Prognostic features of endometrial cancer metastasis to the central nervous system.

OBJECTIVES: Central nervous system metastases (CNSm) secondary to endometrial cancer (EC) are rare. As a result, prognostic factors for this patient population are not well described.
METHODS: EC patients with CNSm were identified retrospectively from two academic centers. EC patients without CNSm (non-CNSm) were used as controls. Chi-square and Fisher's exact tests were used for analysis of categorial variables. Wilcoxon tests were used for quantitative measures. Overall survival (OS) was compared with Log-rank test. Cox proportional hazard models were used to estimate hazard ratios for OS.
RESULTS: 22 EC patients with CNSm and 354 non-CNSm patients were included. Compared to non-CNSm EC, the CNSm cohort was younger (58.5 vs 62.0 years, p = 0.018) with lower BMI (27.7 vs. 33.7 kg/m2, p = 0.005), and had more advanced stages (p = ≤ 0.001), grade 3 tumors (81.8% CNSm vs 25.1% non CNSm, p≤0.001) and serous histology (22.7% vs 8.5%, p = 0.010). Median survival after CNSm diagnosis was 9 months (95% CI 4, NA). CNSm was a strong poor prognostic factor (HR death 4.96, p = 0.022). Improved OS was seen with CNS as the only disease site (83m CNSm only vs 30m additional sites, p = 0.007) and less than five CNSm (49m <5 vs. 23m ≥5, p = 0.004). Surgical resection of CNSm (OS 83m surgery vs 33m no surgery, p = 0.003) or multimodal therapy (83m multimodal vs 33m single therapy, p = 0.027) resulted in longer OS.
CONCLUSIONS: CNSm is a poor prognostic factor in EC, however, low volume disease with aggressive treatment may result in more favorable survival outcomes.

Introduction

Endometrial cancer (EC) is the most common gynecologic cancer with 65,000 new cases estimated in 2020 [1]. EC has traditionally been divided into Type 1 and Type 2 tumors, which is based on histology and correlates with clinical outcomes [2]. Type 1 tumors are comprised of endometrioid histology, usually of low to moderate grade, and tend to occur in the setting of obesity and advanced age with a generally favorable prognosis [3]. Alternatively, type 2 tumors include high grade histologies such as grade 3 endometrioid adenocarcinoma, serous carcinoma, clear cell carcinoma and carcinosarcoma which are more prone to metastasis and recurrence, and therefore considered high risk [2, 4]. This model has proven to be imperfect with certain type 1 tumors, not infrequently, exhibiting aggressive clinical features [2].

Spread of EC from the primary site can occur by way of lymphatics to locoregional lymph nodes, local invasion to surrounding tissues, or hematogenous spread to distant organs [5]. Distant metastasis is an uncommon occurrence, present approximately in only 15% of patients at the time of initial diagnosis [1]. Central nervous system metastasis (CNSm) from EC represents a very small subset of distant metastasis with reported incidence between 0.3 to 1.2% [6]. As expected, CNSm from EC has previously been associated with higher tumor grade, advanced stage and high-risk histologic types [7-9]. Interestingly, CNSm from gynecologic cancers appears to be increasing over time, although, it is thought to be secondary to prolonged survival and advancements in imaging techniques [10]. However, due to the rarity of CNSm from EC, there remains limited data to inform patient counseling and appropriate treatment modalities.

In the present retrospective cohort study, conducted at two major NCI designated cancer centers, we sought to describe characteristics and survival outcomes of patients with CNSm from EC. Additionally, we aimed to evaluate potential prognostic factors associated with improved outcomes following CNSm diagnosis. Due to the rare nature of this presentation, we further aimed to complete a review of the literature and provide additional valuable insight on its outcomes.

Materials and methods

This retrospective cohort study includes EC patients with CNSm from H. Lee Moffitt Cancer Center (MCC) and University of Southern California (USC). Institutions were chosen due to preexisting collaborative relationships and were the main practicing sites of involved authors. Institutional Review Board approval was obtained at both institutions which each waived the requirement for informed consent given the retrospective nature of this study. Electronic medical record systems then were queried at both institutions to identify cases of EC with CNSm diagnosed between January 1,1986 (earliest date of scanned records) through June 1, 2016 (initial start date of data collection). Patients with a diagnosis code of EC and billing code for CNS imaging comprised the screening pool. Inclusion criteria required a CNSm to be identified in an official imaging report with or without a confirmatory biopsy. Only carcinomas arising from the endometrium were included in the present study, which was limited to endometrioid, serous, clear cell or carcinosarcoma. Uterine sarcomas were excluded. Patients were excluded if they had a separate non-EC cancer diagnosis, unless the CNSm was biopsy proven EC primary.

Data regarding demographics, clinical characteristics, pathology, treatment data, and survival outcomes of eligible patients were extracted from the medical record. Overall survival (OS) was calculated from date of primary EC diagnosis to date of death or last follow up. Survival after CNSm was calculated from date of CNSm diagnosis to date of death or last follow up. Data from a preexisting database of all EC patients treated at the primary site (MCC) from 2014–2017 was used to create a comparison cohort (non-CNSm). This data set included all stages of EC, all EC histologies and were included regardless of recurrence status. A limited data set was collected for these patients including demographic information, primary disease characteristics, treatment modality and survival outcomes.

Chi-square or Fisher’s exact tests were used when comparing categorial clinical characteristics between EC CNSm cohort and EC non-CNSm cohort or site A and B within CNSm cohort, while Wilcoxon tests were used for comparison of quantitative measures. For analysis within CNSm group only, the associations between categorical variables and interested endpoints were evaluated using Chi-squared tests or Fisher’s exact tests when the expected frequencies were low. Categorical variable levels for OS were compared using the Log-rank test, and the corresponding Kaplan-Meier curves were generated. For these analyses within the CNSm cohort we did not adjust for covariates due to the limited sample size. For analysis between CNSm and non-CNSm cohorts, Cox proportional hazards regression model was used to estimate hazard ratios (95% confidence interval) for OS with adjustment for clinical covariates. For the analysis using the larger non-CNSm EC cohort, which involved a larger sample size, we were able to fit multivariable models adjusting for stage, histology and age.

A review of the literature on the present topic was performed using Pubmed with search terms “endometrial cancer” or “uterine cancer” and “brain metastasis” or “central nervous system metastasis.” Included articles were limited to studies reporting original data describing outcomes specific to patients with CNSm from endometrial cancer. Systemic reviews and articles reporting on multiple cancer types without delineating data specific to uterine cancer patients were excluded. Reference lists from publications of interest were also used as a search method to identify all relevant articles.

Results

Patient characteristics

A total of 22 patients with CNSm from EC were identified (MCC Site A n = 11, USC Site B n = 11). Institutional cohorts were comparable for all clinical and pathologic characteristics aside from race where the USC cohort represented the majority of non-white patients (). For all patients, initial treatment for primary EC was surgery (n = 12, 54.5%), chemotherapy (n = 8, 36.4%), or unknown (n = 2, 9.1%). The majority of patients had advanced stage disease (Stage III/IV: n = 20, 90.9%) and high-grade tumors (grade 3: n = 18, 81.8%). Additionally, 27.2% represented high risk histology (serous n = 5, carcinosarcoma n = 1).

