). Here, we present a patient with a rare uterine NET comprising frequent ganglion-like cells.

Case report

A 62-year-old Japanese woman was receiving medications for cellulitis and deep vein thrombosis of her right and left lower extremities. During follow-up visits for these ailments, contrast-enhanced computed tomography (CT) revealed a solid uterine tumor exhibiting heterogeneous enhancement (Fig. 1A) with multiple swollen intra-pelvic and para-abdominal aortic lymph nodes. The uterine mass exhibited hypointensity and high intensity on T1-weighted (Fig. 1B) and T2-weighted (Fig. 1C) pelvic magnetic resonance imaging, respectively. As the patient also complained of vaginal bleeding, she was admitted to our hospital for further examinations. Blood tests revealed elevated levels of the following tumor markers: carcinoembryonic antigen, 14.8 ng/mL (normal, <5 ng/mL); carbohydrate antigen (CA) 19-9, 1300 U/mL (normal, <40 U/mL); CA125, 68 U/mL (normal, <35 U/mL); and neuron-specific enolase (NSE), 77.4 ng/mL (normal, <16.3 ng/mL). Endometrial biopsy was performed, and the specimen was diagnosed as a leiomyosarcoma. There were para-abdominal aortic lymph node metastases (Fig. 1D), resulting hydronephrosis of both kidneys (Fig. 1E). One month later, the patient underwent total abdominal hysterectomy (TAH), bilateral salpingo-oophorectomy (BSO), and partial omentectomy. However, her renal dysfunction did not improve and her general condition gradually worsened to a level that precluded postoperative chemotherapy or radiation therapy. She died of multiple organ failure 2 months after the discovery of the tumor.

Fig. 1

Radiological features of the uterine tumor and hydronephrosis. Contrast-enhanced computed tomography (CT) showed the large solid uterine tumor with heterogeneous enhancement (A). Magnetic resonance imaging showed the uterine tumor with slightly heterogeneous hypointensity on T1-weighted imaging (B) and hyperintensity on T2-weighted imaging (C). CT showed prominent para-aortic lymph node metastases (D), resulting bilateral hydronephrosis (drip infusion pyelography; E).

Pathologic findings

The resected uterus comprised almost entirely of a milky-whitish tumor with necrosis, measuring 15 × 9 cm in size (Fig. 2A). The tumor was histopathologically classified as a highly malignant cellular neoplasm (Fig. 2B) and was mainly composed of small naked neoplastic cells (Fig. 2C). The following additional histological components were noted: atypical ganglion-like cells with a fibrillary background (Fig. 2D), endometrial adenocarcinoma with squamous differentiation (Fig. 2E), rhabdoid-like cells (Fig. 2F), atypical spindle cells resembling skeletal muscular cells, and an atypical cartilaginous component. The component comprising atypical ganglion-like cells with a fibrillary background occupied approximately 92% of the uterine tumor. The neoplasm directly infiltrated the parametrium and had metastasized to both ovaries as well as the major omentum.

Fig. 2

Macroscopic and histopathological features of the uterine neuroectodermal tumor. The uterus was almost totally occued by the neoplasm (A, sagittal section of the uterus). Histopathologically, the uterine tumor was a highly cellular neoplasm (B, hematoxylin and eosin [H&E]) mainly composed of small round neoplastic cells (C, H&E) and ganglion-like cells with fibrillary background (D, H&E). Moreover, components of adenocarcinoma with squamous metaplasia (E, H&E) and rhabdoid-like cells (F, H&E) were intermingled in the tumor.

