Abstract
Purpose: Lung metastasis is an independent risk factor affecting the prognosis of ovarian cancer patients. We developed and validated a nomogram to predict the risk of synchronous lung metastases in newly diagnosed ovarian cancer patients.
Methods: Data of ovarian cancer patients from the Surveillance, Epidemiology, and Final Results (SEER) database between 2010 and 2015 were retrospectively collected. The model nomogram was built on the basis of logistic regression. The consistency index (C-index) was used to evaluate the discernment of the synchronous lung metastasis nomogram. Calibration plots were drawn to analyze the consistency between the observed probability and predicted probability of synchronous lung metastases. The Kaplan–Meier method was used to estimate overall survival rate, and influencing factors were included in multivariate Cox regression analysis (P<0.05) to determine the independent prognostic factors of synchronous lung metastases.
Results: Overall, 16059 eligible patients were randomly divided into training (n=11242) and validation cohorts (n=4817). AJCC T, N stage, bone metastases, brain metastases, and liver metastases were evaluated as predictors of synchronous lung metastases. Finally, a nomogram was constructed. The nomogram based on independent predictors was calibrated and showed good discriminative ability. Mixed histological types, chemotherapy, and primary site surgery were factors affecting the overall survival of patients with synchronous lung metastases.
Conclusion: The clinical prediction model has high accuracy and can be used to predict lung metastasis risk in newly diagnosed ovarian cancer patients, which can guide the treatment of patients with synchronous lung metastases.
Keywords: Mitochondrial dysfunction, Mitochondria unfolded protein response (mtUPR), monitoring epidemiology final results, nomogram, receiver operating characteristic curve
Introduction
Ovarian cancer is among the most common malignant tumors in the female reproductive system. Ovarian cancer is the fifth most common cause of cancer-related deaths among American women. In 2018, an estimated 14070 people died of ovarian cancer in the United States [1]. Since the symptoms of ovarian cancer are unclear and there is currently no effective screening method, most patients are already at advanced stages (III and IV) at the time of diagnosis, accompanied by synchronous distant metastases [2,3].
Lung metastasis is the third most common distant metastatic site of ovarian cancer, accounting for 28.42% of distant metastatic sites. The location of distant metastases is an independent prognostic factor for overall survival [4]. Previous studies show that the risk factors for distant metastases are stage, grade, and lymph node involvement [5]. However, the sample size of the study was small. There are few studies on the risk factors of synchronous lung metastases, and most of them are case reports [6,7]. The median interval between the diagnosis of ovarian cancer and recording of metastatic disease was 44 months [5].
Identifying the risk factors for synchronous lung metastases can ensure that high-risk patients are thoroughly investigated at the initial diagnosis.These patients can then be treated as early as possible or provided with appropriate preventive treatment. A large number of studies and realistic evidence is also needed to determine the risk factors for synchronous lung metastases in ovarian cancer patients.
The purpose of the present study was to use Surveillance, Epidemiology, and End Results (SEER) database to characterize the prevalence, related factors, and prognostic factors of synchronous lung metastases in ovarian cancer patients. At the same time, a nomogram to predict the risk of synchronous lung metastases was developed on the basis of clinical factors, which may guide screening.
Methods
Study population
Data were obtained from the SEER database. The SEER *Stat 8.3.5 software (https://seer.cancer.gov/data/) was used to access the database. The site code was restricted to the ovary. Since the details of metastases were not recorded before 2010, patients with primary cancer of the ovary, aged ≥ 18 years at diagnosis, between 2010 and 2015 were analyzed. The exclusion criteria for patient selection included the following: (1) unknown grade; (2) unknown AJCC T, N stage and AJCC T0 stage; (3) unknown metastases information; (4) unknown tumor size; (5) unknown laterality; and (6) unknown therapy information. The flowchart of the subjects’ selection is listed in Figure 1. According to the inclusion and exclusion criteria, 16059 patients with ovarian cancer were finally enrolled in our study. We further randomly divided the patients in a 7:3 ratio to form a training cohort (n=11242) for nomogram construction and a validation cohort (n=4817) for internal verification.
