Background
Struma ovarii is a form of ovarian teratoma defined by the presence of thyroid tissue comprising more than 50% of the tumor [
1]. Approximately 2–5% of all ovarian teratomas and 0.5–1% of all ovarian tumors can be classified as struma ovarii [
2‐
4], and malignant struma ovarii (MSO) accounts for approximately 5–10% of struma ovarii that can be histologically identified as differentiated thyroid carcinoma [
5,
6]. MSO occurs most commonly in women in their 30s and 40s with widely varied clinical manifestations [
7,
8]. There are fewer than 200 cases reported in literature so far [
7] and due to its rarity, there is currently no consensus on management of MSO, and further evidence is needed on the factors affecting prognosis to reach this consensus. Evidence on and the benefits of different surgical options mainly originate from case reports.
DeSimone et al. and Shrimali et al. both suggested that MSO should be treated with a combination of local surgery, a total thyroidectomy, and radioiodine therapy (RAI) for both patients with and without extra-ovarian metastasis [
9,
10]. For women of childbearing age, a more conservative management option is certainly practical, while complete staging surgery should be performed in patients who are not concerned with fertility preservation [
9]. However, McGill et al. and Marti et al. advocated that RAI should be reserved only for patients who have evidence of metastasis [
11,
12]. They also found the recurrent rate was as low as 7.5% in 25-year-olds with well-differentiated thyroid cancer within an ovary, therefore pelvic surgery alone may be adequate in this population [
12]. Nonetheless, other studies reported a much higher recurrent rate, ranging from 22 to 35% [
9,
13]. While the survival rate in patients with MSO is excellent regardless of management strategy employed [
8,
14], the specific pelvic surgical option that should be prioritized and the benefits of aggressive surgery are still uncertain. Moreover, none of these studies emphasized the overall survival (OS) rate in patients with MSO within the ovary with varied differentiated degrees, and the factors associated with recurrent-free survival (RFS) and OS have not been well defined.
Here we aimed to present the clinical and pathological characteristics and treatment options for patients with MSO confined to the ovary, while also investigating the RFS and OS rate, as wells as factors that affect both rates in this population. We included five cases of MSO confined to the ovary from our hospital and comprehensively reviewed another 120 cases documented in literature from MEDLINE from the last 80 years.
Discussion
Our study presented one of the largest cohorts concerning the management of disease in patients with MSO confined to the ovary and was the first to assess factors associated with RFS and OS in this population. The prognosis of MSO within the ovary was explored for varied treatment options, and we did not identify any statistically significant factors to predict RFS and OS. Based on the rarity and promising prognoses of these tumors and their lack of identified prognostic factors, an individualized and less aggressive treatment strategy is recommended.
Overall survival outcome in patients with MSO has been previously well demonstrated in two large cohort studies [
8,
14]. In 2009, Robboy et al. reported that the OS rates of patients with struma ovarii with malignant behavior were 89 and 84% at the 10-year and 25-year time point, respectively [
14]. However, 58 (66%) of their cases were follicular adenomatous or benign tumors; only 27 cases were strictly histologically consistent with MSO, which may have biased the true OS rate in patients with MSO. Goffredo et al. in 2015 published another study on 68 patients and revealed OS rates of 96.7, 94.3, and 84.9% at the 5-year, 10-year, and 20-year mark, respectively [
8]. Their study was the first to present a more accurate OS rate in patients with MSO, but 20% of their cases were MSO with local or distant metastasis. Our current study provided a larger sample size and was restricted to those without extra-ovary spread at the initial diagnosis, and demonstrated a similarly excellent OS.
Overall recurrent rate in our research was 21.8%, with a cumulative recurrent rate of 27.1% at the 5-year and 35.2% at the 10-year mark, which was much higher than that of the 7.5% in the 25-year mark from Marti et al. [
12]. However, it was compatible with the study reported by DeSimone et al., which had an overall recurrent rate of 35% in 24 patients [
9], and Jean et al., which had an overall recurrent rate of 22% in 59 patients [
13]. This may be attributed to the fact that Marti et al. restricted their selected patients to well-differentiated thyroid cancer. Meanwhile, we excluded more than 30 patients who lived with no evidence of disease recurrence but lacked individual detailed follow-up data in the literature [
1,
14,
18‐
21]. This partly overestimated the recurrent rate. Although we reported a relatively higher recurrence rate, the cumulative recurrent rate was much lower than epithelial ovarian cancer [
22]. Furthermore, an interesting result was that tumor recurrence was not a risk factor for OS, which was not consistent with common ovarian cancer [
23]. This could be explained by the fact that the impact of MSO is generally milder and not lethal, as well as by its positive response to adjuvant therapy after recurrence. This reminds us that the treatment strategy of MSO needs to find a balance between pursuing radical treatment and improving long-term quality of life.