Immunohistochemically, the small naked neoplastic cells showed varying degrees of immunoreactivity for vimentin, CD99, CD56, S100, synaptophysin (Fig. 3A), alpha-smooth muscle actin (α-SMA), neurofilament (NF), and chromogranin A (CGA). Both the atypical ganglion-like cells and fine fibrillary background were positive for synaptophysin (Fig. 3B), S100, CD56, CD99, and NF. The atypical ganglion-like cells were also positive for CGA. A few neuronal nuclei (NeuN)-positive atypical ganglion-like cells and glial acidic protein (GFAP)/oligodendrocyte lineage transcription factor 2 (Olig2)-positive fibrillary astrocytes were also detected (Fig. 3C). The endometrial adenocarcinoma with a squamous differentiation component was diffusely positive for cytokeratin (CK) AE1/AE3 (Fig. 3D) and epithelial membrane antigen (EMA), and was focally positive for vimentin. The squamous differentiation component showed p40 immunoreactivity. The rhabdoid-like cells revealed immunoreactivity for vimentin, synaptophysin, CGA, and NF, suggesting small ganglion cells, whereas it was negative for S-100, human melan black-45 (HMB-45), GFAP, Olig2, NeuN, epithelial markers (cytokeratin [CAM5.2], EMA, and pan-CK [AE1/AE3]), and muscular markers (desmin, myogenin, and α-SMA). Nuclear INI1 protein immunoreactivity was preserved in the tumor, including in the rhabdoid-like cells (Fig. 3E). α-SMA-positive atypical spindle cells were intermingled with the epithelial and neuronal components. The MIB-1 labeling index was >50% in the small round neoplastic cells (Fig. 3F) and approximately 20% in the ganglion-like cells with fibrillary background. No neoplastic cells were positive for melanoma (HMB-45 and melan-A) or skeletal muscle (desmin and myogenin) markers. Based on these features, the pathological diagnosis was uterine NET with frequent ganglion-like cells.

Fig. 3

Immunohistochemical features of the uterine neuroectodermal tumor. Small round neoplastic cells (A) and ganglion-like cells with fibrillary background (B) showed immunoreactivity for synaptophysin. Furthermore, glial acidic protein (GFAP)-positive astrocytes were also intermingled (C, GFAP), suggesting a central-type neuroectodermal tumor. The adenocarcinoma component was immunopositive for cytokeratin AE1/AE3 (D, AE1/AE3). Nuclear INI1 protein expression was preserved in the tumor, including in the rhabdoid-like cells (E, INI1). The tumor showed a high MIB-1 labeling index at the area of small round neoplastic cells (F, Ki-67).

Widespread dissemination of the uterine NET was found on autopsy. The uterine neoplastic cells had metastasized or disseminated to the lungs, liver, appendix vermiformis, urinary bladder, ureters, Douglas' pouch, peritoneum, mesenterium, and lymph nodes (para-aortic, peri-tracheal, and peri-pancreatic). The metastatic cells were mainly comprised of NET with ganglion-like cells and a fibrillary background; however, no metastases of the carcinomatous or sarcomatous components were noted. Both kidneys showed mild hydronephrosis that was secondary to tumor spreading. No remarkable changes were noted in the heart, alimentary tract, pancreas, gallbladder, thyroid gland, or adrenal glands.

Discussion

Uterine NET is rare; only 69 patients with this tumor type have been reported in the English-language literature to date (Table 1). Clinically, uterine NET usually occurs in postmenopausal women and presents with vaginal bleeding (Euscher et al., 2008; Prat et al., 2014). Indeed, 78.7% of the patients with uterine NETs listed in Table 1 experienced vaginal bleeding, and 72.9% of them were over 40 years old. Approximately 50% of these uterine neoplasms are found to have metastasized to the extra-uterine tissues/organs at diagnosis (Prat et al., 2014). The major metastatic sites of uterine NETs are the lymph nodes via the lymphatic system (Daya et al., 1992; Odunsi et al., 2004; Shah et al., 2009; Park et al., 2007; Elizalde et al., 2016) and lungs/liver via the vasculature (Bartosch et al., 2011; Gersell et al., 1989; Hendrickson and Scheithauer, 1986; Shah et al., 2009; Sinkre et al., 2000; Yi et al., 2015), as was also observed in our patient. Although the standard treatment for uterine NETs normally involves surgery (TAH + BSO) with or without chemotherapy and/or radiotherapy (Elizalde et al., 2016), we recommend that lymph node dissection also be performed when possible. However, the necessity of omentectomy in patients with uterine NETs remains unconfirmed because it has been performed in too few patients who underwent TAH + BSO (Table 1).

Table 1

Clinicopathological feature of 69 uterine neuroectodermal tumor cases.

