Discussion
The current WHO Classification of Tumors of Female Genital Organs 2014 introduced new terminology for two separate categories of ovarian serous carcinomas, namely low- and high-grade ones, the former being described as being on a continuous spectrum from cystadenoma to carcinoma via borderline tumor [
1]. The important changes in classification of this entity include classifying SBT/APSTs with any invasive foci (previously known as “invasive implants”) as peritoneal LGSC and identifying a histologic variant as an SBT/APST with significant micropapillary features (size > 5 mm, currently synonymous with “non-invasive LGSC”) because its biological behavior more closely resembles that of ovarian LGSC than of conventional SBT/APST [
2]. In a meta-analysis of 97 studies including 4129 patients with SBT/APST, Seidman et al. reported overall survivals of virtually 100% for Stage I tumors and 95.3% for advanced stage tumors with non-invasive implants, whereas the survival for tumors associated with invasive peritoneal disease, that is, LGSC, was only 66% after a 7.4-year follow-up [
3]. A nationwide cohort study in Denmark of 1487 women with SBT/APST with or without micropapillary features over a 25-year period (1978–2002) found that overall survival of patients with tumors confined to the ovaries was not different from that of the general population [
4]. Further review of the histopathological features of these patients’ tumors by a panel of pathologist resulted in classifying 867 of them as conventional SBT/APSTs and 75 as non-invasive LGSCs. In the entire cohort, 41 women (4%) developed invasive serous carcinoma (38 low grade, three high grade; median time to diagnosis, 10 years; range, up to 25 years), such progression occurring significantly more frequently in patients with non-invasive LGSCs than in those with conventional SBT/APSTs [
5]. In the present case, we speculate that recurrence was attributable to microscopic residual tumor after debulking surgery, possibly related to seeding caused by intraoperative rupture, developing in an indolent manner during the 13-year disease-free period. Given that distant metastasis must have occurred via lymphovascular permeation and that no peritoneal foci were detected at the time of recurrence, we consider it more likely that there was unrecognized lymph node involvement because lymph-node biopsy was not done during the original surgery. An alternative explanation is undersampling of the primary tumor (6 sections in all per a 13-cm tumor), in that the absence of invasive foci or micropapillary components in the tissue sections examined histologically did exclude their presence in unsampled parts of the primary tumor. Though additional sampling may not have changed the overall diagnosis, it would have provided credence to the assertion that no invasion or micropapillary features were identified.
Immunohistochemistry is essential for histologic diagnosis of suspected metastases, especially with unknown primary tumors. In women with a history of gynecologic tumor and proliferative cells in a papillary arrangement in a metastasis, PAX8 is very helpful in that it is a highly specific marker for neoplasms originating in the Müllerian duct system [
6], whereas strong WT1 positivity points toward extra-uterine serous differentiation [
7]. In our patient, another possibility, thyroid cancer, was excluded by the finding of negativity for both thyroglobulin and TTF1, the former being highly specific to follicular neoplasms with the exception of poorly differentiated and anaplastic types [
8]. The radiologically-determined location of the tumor was consistent with aggressive meningioma; however, this possibility was ruled out because the morphology was not typical of either conventional or papillary subtypes and both of these are usually negative for PAX8 and WT1. Negative calretinin staining ruled out mesothelioma. The presence of wild-type expression of P53 excluded high-grade serous carcinoma (HGSC) [
9]. Thus, we made a final diagnosis of metastatic LGSC transformed from SBT/APST originating in the right ovary.
To the best of our knowledge, no patient with initially diagnosed SBT/APST confined to the ovary and subsequently presenting with a lesion in bone as the first metastatic site has previously been reported. Intraparenchymal metastasis, that is, neither peritoneal nor lymph node involvement, is quite uncommon in early low-grade serous ovarian tumors, whereas metastases can occur in various extraperitoneal sites in advanced stage LGSCs [
10‐
12]. Xing et al. recently reported a 59-year-old woman with a history of breast cancer who developed metastatic SBT/APST in the brain, this being considered to have originated in serous cystadenofibromas of both ovaries excised 5 years after a brain metastasis of unknown primary had been diagnosed [
13]; thus, pathologists should take into consideration that SBT/APSTs lacking any LGSC features can metastasize to parenchymal organs in an occult fashion. Given that time to relapse after initial diagnosis of SBT/APST or LGSC has been reported to be as long as 28 years [
14], long-term surveillance should be continued in women with SBT/APST and LGSC, which unpredictably manifest as overt metastases after a long disease-free period.
Somatic mutations in KRAS or BRAF occur separately or not all (never simultaneously) in about 50% of SBT/APSTs and LGSCs [
15,
16], whereas P53 mutations occur almost exclusively in HGSCs [
9]. Wong et al. reported detection of BRAF or KRAS mutations in nine (30%) and five (17%), respectively, of 30 SBT/APSTs, but in only one (2%) and eight (19%), respectively, of 43 LGSCs. BRAF mutations occurred less frequently at advanced stages [
17]. The same group detected KRAS mutations with PCR and Sanger sequencing in 10 (43%) of 23 women with recurrent LGSCs (six in primary SBT/APSTs, five in recurrent LGSCs, and one in both), whereas only one patient had BRAF V600E mutation in a SBT/APST [
18]. Because advanced stage and recurrent LGSCs are usually chemoresistant, clinical trials have been performed on new chemotherapeutic agents correlated to genetic profiles of individual tumors. LGSCs harboring BRAF or KRAS mutations often respond to mitogen-activated protein/extracellular signal-regulated kinase inhibitors such as selumetinib via down-regulation of the MAPK pathway. However, there is reportedly no correlation between BRAF/KRAS mutations and objective response [
19]. Bevacizumab in conjunction with chemotherapy and hormonal therapy is one recommended combination for BRAF/KRAS wild-type SBT/APST or LGSCs, as in the present patient [
20].