In comparison with the non-CNSm EC cohort, the CNSm cohort was found to be significantly younger at time of initial EC diagnosis (58.5 vs 62.0 years, p = 0.018), had lower BMI (27.7 vs. 33.7 p = 0.005) and were more likely to be non-white (p < 0.0001). Advanced stage disease (stage III/IV 86.3% vs 16.4%, p ≤ 0.001), high-grade tumors (81.8% vs 25.1%, p ≤ 0.001) and serous histology (22.7% vs 8.5%, p = 0.01) were all more prevalent in the CNSm cohort (). An additional comparison was completed between the CNSm cohort and patients within the non-CNSm cohort who developed recurrences (n = 30), in attempt to capture the more aggressive cases. In this comparison, the CNSm cohort and the recurrent non-CNSm cohort had similar rates of serous histology (22.7% CNSm vs 23.3% recurrent non-CNSm, p = 0.205) and grade 3 tumors (81.8% CNSm vs 63.3% recurrent non-CNSm, p = 0.370). However, the CNSm cohort continued to have younger age at diagnosis (58.5 years CNSm vs 62.5 years recurrent non-CNSm) and higher rates of stage 3 or 4 disease at time of diagnosis (86.4% CNSm vs 56.6% recurrent non-CNSm, p = 0.029).

CNS metastasis

In EC patients with CNSm, the median interval between primary EC treatment completion and CNSm diagnosis was 17.5 months (IQR 1.8–26). Routine CNS imaging to screen for CNSm during treatment or surveillance of EC was not a regular practice at either institution in the absence of symptoms. The majority of patients (82%) were diagnosed following new onset of CNS symptoms (). Diagnosis was made by magnetic resonance imaging (MRI) in 45.4%, computerized tomography scan (CT) in 27.3%, and both MRI and CT in 9.1%. The number of CNSm varied with solitary metastasis identified in 10 patients (45.5%), 2–5 metastases in 4 patients (18.2%), and diffuse metastases (>15 metastases) in 6 patients (27.3%). Location of CNSm included cerebrum alone (31.8%), cerebellum alone (31.8%), both cerebrum and cerebellum (27.3%), and meningeal involvement (4.5%). Diameter of the largest CNSm per patient varied including <1 cm (n = 1, 4.6%), 1–3 cm (n = 7,31.8%), and 3–5 cm (n = 10, 45.5%), data on largest diameter CNSm was missing for four patients.

Treatment of CNS metastasis

Treatment modalities for CNSm are displayed in which included whole brain radiation (WBRT) alone (n = 7, 31.8%), surgical resection followed by localized radiation (n = 5, 22.7%), stereotactic radiosurgery alone (n = 3, 13.6%), surgical resection alone (n = 1, 4.6%), surgical resection followed by WBRT (n = 1, 4.6%), and surgical resection followed by placement of chemotherapy wafers (n = 2, 9.1%). Three patients (13.6%) received no therapy for CNSm. Six patients (27.3%) who received treatment for CNSm developed recurrence/progression in the CNS following therapy. Of these six patients, three underwent stereotactic radiosurgery alone, two underwent surgical resection followed by radiation to surgical bed, and one underwent surgical resection followed by placement of chemotherapy wafer.

Treatment modality was found to be associated with number of CNSm. Surgical resection occurred more often in patients with solitary brain metastasis versus multiple metastases (77.78% vs 20%, p = 0.023). WBRT was administered more often in women with greater than five CNSm (100% ≥5 vs 16.7% <5, p = 0.0015). Additionally, multimodal therapy was used more frequently in women with solitary brain metastases compared to multiple (70% vs 8.3%, p = 0.005).

Survival

OS varied widely in the CNSm EC cohort with a median of 49 months (95% CI: 30, NA). Median survival after CNSm diagnosis was 9 months (95% CI 4, NA). Five patients (22.7%) survived longer than 1 year after CNSm diagnosis (72, 14, 18, 41, 37 months).

We identified improved OS when CNS was the only site of metastasis compared to those with additional non- CNS metastasis (83m CNSm only vs 30m additional sites, p = 0.007) (Table 3). Those with CNSm as the only site of disease also had improved survival following CNSm diagnosis compared to those with concurrent disease outside CNS (p = 0.0096) (). Out of seven patients with CNSm and no evidence of disease outside of the CNS, only one died during the follow-up period. Those with solitary CNSm had improved OS compared to those with multiple CNSm (83m solitary vs 28m multiple, p = 0.04) (). Patients with solitary CNSm also had improved survival following CNSm diagnosis (p = 0.0034) (). Overall, a small number of CNSm was associated with improved OS. OS for patients with less than five CNSm was more favorable compared to those with five or greater CNSm (49m <5 vs. 23m ≥ 5, p = 0.0045) (). Similar survival benefit was found for survival following CNSm diagnosis (). No significantly associated survival benefit was found with increasing diameter of CNSm (6.5months <3cm vs 14months >3cm, p = 0.327).

Table 3

Differences in overall survival based on number of CNSm and disease burden.

Comparison Groups	Median Overall Survival	P Value
CNSm only site of metastasis	83 months	0.007
CNSm + other sites of metastasis	30 months
Solitary CNSm	83 months	0.040
Multiple CNSm	28 months
<5 CNSm	49 months	0.005
≥5 CNSm	23 months

Abbreviations: CNSm: central nervous system metastasis, +: in addition to, <: less than, ≥:greater than or equal to

Survival based on central nervous system metastasis characteristics.

Survival time displayed in number of months. (1.1) Survival following CNS metastasis with CNS metastasis as only site of recurrence versus those with multiple sites of disease. (1.2) Survival following CNS metastasis diagnosis in patients with solitary CNS metastasis compared to those with multiple CNS metastases. (1.3) Survival following CNS metastasis diagnosis in patients with less than five CNS metastases compared to those with five or more CNS metastases.

Compared to the non-CNSm EC cohort, CNSm patients had a significantly higher risk of death (HR 11.3, 95% CI 3.4–36.9, p < 0.001). This remained significant after controlling for stage (HR 4.53, 95% CI1.18, 17.32, p = 0.027), histology (HR 13.51, 95% CI 3.88, 47.04, p = <0.001) or age (HR 10.4, 95% CI 3.18–34.01, p = <0.001). Controlling for all three variables simultaneously resulted in persistent significantly worse prognosis for CNSm (HR death 4.96, p = 0.022). CNSm cohort OS was also compared specifically to patients within the non-CNSm EC cohort who recurred. Although the CNSm cohort trended towards higher risk of death, this was not significant compared to the recurrent patients from non-CNSm cohort (HR 2.28, 95% CI 0.578–8.974, p = 0.239).

An analysis comparing survival outcomes based on treatment approach was completed. Three patients received no treatment for their CNSm, all of whom survived ≤1 month following CNSm diagnosis. Overall survival from EC diagnosis was 29 months, 5 months and 1 month for these three patients. With these three patients excluded, patients who underwent surgical resection of CNSm had longer median OS (83m vs 33m, p = 0.003) and longer median survival after CNSm diagnosis (NR vs. 4m, p = 0.002) compared to those who did not undergo surgical resection of CNSm (). Those who received multimodal treatment for CNSm with any combination of modalities had improved median OS (83m vs 33m, p = 0.027) and improved survival after CNSm diagnosis (NR vs 4m single modality, p = 0.002) (). Additionally, those who received WBRT for treatment of CNSm had decreased OS (23m vs 83m, p = 0.014) and decreased survival following CNSm diagnosis (4m vs 14m, p = 0.022) compared to those who did not receive WBRT ().