Case no	Age (y.o)	Symptom	FIGO stage	Surgery	Postoperative therapy		Prognosis		Pathological findings					Reference
					CT	RT	Alive/died	Follow-up (month(s))	Tumor size (cm)	Component			Metastasis
										Major component	Minor component	Ganglion cells
1	58	Vaginal bleeding	IIIc	+ (unknown detail)	−	−	DOD	2	ND	NET	−	−	ND	Euscher et al. (2008)
		Palpable mass
2	31	Back paine	IV	−	+ (unknown regimen)	+	DOD	20	ND	NET	−	−	ND	Euscher et al. (2008)
3	72	Vaginal bleeding	Ia	ND	ND	ND	DOD	11	ND	NET	−	−	ND	Euscher et al. (2008)
4	48	ND	IIIc	ND	ND	ND	ND	ND	ND	EM carcinoma	NET	−	ND	Euscher et al. (2008)
5	81	Vaginal bleeding	ND	ND	ND	ND	ND	ND	ND	NET	–	−	ND	Euscher et al. (2008)
6	66	Pelvic mass	IIIc	+	Letrozole	−	NED	41	ND	NET	High grade sarcoma	−	ND	Euscher et al. (2008)
				(unknown detail)
7	53	Vaginal bleeding	ND	ND	ND	ND	DOD	22	ND	NET	−	−	ND	Euscher et al. (2008)
8	51	Vaginal bleeding	ND	ND	ND	ND	DOD	12	ND	NET	−	−	ND	Euscher et al. (2008)
9	31	Vaginal bleeding	ND	ND	ND	ND	DOD	26	ND	NET	EM hyperplasia with atypia	−	ND	Euscher et al. (2008)
10	64	Endocervical polyp	IIb	TAH, BSO	+	+	NED	36	ND	NET	−	−	ND	Euscher et al. (2008)
					(unknown regimen)
11	64	Vaginal bleeding	ND	ND	ND	ND	ND	ND	ND	Adenosarcoma	NET	−	ND	Euscher et al. (2008)
		Pain
12	69	Vaginal bleeding	IV	ND	ND	ND	ND	ND	ND	NET	Rhabdomyosarcoma	−	ND	Euscher et al. (2008)
13	62	Uterine fibroids	IIIc	ND	ND	ND	DOD	22	ND	NET	−	−	ND	Euscher et al. (2008)
14	55	Vaginal spotting	Ib	TAH, BSO	+	−	NED	38	ND	NET	−	−	ND	Euscher et al. (2008)
					(unknown regimen)
15	52	ND	IV	ND	ND	ND	NED	6	ND	NET	−	−	ND	Euscher et al. (2008)
16	58	Vaginal pressure	IV	−	+	−	NED	6	ND	Carcinosarcoma	NET	−	ND	Euscher et al. (2008)
					(unknown regimen)
17	57	ND	IIIc	−	+	−	NED	35	ND	Carcinosarcoma	NET	+	ND	Euscher et al. (2008)
					(unknown regimen)
18	12	Vaginal bleeding	IV	TAH, LSO	Cyclophosphamide	−	DOD	25	ND	NET	−	+	Lung	Hendrickson and Scheithauer (1986)
					Doxorubicin
					Adriamycin
					Dactinomycin
					Vincristine
19	57	Vaginal bleeding	IIIc	TAH, BSO, PALND	Cisplatin	+	DOD	24	ND	NET	−	+	Lung	Hendrickson and Scheithauer (1986)
					Vinblastine								Retroperitoneum
					Bleomycin
20	17	Vaginal bleeding	IIIc	TAH, PLND, left uretectomy,	Vincristine	−	NED	10	ND	NET	−	+	−	Rose et al. (1987)
		Pelvic mass		bilateal ovarian wedge biopsy	Cisplatin
					Daunorubicin
					Dactinomycin
					Cyclophosphamide
					Etoposide
21	67	Vaginal bleeding	IIIc	STAH, BSO	Cisplatin	+	DOD	6	ND	NET	−	ND	−	Daya et al. (1992)
		Enlarged uterus			Doxorubicin
					Carboplatin
					5-FU
22	68	Vaginal bleeding	IVb	TAH, BSO, PLND	−	+	DOD	12	7.5	NET	−	+	LNs (supraclavicular)	Daya et al. (1992)
23	69	Vaginal bleeding	I	TAH, BSO, PLND	−	+	NED	72	2	NET	EM stromal sarcoma	ND	−	Daya et al. (1992)
24	68	Vaginal bleeding	I	TAH, BSO	−	−	NED	60	2	NET	EM carcinoma	ND	–	Daya et al. (1992)
25	72	Vaginal bleeding	Ib	TAH, BSO	−	−	DOD	8	6.5 × 3.5 × 3.0	NET	–	−	ND	Molyneux et al. (1992)
26	54	Vaginal bleeding	IIIa	TAH, BSO, PLND	Cyclophosphamide	−	AWD	3	8.5 × 8.0 × 6.5	NET	Carcinosarcoma	−	−	Fukunaga et al. (1996)
					Cisplatin
					adriamycin
					5-FU
27	78	Vaginal bleeding	Ib	TAH, BSO, PLND	−	−	NED	9	6	NET	Cartilaginous component	−	−	Fraggetta et al. (1997)
