Pathogenesis of peritoneal high-grade serous carcinoma after risk-reducing surgery: a systematic review.

Conception: TAG, ABB, JAH, JMJP, JAWML, MS, MPS; data collection: TAG, ABB; data analysis: TAG, ABB, MPS; writing: TAG, ABB, JAH, JMJP, JAWML, MS, MPS.

A total of 4,123 studies were identified of which 442 were duplicates. After title/abstract screening and full‐text screening, 42 studies were included for review. Via a cross‐reference search, six other studies were included. One additional study was identified via email alerts based on our search strategy and was also included. In summary, the total of included studies was 49 (Table  1 ). Of these studies, 26 were rated as ‘high’ quality, 18 as ‘moderate’ quality and five were of ‘low’ quality (supplementary material, Table  S3 ). Overview of pathogenic mechanisms leading to peritoneal HGSC Observation of a clonal relationship between 5 and 7 STICs and metachronous peritoneal HGSCs. Observation of a clonal relationship between 27/29 STICs and concurrent peritoneal HGSCs. Observation of WGS and WES data showing tubal lesions preceding concurrent HGSCs. A strong correlation between ploidy in STIC lesions and concurrent HGSC was shown. Molecular profile similarity FT/STIC – HGSC [ 34 , 35 , 36 ] Lack of difference in DNA expression of HGSC with and without concurrent STIC suggests identical origins. Observation that HGSC has a molecular profile closer to FT epithelium than OSE. Observation that mouse models using FT specific promotors, Müllerian tissue specific promotors or injected FT cells developed HGSC. Evidence that salpingectomy at early age in mouse models using a Müllerian tissue specific promotor prevented HGSC formation in mice. Observation that changes in telomere length correspond with STIC being a precursor of HGSC. Tubal lesion morphology with loose cells was more often observed concurrent with HGSC. Several cell processes (L1CAM, PR, norepinephrine) might play a role in quiescence and pro‐survival/pro‐dissemination cell behaviors in ultra‐low attachment conditions. Evidence that follicular fluid (released at ovulation) enhanced migration of FT cells in mice. Observation that netrin (involved in pro‐survival signaling cells) overexpression elevated cell survival in dormant culture conditions and contributed to greater spread of disease in mice. Indication of a tubal origin beyond STIC as in 10/13 cases STIL + p53 signatures share identical TP53 mutations with their concurrent HGSC. Observation that 3/20 cases showed fimbria adherent to the ovary at histopathological analysis of RRSO specimen. Observation that in 19/30 samples TP53 alterations were identical in SEIC and HGSC, indicating a clonal relationship. Observation of two cases with a clonal relationship between STIC and USC. Detection of a SEIC in 2/98 STIC cases. In 9/60 HGSC cases a SEIC was observed and in 4/28 USC cases a STIC was found. Metastasis of HGSC to the endometrium in a mouse model was observed. Observation that cortical inclusion cysts are not found in premenarchal girls and frequency increases with age. Observation of a small p53 signature in the OSE of one ovary. Development of (epithelial) metastatic tumors (after salpingectomy) in mouse models with a Müllerian mesenchymal promotor. Observations of simple mesothelial cysts were found in 55/172 investigated ovaries. A transition from benign epithelium to malignant epithelium of the cyst was observed. Observation of identical KRAS and wild type p53 mutations in the serous borderline tumor and high‐grade component of the tumor in 2/6 cases, indicating a clonal relationship. Observation of HGSC arising in association with SBT and LGSC in one case. Evidence of HGSC and LGSC coexisting in three cases. Three cases of an intra‐epithelial carcinoma in the ovary, one in OSE, two in ovarian serous cystadenofibroma. Analyses of immunohistochemistry (PAX8) and gene expression profiles indicating the fallopian tube as a more common HGSC origin; however, an ovarian origin cannot be excluded. Observation of a lower latency and penetrance of ovarian origin compared to fallopian tube origin in mice. Evidence that inflammation and adipocytes have a role in the migratory process of HGSC cells in mice. We grouped the findings according to their pathophysiological mechanism: mechanisms following FT origin, endometrial origin, ovarian origin, and micro‐environment (Figure  2 ). Within these pathways, six, three, six, and one subarea(s) were identified, respectively (Table  1 ). Sampling error was identified as a confounding factor. Systematic overview of potential pathways that lead to the development of peritoneal HGSC. Created with BioRender.com. Twenty‐three of the identified studies provided evidence