In our study, we also noted there was no potential risk factor for RFS and OS in patients with MSO confined to the ovary. Patients’ age at diagnosis, tumor size, presence of ascites, surgical options, pathological subtypes, RAI, and chemotherapy failed to predict either RFS or OS. Age over 55 years old was a strong predictor of relapse [
24] and disease-specific death [
16] in well-differentiated thyroid carcinoma. It is thought that older patients had a worse response to therapy (i.e. RAI) [
25] and were not amenable to an aggressive therapeutic approach after the initial treatment [
26]. However, all of our patients were diagnosed as MSO within the ovary at initial presentation and only 20.0% were initially administered RAI. A large cohort that includes MSO both with and without metastasis is needed to evaluate whether age at diagnosis can predict prognosis.
Several studies have investigated the risk factors for poor prognosis. For example, tumor size varying from 2 to 12 cm was proposed to predict an adverse clinical course or help classify “low-risk” and “high risk” patients [
2,
14,
17]. Robboy et al. [
14] suggested that a size of the strumal component ≥6 cm was related to disease recurrence while Shaco-Levy et al. [
17] proposed that an overall size ≥10 cm and a strumal component ≥80% were associated with rapid disease progression and death. In this study, univariate analysis revealed that tumor size failed to predict either RFS or OS. However, most literature only provided the size of the struma instead of thyroid carcinoma because the carcinoma component is usually multifocal and blends together with the struma. It might be hypothesized that the size of the carcinoma component, rather than the strumal component, might affect the prognosis in patients with MSO, but further relevant research is needed.
Ascites more than one liter was also defined as a feature that can predict an adverse clinical course in previous studies [
14,
17]. In our study, multivariate analysis showed that ascites was likely to predict OS (
p = 0.054). This may be explained by that patient presented with ascites may in a more advanced tumor stage, while a more advanced stage of thyroid carcinoma or typical epithelial ovarian carcinoma both predict a poor prognostic [
27,
28]. However, patients with benign struma ovarii can also present with ascites, such as in patients with pseudo-Meigs’ syndrome [
29]. A previous study also suggested that histologic features used to predict clinical outcomes in thyroid tumors were not applicable to MSO [
30]. We did not include these factors because most of the cases in our study had no detailed information about such histologic features. In addition, different subtypes of MSO showed similar RFS and OS in our study, which was not compatible with the previous study that found that recurrence of papillary carcinoma occurred earlier [
14]. The lack of correlation between morphology and clinical outcome in MSO is striking, making the behavior of these tumors particularly unpredictable and the decision of treatment course more uncertain. Interpretation of the excellent prognosis of MSO may help to optimize disease management in this population.
Currently, the surgical options reported in the literature include an ovarian cystectomy, USO, BSO, TAH/BSO, and debulking surgery. However, the priority of these different surgical approaches has not been well evaluated in large cohorts, and evidence mainly relies on case reports. Some researchers argue for comprehensive staging surgery in postmenopausal patients or those who do not need to preserve fertility, otherwise USO should be preferred [
9,
13]. We found that while both the RFS and OS rates were promising, there was no specific surgical option that could promise a more favorable prognosis. Ovarian cystectomies may cause the intraoperative rupture of tumor cysts, which may inevitably lead to the dissemination of tumor cells. This was not compatible with the tumor-free principle. Since nearly half of our patients were younger than 45 years old, surgical resection exceeding USO, especially debulking surgery, can increase trauma and blood loss and impact long-term quality of life. In reproductive women affected with MSO confined to the ovary, we strongly recommend USO as the preferred surgery. In perimenopausal or postmenopausal women, USO or BSO is more advisable for their feasibility prior to TAH/BSO and debulking surgery, regardless of their similar outcomes.