28	62	Vaginal bleeding	Ib	TAH, BSO	Vincristine	+	DOD	18	4 × 2	NET	−	−	Terminal ileum	Sørensen et al. (1998)
					Cyclophosphamide								Cecum
					Cisplatin
					tenisopid
29	36	Enlarged uterus	Ib	RH, BSO, PLND	−	+	ND	ND	11	NET	−	−	−	Taïeb et al. (1998)
30	47	ND	IIb	TAH, BSO, LND	+	+	DOD	18	7.8	NET	Endometrioid carcinoma		Pelvis	Sinkre et al. (2000)
					(unknown regimen)
31	67	ND	IIIc	TAH, BSO, LND	+	−	DOD	3	4.5	NET	EM carcinoma		Peritoneum	Sinkre et al. (2000)
					(unknown regimen)
32	71	ND	IIIc	TAH, BSO, LND	+	−	DOD	4	6	NET	EM carcinoma		Lung	Sinkre et al. (2000)
					(unknown regimen)								Peritoneum
33	16	Vaginal bleeding	Ic	TAH, BSO, omentectomy	Vincristine	+	NED	48	ND	NET	−	−	−	Karseladze et al. (2001)
					Cyclophosphamide
					Doxorubicin
34	48	Vaginal bleeding	IIIc	TAH, BSO	−	−	NED	6	ND	NET	EM carcinoma	−	−	Ng et al. (2002)
		Pelvic mass
35	68	Vaginal bleeding	I	TAH, BSO	−	−	NED	10	ND	NET	−	−	ND	Venizelos et al. (2004)
36	66	Vaginal bleeding	Ia	TAH, BSO, omentectomy	ND	−	NED	24	4 × 3.5 × 2	NET	−	−	–	Odunsi et al. (2004)
				PLND, PALND
37	65	Vaginal bleeding	IIIc	TAH, BSO, PLND, PALND,	Cisplatin	+	AWD	12	7	NET	−	−	Vagina	Odunsi et al. (2004)
				omentectomy, upper vaginectomy	Adriamycin								Obturator lymph nodes
					Etoposide
38	15	Abdominal pein	I	TAH, PLND	Carboplatin	−	NED	12	6 × 7	NET	−	−	−	Peres et al. (2005)
		Pelvic mass			Etoposide
39	43	Vaginal bleeding	IIIc	TAH, BSO, PLND	Cytoxan	−	NED	2	13.3	NET	−	−	Left adnexa	Varghese et al. (2006)
		Uterine enlargement			Adriamycin
					Vincristine
					Etoposide
40	58	Vaginal bleeding	IV	TAH, BSO, right PLND,	Carboplatin	−	DOS	11	12	NET	EM carcinoma	−	Lung	Bartosch et al. (2011)
		Abdominal pain		segmental enterectomy,	Paclitaxel
				total coloectomy
41	26	Vaginal bleeding	IV	TAH, BSO, PLND, omentectomy	Cisplatin	−	NED	48	5.8 × 4.2	NET	−	−	−	Novo et al. (2015)
					Etoposide
					Avastin
42	50	Abdominopelvic pain	ND	TAH, BSO, omentectomy	Carboplatin	−	NED	16	15	NET	−	−	−	Dizon et al. (2013)
					Etoposide
43	63	Vaginal bleeding	IIIc	TAH, BSO, LND	Ifosfamide	−	DOD	7	5.0 × 4.5 × 3.0	NET	Rhabdomyosarcoma	−	Pelvis	Dundr et al. (2010)
					Cisplatin								Mesenterium
													Peritoneum
44	80	Abdominal pain	Ib	TAH, BSO, LND	−	+	AWD	6	5.0 × 4.0 × 3.0	NET	EM carcinoma	−	Intraabdominal metastasis	Dundr et al. (2010)
45	79	Vaginal bleeding	Ib	TAH, BSO, LND	−	−	NED	29	4.5 × 3.0 × 3.0	NET	EM carcinoma	−	−	Dundr et al. (2010)
46	78	Vaginal bleeding	IIIa	TAH, BSO, LND	−	−	NED	8	7.5 × 7.0 × 5.5	NET	−	−	−	Dundr et al. (2010)
47	32	Abdominal pain	IIIa	TAH, BSO, PLND, PALND,	Cisplatin	+	AWD	38	3	NET	−	−	−	Celik et al. (2009)
				omentectomy, appendectomy	Ifosfamide
					Adriamycin
					Vincristine
48	66	Vaginal bleeding	IVb	TAH, BSO	Cisplatin	+	DOD	24	6 × 4	Carcinosarcoma	NET		Lung	Gersell et al. (1989)
		Pelvic pain			Cyclophosphamide								LNs (left supraclavicula, right axillary)
					Doxorubicin
					Dexamethasone
49	32	Vaginal bleeding	IV	TAH, BSO, PLND	Holoxan	−	AWD	24	9 × 6.5	NET	−	−	Peritoneal seeding	Aminimoghaddam et al. (2015)
		Abdominal pain			Mens
					Cisplatin
					Paclitaxel
					Carboplatin
