Article Text
PDF
Abstract
Objective Lymphovascular space invasion (LVSI) is a known prognostic factor for oncological outcome in endometrial cancer patients. However, little is known about the prognostic value of LVSI among the different molecular subgroups. The aim of this study was to determine the prognostic dependence of LVSI from the molecular signature.
Methods This study included endometrial cancer patients who underwent primary surgical treatment between February 2004 and February 2016 at the Karolinska University Hospital, Sweden and the Bern University Hospital, Switzerland (KImBer cohort). All cases had complete molecular analysis performed on the primary tumor according to the WHO Classification of Tumors, 5th edition. LVSI was reviewed by reference pathologists for all pathology slides.
Results A total of 589 endometrial cancer patients were included in this study, consisting of 40 POLEmut (polymerase epsilon ultramutated), 198 MMRd (mismatch repair deficient), 83 p53abn (p53 abnormal), and 268 NSMP (non-specific molecular profile) cases. Altogether, 17% of tumors showed LVSI: 25% of the POLEmut, 19% of the MMRd, 30% of the p53abn, and 10% of the NSMP cases. There was a significant correlation of LVSI with lymph node metastasis in the entire study cohort (p<0.001), remaining significant in the MMRd (p=0.020), p53abn (p<0.001), and NSMP (p<0.001) subgroups. Mean follow-up was 89 months (95% CI 86 to 93). The presence of LVSI significantly decreased recurrence-free survival among patients with MMRd, p53abn, and NSMP endometrial cancer, and overall survival in patients with p53abn and NSMP tumors. In patients with NSMP endometrial cancer, evidence of substantial LVSI remained a significant independent predictor of recurrence in multivariable Cox regression analysis including tumor stage and grade (HR 7.5, 95% CI 2.2 to 25.5, p=o.001).
Conclusion The presence of LVSI was associated with recurrence in each subgroup of patients with MMRd, p53abn, and NSMP endometrial cancer, and LVSI remained an independent predictor of recurrence in NSMP endometrial cancer patients.
- uterine cancer
- pathology
- lymphatic vessels
- lymphatic metastasis
Data availability statement
Data are available upon reasonable request. In accordance with the journal’s guidelines, we will provide our data for independent analysis by a selected team by the Editorial Team for the purposes of additional data analysis or for the reproducibility of this study in other centers if such is requested.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, an indication of whether changes were made, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Lymphovascular space invasion is a well-known parameter for risk stratification in endometrial cancer as it is associated with lymph node metastasis and worse oncological outcome. Evidence of the prognostic impact of lymphovascular space invasion separately assessed in each molecular subgroup is scarce.
WHAT THIS STUDY ADDS
In patients with NSMP endometrial cancer, presence of lymphovascular space invasion remained a significant independent predictor of recurrence. In the subgroup of stage I endometrioid NSMP and MMRd endometrial cancer patients the presence of lymphovascular space invasion is an independent predictor of recurrence, confirming the European Society of Gynecological Oncology/European Society for Radiotherapy and Oncology/European Society of Pathology (ESGO/ESTRO/ESP) risk stratification.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study provides additional information on the gap between the prognostic relevance of the ‘old’ pathological risk factors and the ‘new’ molecular subgroups.
INTRODUCTION
Endometrial cancer is the most common gynecological tumor in high-income countries with a 5-year survival rate of 80%.1 Despite its generally favorable diagnosis, about 18% of endometrial cancer patients experience recurrence and in these patients mortality remains high.2 3 In order to improve outcome, management of endometrial cancer has become considerably more personalized over the past years with the aim to individually tailor surgical and adjuvant therapy and to de-escalate treatment in patients with low risk of relapse.4 5
There is clear evidence that traditional pathological factors, such as histological subtype, stage, grading, and lymphovascular space invasion (LVSI), are prognostic and should be taken into consideration for risk stratification in endometrial cancer.6 Of these, LVSI is well-known to be associated with lymph node metastasis and worse oncological outcome.7–9 Beside these traditional factors, the understanding of endometrial cancer at the molecular level has changed fundamentally during the last decade. The Cancer Genome Atlas collaborative endometrial project has evolved a genomic classification that was further developed by the simplified Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE), identifying the four molecular subtypes: polymerase epsilon ultramutated (POLEmut), mismatch repair deficient (MMRd), p53 abnormal (p53abn), and non-specific molecular profile (NSMP).4 10 The prognostic significance of this molecular classification has been investigated by numerous research groups.11–14 In 2021 the European Society of Gynecological Oncology (ESGO), the European Society for Radiotherapy and Oncology (ESTRO), and the European Society of Pathology (ESP) integrated classic pathological prognostic factors such as LVSI together with the molecular subgroups into the risk classification of their endometrial cancer guidelines.15 According to this ESGO/ESTRO/ESP risk stratification, the presence of LVSI affects adjuvant treatment recommendations in stage I endometrioid MMRd or NSMP endometrial cancer patients. However, assignment of patients to specific risk groups based on such integration should require prognostic independence of the factors considered. A recently published review has shown that LVSI has a prognostic value for worse oncological outcome independent of the molecular classification.16 Nevertheless, evidence on the prognostic impact of LVSI separately assessed in each molecular subgroup is needed to tailor adjuvant treatment optimally in these patients.15 16 The aim of this study is to evaluate the impact of LVSI on oncological outcome in each molecular subgroup.
