Low-grade endometrial stromal sarcoma recurring after 9 years. (Case Reports).

ABSTRACT

A 6l-year-old Japanese woman had recurrent low-grade endometrial stromal sarcoma (ESS) after primary treatment 9 years earlier. Initial and recurrent tumors showed the same configuration that the polycystic part showed in the solid tumor ultrasonographically. The central part of both the initial and recurrent tumors showed typical histologic findings, but bizarre cells were detected in the peripheral layer of the recurrent tumor. They resembled glandular cells, Comet cells of high-grade ESS, fibroblasts, decidual cells, and predecidual endometrial stromal cells at the middle to late secretory phases. The change and differentiation of low-grade ESS by therapy and environment seem to be the same as the differentiation in mesenchymal stem cells. We speculate that recurrence and prognosis of low-grade ESS are related by extrauterine development but not by mitotic activity or DNA content.

**********

MALIGNANT TUMORS composed exclusively of cells originating from the endometrial stroma are rare. These neoplasms have a variety of histopathologic characteristics and clinical courses, ranging from indolent local growth to rapid extensive progression. (1,2) Recurrent cases of low-grade endometrial stromal sarcoma (ESS) are especially rare. We report a case of recurrent low-grade ESS that regrew in the left pelvic wall, where the primary tumor had been surgically resected 9 years earlier. The histologic activity of the recurrent tumor was not changed from that of the primary tumor, and tumor aggressiveness was not detected by DNA content.

CASE REPORT

In 1987, a 61-year-old Japanese woman (gravida 5, para 2) had vaginal spotting, hot flushes, and a palpable lower abdominal mass. Examination revealed a goose-egg sized uterine corpus with a huge, soft, slightly cystic tumor on the left side. Ultrasonography and computed tomography showed a 16.1 x 10.5 x 10.5 cm solid pelvic mass with polycystic portions originating from the uterine corpus (Fig 1, left). Blood values were within normal limits except for hemoglobin (8.2 g/dL) and CA125 (52 U/mL).

Abdominal exploration revealed a large tumor arising from the left side of the corpus wall. The tumor, which adhered to the pelvic wall, contained a yellow-white gelatinous substance and was encapsulated by a thin membrane. The patient had total abdominal hysterectomy and bilateral salpingo-oophorectomy for a preoperative diagnosis of myoma uteri. The postoperative diagnosis was low-grade ESS, for which the patient was treated with 3 courses of CAP (cyclophosphamide, Adriamycin [doxorubicin], and Platinol [cisplatin]) at doses of 300, 30, and 50 mg/[m.sup.2], respectively.

Nine years later, digital examination, transvaginal ultrasonography, and magnetic resonance imaging (MRI) revealed a left pelvic mass. The MRI showed a solid mass measuring 4.7 x 3.4 x 3.0 cm with polycystic sections, some of which were located at the same position as the initial tumor (Fig 1, right). Laparotomy confirmed recurrent low-grade ESS in the left pelvic wall with a small amount of peritoneal effusion. The tumor surface was covered with a thin membrane macroscopically. Blood tests yielded normal values. Complete resection was done.

Smears of ascitic fluid showed a few atypical cells with the same configuration as those seen in the recurrent tumor, as well as a few larger atypical cells. Thus, the tumor was considered clinical stage IIIA (according to the new FIGO classification of corpus cancer). Postoperative chemotherapy consisted of 3 courses of cisplatin (50 mg/[m.sup.2]) and etoposide (300 mg/[m.sup.2]). Medroxyprogesterone acetate (400 mg/day orally) was administered for 2 years. There was no evidence of recurrence during the first 50 months of follow-up, since treatment of the second tumor.

Pathology

On histologic analysis, the initial tumor was a malignant spindle cell neoplasia manifesting characteristics of low-grade ESS. Sections of the mass showed densely cellular proliferation punctuated by frequent vascular spaces with vessel permeation and edematous change. Although sex cord-like arrangements, hemorrhage, and hyalinized stroma were observed, epithelioid differentiation was not detected. These atypical cells were faintly stained with eosin and had elongated inconspicuous cytoplasm and oval to round nuclei with small and round nucleoli (Fig 2). The assessed mitotic activity was 1 to 2 mitoses per 10 high-power field (HPF). Extrauterine development of the tumor was observed only at the hilus of the left ovary.

On the macroscopic and microscopic observation, the composition of the recurrent tumor was similar to that of the initial tumor, consisting mainly of a dense array of monotonous spindle atypical cells. In one portion, larger atypical cells had enlarged oval nuclei and abundant cytoplasm. These atypical cells showed more anisonucleosis in the peripheral layer of the tumor than the uniform smaller low-grade ESS cells (Fig 3). However, mitotic activity was 0 to 1 per 10 HPF.