Figure 1. Flowchart of patients’ selection.
Data regarding clinical characteristics including age, race, marital status, insurance status, year of diagnosis, household income at diagnosis, histological type, grade, laterality, clinical AJCC T, N stage, tumor size, metastatic status, and therapy information were collected from the SEER database. Since all information from the SEER database was identified and no personal identifying information was used in this analysis, informed consent was not required. The present study complied with the 1964 Helsinki Declaration, its later amendments, and comparable ethical standards.
Statistical analysis
Statistical analysis was performed using the SPSS 21 software. Categorical data were presented as frequency (%) and analyzed using the chi-squared test. The Kolmogorov–Smirnov test was used to verify the normality of variables. Normally distributed variables were expressed as mean ± standard deviation, while non-normally distributed variables were expressed as median (interquartile range). Hazard ratios and 95% confidence intervals (CIs) were calculated. Univariate and multivariate logistic regression analyses were used to determine the risk factors of synchronous lung metastases in patients with ovarian cancer. Factors with a P-value less than 0.05 were incorporated into the multivariable logistic regression model.
A synchronous lung metastases nomogram was formulated on the basis of the results of multivariate logistic analysis using the rms package in R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria; www.r-project.org). Receiver operating characteristic (ROC) curves were drawn. Finally, we evaluated the stability of the prognostic nomogram and the synchronous lung metastasis nomogram by internal validation with 1000 bootstrap samples. The nomograms were validated both internally and externally. The C-index (Harrell’s concordance index) was used to assess the exact predicted values of nomograms. Calibration plots were drawn to analyze the consistency between the observed and predicted probabilities. Overall survival was estimated by the Kaplan–Meier method, and the difference between distinct groups was compared using the log-rank test. A multivariable Cox regression model, incorporating the significant factors in the Kaplan–Meier method (P<0.05) was conducted to analyze the independent prognostic factors for synchronous lung metastases.
Results
Patients’ basic information
According to the inclusion and exclusion criteria, data of 16059 of the 35333 ovarian cancer patients registered between 2010 and 2015 were collected from the SEER database. The patients were divided into training (n=11242) and verification (n=4817) groups. The basic information of the patients is listed in Table 1. The median age of the patients was 59 years. Among these patients, 13223 (82.3%) were white, 1057 (6.6%) were black, and 1711 (10.7%) were of other races. A total of 3377 (21.0%) patients were unmarried, 8549 (53.2%) were married, and 3486 (21.7%) were separated. The number of insured and uninsured patients was 861 (3.5%) and 15337 (95.5%), respectively. The median household income was 6255. The number of patients with tumor diameters <2 cm, 2–5 cm, >5 cm was 1311 (8.1%), 2678 (16.7%), and 12076 (75.2%), respectively. A total of 4947 (30.8%) patients had tumors on the left, 5109 (31.8%) patients with tumors on the right, and 6003 (37.4%) patients with tumors on both sides. The number of well differentiated, moderately differentiated, poorly differentiated, and undifferentiated histology tumors was 2011 (12.5%), 2758 (17.2%), 6395 (39.8%), and 4895 (30.5%), respectively. The number of T1, T2, and T3 stage tumors was 5500 (34.2%), 2552 (15.9%), and 8007 (49.9%), respectively. The numbers of N0 and N1 stages were 12514 (77.9%) and 3545 (22.1%), respectively. Bone metastases occurred in 54 (0.3%), brain metastases in 15 (0.1%), liver metastases in 572 (3.6%), and lung metastases in 411 (2.6%) patients. The number of histology-type serous, endometrioid, mucinous, clear cell, carcinosarcoma, malignant Brenner, carcinoma, NOS, mixed, and other was 8644 (53.8%), 2367 (14.7%), 1071 (6.7%), 1124 (7.0%), 515 (3.2%), 18 (0.1%), 516 (3.2%), 1140 (7.1%), and 664 (4.1%), respectively. The chi-square test for all variables between the two groups yielded P>0.05.