METHODS
Patient Cohort and Clinicopathological Data
This retrospective cohort study includes endometrial cancer patients who underwent primary surgical treatment between February 2004 and February 2016 at the Karolinska University Hospital, Sweden and the Bern University Hospital, Switzerland (the KimBer cohort).13 17 All pathology slides were reviewed by reference pathologists as previously described.13 LVSI was described as present or absent in the cohort from Sweden (n=339) and absent, focal, or substantial in the patients from Switzerland (n=250). Patients were considered as LVSI positive if their tumor either had evidence of LVSI in the Swedish population or substantial LVSI in the Swiss population. Follow-up data on recurrence and survival were available through standardized databases and follow-up controls. Molecular analysis was performed on the primary tumor according to the WHO Classification of Tumors, 5th edition.18 Immunohistochemistry for p53 and MMR proteins was performed on a tissue microarray as published previously.13 19 All tumors were analyzed for mutations of POLE gene (NM.006231) exons 9–14 by Sanger sequencing. Recurrence-free survival was defined as time from primary staging surgery to first recurrence or death of any cause. Overall survival was defined as time from primary staging surgery to death of any cause. Patients who were alive were censored at the date of their last follow-up.
Statistical Analysis
Statistical calculations were performed using the Statistical Package for Social Sciences (SPSS Statistic version 28.0.1.1, International Business Machines (IBM), Armonk, NY). Categorical variables were reported as frequencies and percentages, while continuous variables were reported as means and SD. Patients, tumor, and treatment characteristics were analyzed using χ2 statistics in the case of categorical and analysis of variance (ANOVA) for continuous variables. Survival curves were generated using the Kaplan-Meier method and compared using the log-rank test. Cox regression analyses were conducted to assess the relationship between the risk of recurrence and death with LVSI and other prognostic factors. A p value <0.05 was considered statistically significant.
RESULTS
The entire study cohort included 589 molecularly classified endometrial cancer patients: 6.8% were classified as POLEmut, 33.6% as MMRd, 14.1% as p53abn, and 45.5% as NSMP. Table 1 provides a detailed description of the main clinicopathological characteristics.
Demographic and clinicopathological characteristics of the study cohort
Association of LVSI with Clinicopathological Characteristics and Oncological Outcome
LVSI was assessed in all 589 patients and described as present (not specified) in 73 (12.4%), focal in 36 (6.1%), and substantial in 25 (4.2%) of the tumor samples. LVSI positive tumors were significantly associated with negative prognostic factors (details are provided in Table 2). Mean follow-up was 89.4 months (95% CI 85.7 to 933.1). Mean recurrence-free survival was significantly longer for patients without evidence of LVSI (149.4 months, 95% CI 138.7 to 160.1) compared with LVSI positive (90.6 months, 95% CI 75.7 to 105.5) patients (log-rank, p<0.001). Similar associations were seen in overall survival, amounting to 156.0 months (95% CI 145.8 to 166.3) and 104.9 months (95% CI 90.7 to 119.0), for LVSI negative and positive patients, respectively (log-rank, p<0.001). In the subgroup of patients without lymph node assessment, LVSI was significantly associated with recurrence (HR 3.3, 95% CI 1.5 to 7.3, p=0.003) and death (HR 2.2, 95% CI 1.1 to 4.2, p=0.037) in univariable Cox regression analysis. On the other hand, in patients with histologically confirmed node negative endometrial cancer there was no correlation of LVSI with recurrence (HR 2.0, 95% CI 0.9 to 4.3, p=0.11) or death (HR 1.6, 95% CI 0.7 to 3.3, p=0.27). In multivariable Cox regression analysis including stage and grading, LVSI was no longer associated with a higher risk of recurrence (HR 1.5, 95% CI 0.9 to 2.3, p=0.10) or death (HR 1.2, 95% CI 0.8 to 1.8, p=0.36) in the overall study cohort.
