http://orcid.org/0000-0001-9456-5392
Abstract
Background: The International Federation of Gynecology and Obstetrics (FIGO) ovarian cancer staging system includes no sub-stage for lymph nodes (LN) as only distant disease manifestation. We explore the prognostic implication of LN as only stage IV classifier in serous ovarian cancer.
Method: This is a nation-wide, population-based study on 551 women with serous stage IV cancers diagnosed between 2009–2014. We compare overall survival (OS) in women with LN as only distant metastatic site to those with pleural metastases only and to patients with other/multiple stage IV manifestations. Cox regression models were used for uni- and multivariable estimations.
Results: Of 551stage IV cases, distant metastatic site was registered in 433. Median OS for women with LN (n = 51) was 41.4 months, compared to 25.2 and 26.8 months for patients with pleural (n = 195) or other/multiple (n = 187) distant metastases (p = .0007). The corresponding five-year survival rates were 32, 11 and 22%, respectively. Multivariable analyzes confirmed shorter survival for women with pleural (HR 2.99, p = .001) or other/multiple distant sites (HR 2.67, p = .007), as compared to LN cases. LN only patients lived 9.1 months longer after primary than after interval surgery, but this difference was not significant (p = .245).
Conclusion: Women with stage IV serous ovarian cancer having lymph nodes as only distant metastatic site live longer than other stage IV patients.
Introduction
Epithelial ovarian, fallopian tube or primary peritoneal cancers (EOC) have the highest mortality rate of all gynecologic malignancies. This can be ascribed to the high proportion of patients presenting with already disseminated disease, with 12–33% being diagnosed in International Federation of Gynecology and Obstetrics (FIGO) stage IV [Citation1,Citation2]. The different EOC subtypes exhibit diverse genetic traits and clinical behavior, with the high-grade serous (HGS) carcinomas being the most aggressive comprising the vast majority of late stage cases [Citation3].
Standard treatment for EOC confined to the abdominal cavity (up to stage IIIC) is cytoreductive surgery aiming at no visible disease and platinum-taxane combination chemotherapy. However, a standard approach for stage IV disease, with hematogenous metastases and/or spread outside the abdomen, is not so easily defined and the treatments offered differs between countries and institutions [Citation4–6]. For example, there is an ongoing discussion regarding the timing of surgery for stage IV disease, where some favors interval over primary debulking [Citation7,Citation8].
The prognostic factors that best predict survival from HGS carcinomas are age, performance status, residual tumor after surgery and most importantly, stage at diagnosis, the latter indicating pattern of spread, consequently implicating tumor biology. In 2013, FIGO revised the staging system with subgrouping of stage IV cases into stage IVA and B [Citation9]. Thus, the current FIGO stage IVA includes women with cytologically proven pleural metastases as only site of non-regional spread of disease, whereas stage IVB encompass all other stage IV cases.
Based on clinical observation, we believe patients with non-regional lymph nodes (i.e., outside the abdomen and retroperitoneum) as only site of distant metastases to live longer than those with either invasion of the pleura or hematogenous spread of disease. This would also be consistent with the substaging in stage III, where isolated lymphatic spread outside the pelvis renders a stage IIIA1 diagnose, while peritoneal metastases to the upper abdomen is classified as stage IIIA2–IIIC.
The aim of our study was to explore whether women with non-regional lymph nodes (LN) as only site of serous stage IV ovarian cancer can expect a longer survival than that otherwise seen in stage IV disease. In parallel, we wanted to evaluate the prognostic implication of pleural metastases as only stage IV classifier. To do this, we used the population-based Swedish Quality Registry for Gynecological Cancer (SQRGC).