Survival after central nervous system metastasis based on treatment approach.

Survival after central nervous system metastasis based on treatment approach in endometrial cancer patients with central nervous system metastasis. Survival time displayed in number of months. (2.1) Survival following CNS metastasis diagnosis for patients undergoing surgical resection of CNSm versus those who did not undergo surgery. (2.2) Survival following CNS metastasis diagnosis for patients undergoing multimodal therapy for treatment of CNSm compared to those who underwent single modality therapy. (2.3) Survival following CNS metastasis diagnosis for patients undergoing whole brain radiation therapy (WBRT) compared to those who did not undergo WBRT.

Discussion

Endometrial cancer is described as a “neurophobic disease” that is rarely found to metastasize to the CNS [11]. Neuroimaging is not a routine part of the initial work-up unless there are concerning symptoms. Given the paucity of CNSm cases in EC, it is difficult to identify those at risk, and a diagnosis of CNSm may be delayed due to low suspicion. Additionally, following diagnosis of CNSm it is challenging to appropriately counsel patients on oncologic expectations because prognostic features specific to endometrial cancer are not well-reported. In the present retrospective study, we aimed to describe this unique patient population and evaluate for features associated with development of CNSm and survival thereafter. Multiple risk factors were identified to be significantly more prevalent in EC patients with CNSm compared to those without CNSm including lower BMI, non-white race, advanced stage disease, high-grade tumors, and high-risk histology. Despite the poor prognosis identified in patients with CNSm from EC, the present study found that patients with low disease burden and those who received multimodal therapy had improved survival.

A review of the literature was also completed to correlate our findings compared to those previously described in patients with CNSm from uterine cancers, results summarized in Table 4 [7–10, 12–22]. Fifteen publications, in addition to the present study, were identified describing 274 patients in total, with 240 representing endometrial cancers. Consistent with our results, the majority were advanced stage (Stage III/IV: 189/260, 72.7%) and poorly differentiated tumors (Grade 3: 101/156, 64.7%) with a large representation of high-risk histology endometrial subtypes (56/240, 23.3%) ().

Table 4

Summary of studies on central nervous system metastasis in endometrial cancer.

Author, Year	No. Patients	Histology	Grade	Stage	CNSm Treatment	mOS CNSm
Kuznicki et al, 2021 (Current Study)	22	16 Endometrioid5 Serous1 Carcinosarcoma	1 : 22 : 23 : 18	I : 1II : 1III : 10IV : 10	Surgery: 1WBRT: 7Radiosurgery: 3Surgery + localized RT: 5Surgery + WBRT: 1Surgery + chemo wafers: 2	4m (1–72)
Bhambhvani et al, 2020 [12]	30	16 Endometrioid7 Serous5 Carcinosarcoma1 Glassy Cell1 Clear Cell	1 : 12 : 33 : 24	I : 6II : 1III : 11IV : 9	Surgery : 11Radiosurgery : 28	6.8m (1–58.2)
Ogino et al, 2020 [13]	12	*	*	*	Radiosurgery +/- additional treatment: 12	4.5m (1–11)
Zhang et al, 2019 [9]	24	11  11 Endometrioid4 Serous3 Carcinosarcoma1 Adenosquamous1 Clear Cell1. Leiomyosarcoma1 Pleiomorphic sarcoma1 Small cell neuroendocrine1 Mixed adenocarcinoma	1: 22: 23: 20	I: 3II: 2III: 12IV : 7	*	*
Moroney et al, 2019 [14]	12	9 Endometrioid2 Serous1 Adenosquamous	1 : 12 : 33 : 8	I : 4II : 1III : 3IV : 4	RT : 4Surgery+ RT : 2RT + chemo : 2Surgery + RT + chemo : 2None: 2	*
Cybulska et al, 2018 [7]	23	23 Endometrioid	1 : 92 : 143: 0	I : 15II : 2III : 3IV : 3	RT : 7Surgery + chemo : 2Surgery + chemo + RT: 4Surgery + RT : 3RT + chemo : 1None : 6	5.1m (2.2–8.1)
Uccella et al, 2016 [21]	18	12 Endometrioid3 Serous1 Adenosquamous2 Undifferentiated	1 : 22 : 23 : 14	I : 6III : 7IV : 5	WBRT : 5Radiosurgery : 1Surgery + WBRT : 8WBRT + chemo : 1None : 3	6.5m (0–118)
Shin et al, 2016 [15]	6	5 Endometrioid1 Small cell Carcinoma	*	*	Radiosurgery +/- additional therapy: 6	7.5m
Gressel et al, 2015 [10]	22	12  13 Endometrioid13  4 Serous14  2 Carcinosarcoma1 Squamous cell15  1 Adenosquamous16  1 Leiomyosarcoma	*	I: 1II: 2III: 6IV: 12	Surgery : 2RT : 15Surgery + RT: 2None: 4	4m (0–123)
Kim et al, 2015 [16]	19	17  6 Carcinoma18  11 Adenocarcinoma19  2 Leiomyosarcoma	*	II: 3III: 8IV: 8	Surgery: 9RT: 14Chemo: 9None: 2	23.3m (17.8–28.8)
Shepard et al, 2014 [20]	6	20  *	*	I: 2III: 1IV: 2	Radiosurgery +/- additional therapy: 6	8.3m (4–16)
Chura, 2007 [17]	20	11 Endometrioid1 Serous2 Adenosquamous3 Carcinosarcoma3 Undifferentiated	1 : 32 : 63: 11	I : 3III : 8IV : 9	WBRT: 7WBRT + Chemo: 4WBRT + Radiosurgery : 1Surgery + WBRT : 1Surgery + WBRT + chemo : 3None : 4	2.0m (0.1–39.2)
Orru, 2007 [19]	3	1 Endometrioid2 Adenocarcinoma, other	1: 02: 03: 3	III : 3	WBRT: 1Surgery + WBRT : 2	*
Mahmoud-Ahmed, 2001 [22]	10	7 Adenocarcinoma3 Adenosquamous	*	II : 1III : 4IV : 4	WBRT: 4Surgery : 2Radiosurgery + WBRT: 1Surgery + WBRT: 2Surgery + radiosurgery + WBRT: 1	3.3m
Totals	274	147 Endometrioid30 Serous21 Adenocarcinoma NOS18 Carcinosarcoma11 Adenosquamous6 Carcinoma NOS4 Leiomyosarcoma3 Clear cell3 Undifferentiated carcinoma NOS2 Small cell neuroendocrine1 Squamous1 Glassy Cell1 Pleiomorphic sarcoma1 Undifferentiated sarcoma	1: 202: 353: 101	I: 54II: 17III: 102IV: 87

Abbreviations: No. number, CNSm central nervous system metastasis, mOS CNSm median overall survival following CNS metastasis diagnosis, RT radiation, chemo chemotherapy, WBRT whole brain radiation therapy

*missing data

mOS : mean/median (range) as reported in the primary reference.