Association of lymphovascular space invasion with clinicopathological characteristics
Prognostic Value of LVSI According to the Different Molecular Subgroups
POLEmut
Tumors of 40 patients were classified as POLEmut endometrial cancer. Of these, 82.5% presented with stage I disease, 60% with low-grade tumors, and 82.5% with endometrioid histology. LVSI was present in 25% and lymph node involvement in 7.5% of all POLEmut endometrial cancer samples. In this subgroup, presence of LVSI was not associated with tumor size (p=0.28), high-grade histology (p=0.06), or lymph node involvement (p=0.66). During mean follow-up of 81.6 months (95% CI 69.5 to 93.7), five patients developed recurrence and four patients died. There was no significant association of LVSI with recurrence rate (p=0.58) in this subgroup. Due to the low number of events, no further calculations on oncological outcome were performed in the POLEmut subgroup.
MMRd
MMRd endometrial cancer was present in 198 patients. In this subgroup, 74.7% had stage I disease at primary diagnosis, 75.3% presented with low-grade tumors, and 86.4% with endometrioid histology. Altogether, 18.7% of the MMRd endometrial cancer patients were LVSI positive and 10.6% presented with lymph node metastases. In MMRd endometrial cancer patients LVSI was associated with larger tumor size (p=0.003), high-grade histology (p=0.020), advanced tumor stage (p=0.003), and lymph node involvement (p=0.020). In this subgroup, 37 patients suffered a recurrence and 50 patients died during mean follow-up of 88.9 months (95% CI 82.1 to 95.6). Recurrence-free survival was significantly lower in MMRd patients showing LVSI (112.6 months, 95% CI 91.6 to 133.5) compared with LVSI negative (182.4 months, 95% CI 169.3 to 195.4) MMRd patients (log-rank, p=0.046) (Figure 1A). On the other hand, overall survival did not differ significantly between LVSI positive (128.7 months, 95% CI 111.1 to 146.3) and LVSI negative (148.6 months, 95% CI 130.9 to 166.3) patients (log-rank, p=0.72) (Figure 1B). In multivariable Cox regression analysis including stage and grading, LVSI was no longer an independent factor for recurrence (HR 1.4, 95% CI 0.7 to 3.0, p=0.37).
Kaplan-Meier survival curves according to LVSI status in the MMRd subgroup for (A) recurrence-free survival (log-rank, p=0.046) and (B) overall survival (log-rank, p=0.72). LVSI, lymphovascular space invasion; MMRd, mismatch repair deficient.
p53abn
This subgroup includes 83 endometrial cancer patients. Compared with the other molecular subgroups, this category contains the most advanced stages (31.4%), high-grade tumors (67.5%), non-endometrioid histological subtypes (54.2%), LVSI positivity (30.1%), and lymph node involvement (19.3%). In p53abn patients, LVSI positivity was associated with larger tumor size (p=0.003), advanced tumor stage (p<0.001), and lymph node metastasis (p<0.001). p53abn endometrial cancer patients showed the highest recurrence (33.7%) and death (38.6%) rates among all molecular subgroups. Mean follow-up was 94.0 months (95% CI 83.1 to 105.0). Recurrence-free survival was worst compared with the other molecular subgroups, and significantly worse in p53abn LVSI positive (67.8 months, 95% CI 43.6 to 92.0) compared with p53abn LVSI negative (110.6 months, 95% CI 91.8 to 129.4) endometrial cancer patients (log-rank, p=0.037) (Figure 2A). The same correlations were seen for overall survival with a significant difference between LVSI positive (78.3 months, 95% CI 55.6 to 100.9) and LVSI negative (149.4 months, 95% CI 122.2 to 176.6) patients (log-rank, p=0.039) (Figure 2B). In multivariable Cox regression analysis including stage and grading, there was no longer a significant correlation with LVSI and recurrence (HR 0.9, 95% CI 0.4 to 2.1, p=0.81) or death (HR 1.2, 95% CI 0.5 to 2.6, p=0.73).