50	29	Abdominal swelling and pain	IVb	STAH, BSO, PLND, omentectomy,	Docetaxel	+	AWD	18	3.0 × 2.5 × 2.0	NET	−	−	Liver	Yi et al. (2015)
				appendectomy,	Carboplatin
				metastatic nodule resection	Vincristine
					Adriamycin
					Cyclophosphamide
					Ifosfamide
					Etoposide
51	63	Constipation	ND	TAH, BSO	Cyclophosphamide	−	NED	24	13.0 × 10.0	NET	−	−	−	Shimada et al. (2014)
					Vincristine
					Adriamycin
52	25	Vaginal bleeding	ND	TAH, BSO	Vincristine	+	NED	18	7.6 × 4.0 × 5.9	Rhabdomyosarcoma	NET	−	Vagina	Cate et al. (2013)
					Adriamycin
					Cyclophosphamide
					Ifosfamide
					Etoposide
53	12	Vaginal bleeding	ND	−	Etoposide	−	NED	36	12	Rhabdomyosarcoma	NET		−	Stolnicu et al. (2012)
					Cisplatin
					Bleomycin
54	56	Vaginal bleeding	Ib	TAH, BSO, PLND	Ifosfamide	−	NED	41	4.0 × 3.5 × 2.0	NET	−	−	−	Ren et al. (2011)
					Etoposide
					Cisplatin
55	59	vaginal bleeding	IIIc	TAH, BSO, PLND, PALND,	Carboplatin	+	AWD	12	1.1	NET	−	−	Lymph nodes (mediastinal, paraaortic)	Shah et al. (2009)
				omentectomy	Paclitxel								Vaginal cuff
					Cisplatin								pelvic wall
													Lung
													liver
56	30	Vaginal bleeding	IVb	−	Doxorubicin	+	DOD	16	18 × 20 × 21	NET	−	−	Lymph nodes (paraaortic, pelvic)	Park et al. (2007)
					Ifosfamide								Omentum
					Vincristine								Thoracolumbar spine
					Carboplatin								Right humerus
					Etopiside								Left lower rib
					Docetaxel								Left femur
					Irinotecan
					Celecoxib
57	22	Vaginal bleeding	I	TAH, BSO, PLND, PALND,	Cisplatin	−	NED	10	7.6 × 6.1	NET	−	−	−	Akbayir et al. (2008)
		Adnexal mass		omentectomy	Doxorubicin	−
58	24	Fever	II	TAH, BSO, omentectomy	Vincristine	−	AWD	1	9 × 10	NET	−	−	Residual tumor	Mittal et al. (2007)
		Lower abdominal pain			Adriamycin
					Cyclophosphamide
					Ifosfamide
					Etoposide
59	26	Pelvic mass	III	Modified TAH, PLND,	Vincristine	+	NED	16	7.0 × 5.0	NET	−	−	−	Blattner et al. (2007)
		(found at cesarean section)		bilateral ovarian transposition	Doxorubicin
					Cytoxan
					Mensa
					ifosfamide
					Etoposide
60	50	Vaginal bleeding	IIIc	TAH, BSO, PLND,	+	+	NED	6	10 × 8	NET	Adenosarcoma	−	Vaginal vault	Bhardwaj et al. (2010)
				omentectomy	(unknown regimen)
61	51	Vaginal bleeding	III	TAH, BSO	+	−	ND	ND	ND	NET	EM carcinoma	−	ND	Chiang et al. (2017)
					(unknown regimen)
62	50	Vaginal bleeding	III	TAH, BSO	−	−	ND	ND	ND	NET	EM carcinoma	−	ND	Chiang et al. (2017)
63	31	Vaginal bleeding	III	TAH, BSO	−	−	ND	ND	ND	NET	Carcinosarcoma	−	ND	Chiang et al. (2017)
64	26	Vaginal bleeding	I	TAH, BSO	−	−	ND	ND	ND	NET	−	−	ND	Chiang et al. (2017)
65	64	ND	III	TAH, BSO	+	−	ND	ND	ND	NET	−	−	ND	Chiang et al. (2017)
					(unknown regimen)
66	ND	ND	ND	TAH, BSO	ND	ND	ND	ND	ND	NET	−	−	ND	Chiang et al. (2017)
67	60	Vaginal bleeding	IV	TAH, BSO, PALND	Carboplatin	−	ND	ND	10 × 13	NET	−	−	Pelvis	Elizalde et al. (2016)
		Abdominal pein			Etoposide								LNs (para-aortic, retropetitoneal)
68	31	Vaginal bleeding	IIIc	+	Cisplatin	+	NED	24	ND	NET	−	−	/	Tsai et al. (2012)
		Abdominal pain		(unknown detail)	Etoposide
69	62	Vaginal bleeding	IVb	TAH, BSO, omentectomy	−	−	DOD	2	15 × 9	NET	Carcinosarcoma	+	Peritoneum	present case
													Liver
													Appendix vermiformis