Regarding initial diagnosis of CNSm, prior studies have reported 89–100% of CNS metastasis diagnoses were made following work-up for new neurologic complaints [7, 10, 21]. New neurologic symptoms were present in 83% of patients in the present study at the time of CNSm diagnosis with the most common symptoms being headache and dizziness. Early diagnosis of CNSm following prompt investigation of new neurologic symptoms would potentially allow for multimodal intervention at a time of lowest volume CNSm [15, 22, 23]. Timing of CNSm diagnosis appears to correlate with timing of disease recurrence with all but four patients presenting with CNSm within three years of original EC diagnosis. Additionally, 36% of patients in the present study were diagnosed with CNSm at the time of their first recurrence. This suggests CNSm can occur early in the course of disease recurrence and CNS imaging should be completed in the setting of new neurologic symptoms regardless of timing from EC diagnosis. Identifying CNSm as early as possible is clinically ideal in order to expedite treatment and avoid potentially catastrophic consequences of untreated CNSm including hemorrhage or herniation. Interestingly, only one patient in the present study was identified as having leptomeningeal metastasis in the absence of a parenchymal lesion. If neurologic symptoms persist in the absence of gross parenchymal metastasis on brain imaging, a consideration may be made for addition of spinal imaging if this has not been completed.

Reported survival following CNSm from EC is poor, consistent with our finding of a median survival after CNSm of 3.5 months (). We identified CNSm in EC as an independent poor prognostic factor (HR 0.09, p = <0.001) which remained significant after controlling for age, stage, and histology compared to non-CNSm EC patients. Although median survival is not encouraging, there was a wide variability in survival following CNSm diagnosis (range 1–72 months), and there seem to be prognostic features associated with improved outcomes. Additionally, compared to recurrent non-CNSm EC patients, we were unable to demonstrate a significant difference in survival for CNSm patients (HR 2.28, 95% CI 0.578–8.974, p = 0.239). Volume of disease appears to be a driving factor for prognosis after CNSm diagnosis along with aggressive treatment [21]. We demonstrate improved survival following CNSm diagnosis in the setting of solitary CNSm (median 12.5 months), less than five CNSm (median 6.5 months), absence of extracranial disease (median 14 months), surgical resection of CNSm (median NR) and multimodal therapy for CNSm (median NR). Other studies also support that good performance status at the time of CNSm treatment is significantly associated with improved survival [13, 18]. Given these data, prompt multi-modality treatment approach with neurosurgical and radiation oncology input in the appropriate patient may allow for optimized outcomes following CNSm diagnosis. Overall, this information is valuable to guide patient selection for treatment as well as counseling on expectations and prognosis following CNSm diagnosis.

Although there have been consistent reports that any treatment for CNSm is superior to no treatment, it remains in question whether treatment type affects outcomes independent of disease burden [7]. A survival advantage has been previously reported for patients with CNSm from EC receiving a combination of surgery plus radiotherapy compared to those who did not receive this type of multimodal treatment [21] Unfortunately, studies evaluating treatment type and prognosis have had difficulty discerning whether the aggressive treatment was causally related to improved survival or if the ability to resect CNSm was due to an intrinsic indolence of disease (i.e. small volume) leading to improved outcomes. Due to the close relationship between disease volume and treatment of choice, the present study is met with similar limitations in determining causality of improved outcomes when considering these two variables.

Additional limitations met by this study include a small CNSm cohort and retrospective nature of the study. Although we included patients from two large NCI cancer centers, the total number of patients remains relatively low and as such, the ability for multivariable analysis was limited. We also note that the non-CNSm comparison cohort is quite diverse in that it includes all patients treated for EC over the stated time period. We attempted at add a clinically relevant comparison by selecting out those patients in the non-CNSm cohort who had recurred with similar survival outcomes seen, however this resulted in a significant reduction in number of patients available for comparison. Nonetheless, this study represents one of the largest series in the literature describing CNSm from EC. The patient population was carefully selected to exclude uterine sarcomas to describe outcomes specific to EC more accurately. Population diversity was maintained by including patients from two academic cancer centers with minimal missing data points.

Our data demonstrates that low number of CNSm and multimodal treatment approach most closely predict improved outcomes after CNSm diagnosis. Prompt work-up of new neurologic findings followed by aggressive treatment may be a reasonable treatment option for those with low volume CNSm or minimal extra-CNS disease burden. Given the poor outcomes otherwise, palliative treatment alone and/or hospice care may be considered for EC patients with CNSm who lack the favorable prognostic features identified in this study. Additional molecular profiling of these tumors may allow for elucidation of a mechanism behind the propensity for CNS metastasis in order to aid in early diagnosis and potentially prevention in patients at risk.

Available data for central nervous system cohort.

(XLSX)

Click here for additional data file.

Available data for endometrial cancer without central nervous system cohort.

(XLSX)

Click here for additional data file.

23 Mar 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Manish S. Patankar, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

Additional Editor Comments:

The manuscript has been reviewed by two reviewers. We encourage you to consider the critique carefully and make the necessary changes to the manuscript.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Introduction

Line73 : why those two centers ? And how did you decide on those ?

Line74 : can the authors be more specific about the type of outcome they are setting to describe ?

Material and method

Line80-85: few questions :

1- why the years from 1986-2016 for the recurrence ?

2- are those years symmetrical in both institutions ?

3-how did you identify the CNS recurrence in EMR ? Is it through billing of treatments given to those CNS recurrence?

4- how did you match those CNS recurrences with the EC diagnosis? Is it by ICD ?

Line90-93: regarding the control group :

1- why only one center ?

2- why only the years 2014-2017?

3- more importantly, are those control group include patients who have had primary EC without recurrences ? Or is it all EC with recurrences but not CNS met ? Or a mixture of both ( I mean , those with recurrences mixed with those without) ?

Line96: you mentioned that because of he small number of patients only univariate analysis are done , yet in line 102 you mentioned that OS with adjustments for clinical variables . Can you please clarify this discrepancy ?

Line 103-109: are the authors trying to do literature review also ? If so can they put the literature review an an objective ?

Results

Table -1

Looking at the initial stage , we find that majority of cancers are initially are stage 3-4 , which is contrary to the national data ? Can the authors describe this discrepancy or phenomenon ?

If those patients ( who has CNS met ) are initially found to be advanced stage , then a question can be asked if those patients have already had CNS met and it just became a matter of time to surface clinically ?

This can Cleary be seen in table -2 where the discrepancy of the initial stage is clear .

Which brings us to the initial question posted in the material and method about the matter of the control group . It seams that the control group are all EV patients . Which by itself create a huge imbalance between those two groups . And advise is to select the control group to be those EC who had recurrences without CNS met .

Line 131:

The authors mentioned that the majority of patients had syndromes of CNS recurrences, so does that mean that patients are screened for CNS recurrences?

Line 158: the OS is very short for CNS 9 m compared to 49 for EC. Again this is a major discrepancy between the two groups . I would advise the authors to make the comparison to EC who had recurrences without CNS met

Line 161-172: this is good data. I would advise the authors to put this data in a table format for the reader to see the number of patients ..etc

Line 180-181 : the authors here states that HR of death is higher in CNS met : two questions :

1- in retrospective studies , traditionally we analyze the Odd ratio of death ( as Hazard ratio is soared to prospective data) , can the authors deliberately explain please ?

