Abstract
Background/Aim: Residual disease (RD) after primary debulking surgery (PDS) is a prognostic factor for survival in advanced ovarian cancer (AOC). This study aimed to examine whether the tumor extent affects overall survival (OS) and progression-free survival (PFS) in AOC patients treated with PDS. Patients and Methods: A total of 118 patients treated with PDS were included. Age, ECOG score, AOC International Federation of Gynecology and Obstetrics (FIGO) stage, CA-125, RD, peritoneal cancer index (PCI), preoperative imaging (CT-PCI) and macroscopic visualization at the surgery start (S-PCI) were analyzed. Tumor extent was quantified using the PCI, and by CT-PCI and S-PCI. Cox regression, Kaplan-Meier and receiver operating curves (ROC) were performed for survival analyses. Results: S-PCI correlated with both OS (1.067, 95%CI=1.018-1.119, p<0.007) and PFS. Patients exhibiting S-PCI≥18.5, adjusted to age, performance status, and RD, had a two-fold risk of dying (HR=2.070, 95%CI=1.061-4.038, p=0.033) compared those with PCI<18.5. CT-PCI correlated with OS in crude data (1.037, 95%CI=1.005-1.071, p=0.025), but this was not sustained in multivariate analyses. RD of any size doubled the risk of dying (2.177, 95%CI=1.235-3.838, p=0.007). Conclusion: The tumor extent at the beginning of surgery seemed to affect OS in patients with AOC, regardless of the extent of RD at the end of the surgery. PCI above 18.5 doubled the risk of dying of the disease. No difference in major complications was noted in the two groups of patients. CT-PCI seemed to play a prognostic role for PFS; however, it is still to be investigated as a prognostic factor for OS.
Epithelial ovarian cancer is the gynecological cancer with the highest mortality rate due to the fact that most cases are diagnosed at an advanced stage when the disease has already spread in the abdominal cavity (1). Due to ascites formation, and tumor cell exfoliation into the peritoneal space and along the lymphatic pathways, the disease quickly spreads in the whole abdominal cavity. Advanced ovarian cancer (AOC) is characterized by peritoneal carcinomatosis, generally found in the omentum and peritoneal surfaces from the diaphragm to the pouch of Douglas (stage III), malignant pleural infiltration, or visceral metastases (stage IV) (2-4).
The standard of treatment for AOC is upfront surgery with the aim of macroscopic radicality, followed by platinum-based chemotherapy (5). The most important prognostic factor in AOC is macroscopic radicality, with no residual tumor (6, 7). The preoperative evaluation of the tumor burden is crucial to make a proper judgment regarding the upfront surgery. The extent and distribution of peritoneal carcinomatosis have been widely studied, first in colorectal cancer and later in ovarian cancer, and numerically quantified as the peritoneal cancer index (PCI) (8). Sugarbaker et al. first described the PCI in colon cancer with peritoneal spreading, and found that certain PCI scores are correlated with the probability of complete cytoreductive surgery (CCS), and survival (9-11). The role of PCI scoring as a predictive factor for surgical outcome and survival (12) in ovarian cancer patients is not clear due to heterogeneous results (12-15).
The main challenge for the gynecologic oncologist in the therapeutic decision-making process is the preoperative evaluation of the tumor extent by a non-invasive method and designing a proper treatment plan for the patient. Laparoscopy, magnetic resonance imaging and computed tomography (CT) have been used to make a correct selection of patients for upfront cytoreductive surgery (16-19).
In AOC, CT of the abdomen and thorax are included in the preoperative standard imaging assessment of the tumor extent (20). The appropriate imaging includes an assessment of the thorax in a search for distant metastasis, pleural effusion, and enlarged lymph nodes. Pleural effusion may indicate pleural carcinomatosis or it may be reactive to the presence of carcinomatosis on the abdominal diaphragmatic surface (21).
