Significant risk factors for malignant transformation of ovarian endometrioma during dienogest treatment: a case report and retrospective study

Background To determine the prevalence of and risk factors for malignant transformation of ovarian endometrioma during dienogest therapy, which is very rare, we examined multiple cases of malignant transformation of ovarian endometrioma during dienogest therapy and performed a multivariate analysis of the records in our hospital. Methods The medical records of 174 patients who underwent DNGT for the treatment of OMA from June 1, 2011, to May 31, 2018, were reviewed retrospectively with the approval of the Human Ethical Committee of the University of Teikyo Hospital. And we provided one representative case of MT with obtaining written informed consent. To assess the effects of six representative factors, including advanced age, parity, surgical history, and endometrial cyst characteristics (including 3 factors), on the possibility of malignant transformation, we performed a multivariate logistic regression analysis. Results Of the 174 cases, 4 were diagnosed with malignant transformation, and these cases are reported. In the multivariate analysis, advanced age (P = 0.0064), nullipara (P = 0.0322), and enlargement (P = 0.0079) showed significant differences for malignant transformation occurrence. All 4 malignant transformation cases were among the 19 patients who had all of these 3 factors. Conclusions For a more accurate determination of the treatment approach, a larger sample size will be needed to determine the risk factors for malignant transformation during dienogest therapy.


Background
Endometriosis is among the most common gynecological diseases and is known to affect 2-10% of women of reproductive age [1,2]. This disease most commonly affects the ovaries, forming a cystic mass called an ovarian endometrioma (OMA) [3], and is diagnosed by an outpatient ultrasound examination. In many cases, magnetic resonance imaging (MRI) is needed for a definitive diagnosis [4]. Because endometriosis is frequently associated with infertility and pelvic pain [5], laparoscopic surgery or drug treatment is often performed for symptom relief [6]. The use of dienogest therapy (DNGT), a drug-based treatment, has recently expanded because DNGT has a curative effect in terms of pelvic pain [7] and is expected to reduce the size of an OMA [8]. This therapy can also reduce postoperative lesion recurrence and extend the pain-free period [9]. Thus, dienogest is a new and effective drug. However, several adverse effects, including uterine bleeding, have been reported [10]. Among these side effects, some cases of malignant transformation (MT) during DNGT have been reported [11], although MT during DNGT is very rare. However, a detailed analysis of this risk factor has not been performed. For this reason, in addition to the present report of several cases of MT during DNGT in our hospital, we have performed a multivariate analysis of the risk factors for MT during DNGT.
Recent studies have indicated that endometrioid carcinoma and clear cell carcinoma are associated with OMA [12]. Ovarian endometrioma is a subtype of endometriosis, affecting 17-44% of women with endometriosis [13]. In most cases, this disease is diagnosed as endometrioma by an outpatient ultrasound examination [4], and laparoscopic surgeries are often performed to treat pelvic pain and subfertility [6]. In a previous report, these cancers reportedly developed in 0.72% of women with a previous diagnosis of OMA who had been managed expectantly [14]. Conversely, some reports showed that a certain ratio of patients with ovarian cancers have a history of endometriosis [15,16]. However, no significant relationship between previous treatments with hormone drugs (or surgery) and the risk of MT has been validated [17,18]. Therefore, our analysis may provide new insight into preventing future cases of MT of OMA.

Data collection
This study was reviewed and approved by the Human Ethics Committee of the University of Teikyo Hospital (trial registration number 18-233). The de-identified medical records of 185 female patients who received dienogest for the treatment of OMA between June 1, 2011, and May 31, 2018, were reviewed retrospectively with the approval of the Human Ethical Committee of the University of Teikyo Hospital. And we provided one representative case of MT with obtaining written informed consent. In this study, we included patients with and without surgery for OMA. Of the 185 patients, 11 patients were excluded because of insufficient data. Among the 174 remaining cases were 4 cases with MT of OMA, and the following data were collected: (1) the size of the OMA before and after starting DNGT (when the patients had bilateral OMA, we adopted the larger OMA from the right or left side), (2) the patient's parity (nulliparous or multiparous patient), (3) the patient's age at the time of starting DNGT, (4) the location of the endometrial cyst before starting DNGT (unilateral or bilateral), and (5) whether surgery was performed. In this study, we focused mostly on the change in the cyst size. In addition, 174 cases were classified into the following three groups: (1) "reduction," defined as cases in which the cyst size showed a decline of 30% or more; (2) "enlargement," defined as cases in which the cyst size showed an increase of 30% or more or in which a new cyst appeared after surgery; and (3) "no change," which included the remaining cases.
Usually, when treating patients with endometriosis, we use the revised American Society for Reproductive Medicine (rASRM) score to classify the severity [19]. However, due to insufficient data, we could not include the rASRM score in the multivariate analysis. Because not all patients underwent surgery, we were able to obtain data regarding the rASRM score for only 73 cases (42%; n = 73 of 174). When an ovarian endometrioma is detected, the severity of endometriosis is regarded as moderate or severe, meaning that the severity of endometriosis is at least stage 3 among four stages. Therefore, all patients in this analysis had a certain level of endometriosis severity. Additionally, we did not analyze the treatment period of dienogest, because DNGT was generally used for the long term and was continued until menopause in some cases [20].