on mechanisms occurring after the initiation of HGSC in the FT (Table  1 , Figure  2 ‐IV; supplementary material, Table  S4 ). Van den Berg et al showed a clonal relationship between STIC at RRSO and metachronous peritoneal HGSC in 5/7 patients. In the other two patients, different TP53 mutations were found [ 31 ]. Furthermore, Cheng et al and Labidi‐Galy et al performed whole genome sequencing (WGS) and whole exome sequencing (WES) analyses, respectively, on tubal lesions and concurrent HGSC samples. With phylogenetic analysis, they showed a clonal relationship based on TP53 mutations. This indicated that the tubal lesions preceded omental HGSC [ 8 , 32 ]. Kuhn et al showed a clonal relationship between STIC and concurrent HGSC in 27/29 cases [ 33 ]. Additionally, three studies showed a similar molecular profile in FT epithelium/STIC and concurrent HGSC [ 34 , 35 , 36 ]. One study showed that STIC had a shorter telomere length compared to HGSC, which indicated STIC being a precursor for HGSC [ 43 ]. Six mouse models showed HGSC development when using a FT‐specific promotor, a Müllerian tissue‐specific promotor, or after injecting FT cells [ 37 , 38 , 39 , 40 , 41 , 42 ]. Metachronous HGSC formation was prevented when salpingectomy was performed in two mouse models that used a Müllerian tissue‐specific promotor. Kim et al and Perets et al showed that HGSC formation was not prevented when oophorectomy was performed, which in Perets et al was also not the case after hysterectomy [ 40 , 41 ]. However, removal of the ovary without the FT did reduce the occurrence of peritoneal metastases [ 40 ]. Eight studies showed that FT epithelium might be able to migrate to the abdominal cavity and survive in ultra‐low attachment conditions in an in vitro simulation of peritoneal fluid, in which cells cannot attach to surrounding stroma [ 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 ]. One of these studies showed that a STIL/STIC morphology with loosely adherent or detached cells more often occurred with HGSC [ 51 ]. Another study showed that upregulation of L1CAM could play a role in the migration of FT cells, with increased immunohistochemical expression of dissociating cells [ 44 ]. A potential role for follicular fluid in the dissemination of tubal cells is described in three mouse models [ 47 , 48 , 49 ]. Chu et al showed that hepatocyte growth factor in the follicular fluid plays a crucial role in migration and invasion [ 47 ]. Hsu et al confirmed this and additionally showed that follicular fluid enables cells to survive without being attached to the extracellular matrix. This is vital in metastasizing and means circulating cells can survive [ 48 , 49 ]. The described effect of the follicular fluid could be an explanation of why removal of the ovary reduced the spread of the disease in the mouse model study of Perets et al [ 40 ]. Two other studies showed that STILs/p53 signatures shared identical TP53 mutations with concurrent HGSC, indicating that a STIL or p53 signature may directly transform into HGSC, without first developing into STIC: the precursor escape mechanism [ 32 , 52 ]. Additionally, one of these studies showed that STILs and p53 signatures were found after the whole FT was assessed with deeper sections, but STIC was not [ 52 ]. Lastly, one study showed FT removal could be incomplete, as parts of the distal fimbriae could be adhered to the ovary and stay intra‐abdominal, which might cause HGSC after salpingectomy [ 53 ]. Seven studies indicated a potential role for the endometrium (Table  1 , Figure  2 ‐II; supplementary material, Table  S5 ). In 9/60 HGSC cases, a serous endometrial intraepithelial carcinoma (SEIC) was found, a precursor for USC [ 57 ]. In two other studies, 19/30 SEICs and their concurrent extrauterine carcinoma had an identical TP53 mutation, indicating a clonal relationship [ 54 , 55 ]. In all cases where no clonal relationship was proven, this relationship could not be ruled out, as only exons 5 to 8 were sequenced and these cases had only one or no TP53 mutation. However, the chance of finding a clonal relationship in the cases with already one TP53 is rather limited [ 55 ]. Jia et al confirmed this, as in multiple patients a simultaneous clonal relationship between peritoneal HGSC and serous tubal (intraepithelial) carcinoma as well as ovarian/peritoneal HGSC and SEIC was found, with each pair having a different TP53 mutation. In most cases where no clonal relationship between SEIC and peritoneal HGSC was observed, a clonal relationship between serous tubal (intraepithelial) carcinoma and peritoneal HGSC was found [ 54 ]. In cases without STIC and with completely embedded FTs, ploidy analysis between SEIC and concurrent peritoneal