y.o., years old; FIGO, International Federation of Gynecology and Obstetrics; TAH, total abdominal hysterectomy; BSO, bilateral salpingo-oopholectomy; PLND, pelvic lymph node dissection; PALND, paraaortic lymph node dissection; CT, chemotherapy; RT, radiation therapy; DOD, die of disease; NED, no evidence of disease; AWD, alive with disease; ND, no data; NET, neuroectodermal tumor; EM, emdometrial

As for the prognosis of patients with uterine NETs, Euscher et al. (2008) reported a mortality rate of 47% in their largest uterine NET series; furthermore, the 2-year survival rate of postmenopausal patients with uterine NET was reported to be approximately 30% (Elizalde et al., 2016; Prat et al., 2014). Consistent with previous reports, our patient was also a postmenopausal woman with minimal vaginal bleeding, and had a uterine tumor with lymphadenopathy at the time of diagnosis. She died 2 months after the uterine mass was diagnosed despite undergoing TAH, BSO, and omentectomy; however, lymph node dissection was not possible. As such, our patients' uterine NET was consistent with previously reported tumors that had poor prognoses. Of the 69 patients with uterine NETs previously reported in the English-language literature (Table 1), 36% died of their uterine tumors after a mean post-surgical duration of 14.1 months (range, 2–26 months), 50% were free of disease after a maximum follow-up period of 72 months, and 14% were alive with disease after a maximum follow-up period of 38 months. Furthermore, the mean follow-up duration from diagnosis to death in the non-surviving patients was 14.5 ± 8.4 months. Taken together, uterine NETs may not necessarily have as poor a prognosis as previously thought (Elizalde et al., 2016; Euscher et al., 2008; Novo et al., 2015; Prat et al., 2014).