Whether RAI should be performed postoperatively has always been argued over [
2,
9‐
13]. Most authors advocated for routine RAI to lower the recurrent rate after primary surgical resection of tumors [
9,
10,
13]. However, Marti et al. found that pelvic surgery alone may be sufficient in patients without extra-ovarian metastasis [
12]. Two large studies from Marti et al. and McGill suggested RAI should be reserved for cases with evidence of metastasis [
11,
12]. Research published by Goffredo et al. found that only 9.2% of the 68 patients received RAI, but their survival outcomes were excellent [
8]. In our study, the overall recurrent rate was significantly lower in patients treated with RAI than in those who received no RAI (3.8% vs. 26.1%,
P = 0.014), yet its efficacy in improving RFS was not proven in univariate and multivariate analysis. Nonetheless, total thyroidectomies must be conducted before RAI, leading to a need for lifelong thyroxine supplements. Moreover, in premenopausal women, especially reproductive aged women, the impact of RAI on ovarian function cannot be neglected. Previous research demonstrated that RAI would lead to decreased ovarian function; the impact was more obvious in patients receiving multiple RAI therapies and in those older than 35 years [
31]. Furthermore, ovarian function in patients with MSO has already been impaired by surgery. Knowing that the benefits of RAI on lowering recurrent rate is uncertain and it does not optimize OS, the use of RAI in patients with MSO confined to the ovary remains controversial. Physicians and patients must have clear communication before RAI and an individualized therapeutic plan taking into consideration patients’ personal intentions might be more practical.
Molecular profiling has been shown to impact clinical outcomes and help assess risk stratification of thyroid cancer. For example, BRAF
V600E mutation is associated with aggressive histologic features and metastases, and coexistent BRAF
V600E and telomerase reverse transcriptase (TERT) promoter mutations have a synergistic effect on increasing the risk of recurrence [
32]. Published researches had also identified different somatic mutations in malignant struma ovarii, mainly in BRAF and RAS, as well as RET/PTC rearrangement [
33‐
36]. Although MSO is histologically and genetically similar to primary thyroid carcinoma, currently no study has reported a molecular pattern that can predict more aggressive behavior in MSO. The reason may be that BRAF or RAS mutations and TERT promoter mutations were detected alone in 53 MSO cases with available data of molecular profile [
35]. However, the fatal forms of non-ATC are generally PTC variants harboring BRAF or RAS mutations plus other genomic alterations such as TERT promoter [
37]. Therefore, data were insufficient to evaluate the impact of molecular profiles on RFS or OS in patients with MSO. These researches indicated the potential significance of molecular features in guiding postoperative treatment, and further study should be conducted.
The follow-up strategy of patients with MSO confined to the ovary has not yet been well established. Serum thyroglobulin (TG) concomitant assessment of serum TG antibody (TGAb) and imaging should be the mainstays of MSO follow-up. Prophylactic total thyroidectomy to exclude a primary thyroid carcinoma and potentiate RAI therapy has been recommended, which allow for TG monitoring of possible metastases, remained mass or recurrence [
9,
10]. However, our study revealed that total thyroidectomy followed by RAI is not mandatory for patients with MSO confined to the ovary. We recommend that all patients with MSO should have serum TG and TGAb assessments every 6–12 months refer to the guideline of thyroid cancer [
38]. Meanwhile, the follow-up schedule may be individualized depends on the disease condition and initial treatment efficacy. We recommend that pelvic MRI can be performed every 1–2 years to exclude recurrent disease. Other imaging studies (WBS, FDG–PET, ultrasonography, and CT) should be ordered if locoregional or distant recurrences are presented, increasing serum TG or TGAb levels, or patients have suspected diseased-related clinical symptoms. For patients preserving thyroid, neck ultrasound should be taken regularly and elevated TG above baseline should prompt further evaluation for recurrent disease. We also recommend the monitoring of MSO is at least 20 years since the median recurrence time in our study was 14 years and cases of late recurrence have been reported [
9].
This study has several limitations. First, most cases from the literature review increased the heterogeneity of this study, weakening the validity. Second, follow-up time among the patients significantly varied, and in some of the patients it was not long enough, which may influence the exact outcomes. Third, surgeries being performed by surgeons in different institutions may also further impact the prognosis, even if the same surgical approaches were used. An ideal method would be conducting a prospective, randomized cohort to determine the optimal surgical option and examine the role of RAI in patients with MSO confined to the ovary. However, it is unrealistic due to its low feasibility. Last, we excluded many records mainly due to language and lack of institutional access. Further research is needed to determine optimal disease management.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.