Patient and methods
Swedish Quality Registry for Gynecological Cancer (SQRGC)
Reporting to the Swedish National Cancer Registry (NCR) is mandatory for both clinicians and pathologists. NCR has over 95% coverage and 99% of reported malignancies are morphologically verified [Citation10]. However, clinical data including treatment and follow-up is lacking. Hence, the SQRGC was established in 2008, in detail described elsewhere [Citation11]. Reporting to the SQRGC is performed prospectively by all hospitals and clinics in Sweden. Registration is web-based and SQRGC data can be accessed, through the personal identification number allocated to all citizens, which are linked to both the NCR and to the National Cause of Death Registry enabling coverage control and life-long follow-up of patients.
The capture rate of ovarian cancers in the SQRGC has been shown to be 94–97% of cases reported to the NCR [Citation12]. Validity of reported SQRGC data has also been assessed, with 70–100% agreement between registered data and original case files [Citation12]. Patient consent is obtained before registration and regional registrars continuously monitor SQRGC files.
Patient population
We included all women with serous stage IV, non-uterine adenocarcinoma of the ovary, fallopian tube, peritoneum or undesignated primary site registered in the SQRGC between 2009 and 2014. Inclusion was restricted to morphologically verified invasive serous cases by using the International Classification of Diseases for Oncology (ICD-O-3) coding (8441/3, 8460/3, 8461/3 and 9014/3). Topographical codes were C56.9 (ovary), C57.0 (fallopian tube), C48.1 and C48.2 (peritoneum), C76.2 (abdomen NOS) and C76.3 (pelvis NOS). Tumors were classified according to WHO standards [Citation13] and the Silverberg grading system [Citation14]. The Malpica serous two-tier criteria were not widely practiced in Sweden before 2014.
We extracted SQRGC data on diagnosis, age, site of distant metastases, performed surgery, postoperative residual tumor size, type of first line chemotherapy and survival. Patients were followed up until October 31 2016 or to death, whichever came first. The ethical review board at Gothenburg University (Dnr193-14) approved the study and deemed a separate study-specific consent as not needed.
Statistical analyses
Overall survival (OS) was defined as the time from diagnosis to death or last follow up. It was assessed by the Kaplan–Meier method and survival curves were compared by the log rank test.
We performed Cox uni- and multivariable regression models for estimation of the different covariates influence on mortality. Cases with non-regional LN as only registered site of stage IV disease (LN only) were thus compared to cases with exclusively pleural metastases (current stage IVA) and to all other stage IV cases (other/multiple). In building the multivariable model, the variables were added stepwise for analysis of respective effect. For the final analyses, we adjusted for age and given treatment (primary- versus interval debulking surgery versus chemotherapy alone). The proportional hazards assumption of the Cox models was tested using Schoenfeld’s test.
In separate analyses restricted to surgically treated patients, we estimated survival after primary- and interval debulking surgery, respectively. Covariates in this multivariable model were; age, site of distant metastases and residual tumor after surgery.Cases with missing data for any covariate were excluded from the corresponding calculations. Comparisons were two-sided with a 5% significance level. We used the R statistical software for Macintosh version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria) and the ‘survival’ package 2.39.5 for estimation of survival and Stata statistical software version 13.1 (StataCorp, College Station, TX, USA) for the regression models.
Results
Patient characteristics
We found 834 registered invasive stage IV cases in the SQRGC, of which 573 (68.7%) were diagnosed as serous. Other subtypes and cases with adenocarcinoma NOS or undifferentiated tumors were excluded. The final study population thus consisted of 551 histologically verified serous carcinoma cases.
depicts the study inclusion flow chart and displays the patients characteristics. Median age for the entire cohort was 67 years and did not differ significantly between cases with different metastatic sites. Sixty-eight percent of cases were classified as of ovarian origin and in 14% the primary site had not been established (pelvis/abdomen NOS). Only 7% were diagnosed as primary fallopian tube cancers. Ninety-five percent of cases were of poor or moderate differentiation.
Figure 1. Study inclusion flowchart.