HGSC showed increased ploidy in peritoneal HGSC compared to SEIC, indicating neoplastic progression from SEIC [ 55 ]. Additionally, Mingels et al showed that STIC was found significantly more often simultaneously with SEIC than with benign endometrium at HGSC debulking surgery, hinting at the involvement of the whole Müllerian tract [ 57 ]. To differentiate between tubo‐ovarian origin and uterine origin, WT1 staining on concurrent STIC and SEIC/USC was performed in two studies [ 56 , 58 ]. In these studies, 5/6 cases had a similar staining pattern. Of these, 3/5 showed positive staining, 1/5 showed weakly positive staining, and 1/5 showed negative staining. These ambiguous results did not point toward one (tubo‐ovarian or uterine) origin of these lesions, but rather toward multiple origins. In 2/5 of the cases stained for WT1, a clonal relationship was confirmed with clonality analysis [ 56 ]. In the other cases, clonality analysis failed or was not performed. Lastly, a mouse model with a fallopian tube‐specific promoter showed metastasis toward the endometrium [ 42 ]. Ten studies investigated the ovarian origin of peritoneal HGSC (Table  1 , Figure  2 ‐III; supplementary material, Table  S6 ). Ovarian surface epithelium (OSE) as the tissue of origin was investigated in several studies. Folkins et al assessed the ovaries of 75 BRCA1/2 PV carriers and found one p53 signature on the OSE, which consisted of flat cells without cilia [ 61 ]. Immunohistochemistry to determine the origin of the mutated cells was not performed. Silva et al showed a transition of morphologically normal mesothelial inclusion cysts toward malignant epithelium and eventually to HGSC in 55/172 cases. In five of these cases, immunohistochemistry was performed, showing positive PAX 8 and WT1 staining in morphologically benign and malignant epithelium; however, increased mitotic activity and aberrant p53 staining were only observed in the malignant epithelium [ 64 ]. Two mouse models investigated a possible ovarian origin; Shi et al developed a mouse model with a Müllerian mesenchymal‐specific promoter and showed tumors with both high‐grade and low‐grade morphology. However, developed tumors had negative immunohistochemical staining for PAX8, which is uncommon for HGSC [ 62 ]. Kim et al showed that, after salpingectomy, ovaries alone could develop tumors that had positive staining for the epithelial‐specific marker E‐cadherin in a mouse model. PAX8 staining was not performed [ 63 ]. Some case reports showed close proximity of serous borderline tumors or low‐grade serous carcinomas with HGSC and showed a clonal relationship based on KRAS in two cases. p53 immunohistochemistry was aberrant in 4/7 cases [ 65 ]. However, since analysis of the TP53 gene in two different cases showed no mutations, it is very unlikely that the high‐grade component is HGSC [ 65 , 66 , 67 ]. A precursor similar to STIC is reported in three cases. One of these intraepithelial carcinomas was located on the OSE [ 68 ]. The other two were present in cystadenofibromas [ 69 ]. Banet and Kurman investigated the presence of tubal epithelium in inclusion cysts of autopsy cases with morphology and subsequently immunohistochemistry. Inclusion cysts with tubal epithelium (ciliated cells with PAX8 positivity) only occurred in postmenarchal women and increased with age in contrast to mesothelial inclusion cysts (flat cells with calretinin positivity). This supports the hypothesis that tubal epithelium can only arise in the ovary via metaplasia or implantation, which would be associated with ovulation [ 60 ]. Lastly, three studies hypothesized a dual origin: development from either the FT or the ovary. Two of these studies investigated the expression of genes like PAX8 and calretinin. Based on their observations, they hypothesized that the FT is more often the origin of HGSC [ 70 , 76 ]. The third study evaluated tumor‐forming properties of OSE and FT epithelium with organoids and mouse models. They showed that both cell types could develop into peritoneal tumors, but that the OSE had longer latency and lower penetrance. However, since there was limited PAX8 staining in the tumors that developed from OSE cells, the probability that these tumors are HGSC is low [ 71 ]. Beyond these three pathways, the role of the abdominal environment in HGSC development in mice was investigated in several studies (Table  1 , Figure  2 ‐I; supplementary material, Table  S7 ). Inflammation and adipocytes might have a role in the migration process of HGSC cells. Jia et al showed in a mouse model that wound repair of the ovary increased implantation of cells on tissues surrounding the ovary, but not on the ovary itself [ 73 ]. Morrisson et al showed in a