In a previous study, analyzing the same study population as in the present study, a positive correlation between the S-PCI and RD was found. Similarly, a correlation was found, between preoperative assessment of the tumor extent (CT-PCI) and both, S-PCI, and residual disease (22). Besides these positive relationships and the important prognostic role of RD on survival, we examined whether S-PCI and CT-PCI can be correlated to patients’ survival. This in turn might imply that the preoperative tumor extent in AOC is a prognostic factor for patients treated with upfront cytoreductive surgery. Finally, the aim of this study was to examine whether the preoperative tumor extent as indicated by PCI, could be a prognostic factor as well. The extent of surgery, expressed as the Complexity Surgical Score (CSS) described by Aletti, and severe complications, defined as Clavien-Dindo ≥3 and their possible influence on OS were also analyzed.
Patients and Methods
Ethical consideration. All patient data were handled according to the Word Medical Association’s Declaration of Helsinki 2008 and in compliance with the Swedish national law. The Swedish Ethical Review Authority with apl.no.2019/00450 approved this study.
Patient characteristics. In total, 194 consecutive patients with primary FIGO stadium III and IV AOC, treated in a single tertiary center between 2016 and 2018, were retrospectively analyzed. The same cohort has been used in a previous study that investigated whether the CT-PCI correlated with S-PCI and to discover its role in predicting residual disease at any size (22).
The inclusion criteria for this study were as follows: 1) AOC; 2) upfront cytoreductive surgery with curative intention; 3) preoperative CT of the thorax and abdomen available; 4) follow-up data until January 2021. The exclusion criteria were: 1) interval debulking surgery; 2) palliative treatment; 3) early stages of ovarian cancer (FIGO stage I and II); 4) incomplete follow-up data (Figure 1).
Study population. Flow chart of sampling.
The clinical data were collected from patient medical records. The PCI was calculated retrospectively based on medical and histopathological reports (22).
All patients underwent upfront surgery at the Gynecology Department of Skåne University Hospital, Lund, Sweden. The 13 abdominal regions were assessed for tumor content and scored on a range from 0 to 3, depending on the tumor size: 0 defines no visible tumor, while 1, 2, or 3 describe lesions with maximum diameters of 0.5, 5.0, and >5 cm or lesions confluence. A final sum ranging from 1 to 39 points was calculated, using the European Society of Gynecological Oncology (ESGO) Ovarian Cancer Operative Report (23).
The complexity of surgery was estimated by using the Aletti Surgical Complexity Score (SCS) (24). The Clavien-Dindo classification for postoperative complications was used and all severe complications defined as Clavien-Dindo ≥3 were registered (25). The patients were categorized into a group with CCS and a group with residual disease of any size (non-CCS), and the surgical outcome was examined. The date of first relapse and survival data [progression-free survival (PFS) and overall survival (OS)] were extracted from patients’ medical records.
Image analysis. All eligible patients underwent CT in the supine position with intravenous and oral contrast. By convention, all digital CT-images were reformatted in the coronal and sagittal planes. One of two radiology specialists scored the CT-PCI using the same concept as for S-PCI, using the ESGO Ovarian Cancer Operative Report. The amount of ascites was quantitatively estimated by one of two radiologists concurrent with the CT-PCI evaluation. The radiologists were blinded to the S-PCI and surgical outcome (22).
Statistical analyses. IBM SPSS Statistics 26 was used (IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Armonk, NY, USA: IBM Corp). Age, CA-125, CT-PCI, S-PCI, estimated blood loss, and operation time were analyzed as continuous variables. The normality of the data was tested using the one-sample Kolmogorov-Smirnov test. FIGO stage (III vs. IV), ECOG performance status (<2 vs. ≥2), surgical outcome (CCS vs. non-CCS), ascites volume (<1,000 ml vs. ≥1,000 ml) and complications were dichotomized (<3 vs. ≥3) The surgical outcomes were indicated as CCS with no macroscopic tumor left or non-CCS with residual tumor left of any size. The SCS was categorized into three groups: low (≤3), intermediate (4-7), and high (≥8). PFS was defined from the date of diagnosis (date of surgery) to the date of progression. OS was defined from the diagnosis date to the death date or last follow-up. The student’s t-test and Mann-Whitney U-test were used to analyze the association between PCI and clinical factors. The survival analyses were performed by using univariate and multivariate analyses. The median survival and the rate of survival were calculated with the Kaplan-Meier method using multivariate Cox regression analyses for the CCS and non-CCS groups and the different PCI cut-off values for CT-PCI and S-PCI, respectively. The Log-Rank test was used to compare the statistical significance of the differences between the Kaplan-Meier curves.