Table 1. Characteristics of patients with serous stage IV non-uterine cancer diagnosed in Sweden between 2009–2014. Total cohort (n = 551) and subgroups including only cases with specified distant metastatic sites (n = 433).
The stage IV rendering metastatic site was registered in 433 cases (79%). Among these, the by far most frequent site of distant metastases was pleura, with 57% having pleural involvement. Non-regional LN were registered in 20% and liver, lung or spleen metastases in 31%. Twenty-one percent of the women had multiple extra-abdominal metastatic sites. Fifty-one cases had non-regional LN as the only registered distant metastatic site (LN only group) and 195 had exclusively pleural metastases.
Primary treatment
Among the studied 551 serous stage IV cases, primary treatment modality was registered in 492 (89%). Of these, 395 (80%) underwent cytoreductive surgery, of which 65% at diagnosis (primary debulking surgery, PDS) and 35% after three to four courses of neoadjuvant chemotherapy (interval debulking surgery, IDS). In 13 cases, upfront surgery was not followed by chemotherapy.
Surgical treatment differed somewhat between subgroups, with 82% of LN only cases undergoing PDS or IDS compared to 78 and 73% of patients with pleural or other/multiple distant sites, respectively. Details on surgical outcome were registered in 361 of the 395 operated cases. In 36% of surgeries (PDS or IDS), a complete macroscopic resection (surgeon-reported no visible residual tumor at the end of surgery) was achieved. The complete resection rate was higher among LN only cases than in patients with registered pleural or other/multiple distant metastases (54, 29 and 34%, respectively).
Of 389 cases with registered specifics on given chemotherapy, 78% received carboplatin-paclitaxel combination treatment and a further 21% had other platinum-based chemotherapy. Bevacizumab was not incorporated into routine treatment in Sweden until 2014. Fifteen percent of patients received chemotherapy only and 4% did not have any anti-tumoral treatment at all. The registered reasons for withholding treatment were: death or poor performance status (3%), serious comorbidity (one case), patient’s preference (one case) or not stated (one case).
Type of administered chemotherapy did not differ between groups with different stage IV metastatic sites, but the proportion of patients not receiving any anti-tumoral treatment at all was lower in the LN only group, where all underwent treatment. For the entire cohort, median follow up was 24.7 months (range 0.2–86.1 months), see .
Survival analyses
The five-year OS for the entire serous stage IV cohort was 19%; 32% for the LN only group, 11% for pleura only and 22% for patients with other/multiple metastatic sites. Consequently, the LN only group lived considerably longer than cases with either pleura or other sites of stage IV disease, with a median survival of 41.4 months compared to 25.2 and 26.8 months, respectively (p = .007), see and .
Figure 2. Survival according to site of stage IV metastatic disease. Log-rank test, p = .0007.
Table 2. Estimated overall survival in relation to metastatic pattern and performed surgery.
Univariable Cox analyses showed age, cytoreductive surgery, platinum combination chemotherapy and site of stage IV metastatic disease to be associated with survival ().
Table 3. Uni- and multivariable Cox regression analyses on survival. Entire serous stage IV cohort. 400 cases with data complete for covariates age, metastatic site and primary treatment was included in the multivariable model.
The multivariable analyses included 400 cases with complete data for all significant covariates from the univariable analyses except for grade and type of chemotherapy, since these had no confounding effect on stage IV metastatic site. For the metastatic site comparisons, the follow-up time was split into two time periods (0–2.5 and 2.5–and above years) to uphold the assumption of proportional hazards of the Cox regression model. Compared to LN only cases, patients with pleural metastases had the poorest short-term (up to 2.5 years) survival with hazard ratio (HR) 3.16 (95% CI 1.17–5.78) in the univariable analysis and 2.99 (95% CI 1.53–5.85) in the multivariable model. The corresponding hazard ratios for patients with other/multiple metastatic sites were 2.91 (95% CI 1.59–5.57) and 2.67 (95% CI 1.31–5.44), respectively. Of 156 patients still alive at 2.5 years, 29 had LN as only stage IV classifier. We found no significant difference in survival between the three metastatic site groups after 2.5 years ().