2- the authors mentioned that those HR is after controlling for factor , but the number of patients are too small and the authors said in their material and method that the small number will not allow them to control and be only doing univariate analysis, can the authors please explain this ?

Discussion

Line 262 : the authors mentioned here that they excluded patients with sarcoma, but in table 1 they have sarcoma as a histology . That patient with sarcoma need to be excluded from analysis .

Line 265-266: how is your data super early work up of neurological findings , did the authors study those who did not have an early work up ?

I advise the authors to explore the limitations and the shortcomings of their study and not only discuss its strength

Reviewer #2: Kuzniki et al have reviewed the data from two institutions over three year period from 2014 – 2017. Overall the manuscript is well written and adds valuable information about the incidence, risk-factors and outcomes of a rare event – CNS metastasis in Endometrial Cancer. Information is timely, important as well as well-presented.

Can the authors address the following minor issues to improve the manuscript –

1. More than 50% of the patients are listed as “other” race from the USC cohort. Is there anyway to figure out the actual race/ethnicity in these patients?

2. Table 1 mentions that 36.4% of the patients with CNSm, the first site of recurrence was actually the brain. I am not sure if this is highlighted somewhere else in the manuscript. I think its an important clinical finding that should be highlighted if possible in the results and discussion.

3. Line 145 mentions that only 4.5% patients had leptomeningeal spread. Again, I think its important to highlight in discussion that some patients

4. Please clarify the lines – 161 and 162: Do the authors mean that the survival was better for patients with recurrence site as CNS – but with solitary nodules? I cannot imagine that the patients with multiple CNS Mets did well even if CNS was the only site of Mets.

5. Lines 185-187 are very confusing please clarify

6. Line 215-216 is a better choice for the conclusion of the abstract. Consider using that instead of the one used right now.

7. Table 3 is nicely done but not discussed anywhere except line 232 as a passing comment. Authors should consider comparing their study to published literature.

Congratulation on achieving the difficult task of pooling data across two institutions and thank you for submitting this important study to PLOS One.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

23 Apr 2022

Thank you for the opportunity to re-submit our revised manuscript entitled “Prognostic Features of Endometrial Cancer Metastasis to the Central Nervous System.” We appreciate the reviewers’ constructive feedback. Please find our point-by-point response to reviewer comments and recommended manuscript changes. Reviewers’ comments were very encouraging, and we hope this improved manuscript will be considered for publication in PLOS ONE. All de-identified patient data has also been submitted to PLOS ONE site with this submission.

Reviewer #1:

Introduction

Line73: why those two centers? And how did you decide on those?

- Author Response: We thank the reviewer for this comment. These sites were chosen as they are the practicing centers of two collaborating senior authors.

-Manuscript edit: Lines 93-94: Institutions were chosen due to preexisting collaborative relationships and were the main practicing sites of involved authors.

Line74: can the authors be more specific about the type of outcome they are setting to describe?

- Author Response: We thank you for this comment and opportunity to elaborate. We sought to describe survival outcomes (reported as overall survival from original endometrial cancer diagnosis and survival following diagnosis of central nervous system metastasis.) We have added clarification within the manuscript as written below.

- Manuscript edit: Line 86-90: “In the present retrospective cohort study, conducted at two major NCI designated cancer centers, we sought to describe characteristics and survival outcomes of patients with CNSm from EC. Additionally, we aimed to evaluate potential prognostic factors associated with improved outcomes following CNSm diagnosis. Due to the rare nature of this presentation, we further aimed to complete a review of the literature and provide additional valuable insights on its outcomes .”

Material and Method

Line80-85: few questions:

1- why the years from 1986-2016 for the recurrence? (see response to #2)

2- are those years symmetrical in both institutions?

- Author Response: Thank you for allowing us to elaborate. The beginning time point 1986 was the earliest year that we were able to find scanned documents for patients reflecting medical evaluation and imaging results to evaluate for central nervous system metastasis. The end time point 2016 was the year when our data collection for this study began at Moffitt Cancer Center. With the addition of patients from USC the search dates were matched to those used at Moffitt for consistency.

- Manuscript edit: Line 96-99: “Electronic medical record systems then were queried at both institutions to identify cases of EC with CNSm diagnosed between January 1,1986 (earliest date of scanned records) through June 1, 2016 (initial start date of data collection).”

3-how did you identify the CNS recurrence in EMR ? Is it through billing of treatments given to those CNS recurrence? (see response to #4)

4- how did you match those CNS recurrences with the EC diagnosis? Is it by ICD ?

- Author Response: We thank the reviewer for this comment and have added appropriate clarification in methods. Patients of interest were identified by selecting patients who had both a diagnosis code for endometrial cancer and a billing code for central nervous system imaging (i.e. brain or spine CT or MRI).

- Manuscript edit: Line 99-100: “Patients with a diagnosis code of EC and billing code for CNS imaging comprised the screening pool.”

Line90-93: regarding the control group :

1- why only one center ? (see answer to #2)

2- why only the years 2014-2017?

- Author Response: The authors thank the reviewer for this comment. The control group was extracted from a preexisting endometrial cancer database at Moffitt cancer center which encompassed patients treated from 2014-2017, this database only included patients from a single center. Our manuscript now reflects this clarification.

-Manuscript edits: Lines 108-110 “Data from a preexisting database of all EC patients treated at the primary site (MCC) from 2014-2017 was used to create a comparison cohort (non-CNSm).”

3- more importantly, are those control group include patients who have had primary EC without recurrences? Or is it all EC with recurrences but not CNS met? Or a mixture of both (I mean , those with recurrences mixed with those without) ?

- Author Response: We thank you for this thoughtful comment. The control group consists of all patients with EC treated at MCC from 2014-2017 including those both with and without recurrences. Patients with CNSm were not specifically filtered out of this group as the medical record numbers of the patients in the control group were not provided to the study team from the database. Given that patients presented with CNSm both at initial endometrial cancer diagnosis and at the time of recurrence, we felt it was clinically appropriate to compare the CNSm cohort to all patients with an EC diagnosis, regardless of whether they experienced a recurrence. In response to this comment and we included an additional comparison between the CNSm cohort and recurrent patients from the non-CNSm cohort.

- Manuscript edits:

- Line 110-111 “This data set included all stages of EC, all EC histologies and were included regardless of recurrence status.”

- Line 149-156: “An additional comparison was completed between the CNSm cohort and patients within the non-CNSm cohort who developed recurrences (n=30), in attempt to capture the more aggressive cases. In this comparison, the CNSm cohort and the recurrent non-CNSm cohort had similar rates of serous histology (22.7% CNSm vs 23.3% recurrent non-CNSm, p=0.205) and grade 3 tumors (81.8% CNSm vs 63.3% recurrent non-CNSm, p=0.370). However, CNSm cohort continued to have younger age at diagnosis (58.5 years CNSm vs 62.5 years recurrent non-CNSm) and higher rates of stage 3 or 4 disease at time of diagnosis (86.4% CNSm vs 56.6% recurrent non-CNSm, p=0.029.”