Touch smear cell diagnosis of a section of the recurrent mass showed the same atypical cells as in the initial tumor. Moreover, ascitic fluid obtained at the second laparotomy showed atypical cells having a paving stone--like cellular arrangement, different from mesothelial cells, that originated from larger atypical cells and were the same as those in the peripheral layer. These atypical cells resembled predecidual stromal cells at the middle to late secretory phases (Fig 4, left). Also observed were a few bizarre cells with eccentric nuclei and streaming cytoplasm, resembling glandular cells or Comet cells of high-grade ESS (Fig 4, right).

Immunohistochemical analysis revealed similar appearance between the primary and the recurrent tumors, with the following immunoreactivity findings: vimentin (++), NSE (+), desmin (-), keratin (-), actin (-), S-100 (-), chromogranin A (-), CA125 (-), and estrogen receptor (+), though the primary tumor was negative for progesterone receptor (PR) and the recurrent tumor was positive for PR. In addition, only larger atypical cells of the peripheral layer in the recurrent tumor were positive for vimentin and desmin. Flow cytometry of these tumors revealed a diploid DNA pattern, with only 2.4% to 3.0% of the cells in the proliferation index (S + G2 + M phases) (Fig 5).

DISCUSSION

Low-grade ESS is associated with a good prognosis in most early stages (5-year survival rate, 100%), and most require no further therapy after complete surgical resection. On the other hand, high-grade ESS presents a poor prognosis, with a 5-year survival rate of 55%. (1)

Recurrence of low-grade ESS is rare. The pelvic cavity is the main site of recurrence (as in our patient's case), followed by vagina and lung. (3,4) In other reports, the median time to recurrence was 34 months, and the rate was 16% to 43% among patients who had bilateral salpingo-oophorectomy (BSO) at initial treatment, versus 68% to 100% in patients who did not have BSO. (5-7) Although our patient was postmenopausal and had BSO and chemotherapy as initial treatment, the tumor recurred after 9 years of follow-up.

The most common symptoms of low-grade ESS (as well as general uterine tumors) are menorrhagia, vaginal bleeding, pelvic or abdominal pain, metrorrhagia, and abdominal fullness or distention. (2)

Huang et al (8) reported that a uniform solid mass with polycystic sections was the ultrasonographic characteristic of low-grade ESS. Our patient's initial and recurrent tumors showed these ultrasonographic and MRI findings. If uterine tumors are assumed to be low-grade ESS or degenerated myoma uteri, the ultrasonographic characteristics are useful for preoperative diagnosis.

When a huge uterine tumor is found, the differential diagnosis includes intravenous leiomyomatosis of the uterus, leiomyomatosis peritonealis disseminata, extrauterine endometriosis, adenomyosis, and low-grade ESS. We can distinguish low-grade ESS from the first 2 diseases, which have mainly benign leiomyoma cells, and from the last 2, which have active endometrial cells. Although the morphologic features of undifferentiated carcinoma are occasionally similar to those of low-grade ESS, we distinguish them by their nucleic configurations.

In general, extrauterine low-grade ESS may develop from the ovary, fallopian tube, pelvis, colon, or retroperitoneum. Recurrence and distant metastasis of low-grade ESS are considered to be related to direct extrauterine development or vessel permeation, as in high-grade ESS. (9) Because our patient had already had a total hysterectomy with BSO 9 years earlier, it was difficult to determine whether the pelvic tumor was a new tumor or a recurrent tumor by means of a deficient hormonal condition. We concluded at the second laparotomy that the pelvic tumor was recurrent because (1) invasion into the hilus of the left ovary was suspected histologically at the first exploration, and the left-sided pelvic recurrent tumor appeared at the same position; and (2) the morphologic and immunohistochemical features of the two tumors were similar.

Rare cases of recurrent low-grade ESS have shown atypical cells that were altered high-grade ESS and leiomyomatous differentiation, reflecting the heterogeneity of this tumor. (4,10) We believe ours may be the first case in which low-grade ESS cells might have changed to these larger atypical cells, which were positive for vimentin and desmin immunohistochemically. They resembled Comet cells of high-grade ESS, fibroblasts, decidual cells, and predecidual endometrial stromal cells at the middle to late secretory phases. The change and differentiation of low-grade ESS by the therapy and environment seem to be the same as the ability for differentiation in mesenchymal stem cells. It suggests that some cases of recurrent low-grade ESS show good responses to progesterone therapy.

The distinction between high-grade ESS and low-grade ESS is whether the level of mitotic activity is greater than 10 mitoses per 10 HPF. (1,11) Huang et al (8) reported that the prognosis of low-grade ESS was provided whether the number of mitoses was less than 3 per 10 HPF or more than 3 per 10 HPF. However, in our case of recurrent low-grade ESS, there were fewer than 3 per 10 HPF. Evans (12) reported that extrauterine development of low-grade ESS was the only prognostic factor, and Chang et al (9) suggested that the mitotic index and cytologic atypia in early ESS could not predict a recurrence. Despite our patient's favorable prognostic factors, the disease took a recurrent course. Therefore, extrauterine development of low-grade ESS at surgery seems to be a better prognostic indicator of recurrence than the level of mitotic activity.