mouse model that treadmill running reduced peritoneal metastasis as this modulated gene expression of intraperitoneal fat tissues. These modulate processes such as lipid formation and inflammation [ 74 ]. Lastly, Nieman et al showed by in vitro co‐culture and a mouse model that omental adipocytes might be able to support migration of tumor cells to the omentum via adipokines, and facilitate cancer growth by providing fatty acids to the cancer cells [ 75 ]. Sampling error at the time of assessing the pathological specimen is investigated in several studies and identified as a confounding factor. Investigation of deeper sections of FT specimens that were initially classified as normal revealed additional detection of p53 signatures, STILs, STICs, and HGSCs [ 52 , 77 , 78 ]. In one study, additional sectioning of the ovary revealed a small HGSC focus in the ovarian tissue and no malignant finding in the FT [ 79 ] (Table  1 ; supplementary material, Table  S8 ).

This systematic review shows several potential pathophysiological pathways and influencing factors involved in the development of peritoneal HGSC after bilateral salpingectomy or salpingo‐oophorectomy. Most evidence was available on mechanisms following an FT origin. Included studies show a clonal relationship between tubal precursor lesions and metachronous peritoneal HGSC. Additionally, precursor cells were shown to be able to migrate and survive in ultra‐low attachment conditions in the abdominal cavity. Furthermore, serous precursor lesions are identified in the endometrium. These precursor lesions had a clonal relationship with extra‐uterine serous carcinoma and showed carcinomatous progression from SEIC based on ploidy. Evidence on the ovaries as the source of peritoneal HGSC was limited. However, the follicular fluid produced by the ovary as well as adipocytes and immune factors seem to play a role in the dissemination process of (pre) malignant cells. The majority of the included studies provided evidence on the FT origin of peritoneal HGSC. This corresponds to the paradigm shift that has occurred regarding the origin of HGSC [ 80 , 81 ]. The identified studies strongly suggest a pathway in which an HGSC precursor originates in FT cells, sheds from the FT, and survives isolated in the abdominal cavity with the help of several substances such as follicular fluid. Subsequently, it could attach to the peritoneum and evolve into peritoneal HGSC. However, it is still unclear at what point the precursor lesion becomes invasive and which precursor lesion(s) eventually develop into peritoneal HGSC. It is clear that women with STIC have an increased risk of developing peritoneal HGSC, but it is unknown whether and to what extent patients with STIL and/or p53 signatures are at risk [ 17 ]. Additionally, it remains unknown whether specific morphologic characteristics are predictive for STIC cells to develop into peritoneal HGSC. Serous carcinoma also occurs in the endometrium as USC, which has SEIC as its designated precursor lesion. Data from the included studies showed clonal relationships between SEIC and HGSC, and increased ploidy confirms the progression from SEIC to HGSC. It has also been shown that STIC coincides with USC/SEIC [ 59 , 82 , 83 ], indicating a connection between these entities. Next to the study by Jarboe et al , a clonal relationship between STIC and USC was also shown by Steenbeek et al [ 56 , 84 ]. This could indicate the involvement of the whole Müllerian tract in the development of peritoneal HGSC in which precursor lesions from both the endometrium and the fallopian tube could lead to (peritoneal) HGSC or USC. There might even be spread of cells from FT to endometrium and vice versa. This raises the question of whether all pelvic serous carcinomas, USC and HGSC, are one entity. Additionally, it sparks the debate of whether a precursor lesion is a sign of a whole tract being affected rather than one isolated abnormality. de Jonge et al showed that women with a BRCA1/2 ‐PV have a relative increased risk for endometrial cancer [ 85 ]. However, evidence on the role of the endometrium in peritoneal HGSC is still scarce. We did not identify large, high‐quality studies providing evidence for an ovarian origin of peritoneal HGSC. Most mouse studies that investigated the OSE confirmed transformation from the mesothelial lining toward carcinoma. However, 2/3 studies used a not fully mesenchymal‐specific promotor, meaning that expression of this gene is also reported in Müllerian epithelium [ 61 , 62 , 63 , 86 ]. Additionally, in two mouse studies these tumors showed lack of or limited PAX 8 staining [ 62 , 71 ]. In the other study PAX8 was not performed [ 41 ]. This is interesting as this stain is