Kaplan-Meier survival curves according to LVSI status in the p53abn subgroup for (A) recurrence-free survival (log-rank, p=0.037) and (B) overall survival (log-rank, p=0.039). LVSI, lymphovascular space invasion; p53abn, p53 abnormal.
NSMP
NSMP endometrial cancer patients form the largest subgroup consisting of 268 patients. Altogether, 78.8% of all NSMP tumors were stage I at primary diagnosis with the highest proportion of low-grade (87.7%) and endometrioid (94.0%) histologies among all molecular subgroups. LVSI positivity was present in 9.7% of all tumors and 9.0% of the patients had lymph node involvement. In NSMP endometrial cancer patients, the presence of LVSI correlated with larger tumor size (p<0.001), non-endometrioid (p=0.012) and high-grade (p<0.001) histology, advanced tumor stage (p<0.001), and lymph node involvement (p<0.001). During mean follow-up of 89.8 months (95% CI 84.6 to 95.1), 34 patients suffered recurrence and 62 died. Mean recurrence-free survival was significantly worse in LVSI positive (58.5 months, 95% CI 39.9 to 77.1) compared with LVSI negative (140.1 months, 95% CI 129.3 to 150.9) NSMP endometrial cancer patients (log-rank, p<0.001) (Figure 3A). Similar results showed up for overall survival with a mean of 63.7 months (95% CI 46.0 to 81.5) in the LVSI positive and 152.7 months (95% CI 141.1 to 164.3) in the LVSI negative subgroups (log-rank, p<0.001) (Figure 3B). Evidence of substantial LVSI was a significant independent predictor of recurrence (HR 7.5, 95% CI 2.2 to 25.5, p=0.001) and death (HR 4.5, 95% CI 1.8 to 11.8, p=0.002) in multivariable Cox regression analysis including tumor stage and grade.
Kaplan-Meier survival curves according to LVSI status in the NSMP subgroup for (A) recurrence-free survival (log-rank, p<0.001) and (B) overall survival (log-rank, p<0.001). LVSI, lymphovascular space invasion; NSMP, non-specific molecular profile.
According to the ESGO/ESTRO/ESP Risk Classification
In the subgroup of stage I endometrioid NSMP and MMRd endometrial cancer patients, the presence of LVSI was significantly associated with recurrence (HR 2.8, 95% CI 1.1 to 7.2, p=0.039) but not with death (HR 1.0, 95% CI 0.3 to 3.0, p=0.91) in univariable Cox regression analysis. In multivariable Cox regression analysis including myometrial invasion and grading, LVSI (irrespective of whether focal or substantial) remained the only independent predictor of recurrence (HR 6.2, 95% CI 1.8 to 21.5, p=0.012). In patients with International Federation of Gynecology and Obstetrics (FIGO) stage IA p53abn endometrial cancer and/or tumors with non-endometrioid histology, the presence of LVSI was not associated with recurrence-free (log-rank, p=0.61) or overall (log-rank, p=0.69) survival. In addition, the subgroup of patients with high-risk tumors according to the ESGO/ESTRO/ESP risk classification showed no association between the presence of LVSI and recurrence (HR 2.0, 95% CI 0.9 to 4.1, p=0.06) in univariable Cox regression analysis.
DISCUSSION
Summary of Main Results
In total, 17% of the tumors showed LVSI with the largest proportion in the p53abn subgroup. The presence of LVSI significantly decreased recurrence-free survival among patients with MMRd, p53abn, and NSMP endometrial cancer and overall survival in patients with p53abn and NSMP tumors. Only in patients with NSMP endometrial cancer, evidence of LVSI remained a significant independent predictor of recurrence in multivariable Cox regression. Furthermore, the presence of LVSI was an independent predictor of recurrence in the subgroup of patients with stage I endometrioid NSMP and MMRd endometrial cancer in our cohort, which is consistent with the ESGO/ESTRO/ESP risk classification. By contrast, LVSI positivity was not associated with worse oncological outcomes in patients with stage IA p53abn and/or non-endometrioid endometrial cancer.