Measurements of DNA content of the 2 tumors in our case indicated normal diploid patterns, as others reported, gathering the good nature of low-grade ESS. (3,13) Moreover, the proliferation index was less than 3% from both tumors in our case. These data suggest that low-grade ESS has a clinical propensity for slow growth and late recurrence in recurrent cases, but DNA content does not accurately reflect the course of this disease. Measurement of estrogen and progesterone receptors has been investigated in a small number of cases of low-grade ESS. These patients tend to have steroid receptors and usually respond to hormonal therapy. (14) Recent studies have raised the possibility of steroid receptors as a prognostic factor. (15,16) Further analysis of a large number of cases would be needed to determine their role. Complete resection of low-grade ESS is the major therapeutic modality, and hormonal treatment should be given expectantly for relapse. This combination therapy with occasional radiotherapy as a loca l control could be attempted in refractory cases. (6)

This case of low-grade ESS with recurrence 9 years later in an atypical anatomic site suggests three key implications. First, recurrence and prognosis of low-grade ESS are thought to be best predicted by extrauterine development, not by mitotic activity or DNA content. Second, it is useful to diagnose low-grade ESS preoperatively using ultrasonography. Third, this recurrent tumor shows various histologic characteristics of differentiation of endometrial stromal cells in proliferative and secretory phases.

References

(1.) Norris HJ, Tayler HB: Mesenchymal tumor of the uterus, a clinical and pathological study of 53 endometrial stromal sarcomas. Cancer 1966; 19:755-766

(2.) Krieger PD, Gusberg SB: Endolymphatic stromal myosis. a grade 1 endometrial sarcoma. Gynecol Oncol 1973; 1:299-313

(3.) Styron SL, Burke TW, Linville WK: Low-grade endometrial stromal sarcoma recurring over three decades. Gynecol Oncol 1989; 35:275-278

(4.) Kato Y, Terashi K, Saotome T, et al: A case of endometrial stromal sarcoma (low-grade malignancy) metastasizing to the lung [in Japanese]. Clin Gynecol Obstet 1991; 45:243-246

(5.) Piver MS. Rutledge FN, Copeland L, et al: Uterine endolymphatic stromal myosis. a collaborative study. Obstet Gynecol 1984; 64:173-178

(6.) Berchuk A, Rubin SC, Hoskins WJ, et al: Treatment of endometrial stromal tumors. Gynecol Oncol 1990; 36:60-65

(7.) Gloor E, Schnyder P, Cikes M, et al: Endolymphatic stromal myosis. surgical and hormonal treatment of extensive abdominal recurrence 20 years after hysterectomy. Cancer 1982; 50:188-193

(8.) Huang KT, Chen CA, Tseng GC, et al: Endometrial stromal sarcoma of twenty cases. Acta Obstet Gynecol Scand 1996; 75:551-555

(9.) Chang KL, Crabtree GS, Lim-tan SK, et al: Primary extrauterine endometrial stromal neoplasms. a clinicopathologic study of 20 cases and a review of the literature. Int Gynecol Pathol 1993; 12:282-296

(10.) Goff A, Rice LW, Fleischhacker D, et al: Uterine sarcoma and endometrial stromal sarcoma. lymph node metastasis and sites of recurrence. Gynecol Oncol 1993; 50:105-109

(11.) Morimoto N, Ozawa M, Kato Y, et al: Diagnostic value of mitotic activity in endometrial stromal sarcoma. report of two cases. Acta Cytol 1982; 26:695-696

(12.) Evans HJ: Endometrial stromal sarcoma and poorly differentiated endometrial sarcoma. Cancer 1982; 50:2170-2182

(13.) Goldfarb S, Richart RM, Okagaki T: Nuclear DNA content in endolymphatic stromal myosis. Am J Obstet Gynecol 1970; 106:524-529

(14.) Katz L, Merino M, Sakamoto H, et al: Endometrial stromal sarcoma. a clinicopathologic study of 11 cases with determination of estrogen and progestin receptor level in three tumors. Obstet Gynecol 1987; 26:87-97

(15.) Sutton GP, Stehman FB, Michael H, et al: Estrogen and progesterone receptors in uterine sarcoma. Obstet Gynecol 1986; 68:709-714

(16.) Mansi JL, Ramachandra S, Wiltshaw E, et al: Endometrial stromal sarcoma. Gynecol Oncol 1990; 36:113-118

RELATED ARTICLE: KEY POINTS

* Recurrence and prognosis of low-grade endometrial stromal sarcoma are thought to be best predicted by extrauterine development not by mitotic activity or DNA content.

* Diagnosing low-grade endometrial stromal sarcoma can be done preoperatively by ultrasonography.

From the Department of Obstetrics and Gynecology, Shiseikai Daini Hospital, Tokyo, Japan.

Reprint requests to Kaoru Yokosuka, MD, Kamisoshigaya, 5-19-1, Setagaya-ku, Tokyo, l57-8550, Japan.