Statistical analysis
We performed a multivariate logistic regression analysis to assess the influence of the following six factors on the possibility of MT: (1) "advanced age," defined as patients who started DNGT at 41 years of age or older; (2) nullipara; (3) "large cyst," defined as patients whose cyst sizes were 8 cm or larger at the start of DNGT; (4) "surgical history," defined as patients who underwent surgery for OMA; (5) bilateral cysts; and (6) enlargement. These statistical analyses were performed using JMP version 12 for Windows (SAS Institute, Inc., Tokyo, Japan). The data are presented as the means ± standard deviations. A P value less than 0.05 was considered statistically significant.

Case presentation
A 43-year-old Japanese woman, gravida 0, para 0, was referred to our hospital because of aggravated dysmenorrhea and hypermenorrhea. An ovarian cyst had been noted in this patient 3 years before her visit to our hospital, and she had dysmenorrhea beginning at approximately 30 years of age. MRI that was performed before her visit to our hospital suggested a left endometrioma that was 7 cm in diameter and a right endometrioma that was 3 cm in diameter, with no solid component in either endometriotic cyst. Bilateral laparoscopic cystectomy (LC) was performed with the patient under general anesthesia, and the diameters of the right and left endometriomas were 3 cm and 7 cm, respectively ( Fig. 1a and b). After both ovaries were subjected to a surgical procedure, including complete Douglas pouch obliteration, to release them from their severe pelvic adhesion, the left endometrial cyst was completely excised, and the right endometrial cyst was ablated. The rASRM score was 81 points, and this case was diagnosed as stage IV endometriosis. The ovarian cyst did not contain a solid component (Fig. 1c), and pathological examination showed a left endometriotic cyst and no malignancy (Fig. 1d). After surgery, DNGT was chosen for the prevention of recurrence, and the patient was monitored by transvaginal ultrasound every 6 months.
At 1 year and 1 month after surgery (that is, 1 year after the start of the administration of DNGT), when the patient was 44 years old, a recurrent left-sided OMA that was 6 cm in diameter with a solid component inside the left ovarian tumor was detected by outpatient transvaginal ultrasound examination. Contrast-enhanced MRI showed that in addition to a 3-cm right endometrioma, there was a 7-cm left endometrioma with a solid component inside the cavity in which MT was strongly suspected (Fig. 1e). In addition, contrast-enhanced computed tomography exhibited numerous para-aortic lymph node enlargements, including a maximum-sized node that was 10 mm in diameter. The serum cancer antigen 125 level, which was 51.4 U/ml at the time of the first visit to our hospital, was increased to 515 U/ml. Soon after the imaging diagnosis was performed, complete abdominal surgery, including removal of the uterus, bilateral ovaries, omentum, pelvic lymph nodes, and para-aortic lymph nodes, was performed. The final diagnosis was clear cell carcinoma of the left ovary (Fig. 1f, g, h), International Federation of Gynecology and Obstetrics (FIGO) stage 1C1, pT1C1N0M0. After the necessity of chemotherapy was explained to her, she received chemotherapy of 175 mg/m 2 paclitaxel and carboplatin (TC; area under the curve, 6 mg/ml/min) for six cycles and was without recurrence during follow-up.