The receiver operating curve (ROC) was used to calculate a cut-off for OS for both S-PCI and CT-PCI. A confidence interval and a p-value were generated in order to calculate the statistical significance. A p-value of <0.05 was considered significant.
Results
Descriptive data. A total of 118 patients were identified as eligible for the study (Figure 1). They were candidates for upfront cytoreductive surgery, followed by adjuvant chemotherapy in the University Hospital of Skåne, Lund, from January 2016 to December 2018.
The patients’ characteristics are presented in Table I.
Patient characteristics.
Survival analysis of the cohort. CT-PCI, S-PCI, ascites volume above 1,000 ml, and surgical outcome (CCS vs. non-CCS) were associated with impaired PFS. Age, ECOG, CT-PCI, S-PCI, and surgical outcome (CCS vs. non-CCS) were related to OS in the univariate analyses (Table II).
Univariate analyses to predict progression-free survival (PFS) and overall survival (OS).
Impact of PCI on PFS and OS. PFS seemed to correlate with both CT-PCI and S-PCI in both univariate and multivariate analysis (Table III). CT-PCI had some correlation with OS in univariate analysis, but the results were not sustained in multivariate analysis (Table IV). The tumor extent at the beginning of surgery, quantified by S-PCI, correlated well with OS, in both univariate and multivariate analyses (Table IV).
Multivariate Cox regression analyses to predict progression-free survival (PFS).
Multivariate Cox regression analyses to predict overall survival (OS).
When a ROC curve was designed to find a cut-off for CT-PCI in relation to OS, a value of 24.5 (AUC=0.6170; 95%CI=0.511-0.719, p<0.031) was found (Figure 2A). The Kaplan-Meier analysis showed a shorter OS (about 28 months) for patients with CT-PCI ≥24.5, while the median OS for patients with CT-PCI <24.5 was 44.3 months (Figure 2B). In univariate Cox regression analyses, a doubled risk of dying of the disease was found for patients with a preoperative CT-PCI above 24.5 (HR=2.06; 95%CI=1.112-3.830, p<0.022). The multivariate Cox regression analyses showed a lower OS as well, with a 50% higher risk of death for patients exhibiting a CT-PCI above 24.5, though this was not statistically significant (HR=1.517; 95%CI=0.759-3.035, p<0.239) (Table IV).
Relationship between computed tomography (CT)-peritoneal cancer index (PCI) and survival. A. Receiver operating curve (ROC) for CT-PCI (24.5) cut-off regarding overall survival (OS) (p<0.031). B. Kaplan-Meier curve for patients having CT-PCI below (green) and above (blue) cut-off of 24.5 regarding OS (p<0.019).
The ROC curve for the S-PCI generated a cut-off value of 18.5 (AUC=0.6934; 95%CI=0.596-0.791, p=0.000) for OS (Figure 3A). The median survival for patients with S-PCI below 18.5 was 46.3 months, whereas the patients having S-PCI above 18.5 survived for a median period of 28.9 months, as shown in the Kaplan-Meier curve (Figure 3B). These results were sustained by both univariate and multivariate Cox regression analyses (Table IV).
Relationship between S-PCI and survival. A. Receiver operating curve (ROC) for S-PCI regarding overall survival (OS) (18.5 cut-off) (p<0.000). B. Kaplan-Meier curve for patients having S-PCI below (blue) and above (green) cut-off of 18.5 regarding OS (p<0.000). S-PCI: Macroscopic visualization at the surgery start; PCI: peritoneal cancer index; CT-PCI: preoperative imaging; CT: computed tomography.
Impact of residual disease on PFS and OS. In multivariate Cox regression analysis, patients with residual disease (non-CCS) had a shorter PFS (HR=1.745; 95%CI=1.056-2.882, p=0.03). With respect to the OS, patients with residual disease were estimated to have a higher risk of death compared to patients without residual disease (HR=2.177; 95%CI=1.235-3.838, p=0.007).
The median PFS (in months) for the entire cohort was 19.7 months. For the CCS and non-CCS group, the PFS was 23.2 months and 14.4 months, respectively (p=0.011) (Figure 4A).