- Lines 207-211: “CNSm cohort OS was also compared specifically to patients within the non-CNSm EC cohort who recurred. Although the CNSm cohort trended towards higher risk of death, this was not significant compared to the recurrent patients from non-CNSm cohort (HR 2.28, 95% CI 0.578-8.974, p=0.239).”

- Lines 278-280: “Additionally, compared to recurrent non-CNSm EC patients, we were unable to demonstrate a significant difference in survival for CNSm patients (HR 2.28, 95% CI 0.578-8.974, p=0.239)”

Line96: you mentioned that because of the small number of patients only univariate analysis are done, yet in line 102 you mentioned that OS with adjustments for clinical variables. Can you please clarify this discrepancy?

- Author response: Thank you for allowing us to clarify our methods of analysis. For clarity, univariate analysis was completed for comparisons within the CNS metastasis cohort (i.e. figure 1 and 2). Due to small numbers, comparisons within the CNSm cohort used only KM curves and log rank test, without adjustments. Alternatively, in comparing the CNSm cohort to the larger non-CNSm EC cohort (as referenced in lines 180-181), a multivariable cox model was used with adjusting for stage, histology or age.

- Manuscript edit: Line 120-125 : “For these analyses within the CNSm cohort we did not adjust for covariates due to the limited sample size. For analysis between CNSm and non-CNSm cohorts, Cox proportional hazards regression model was used to estimate hazard ratios (95% confidence interval) for OS with adjustment for clinical covariates. For the analysis using the larger non-CNSm EC cohort, which involved a larger sample size, we were able to fit multivariable models adjusting for stage, histology and age.”

Line 103-109: are the authors trying to do literature review also? If so can they put the literature review an objective?

- Author Response: Thank you for this feedback. We have added this as an aim in the last paragraph of our introduction to better describe our intentions.

- Manuscript edit: Lines 89-90 : “Due to the rare nature of this presentation, we further aimed to complete a review of the literature and provide additional valuable insights on it outcomes .”

Results

Table -1

Looking at the initial stage , we find that majority of cancers are initially are stage 3-4 , which is contrary to the national data ? Can the authors describe this discrepancy or phenomenon ? If those patients ( who has CNS met ) are initially found to be advanced stage , then a question can be asked if those patients have already had CNS met and it just became a matter of time to surface clinically ? This can Cleary be seen in table -2 where the discrepancy of the initial stage is clear . Which brings us to the initial question posted in the material and method about the matter of the control group . It seams that the control group are all EV patients . Which by itself create a huge imbalance between those two groups . And advise is to select the control group to be those EC who had recurrences without CNS met .

- Author Response: Thank you for this important question and suggestion regarding both table 1 and table 2 highlighting the striking differences between patients who develop CNS metastasis from endometrial cancer and the general endometrial cancer population. As further elaborated in lines 223-227, in our review of the literature, it was found that most patients reported to develop CNSm had an advanced stage endometrial cancer. This likely relates to the intrinsic aggressive biology of endometrial cancers that eventually find their way to the central nervous system. We do concede that our comparison group includes a diverse group of patients with EC including those both with and without recurrences, and therefore our comparison group does not fully address clinical differences in those who recur in the CNSm compared to those that recur elsewhere. Therefore, we have now added a comparison between CNSm and recurrent patients in the non-CNSm cohort.

- Manuscript edit: Lines 304-308. “We also note that the non-CNSm comparison cohort is quite diverse in that it includes all patients treated for EC over the stated time period. We attempted at add a clinically relevant comparison by selecting out those patients in the non-CNSm cohort who had recurred with similar survival outcomes seen, however this resulted a significant reduction in number of patients available for comparison.”

- Line 149-156: “An additional comparison was completed between the CNSm cohort and patients within the non-CNSm cohort who developed recurrences (n=30), in attempt to capture the more aggressive cases. In this comparison, the CNSm cohort and the recurrent non-CNSm cohort had similar rates of serous histology (22.7% CNSm vs 23.3% recurrent non-CNSm, p=0.205) and grade 3 tumors (81.8% CNSm vs 63.3% recurrent non-CNSm, p=0.370). However, CNSm cohort continued to have younger age at diagnosis (58.5 years CNSm vs 62.5 years recurrent non-CNSm) and higher rates of stage 3 or 4 disease at time of diagnosis (86.4% CNSm vs 56.6% recurrent non-CNSm, p=0.029.”

- Lines 207-211: “CNSm cohort OS was also compared specifically to patients within the non-CNSm EC cohort who recurred. Although the CNSm cohort trended towards higher risk of death, this was not significant compared to the recurrent patients from non-CNSm cohort (HR 2.28, 95% CI 0.578-8.974, p=0.239).”

- Lines 278-280: “Additionally, compared to recurrent non-CNSm EC patients, we were unable to demonstrate a significant difference in survival for CNSm patients (HR 2.28, 95% CI 0.578-8.974, p=0.239)”

Line 131:

The authors mentioned that the majority of patients had syndromes of CNS recurrences, so does that mean that patients are screened for CNS recurrences?

- Author Response: Thank you for the opportunity to clarify this important point. Routine screening for CNS recurrence is not practiced at either institution included in this study. A new symptom preceding CNS imaging was confirmed in 82% of patients. This is now further clarified in our manuscript.

- Manuscript edits: Line 160-162: “Routine CNS imaging to screen for CNSm during treatment or surveillance of EC was not a regular practice at either institution in the absence of symptoms”

Line 158: the OS is very short for CNS 9 m compared to 49 for EC. Again this is a major discrepancy between the two groups. I would advise the authors to make the comparison to EC who had recurrences without CNS met. We thank the reviewer for this important comment. We have added a comparison analysis between CNSm and recurrent non-CNSm EC patients as described above.

Line 161-172: this is good data. I would advise the authors to put this data in a table format for the reader to see the number of patients ..etc

- Author Response : We appreciate reviewer’s positive comment and agree this is an important result to highlight. As suggested, we have added Table 3 to summarize the data from this portion of the manuscript that is not depicted in Figure 1.

-Manuscript edit: 202-203: addition of Table 3

Table 3. Differences in Overall Survival Based on Number of CNSm and Disease Burden

Comparison Groups Median Overall Survival P Value

CNSm only site of metastasis

CNSm + other sites of metastasis 83 months

30 months 0.007

Solitary CNSm

Multiple CNSm 83 months

28 months 0.040

<5 CNSm

≥5 CNSm 49 months

23 months 0.005

Line 180-181 : the authors here states that HR of death is higher in CNS met : two questions :

1- in retrospective studies , traditionally we analyze the Odd ratio of death ( as Hazard ratio is soared to prospective data) , can the authors deliberately explain please ?

- Author response: We thank the reviewer for this comment. For the analysis of this study we chose to use a time to event analysis framework (i.e., Cox proportional hazards model). This analysis approach can be used for both retrospective studies and prospective studies in which the outcome of interest is a time to event variable in which some subjects were censored at a particular point in time. Hence, the results are presented in terms of hazard ratios instead of odds ratios, which is the parameter that results from a Cox model.

2- the authors mentioned that those HR is after controlling for factor , but the number of patients are too small and the authors said in their material and method that the small number will not allow them to control and be only doing univariate analysis, can the authors please explain this ?