Table 1. Demographical and clinical characteristics between patient with the training cohort and validation cohort.
| Variables | The training cohort (n=11242) | The validation cohort (n=4817) | Total (n=16059) | P-value | |||
|---|---|---|---|---|---|---|---|
| Number | % | Number | % | Number | % | ||
| Age | 59 | 59 | 59 | 0.360 | |||
| Race | 0.750 | ||||||
| White | 9267 | 82.4 | 3956 | 82.1 | 13223 | 82.3 | |
| Black | 725 | 6.4 | 332 | 6.9 | 1057 | 6.6 | |
| Other (American Indian/AK Native, Asian/Pacific Islander) | 1201 | 10.7 | 510 | 10.6 | 1711 | 10.7 | |
| Unknown | 49 | 0.4 | 19 | 0.4 | 68 | 0.4 | |
| Marital status | 0.363 | ||||||
| Unmarried | 2329 | 20.7 | 1049 | 21.8 | 3377 | 21.0 | |
| Married | 5987 | 53.3 | 2562 | 53.2 | 8549 | 53.2 | |
| Separated | 2473 | 22.0 | 1013 | 21.0 | 3486 | 21.7 | |
| Unknown | 453 | 4.0 | 194 | 4.0 | 647 | 4.0 | |
| Insurance status | 0.577 | ||||||
| Uninsured | 403 | 3.6 | 158 | 3.3 | 561 | 3.5 | |
| Insured | 10724 | 95.4 | 4613 | 95.8 | 15337 | 95.5 | |
| Unknown | 115 | 1.0 | 46 | 1.0 | 161 | 1.0 | |
| Household income | 6204 (5716–8008) | 6325 (5716–8008) | 6255 (5716–8008) | 0.394 | |||
| Year of diagnosis | 0.210 | ||||||
| 2010 | 1783 | 15.9 | 755 | 15.7 | 2539 | 15.8 | |
| 2011 | 1850 | 16.5 | 806 | 16.7 | 2656 | 16.5 | |
| 2012 | 1825 | 16.2 | 816 | 16.9 | 2641 | 16.4 | |
| 2013 | 1873 | 16.7 | 825 | 17.1 | 2698 | 16.8 | |
| 2014 | 1951 | 17.4 | 759 | 15.8 | 2710 | 16.9 | |
| 2015 | 19610 | 17.4 | 856 | 17.8 | 2816 | 17.5 | |
| Tumor size | 0.892 | ||||||
| <2 cm | 906 | 8.1 | 399 | 8.3 | 1311 | 8.1 | |
| 2–5 cm | 1875 | 16.7 | 803 | 16.7 | 2678 | 16.7 | |
| >5 cm | 8461 | 75.3 | 3615 | 75.0 | 12076 | 75.2 | |
| Laterality | 0.628 | ||||||
| Left | 3471 | 30.9 | 1476 | 30.6 | 4947 | 30.8 | |
| Right | 35965 | 32.0 | 1514 | 31.4 | 5109 | 31.8 | |
| Bilateral | 4176 | 37.1 | 18287 | 37.9 | 6003 | 37.4 | |
| Grade | 0.426 | ||||||
| Well differentiated | 1417 | 12.6 | 594 | 12.3 | 2011 | 12.5 | |
| Moderately differentiated | 1904 | 16.9 | 854 | 17.7 | 2758 | 17.2 | |
| Poorly differentiated | 4460 | 39.7 | 1935 | 40.2 | 6395 | 39.8 | |
| Undifferentiated | 34621 | 30.8 | 1434 | 29.8 | 4895 | 30.5 | |
| AJCC T stage | 0.805 | ||||||
| T1 | 3835 | 34.1 | 1665 | 34.6 | 5500 | 34.2 | |
| T2 | 1783 | 15.9 | 769 | 16.0 | 2552 | 15.9 | |
| T3 | 5624 | 50.0 | 2383 | 49.5 | 8007 | 49.9 | |
| AJCC N stage | 0.497 | ||||||
| N0 | 8747 | 77.8 | 37710 | 78.3 | 12514 | 77.9 | |