Relationship between surgical outcome and survival. A. Kaplan-Meier estimate for progression-free survival (PFS) as shown for complete cytoreductive surgery (CCS) and non-CCS group (p<0.011). B. Kaplan-Meier for OS as shown for CCS or non-CCS group (p<0.000). OS: Overall survival.
The median OS (in months) for the entire cohort was 40.4 months, whereas it was 45.5 (95%CI=41.0-49.9) for the CCS group and 30.9 (95%CI=24.33-37.30) for the non-CCS group (p=0.000) (Figure 4B).
The impact of surgical extent and postoperative complications on OS. The SCS was higher in the CCS group, as shown in Table I.
The mean value of the surgical effort was lower for patients with a higher tumor burden; 6.8 for patients with S-PCI ≥18.5 versus 7.5 for those with PCI <18.5 (p=0.000). Small bowel and diaphragmatic carcinomatosis were more frequently represented in patients with S-PCI ≥18.5 (Figure 5). Small bowel carcinomatosis was related to poor survival (HR=1.86, 95%CI=1.07-3.21; p<0.026) in univariate analyses (Table V). This relationship was not sustained by multivariate Cox regression analyses, when adjusted for age, ECOG performance status, and completeness of surgery.
Intraoperative picture (cytoreductive surgery). Small bowel carcinomatosis in a patient with ovarian cancer.
Surgical characteristics of patients with peritoneal cancer index (PCI) below and above 18.5.
Major complications (Clavien-Dindo ≥3) were found in 27.5% of the patients with S-PCI<18.5 and 28.9% of the patients with S-PCI≥18.5 (p=0.539) (Table V). The Kaplan-Meier showed no significant differences in OS regarding complication status, with a median survival of 35 months for patients with major complications vs. 42 months for patients with no major complications (p=0.107).
Discussion
The aim of the study was to examine whether tumor extent, as scored by PCI, preoperatively by CT, and intraoperatively by surgeon’s ocular observation, is correlated to the prognosis of the patient. Our major finding was that regardless of the completeness of the CCS, the tumor extent at the beginning of surgery affected survival in patients with AOC.
There is plenty of evidence that besides FIGO stage, the most important prognostic factor in ovarian cancer is to obtain macroscopic radicality in upfront surgery (26-28). The extent of the tumor at the beginning of the surgery can be expressed as PCI, and this has an impact on the residual tumor as well (7, 29). In gastrointestinal cancer, PCI is accepted as a predictor for surgical outcome and survival in clinical practice, whereas in AOC, the role of PCI is disputed. However, if CT-PCI could predict the extent of the tumor in AOC, this could help the surgical team in decision-making concerning the treatment recommendations (30-33). In a previous study, we found that CT-PCI correlates well with S-PCI and risk of residual disease, which is in concordance with previous studies (14, 15, 22). Mazzei et al. postulated that CT can be used as the only technique to select patients for upfront surgery, if performed with a dedicated protocol and read by an expert radiologist (34).
S-PCI has been shown to be prognostic for survival in patients with AOC, which is in concordance with studies in both colorectal and ovarian cancer (35-37).
In our study, patients exhibiting a high S-PCI had a significantly higher risk of dying from the disease, even after adjusting the data for residual disease, age, and performance status. Hence, patient’s survival can be influenced not only by RD, but also by the tumor extent at the beginning of surgery. Our results are contradictory with those of another retrospective study of 80 patients who underwent CCS followed by chemotherapy by Gasimli et al. Their conclusion was, among many others, that the major prognostic factor is the residual tumor after the cytoreductive surgery, independent of the tumor pattern (38). Jönsdottir et al. presented that PCI above 24 correlated to a higher grade of complications, indicating that neoadjuvant chemotherapy should be considered (15).
The cut-off for OS in our study was 18.5, with significant differences in survival, regardless of the completeness of cytoreductive surgery. In order to find out if these differences in survival can be explained by more extensive surgery followed by higher complication rate, we separately investigated the patients exhibiting a PCI below and above 18.5. Grade 3 and 4 complications, as classified by Clavien-Dindo, occurred in 28.8% of the patients. No differences in complication rate were found between these two groups, nor any relationship between the complication rate and survival. The SCS and complication rate were higher in patients exhibiting PCI <18.5, showing a higher surgical effort due the intraoperative evaluation of complete surgery possibility. Small bowel carcinomatosis is a known risk factor for both residual disease and impaired survival (39, 40). In the PCI ≥18.5 group, 66.7% exhibited small bowel carcinomatosis, which reduced the ambition of the surgery extent, explaining the lower SCS, operation time, and blood loss and no differences in complication rate. In light of these results, the lower OS in patients with PCI ≥18.5 despite complete surgery, cannot be explained by the surgery extent and complication rate.