- Author response: Thank you for allowing us to clarify our methods of analysis. For clarity, univariate analysis was completed for comparisons within the CNS metastasis cohort (i.e. figure 1 and 2). Due to small numbers, comparisons within the CNSm cohort used only KM curves and log rank test, without adjustments. Alternatively, in comparing the CNSm cohort to the larger non-CNSm EC cohort (as referenced in lines 180-181), a multivariable cox model was used with adjusting for stage, histology or age.

- Manuscript edit: Line 120-125 : “For these analyses within the CNSm cohort we did not adjust for covariates due to the limited sample size. For analysis between CNSm and non-CNSm cohorts, Cox proportional hazards regression model was used to estimate hazard ratios (95% confidence interval) for OS with adjustment for clinical covariates. For the analysis using the larger non-CNSm EC cohort, which involved a larger sample size, we were able to fit multivariable models adjusting for stage, histology and age.”

Discussion

Line 262: the authors mentioned here that they excluded patients with sarcoma, but in table 1 they have sarcoma as a histology. That patient with sarcoma need to be excluded from analysis.

- Author response: Thank you for this close review of our tables. Current clinical practice reflects that uterine carcinosarcoma is more related to endometrial carcinomas instead of uterine sarcomas. Uterine carcinosarcomas are therefore staged and treated within endometrial cancer guidelines which is why we have included this patient in our study.

- Manuscript edit: Lines 101-104: “Only carcinomas arising from the endometrium were included in the present study, which was limited to endometrioid, serous, clear cell or carcinosarcoma. Uterine sarcomas were excluded”

Line 265-266: how is your data super early work up of neurological findings, did the authors study those who did not have an early work up ?

- Author response: We thank the reviewer for thoughtful comment. Within our cohort of patients with CNS metastasis from endometrial cancer, nearly all of them were diagnosed following the development of new neurologic symptoms. Clinically, new neurologic symptoms may also be attributed to side effects from oncologic treatment or functional decline from failure to thrive. Given the rarity of CNSm from EC, it may be overlooked in the differential diagnosis of new symptoms. Unfortunately, untreated CNSm can result in catastrophic hemorrhage or herniation and therefore early diagnosis and treatment is clinically ideal if a CNSm is present. The purpose of the above statement is to remind the reader that CNSm from EC, although rare, should not be forgotten during evaluation of new symptoms.

-- Manuscript edit: Lines 266-268: “Identifying CNSm as early as possible is clinically ideal in order to expedite treatment and avoid potentially catastrophic consequences of untreated CNSm including hemorrhage or herniation.”

I advise the authors to explore the limitations and the shortcomings of their study and not only discuss its strength

- Author response: We thank you for this important feedback. We have now elaborately described the study limitations as suggested the reviewer.

- Manuscript edit: Lines 298-308: “Due to the close relationship between disease volume and treatment of choice, the present study is met with similar limitations in determining causality of improved outcomes when considering these two variables. Additional limitations met by this study include a small CNSm cohort and retrospective nature of the study. Although we included patients from two large NCI cancer centers, the total number of patients remains relatively low and as such, the ability for multivariable analysis was limited. We also note that the non-CNSm comparison cohort is quite diverse in that it includes all patients treated for EC over the stated time period. We attempted at add a clinically relevant comparison by selecting out those patients in the non-CNSm cohort who had recurred with similar survival outcomes seen, however this resulted a significant reduction in number of patients available for comparison

______________________________________________________________________________

Reviewer #2:

Congratulation on achieving the difficult task of pooling data across two institutions and thank you for submitting this important study to PLOS One.

-Authors response: reviewer’s comments were very encouraging, and authors appreciate the constructive feedback that has helped authors further enhance our manuscript.

Reviewer’s comments: Can the authors address the following minor issues to improve the manuscript –

1. More than 50% of the patients are listed as “other” race from the USC cohort. Is there any way to figure out the actual race/ethnicity in these patients?

- Author Response: We thank the reviewer for this comment. Information regarding race/ethnicity was collected from electronic medical records. At times “other” was listed by the patient themselves or an employee collecting information on patient’s behalf without further specification. Unfortunately, this cannot be further elaborated on with the information available to our research team.

2. Table 1 mentions that 36.4% of the patients with CNSm, the first site of recurrence was actually the brain. I am not sure if this is highlighted somewhere else in the manuscript. I think its an important clinical finding that should be highlighted if possible, in the results and discussion.

- Author Response: Thank you for this thoughtful comment. We would like to highlight this finding in our manuscript and have added discussion regarding these data as stated below.

- Manuscript edit: Lines 263-266 . “Additionally, 36% of patients in the present study were diagnosed with CNSm at the time of their first recurrence. This suggests CNSm can occur early in the course of disease recurrence and CNS imaging should be completed in the setting of new neurologic symptoms regardless of timing from EC diagnosis.”

3. Line 145 mentions that only 4.5% patients had leptomeningeal spread. Again, I think its important to highlight in discussion that some patients

- Author Response: The authors thank you for this feedback and have added discussion below to highlight this important point.

- Manuscript edits Line 268-272: “Interestingly, only one patient in the present study was identified as having leptomeningeal metastasis in the absence of a parenchymal lesion. If neurologic symptoms persist in the absence of gross parenchymal metastasis on brain imaging, a consideration may be made for addition of spinal imaging if this has not been completed.”

4. Please clarify the lines – 161 and 162: Do the authors mean that the survival was better for patients with recurrence site as CNS – but with solitary nodules? I cannot imagine that the patients with multiple CNS Mets did well even if CNS was the only site of Mets.

- Author Response: We thank you for this opportunity to better demonstrate our data. We have added table 3 in addition to the existing figure 1 so that the reader can better visualize our comparison groups. Survival was improved for those with solitary CNS metastasis compared to multiple CNS metastasis. Survival was also improved if CNS was the only site of disease compared to multiple sites of disease. Although this survival information may seem obvious to a clinician, we feel the publication of this survival data may help oncologists select aggressive treatment for a subset of patients who may do surprisingly well despite CNSm diagnosis. As per your suggestions, follow changes have been made to the manuscript:

- Manuscript edits: Addition of Table 3, as well as edits to Lines 190-202: “We identified improved OS when CNS was the only site of metastasis compared to those with additional non- CNS metastasis (83m CNSm only vs 30m additional sites, p= 0.007) (Table 3). Those with CNSm as only site of disease also had improved survival following CNSm diagnosis compared to those with concurrent disease outside CNS (p=0.0096) (Figure 1.1). Out of seven patients with CNSm and no evidence of disease outside of the CNS, only one died during the follow-up period. Those with solitary CNSm had improved OS compared to those with multiple CNSm (83m solitary vs 28m multiple, p=0.04) (Table 3). Patients with solitary CNSm also had improved survival following CNSm diagnosis (p=0.0034) (Figure 1.2). Overall, small number of CNSm was associated with improved OS. OS for patients with less than five CNSm was more favorable compared to those with five or greater CNSm (49m <5 vs. 23m ≥ 5, p=0.0045)(Table 3). Similar survival benefit was noted following CNSm diagnosis (Figure 1.3). No associated significant survival benefit was found with increasing diameter of CNSm (6.5months <3cm vs 14months >3cm, p=0.327).