Patients receiving chemotherapy only did worse than women also undergoing surgery (HR 2.22, 95% CI 1.44–3.42). For the entire cohort, we found no significant difference in survival between cases undergoing primary or interval cytoreduction.
Acknowledging surgical outcome to be an important factor for survival also in stage IV disease, we performed additional survival analyses on 263 surgically treated patients with SQRGC data complete for covariates age, metastatic site and postoperative residual tumor (Supplementary Table 1). Also in this restricted cohort, multivariable regression showed longer survival for LN only patients compared to women with either pleural metastases (short-term HR 2.69, 95% CI 1.28–5.67) or other/multiple sites of distant disease (short-term HR 2.36, 95% CI 1.20–4.64).
To further address the question of timing of surgery, we did a separate survival comparison of the operated patients in the LN only group. Among these, median survival for cases having PDS (n = 22) or NACT + IDS (n = 20) was 43.1 and 34.0 months, respectively. This difference was not shown of statistical significance (p = .245).
Discussion
To our knowledge, this is the first high-coverage population-based study on survival of women with different metastatic sites of stage IV serous ovarian cancer. Our data indicate that stage IV patients with non-regional lymph nodes as only distant tumor manifestation have significantly longer survival than other stage IV cases. Also, pleural spread (current FIGO stage IVA) does not confer a better prognosis than other/multiple stage IV manifestations.
There are three principally different routes of spread rendering a stage IV diagnose; lymphatic, hematogenous or cross-peritoneal tumor invasion of adjacent tissues (e.g., pleura or abdominal wall). These different patterns of dissemination among stage IV serous cancers potentially reflect inherent tumor biologic differences with possible prognostic implications. Our results suggest that disease disseminating predominately through the lymphatics would have a less aggressive nature than that directly invading the pleura or spreading hematogenously. This would be consistent with previous findings in FIGO stage III disease, in which lymph node metastases as only manifestation in the upper abdomen (stage IIIA1) confers a more favorable outcome than peritoneal disease (stage IIIA2-IIIC) [Citation9,Citation15–17].
Our findings corroborate the results of others demonstrating no prognostic advantage for pleural metastases only as compared to other stage IV disease manifestations [Citation18,Citation19]. In our study, the five-year survival rate for patients with pleura as only stage IV classifier was 11%, as compared to 32% for patients with non-regional LN only and 22% for those with other/multiple distant metastases.
We found metastatic site to be independently associated with survival only for the first 2.5 years after diagnosis. The reason for this non-proportional survival pattern may be attributed to the limited number of patients alive after 2.5 years, especially in the LN only group.
The prognostic impact of residual tumor after cytoreductive surgery is generally well accepted, even in stage IV disease [Citation20–22]. Also in our study, debulking surgery and size of tumor residuals were the strongest prognosticators. However, there is an ongoing debate regarding the timing of cytoreduction in stage IV disease, where some favor neoadjuvant chemotherapy [Citation7,Citation8] while others do not [Citation2,Citation23]. Overall, in our analysis of 551 stage IV serous cases, of which 78% had surgery, we found no significant difference in survival after PDS in comparison to IDS. However, in the LN only group, median survival was 9.1 months longer for patients undergoing PDS compared to NACT + IDS. Although non-significant (possibly due to small sample size), the results are intriguing and a prospective comparison is warranted.