The role of CT-PCI as a prognostic factor for PFS and OS has been sparingly evaluated in ovarian cancer. The CT-PCI score alone or combined with other markers was found to correlate with OS in some studies (12, 17). In the current study, we found that CT-PCI is an independent prognostic factor for the recurrence risk. Our study failed to demonstrate a prognostic effect for CT-PCI regarding OS. The correlation was significant in crude data, but the adjusted hazard ratio of 1.037 was low, which makes the clinical significance of the result questionable. When patients exhibit a preoperative CT-PCI above 24.5 the risk of dying of the disease was two times higher than that of patients exhibiting a CT-PCI below 24.5. When the data were adjusted for age, ECOG performance status, and residual disease, a 50% reduction in the survival period, 28.2 months compared with 44.4 months for CT-PCI <24.5, respectively, was found. This multivariate analysis with a high hazard ratio of 1.51, did not reach statistical significance. Due to the relatively small study population, this result needs to be interpreted with caution. In a study by Avesani et al., a significant correlation between CT-PCI and survival was found, though the study population was different, including FIGO stages I and II as well, making further comparison with our study difficult (41). In malignant peritoneal mesothelioma, a median radiographic PCI of 25 was correlated with OS, though the data seemed not to be adjusted for age, comorbidity, and residual disease (42). We consider that the question whether CT-PCI is a prognostic factor needs further investigation, with a larger prospective study cohort.
These results support the idea that complete removal of the tumor is a very important prognostic factor, as shown by others (27). However, the patient´s age and performance status as well as tumor biology, are important factors to take into account in the decision-making on a treatment plan (43-45). In our study, the patient’s performance status and age were correlated with worse survival as well. An important factor with a major influence on survival is the breast cancer gene mutation (BRCA) status of the patient, which is highly related to both response to platinum-based chemotherapy and OS (46, 47). In a retrospective analysis of 325 patients with high-grade ovarian cancer, of which 107 (39.2%) exhibited BRCA mutations, a higher tumor extent than that in non-mutated cases was found (48). In our study, we had no access to BRCA mutation status for the whole cohort because BRCA mutation testing started routinely in May 2017. Besides the lack of the patients’ BRCA status, a limitation of this study is the retrospective nature of PCI documentation and the relatively small number of patients. The strength of this study is that the study was conducted in a single tertiary center with high CCS rate and highly specialized radiology department, which increases the accuracy of the CT scan evaluation. The centralized documentation system, including the histopathological records, made it possible to verify the S-PCI as a true carcinomatosis, which can be considered as a strength as well.
Conclusion
The tumor extent at the beginning of surgery seemed to affect OS in patients with AOC, regardless of residual disease at the end of the surgery. Patients exhibiting a high PCI had a higher risk of dying of the disease; at PCI above 18.5, the risk of dying was about two times higher compared with that when PCI was below 18.5. There was no difference in major postoperative complications when comparing patients with PCI above and below 18.5. CT-PCI seems to play a prognostic role for PFS; however, whether it acts as a prognostic factor for OS remains to be investigated.
Acknowledgements
The Authors would like to thank Anna Åkesson for the statistical support.
Footnotes
Authors’ Contributions
Mihaela Asp: Conceptualization, Investigation, Writing-Review & Editing, Validation, Visualization, Editing the manuscript, Formal analysis, Literature review. Hanna Sartor: Main contributor to the CT data collection, Validation, Reviewing, Supervision. Johan Bengtsson: interpreted the CT data. Susanne Malander: Validation, Reviewing, Supervision. Päivi Kannisto: Validation, Reviewing, Supervision, Founding acquisition, Project administration, Methodology.
Conflicts of Interest
The Authors declare that there are no conflicts of interest in relation to this study.
- Received March 18, 2022.
- Revision received March 30, 2022.
- Accepted April 1, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.