typically positive in HGSC [ 87 ]. In the case reports showing low grade serous carcinoma (LGSC) with a concurrent high‐grade component, p53 immunohistochemistry was aberrant in 4/7 cases [ 65 ]. However, TP53 analysis performed in two different cases showed a wild type genotype, which raises the question of whether this is a different and rare high‐grade entity [ 66 ]. Banet and Kurman showed that the presence of tubal epithelium in the ovary increases with age [ 60 ]. This is supported by Park et al , who showed an increase from 62.9% to 80.5% of PAX8 positive inclusion cysts in premenopausal versus postmenopausal women [ 88 ]. This also suggests a hormonal effect, which is supported by a study that showed a lower number of inclusion cysts with tubal epithelium in women who used the contraceptive pill compared to women who did not use the contraceptive pill [ 89 ]. Currently, it is unclear whether the tubal epithelium present in inclusion cysts or on the ovarian surface originates from the fallopian tube or is metaplasia from mesothelial ovarian surface epithelium. Another hypothesis that could give an explanation describes the second Müllerian system, in which the most proximal portions of each Müllerian duct halt in their development and become part of the peritoneum. In their regression, they might leave remnants, such as endosalpingiosis, endometriosis, or endocervicosis [ 90 ]. Theoretically, tubal epithelial cells present in the ovary or peritoneum could undergo the same morphologic changes as epithelial cells located on the distal fimbriae of the FT [ 91 , 92 , 93 ]. However, as the ovary contains significantly fewer tubal epithelial cells than the FT, the likelihood of HGSC originating from tubal cells in the ovary is slim. Studies hypothesizing a dual origin showed a slight possibility for OSE to develop into HGSC; however, HGSCs with a FT origin seem more prevalent. [ 70 , 76 ]. Besides, the similarity of OSE and HGSC could also be caused by epithelial‐mesenchymal transition enabling adherence to the peritoneum [ 71 , 94 ]. In summary, the lack of identified premalignant changes and HGSC precursors in the ovary found over the years raises questions regarding the frequency at which ovarian origin could play a role in the development of peritoneal HGSC. Lastly, factors related to surgery and pathological assessment might confound research regarding the pathophysiology of peritoneal HGSC. One included study showed distal fimbriae adherent to the ovary, which might lead to incomplete removal of the fimbriae. This shows that proper execution of salpingectomy is crucial [ 53 ]. Additionally, the role of sampling error in FT specimens should not be overlooked. Even though fallopian tubes are optimally embedded with the Sectioning and Extensively Examining the Fimbria (SEE‐FIM) protocol, STICs and small carcinomas can be missed as the tissue is only examined at increments of 2–3 mm [ 95 ]. Given the small nature of STIC, this might lead to the interpretation that a carcinoma occurred after salpingectomy, whilst it might have been present already at the time of salpingectomy. This review offers a broad perspective on peritoneal HGSC pathophysiology after salpingectomy or salpingo‐oophorectomy, showing multiple possible pathways that can complement each other. A broad variety of studies were included: clinical studies, mouse models, and in vitro research, which led to a comprehensive overview. Due to the large heterogeneity in the design of the included studies, it was not possible to pool results. Furthermore, because STIC is a relatively recently identified lesion, there was variation in the use of the SEE‐FIM protocol and immunohistochemistry for STIC diagnosis across studies. This was taken into account in the quality assessment of the studies (supplementary material, Table  S2 ). Only a limited number of available studies investigated the relation of precursor lesions with subsequently developed HGSC, which impacted the applicability to our research question. The results of our study highlight the challenges still faced in unravelling the pathogenesis of peritoneal HGSC after salpingectomy or salpingo‐oophorectomy. The ongoing clinical studies on the safety of RRS with DO and the efficacy of opportunistic salpingectomy will shed more light on the extent of the role of the fallopian tube and ovary. As the ESGO‐ESMO‐ESP guideline now advises hysterectomy or endometrial biopsy when STIC is found, more pathological data on the endometrium of patients at risk for peritoneal HGSC will become available [ 26 ]. Because a peritoneal HGSC diagnosis is rare, it is important to gather and combine as much pathological and clinical data, preferably by international collaboration.