The proportion of FIGO stage IV ovarian cancer cases has increased since the turn of the century, mostly due to more precise preoperative diagnostics with improved detection of distant metastases [Citation2]. Also attributing to the increased number of stage IV patients are cases with inguinal LN metastases, which before 2014 were classified as having stage IIIC disease, but now are staged as IVB. In accordance with our results, Nasioudis et al. recently demonstrated a similar survival for patients with inguinal LN as only distant metastatic site as that seen in stage IIIC cases with pelvic/paraaortic LN [Citation24]. This Surveillance, Epidemiology, and End Results (SEER) database study (with a population coverage of 27.8%) encompass a vast number of patients enabling inclusion of 151 surgically treated cases with inguinal LN as only stage IV metastatic site. The five-year survival for these patients was 46.3 months, which is strikingly close to the 41.4 months seen for the LN only group of our study. The timing of surgery is however not specified. Also in contrast to our study, the multivariable model used did not include vital prognostic variables as size of tumor residuals or given systemic treatment. In the SQRGC, distant lymph nodes are registered as ‘non-regional’, but their specific location is not noted. Therefore, we do not know the exact proportions of women in the LN only group having e.g., inguinal, cardiophrenic or axillar metastases. However, the different sites for LN metastases may often be accessible to surgery.
Our nation-wide study includes only women with histologically verified serous tumors. It is a population-based study with high coverage, minimizing institution selection bias. SQRGC data is monitored on a regular basis and its validity has been assessed with acceptable results [Citation12].
A limitation is, however, the lack of central pathology review. Although all cases with undesignated primary site were histologically verified as serous, we cannot rule out that some cases may have originated in the uterus. Another weakness is the missing data for site of distant disease in 118 cases. However, our results are in line with others indicating that the registry data is valid.
In addition, although our cohort is comparatively large, the number of patients in the metastatic site subgroups are limited, which may have influenced our results in the later follow-up period. However, also in the time period exceeding 2.5 years, although not statistically significant, cases with pleural metastases had a higher risk for death than patients with other stage IV metastatic sites.
In conclusion, the survival of women with non-regional lymph nodes as only stage IV defining site significantly exceeds that of patients with other stage IV manifestations. We thus suggest revision of the FIGO staging classification with a new definition for stage IVA as having LN metastases as only distant site of disease.
Elisabet_et_al._Suppl_Table_1._Uni-_and_multivariable_analyses_surgically_treated_171025.docx
Download MS Word (18.9 KB)Disclosure statement
All authors declared no conflicts of interest.
Additional information
Funding
References
- Heintz APM, Odicino F, Maisonneuve P, et al. Carcinoma of the Ovary. Int J Gynaecol Obstet. 2006;95S161–SS92.
- Ataseven B, Chiva LM, Harter P, et al. FIGO stage IV epithelial ovarian, fallopian tube and peritoneal cancer revisited. Gynecol Oncol. 2016;142:597–607.
- Köbel M, Kalloger SE, Huntsman DG, et al. Differences in tumor type in low-stage versus high-stage ovarian carcinomas. Int J Gynecol Pathol. 2010;29:203–211.
- Cornelis S, Van Calster B, Amant F, et al. Role of neoadjuvant chemotherapy in the management of stage IIIC-IV ovarian cancer: survey results from the members of the European Society of. Int J Gynecol Cancer. 2012;22:407–416.
- Fagö-Olsen C, Lindberg C, Ottesen B, et al. Differences in regional diagnostic strategies and in intended versus actual first-line treatment of patients with advanced ovarian cancer in Denmark. Int J Gynecol Cancer. 2014;24:1195–1205.
- Shalowitz DI, Epstein AJ, Ko EM, et al. Non-surgical management of ovarian cancer: prevalence and implications. Gynecol Oncol. 2016;142:30–37.
- Meyer LA, Cronin AM, Sun CC, et al. Use and effectiveness of neoadjuvant chemotherapy for treatment of ovarian cancer. J Clin Oncol. 2016;34:3854–3863.
- Vergote IB, Nieuwenhuysen EV, Vanderstichele A. How to select neoadjuvant chemotherapy or primary debulking surgery in patients with stage IIIC or IV ovarian carcinoma. J Clin Oncol. 2016;34:3827–3828.