The histopathology of uterine NET is characterized by a monotonous population of small- to medium-sized round neoplastic cells growing in sheets, nests, and/or cords, with or without fibrillary backgrounds and rosette formations (Euscher et al., 2008; Prat et al., 2014). Some central-type NETs have been reported to show pathological features similar to those of medulloblastoma, medulloepithelioma, glioblastoma, and/or ependymoma (Chiang et al., 2017). Uterine NETs may also include other histologic elements, such as endometrial adenocarcinoma, carcinosarcoma, and/or high-grade sarcoma (Euscher et al., 2008; Prat et al., 2014). Our patient's uterine NET had heterologous carcinosarcoma as a minor component, which has also been described in previous reports (Euscher et al., 2008; Prat et al., 2014). However, frequent ganglion-like cells with a fibrillary background were detected as a major component in our patient, whose NET resembled a ganglioneuroblastoma (McLendon et al., 2016). To the best of our knowledge, this uterine NET subtype is extremely rare, although a patient with a uterine NET comprising foci resembling ganglioneuroma was reported by Hendrickson and Scheithauer (1986).

Immunohistochemical analyses of our patient's tumor showed that the NET component expressed CD99, synaptophysin, NSE, and NF. Although rare, GFAP immunoreactivity is characteristic of CNS-type NETs (Prat et al., 2014). In addition to neuronal markers such as synaptophysin and NF, our patient's tumor also expressed the glial markers GFAP and Olig2. Moreover, an α-SMA immunoreactive spindle cell component and both a vimentin and epithelial marker immunoreactive component were detected, suggesting leiomyosarcoma and endometrial adenocarcinoma, respectively, intermingled as minor components within the neuroectodermal component. EWSR1 rearrangement has been recently reported as a characteristic genetic finding of peripheral-type uterine NETs (Novo et al., 2015); however, we were unable to perform genetic analysis to test for EWSR1 rearrangement.

Surgery (TAH + BSO) with or without chemotherapy and/or radiotherapy is the standard treatment for uterine NETs (Elizalde et al., 2016). As described in Table 1, approximately 92% of patients with uterine NETs underwent surgery, while 72% received chemotherapy and only 36% received radiotherapy. Therapeutic treatment regimens for gynecologic NETs might be selected according to their subtypes, such as NETs resembling medulloblastoma and Ewing sarcoma/peripheral primitive NETs (Chiang et al., 2017). Furthermore, Novo et al. (2015) recently reported a patient with uterine NET treated with surgery and adjuvant chemotherapy using cisplatin, etoposide, and bevacizumab; their patient experienced no recurrence for 48 months. Although our patient was treated with TAH + BSO, she died of multiple organ failure 1 month after surgery owing to the metastasis of multiple tumors that comprised mainly of NET resembling ganglioneuroblastoma (according to autopsy results). In retrospect, treating the ganglioneuroblastoma with total tumor resection followed by chemoradiotherapy with temozolomide should have been considered for our patient, as it was previously reported that 2 patients with cerebral ganglioneuroblastoma treated with this regimen were free of tumor recurrence or progression after 12 and 14 months of follow-up, respectively (Schipper et al., 2012). Interestingly, as shown in Table 1, 42% of the patients with uterine NETs who underwent radiotherapy died of their disease, whereas 32% were free of disease. Although surgery with or without chemotherapy and/or radiotherapy is the standard treatment for uterine NETs (Elizalde et al., 2016), postoperative radiotherapy for such patients might need to be reconsidered. Nevertheless, the accumulation of additional patient data and detailed clinical and pathological analyses are required to devise better treatment modalities for uterine tumors.

Although the pathogenesis of primary uterine NETs remains poorly understood, several possibilities have been suggested, including 1) that they originate from the implantation of aborted fetal tissue in the uterus (Chiang et al., 2017; Fukunaga et al., 1996; Rose et al., 1987; Siddon and Hui, 2010; Young et al., 1981), 2) that they originate from abnormal migrated neural crest cells in the uterus (Chiang et al., 2017; Fukunaga et al., 1996; Rose et al., 1987), and 3) that they are of Müllerian origin (Chiang et al., 2017; Daya et al., 1992; Fukunaga et al., 1996; Gersell et al., 1989; Young et al., 1981). Liao and Choi (1986) reported that malignant mixed Müllerian tumors showed GFAP immunoreactivity; our patient had heterologous carcinosarcoma intermingled within the uterine NET as the minor component. Based on our clinicopathological findings, our patient's tumor appeared to have been of Müllerian origin.

In conclusion, uterine NETs with frequent ganglion-like cells such as the tumor diagnosed in our patient are extremely rare; their pathogenesis is poorly understood and afflicted patients have poor prognoses. Therefore, the accumulation of clinicopathological data from additional patients is needed to establish more effective treatment modalities for patients with these types of tumors.

Author contributions

Taku Homma: Pathological examination, manuscript preparation.

Takehiro Nakao: Patient care, data collection.

Toshiya Maebayashi: Radiology imaging examination.

Toshiyuki Ishige: Pathological examination.

Hiroyuki Hao: Supervisor, manuscript preparation.

Funding disclosure

This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interest

The authors declare no conflicts of interest.