This review shows that there are several mechanisms that may be responsible for the development of peritoneal HGSC after bilateral salpingectomy or salpingo‐oophorectomy. Most available scientific evidence supports dissemination of (precursor) cells originating from the fallopian tube. There seems to be a role for the endometrium but evidence regarding this topic is currently scarce and therefore more research into this pathway is necessary. The ovary as a site of origin is less likely; however the role of this organ cannot be marginalized as follicular fluid might promote the dissemination process. Besides, after salpingectomy residual tubal tissue might be present on the ovary and indirectly cause peritoneal HGSC. Therefore, currently, oophorectomy remains a part of HGSC prevention in women with a hereditary increased risk for HGSC.

Ovarian carcinoma (OC) is the most lethal type of gynecological cancer. Even after intensive treatment with cytoreductive surgery and chemotherapy, prognosis is poor [ 1 , 2 ]. This is due to the late onset of symptoms and the lack of effective screening methods [ 3 , 4 ]. The most common type of OC is high‐grade serous carcinoma (HGSC), which occurs most often in the fallopian tubes, ovaries, and peritoneum. When HGSC occurs on the peritoneum, it will be referred to as peritoneal HGSC [ 5 ]. Serous carcinoma also occurs in the endometrium where it is called uterine serous carcinoma (USC) [ 6 ]. Initially, the ovary was regarded as the primary source of HGSC. However, the discovery of a spectrum of fallopian tube (FT) abnormalities in the fimbriated end has caused a paradigm shift toward the FT epithelium as tissue of origin [ 7 , 8 , 9 ]. This spectrum consists of normal epithelium, p53‐signature, serous tubal intra‐epithelial lesion (STIL), serous tubal intraepithelial carcinoma (STIC), and invasive carcinoma. STIC is regarded as the precursor of HGSC [ 10 ]. STIC is most often found in the distal, fimbriated end of the FT, has a diameter varying from a few cells up to several millimeters, and is morphologically similar to HGSC, apart from stromal invasion which is absent in STIC [ 11 ]. Criteria for STIC diagnosis are poorly reproducible, leaving mainly consensus‐based recommendations [ 12 ]. STIC is found synchronous with invasive carcinomas in at least 11–61% of cases, but is also found isolated in salpingectomy specimens removed for other indications [ 13 ]. As women with a BRCA1/2 pathogenic variant (PV) have an increased lifetime risk to develop HGSC, they are advised to undergo a risk‐reducing salpingo‐oophorectomy (RRSO) around the age of 40 [ 14 ]. The peritoneum stays in situ as removal is not feasible. RRSO reduces HGSC risk by 96% when performed within the current guideline age range [ 15 , 16 ]. Steenbeek et al showed that, if STIC is detected at RRSO, the 10‐year risk for peritoneal HGSC is 27.5%, compared to 0.9% in FTs without detected STIC [ 17 ]. The incidence of isolated STIC in women carrying a BRCA1/2 PV is 3–4%, compared to <0.01% of women in the general population having salpingectomy for other indications [ 12 , 18 ]. Due to the growing evidence on the FT being the site of origin for HGSC and the negative consequences of premature menopause caused by RRSO, the approach for OC prevention has changed. Consequently, three large clinical trials are investigating the safety of a new preventive strategy: risk‐reducing salpingectomy (RRS) with delayed oophorectomy (DO) ( NCT04294927 , NCT04251052 , and ISRCTN25173360) [ 19 , 20 , 21 ]. Salpingectomy in high‐risk women should only be considered within a clinical trial and in combination with DO [ 22 ]. This innovative strategy seems promising to reduce HGSC risk, but prospective data need to confirm this. Additionally, the feasibility and efficacy of opportunistic salpingectomy, an elective bilateral salpingectomy during abdominal surgery for a different indication than OC prevention, is being investigated in women at average risk [ 23 , 24 , 25 ]. Guidelines for follow‐up are limited when STIC is found in a salpingectomy specimen. Execution of oophorectomy is common, and the European Society of Gynaecological Oncology—European Society for Medical Oncology—European Society of Pathology (ESGO‐ESMO‐ESP) guideline advises to consider staging surgery of the peritoneum if STIC is found, based on expert opinion. Lymphadenectomy or adjuvant chemotherapy is not recommended, as there is no data on effectiveness, and these treatments come