- Prat J. Staging classification for cancer of the ovary, fallopian tube, and peritoneum. Int J Gynaecol Obstet. 2014;124:1–5.
- Barlow L, Westergren K, Holmberg L, et al. The completeness of the Swedish Cancer Register: a sample survey for year 1998. Acta Oncol. 2009;48:27–33.
- Dahm-Kähler P, Borgfeldt C, Holmberg E, et al. Population-based study of survival for women with serous cancer of the ovary, fallopian tube, peritoneum or undesignated origin - on behalf of the Swedish gynecological cancer group (SweGCG). Gynecol Oncol. 2017;144:167–173.
- Rosenberg P, Kjølhede P, Staf C, et al. Data quality in the Swedish Quality Register of Gynecologic Cancer – a Swedish Gynecologic Cancer Group (SweGCG) study. Acta Oncol. 2017;22:1–8.
- Tavassoli F, Devilee P, editors. World Health Organization classification of tumours. Pathology and genetics of tumours of the breast and female genital organs. Lyon (France): IARC Press; 2003.
- Silverberg SG. Histopathologic grading of ovarian carcinoma: a review and proposal. Int J Gynecol Pathol. 2000;19:7–15.
- Baek SJ, Park JY, Kim DY, et al. Stage IIIC epithelial ovarian cancer classified solely by lymph node metastasis has a more favorable prognosis than other types of stage IIIC epithelial ovarian cancer. J Gynecol Oncol. 2008;19:223–228.
- Ferrandina G, Legge F, Petrillo M, et al. Ovarian cancer patients with ‘node-positive-only’ Stage IIIC disease have a more favorable outcome than Stage IIIA/B. Gynecol Oncol. 2007;107:154–156.
- Onda T, Yoshikawa H, Yasugi T, et al. Patients with ovarian carcinoma upstaged to Stage III after systematic lymphadenctomy have similar survival to Stage I/II patients and superior survival to other Stage III patients. Cancer. 1998;83:1555–1560.
- Paik ES, Lee YY, Lee EJ, et al. Survival analysis of revised 2013 FIGO staging classification of epithelial ovarian cancer and comparison with previous FIGO staging classification. Obstet Gynecol Sci. 2015;58:124–134.
- Rosendahl M, Hogdall CK, Mosgaard B. Restaging and survival analysis of 4036 ovarian cancer patients according to the 2013 FIGO classification for ovarian, fallopian tube and primary peritoneal cancer. Int J Gynecol Cancer. 2016;26:680–687.
- Ataseven B, Grimm C, Harter P, et al. Prognostic impact of debulking surgery and residual tumor in patients with epithelial ovarian cancer FIGO stage IV. Gynecol Oncol. 2016;140:215–220.
- Bristow RE, Tomacruz RS, Armstrong DK, et al. Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: a meta-analysis. J Clin Oncol. 2002;20:1248–1259.
- du Bois A, Reuss A, Pujade-Lauraine E, et al. Role of surgical outcome as prognostic factor in advanced epithelial ovarian cancer: a combined exploratory analysis of 3 prospectively randomized phase 3 multicenter trials: by the Arbeitsgemeinschaft Gynaekologische Onkologie Studiengruppe Ovarialkarzinom (AGO-OVAR) and the Groupe d'Investigateurs Nationaux Pour les Etudes des Cancers de l'Ovaire (GINECO). Cancer. 2009;115:1234–1244.
- Chiva L, Lapuente F, Castellanos T, et al. What should we expect after a complete cytoreduction at the time of interval or primary debulking surgery in advanced ovarian cancer? Ann Surg Oncol. 2016;23:1666–1673.
- Nasioudis D, Chapman-Davis E, Frey MK, et al. Should epithelial ovarian carcinoma metastatic to the inguinal lymph nodes be assigned stage IVB? Gynecol Oncol. 2017 ;2017147:81–84.