Results in the Context of Published Literature
Consistent with the literature, the largest number of patients with LVSI positive tumors in our study was found in the p53abn subgroup, and patients with NSMP endometrial cancer showed the lowest level of LVSI.10 11 20 In our cohort, LVSI was associated with negative predictive factors such as lymph node metastasis and higher tumor stage, as described previously,7 21 and this association remained present in each subgroup of MMRd, p53abn, and NSMP endometrial cancer patients. In accordance with a recently published review,16 LVSI had a prognostic value independent of the molecular classification in our study, with a decreased recurrence-free survival in each of the subgroups of patients with MMRd, p53abn, and NSMP tumors. In our cohort, LVSI remained an independent predictor of recurrence in the NSMP subgroup, in line with a single center cohort study only investigating NSMP and MMRd endometrial cancer patients.22 In the MMRd subgroup, LVSI was not associated with a higher risk of recurrence on multivariable Cox regression analysis in our study, which is in contrast to previous publications identifying LVSI as an independent predictor for disease-specific survival in MMRd endometrial cancer.22–24 However, these results are not directly comparable with ours, as different testing algorithms for the molecular classification were applied.22–24 For POLEmut endometrial cancer patients we found no association for the presence of LVSI with recurrence, which is in line with previously published literature.25 26 To our best knowledge no study has investigated the prognostic impact of LVSI separately in p53abn endometrial cancer patients. Furthermore, our study results support the assumption of the ESGO/ESTRO/ESP risk classification, that LVSI status has a prognostic impact in early stage endometrioid MMRd or NSMP endometrial cancer patients, but not in p53abn and/or non-endometrioid tumors.27
Strengths and Weaknesses
To the best of our knowledge, this is the first study to separately assess the prognostic impact of LVSI in each molecular subgroup of endometrial cancer patients. The major strengths of the current study include its large sample size, the length of follow-up, and its multicenter design. The most important limitations are the retrospective study design and the missing substratification of LVSI into ‘focal’ and ‘substantial’ in more than half of the study patients. Moreover, LVSI as an important prognostic factor still has some shortcomings, as it is only available post-operatively after examination of the hysterectomy specimen, and there is considerable inter-observer variability in the assessment of LVSI.28
Implications for Practice and Future Research
The optimization of surgical and adjuvant therapy in endometrial cancer based on personalized risk stratification has advanced greatly during the last decade, especially with the introduction of sentinel lymph node mapping and the molecular classification.5 29 Assessment of independent prognostic and predictive factors is needed to obtain an ideally tailored management of endometrial cancer patients. Among the major traditional pathological factors in endometrial cancer is LVSI.7–9 While the prognostic independence of the molecular classification from LVSI seems verified,10 11 30 the opposite has not yet been sufficiently studied. With that in mind, our study results provide further evidence to fill the gap between the prognostic relevance of the ‘old’ pathological risk factors and the ‘new’ molecular subgroups. Furthermore, our findings are of considerable importance in order to confirm the ESGO/ESTRO/ESP recommendations and provide a rationale for the RAINBO study, considering LVSI as risk factor in MMRd and NSMP endometrial cancer.31
CONCLUSIONS
LVSI was associated with worse recurrence-free survival separately in all endometrial cancer molecular subgroups, except POLEmut cases. In addition, LVSI remained an independent predictor of recurrence in NSMP endometrial cancer patients. With our study, we were able to confirm the ESGO/ESTRO/ESP risk stratification, indicating that LVSI is a predictor of recurrence in the subgroup of patients with early stage endometrioid NSMP and MMRd endometrial cancer.
Data availability statement
Data are available upon reasonable request. In accordance with the journal’s guidelines, we will provide our data for independent analysis by a selected team by the Editorial Team for the purposes of additional data analysis or for the reproducibility of this study in other centers if such is requested.
Ethics statements
Patient consent for publication
Ethics approval
Ethical approval was obtained from the local ethics committees in Stockholm and Bern (reference numbers 2016/362 and 2018–00479, respectively).
References
Footnotes
Twitter @FlurinaSaner
Contributors All authors have provided substantial contributions and are in agreement will all aspects of the final manuscript and its submission. FS, SI, MDM: conceptualization, methodology, validation. FS: formal analysis, writing, guarantor. TTR JWC, EE: data curation, validation, investigation, resources. CB, AG, FA-CMS: methodology, data curation. MDM, SI: supervision.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.