with high morbidity [ 26 ]. Currently, follow‐up of women with STIC is not performed consistently, and professionals indicate a need for guidelines regarding additional diagnostics, staging, treatment, and follow‐up [ 27 ]. Understanding the currently unknown pathophysiological mechanisms of peritoneal HGSC development after bilateral salpingectomy or salpingo‐oophorectomy is the first step. Potential hypotheses include misdiagnosis at primary histopathologic assessment, dissemination of tubal lesions, or other sites of origin than the FT. Therefore, the primary aim of this review is to inventory and discuss hypotheses on potential pathophysiologic pathways that can explain the pathogenesis of peritoneal HGSC after salpingectomy or salpingo‐oophorectomy. The secondary aim is to identify gaps in the literature in order to give direction to further research.

This review was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic reviews and Meta‐Analyses) 2020 statement [ 28 ]. After consultation with a librarian, a search strategy was built to search EMBASE and PubMed (supplementary material, Table  S1 ). No specific publication time frame was required. Additional relevant studies were manually identified by recursively including studies referenced in previously identified studies and by setting up an email alert based on our search strategy. We included all papers that focused on or addressed potential pathophysiologic pathways that can explain the development of peritoneal HGSC after a previous salpingectomy or salpingo‐oophorectomy. Articles were excluded when they were not available in English, there was no availability of a full text version, or the format was not an original research paper (e.g., review, abstract, commentary). Identified studies were uploaded into Covidence and duplicates were removed [ 29 ]. Studies were individually assessed by two reviewers (TAG and ABB), first by evaluating the title and abstract and next by assessing the full article text. Reasons for exclusion included: (1) unsuited language – no English text available; (2) unsuited setting – the study did address the desired outcome, but in a context too distant from our research question (e.g., therapeutic options); (3) unsuited outcomes – the investigated outcome did not align with our aims; (4) unsuited study design – the study design prevented answering our research question; (5) unsuited indication – the subject of the screened article was not relevant (e.g., endometriosis); and (6) unsuited intervention – the performed interventions or experiments did not answer our research question (Figure  1 ). Inconsistencies between the reviewers were resolved by discussion or consulting a third researcher (MPS). PRISMA flowchart of included studies. Created with covidence.org. Two reviewers (TAG and ABB) independently extracted the following information from each study: title, year of publication, type of study, supporting hypothesis, study design, country, and arguments in favor and/or against the discussed hypothesis. Discrepancies were discussed until consensus was reached or by consulting a third researcher (MPS). In one case, the author of the included article was approached for clarifications. After extraction, articles were grouped by supporting hypotheses and subareas. To assess the quality, we used a combination of items from the study quality assessment tools on observational cohort and cross‐sectional studies, case–control studies, pre‐post studies, and case series studies of the National Heart, Lung and Blood Institute (NHLBI) [ 30 ]. Our customized tool consisted of nine questions relevant to the range of selected studies (supplementary material, Table  S2 ). Two reviewers (TAG and ABB) assessed all included studies. Discrepancies were discussed until consensus was reached or by consulting a third researcher (MPS). Quality was defined as ‘low’, ‘moderate’, or ‘high’. This review article does not involve any original research with human participants or animals conducted by the authors. As it is based on previously published literature, ethics approval was not required. No patient‐identifiable data are included in this manuscript; therefore, informed consent was not necessary.

Table S1. Search strategy in PubMed and EMBASE Table S2. Customized quality assessment tool Table S3. Quality assessment Table S4. Mechanisms following FT origin Table S5. Endometrial involvement Table S6. Ovarian origin Table S7. Micro‐environment Table S8. Sampling error