Patient characteristics
In the 174 patients with DNGT, the average endometrioma size before the start of therapy, patient age, and observable treatment period were 4.9 ± 2.2 cm, 39.7 ± 5.9 years, and 27.3 ± 21.6 months, respectively. There were 70 patients with bilateral OMAs, 51 patients with rightsided OMAs, and 53 patients with left-sided OMAs. We were able to divide the 174 patients into 57 "reduction" cases, 70 "enlargement" cases, and 47 "no change" cases. All four cases of MT during DNGT were included in the "enlargement" group. Therefore, when we compared the change in cyst size, 57 "reduction" cases and 47 "no change" cases were treated as similar cases in the subsequent multivariate analysis. Table 1 lists the characteristics of the four patients who developed ovarian cancer during DNGT. In this study, we included two patients who were referred to our hospital because the possibility of ovarian cancer during DNGT had been noted in other hospitals. Therefore, 2 of 172 patients who underwent DNGT were diagnosed with MT in our hospital, and the probability reached 1.2%. Of four patients, two patients began DNGT after LC. All four patients were nulliparous. The ages at the time of DNGT ranged from 42 to 44, and the ages at the time of diagnosis of MT ranged from 42 to 46. The duration of DNGT ranged from 9 to 33 months. Two patients had bilateral OMAs, and two patients had a left OMA. Of four cases, three cases of ovarian cancer developed from a left-sided OMA, and one case of ovarian cancer developed from a rightsided OMA after experiencing recurrence or enlargement of the OMA. As shown in Table 1, the ovarian cancers for all four cases were detected in the early stage and were FIGO stage 1C1 (three cases) or 1C3. Complete surgery, including the removal of the uterus, bilateral ovaries, omentum, pelvic lymph nodes, and para-aortic lymph nodes, was performed in two cases (cases 1 and 3). For the remaining two cases, we performed total hysterectomy and bilateral salpingo-oophorectomy (case 4) or left-sided salpingooophorectomy (case 2). For case 2, we performed interval debulking surgery, including the removal of the uterus, right ovary, omentum, pelvic lymph nodes, and para-

Influential factors in malignant transformation
To determine significant factors associated with the possibility of MT, we compared the following six factors using multivariate logistic regression models ( Table 2)

Discussion
Currently, DNGT is being used for patients with symptomatic endometriosis with or without surgery [21]. This progestin agent is reported to decrease the risk of recurrence by inhibiting the growth of endometriotic tissue [21,22]. This mechanism can also theoretically prevent MT [23], and reported cases of MT are very rare [24]. However, in our hospital, we had four cases in which ovarian carcinoma was diagnosed during DNGT for OMA. MT in two of the four cases was detected entirely in our hospital, and the probability of MT after starting DNGT was 1.2% (n = 2 of 172). This probability was similar to the ability of endometriosis to transform into malignancy, which has been reported to be approximately 1% of OMAs in previous studies [14]. In this study, we extracted the details of the patients' clinical histories. Because all four patients were over 40 years old and nulliparous, we hypothesized that there were several risk factors, including the patient's age and parity, for using DNGT.
To verify the significant influence of risk factors on MT, a multivariate analysis of six factors extracted before starting DNGT was performed ( Table 2). Of the six factors, advanced age, nullipara and increased size and recurrence showed significant increases in the probability of MT. In contrast to a previous study in which the association between spontaneous pregnancy and the disease progression of endometriosis was unclear [25], these results were roughly consistent with those of past reports [26,27]. In particular, 19 patients who had all three factors showed a probability of over 20% for the occurrence of ovarian carcinoma (21.1%, n = 4 of 19). Because the sample number was very small, the result for this extracted highrisk group might be coincidental. Therefore, a larger sample size will be needed to determine the risk factors for MT. However, in contrast to previous studies, large cyst [26][27][28] and surgical history [29] did not show a significantly higher probability of MT. "Bilateral cysts" were also not significant. These multivariate analysis results presented the possibility that the characteristics of the patient have a stronger influence on MT than the characteristics and treatment approach of the OMA itself. Because two cases among the four cases of MT exhibited recurrences of OMA after LC and because the other two cases exhibited cyst size enlargement after starting DNGT, careful observation of the cyst size is important, regardless of the use of surgical treatment. However, similar to previous reports [11], in our hospital, we could not prevent the two cases of MT (Table 1, case 1 and case 2) by follow-up ultrasound examination at an interval of 6 months.

Conclusions
Although DNGT seems to be effective in controlling OMA, careful observation of the cyst size is needed. Because the probability for ovarian carcinoma occurrence could exceed 20% in some cases, particularly for nulliparous women of advanced age, a more individualized choice of therapy might be needed. For a more accurate determination of the treatment approach, a larger sample size will be needed to determine the risk factors for MT.  A multivariate analysis of 174 patients with DNGT was performed to examine the influence of six representative factors that could be collected before starting DNGT, as follows: (1) "advanced age," defined as patients who started DNGT at 41 years of age or older; (2) nullipara; (3) "large cyst," defined as patients whose cyst sizes were 8 cm or larger at the start of DNGT; (4) "surgical history," defined as patients who underwent surgery for the OMA; (5) bilateral cysts; and (6) enlargement. The number of patients with each factor, the probability of MT, and the P values are shown. "Advanced age," "nullipara," and "enlargement" were identified as significant risk factors for MT. "Large cyst," "bilateral cysts," and "surgical history" were not significant risk factors Abbreviations: DNGT dienogest therapy, MT malignant transformation, OMA ovarian endometrioma American Society for Reproductive Medicine score; TC: Paclitaxel and carboplatin