5. Lines 185-187 are very confusing please clarify :

- Author Response : Thank you for suggesting to clarify. We were describing outcomes for patients with CNSm from EC that received no further treatment after CNSm diagnosis which appeared confusing. Now, we have re-written these statements to further clarify as recommended.

- Manuscript edits: Lines 213-215: “Three patients received no treatment for their CNSm, all of whom survived ≤1 month following CNSm diagnosis. Overall survival from EC diagnosis was 29 months, 5 months and 1 month for these three patients”

6. Line 215-216 is a better choice for the conclusion of the abstract. Consider using that instead of the one used right now.

- Author Response : We thank you for this feedback and have redirected this comment to the conclusion paragraph of the manuscript to reflect reviewer comment as suggested.

- Manuscript edits: Lines 313-314: . “Our data demonstrates that low number of CNSm and multimodal treatment approach most closely predicts improved outcomes after CNSm diagnosis.”

7. Table 3 is nicely done but not discussed anywhere except line 232 as a passing comment. Authors should consider comparing their study to published literature.

- Author response: We thank the reviewer for this comment. Throughout the discussion we have cited most references from Table 3 (now table 4 in revised manuscript) to compare results from the present study to previously published data. Associated reference numbers have been listed in Table 4 to show how they are referenced within the discussion. Thank you again for highlighting this and hope this further clarifies.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

9 May 2022

Prognostic Features of Endometrial Cancer Metastasis to the Central Nervous System

PONE-D-21-33642R1

Dear Dr. Shahzad,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Manish S. Patankar, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Thank you for Addressing all questions. Manuscript is ready for publications from my point of view

Reviewer #2: Great job making the manuscript better. The authors have adequately addressed the reviewer comments as well as the statistical comments

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

17 Aug 2022

PONE-D-21-33642R1

Prognostic features of endometrial cancer metastasis to the central nervous system

Dear Dr. Shahzad:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Manish S. Patankar

Academic Editor

PLOS ONE

Table 1

Clinical and pathologic characteristics.

	Total	Site A	Site B	p value
	N = 22 (%)	N = 11 (%)	N = 11 (%)
Age at EC Diagnosis	58.5 (52.2; 61.8)	56 (52; 64)	59 (52.5; 60.5)	0.92
BMI	27.7 (25; 31)	27.6 (25.5; 29.6)	28.4 (24.9; 32.9)	0.71
Race				0.001
White	10 (45.5)	9 (81.8)	1 (9.1)
Black	1 (4.6)	0 (0)	1 (9.1)
Hispanic	4 (18.2)	2 (18.2)	2 (18.2)
Asian	1 (4.6)	0 (0)	1 (9.1)
Other	6 (27.3)	0 (0)	6 (54.6)
Initial Stage*				0.52
I	1 (4.6)	1 (9.1)	0 (0)
II	1 (4.6)	0 (0)	1 (9.1)
III	10 (45.5)	6 (54.5)	4 (36.4)
IV	10 (45.5)	4 (36.4)	6 (54.5)
Initial Grade				0.72
1	2 (9.1)	1 (9.1)	1 (9.1)
2	2 (9.1)	0 (0)	2 (18.2)
3	18 (81.8)	10 (90.9)	8 (72.7)
Histology
Endometrioid	16 (72.7)	8 (72.7)	8 (72.7)
Serous	5 (22.7)	3 (27.3)	2 (18.2)
Carcinosarcoma	1 (4.6)	0 (0)	1 (9.1)
LVSI	7 (31.8)	6 (54.5)	1 (9.1)	0.13
CNSm as First Recurrence	8 (36.4)	5 (45.5)	3 (27.3)	0.66
Additional Disease at CNSm Dx	15 (68.2)	7 (63.6)	8 (72.7)	0.99
Presenting Symptom of CNSm
Headache/Dizziness	9 (40.9)	7 (63.6)	2 (18.8)	0.08
Altered Mental Status	3 (13.6)	2 (18.2)	1 (9.1)	0.99
Nausea/Vomiting	4 (18.2)	2 (18.2)	2 (18.2)	0.99
Weakness/Numbness	4 (18.2)	3 (27.3)	1 (9.1)	0.59
Coordination/Gait Issues	5 (22.7)	5 (45.5)	0 (0)	0.04
Dysphagia	2 (9.1)	0 (0)	2 (18.2)	0.48
Number of CNSm**				0.60
1	10 (45.4)	6 (54.5)	4 (44.4)
2–5	4 (18.2)	3 (27.3)	1 (11.1)
>15	6 (27.3)	2 (18.2)	4 (44.4)
Treatment Approach to CNSm				0.21
Surgical Resection alone	1 (4.6)	0 (0)	1 (9.1)
Surgical Resection + localized RT	5 (22.7)	3 (27.3)	2 (18.2)
Surgical Resection + WBRT	1 (4.6)	0 (0)	1 (9.1)
Surgical Resection + Chemo Wafers	2 (9.1)	2 (18.2)	0 (0)
Stereotactic Radiosurgery alone	3 (13.6)	3 (27.3)	0 (0)
No treatment	3 (13.6)	1 (9.1)	2 (18.2)

Continuous variables presented as mean (standard deviation). Significant p value considered <0.05.

Abbreviations: EC endometrial cancer, BMI body mass index, LVSI lymphovascular space invasion, CNSm central nervous system metastasis, Dx Diagnosis, f/b followed by, RT radiation, WBRT whole brain radiation therapy, chemo wafers: chemotherapy wafers placed in surgical bed.

*One patient with unknown stage.

** number of CNSm not available for 2 patients.

Table 2

Endometrial cancer patients with CNS metastasis compared to control cohort without CNS metastasis.

	CNSm (N = 22)	Non CNSm (N = 354)	p value
Age at dx (years)	58.5 [52.2;61.8]	62.0 [55.0;68.8]	0.018
BMI (kg/m2)	27.7 [24.9;31.0]	33.7 [27.3;40.6]	0.005
Race/Ethnicity			<0.001
White	10 (45.5%)	305 (86.2%)
Black	1 (4.6%)	22 (6.2%)
Hispanic	4 (18.2%)	5 (1.4%)
Asian	1 (4.6%)	7 (2.0%)
Unknown/Other	6 (27.3%)	15 (4.24%)
Histology
Endometrioid	16 (72.7%)	262 (74.0%)	0.043
Serous	5 (22.7%)	30 (8.5%)
Stage*
I	1 (4.6%)	272 (76.8%)	<0.001
II	1 (4.6%)	24 (6.8%)
III	10 (45.5%)	45 (12.7%)
IV	10 (45.5%)	13 (3.7%)
Grade*
1	2 (9.1%)	156 (44.1%)	<0.001
2	2 (9.1%)	109 (30.8%)
3	18 (81.8%)	89 (25.1%)
LVSI*
Yes	7 (31.8%)	98 (27.7)	<0.001
No	7(46.4%)	248 (70.1%)
Not Reported	8 (36.4%)	8 (2.3%)

Abbreviations: dx diagnosis of endometrial cancer, BMI body mass index, CNSm central nervous system metastasis, LVSI lymphovascular space invasion.

Continuous variables presented as mean [Interquartile range/IQR]

*Denotes characteristics of primary endometrial tumor