| N1 | 2498 | 22.2 | 1047 | 21.7 | 3545 | 22.1 | |
| Bone metastasis | 0.592 | ||||||
| No | 1120 | 99.7 | 4799 | 99.6 | 16005 | 99.7 | |
| Yes | 36 | 0.3 | 18 | 0.4 | 54 | 0.3 | |
| Brain metastasis | 0.159 | ||||||
| No | 11229 | 99.9 | 4815 | 100.0 | 16044 | 99.9 | |
| Yes | 13 | 0.1 | 2 | 0.0 | 15 | 0.1 | |
| Liver metastasis | 0.681 | ||||||
| No | 10846 | 96.5 | 4641 | 96.3 | 15487 | 96.4 | |
| Yes | 396 | 3.5 | 176 | 3.7 | 572 | 3.6 | |
| Lung metastasis | 0.681 | ||||||
| No | 10959 | 97.5 | 4689 | 97.3 | 15648 | 97.4 | |
| Yes | 283 | 2.5 | 128 | 2.7 | 411 | 2.6 | |
| Histological type | 0.866 | ||||||
| Serous | 6016 | 53.5 | 2628 | 54.6 | 8644 | 53.8 | |
| Endometrioid | 1662 | 14.8 | 705 | 14.6 | 2367 | 14.7 | |
| Mucinous | 758 | 6.7 | 313 | 6.5 | 1071 | 6.7 | |
| Clear cell | 779 | 6.9 | 345 | 7.2 | 1124 | 7.0 | |
| Carcinosarcoma | 361 | 3.2 | 154 | 3.2 | 515 | 3.2 | |
| Malignant Brenner | 14 | 0.1 | 4 | 0.1 | 18 | 0.1 | |
| Carcinoma, NOS | 362 | 3.2 | 154 | 3.2 | 516 | 3.2 | |
| Mixed | 813 | 7.2 | 327 | 6.8 | 1140 | 7.1 | |
| Other | 477 | 4.2 | 187 | 3.9 | 664 | 4.1 | |
| Surgery (primary) | 0.292 | ||||||
| No | 149 | 72.0 | 58 | 70.0 | 207 | 70.0 | |
| Yes | 11093 | 28.0 | 4759 | 30.0 | 15852 | 30.0 | |
| Radiation | 0.493 | ||||||
| No | 11090 | 70.0 | 4751 | 69.7 | 15841 | 70.0 | |
| Yes | 152 | 30.0 | 66 | 30.0 | 218 | 30.0 | |
| Chemotherapy | 0.841 | ||||||
| No | 2753 | 70.1 | 1172 | 70.0 | 3925 | 70.0 | |
| Yes | 8489 | 29.9 | 3645 | 30.0 | 12134 | 30.0 | |
Risk factors for lung metastasis
Univariable logistic analysis showed that factors closely related to the occurrence of lung metastasis included the following: older patient age (OR = 1.015; 95% CI, 1.006–1.025; P=0.001), bilateral tumors (OR = 1.556; 95% CI, 1.179–2.053; P=0.002), lower differentiation grade (poorly differentiated OR = 5.288; 95% CI, 2.583–10.825; P≤0.001; undifferentiated OR = 6.435; 95% CI, 3.139–13.195; P≤0.001), higher AJCC T stage (T2 OR = 4.991; 95% CI, 2.859–8.712; P≤0.001; T3 OR = 8.796; 95% CI, 5.432–14.243; P<0.001), higher AJCC stage N (OR = 2.863; 95% CI, 2.254–3.635; P<0.001), bone (OR = 15.403; 95% CI, 7.355–32.256; P<0.001), brain (OR = 17.443; 95% CI, 5.340–56.981; P<0.001), liver metastases (OR = 10.483; 95% CI, 7.822–14.050; P<0.001), and mucinous (OR = 0.425; 95% CI, 0.190–0.953; P=0.038) and clear cell histological subtypes (OR = 0.248; 95% CI, 0.077–0.794; P=0.019).
Multivariable logistic regression analysis showed that higher T and N stages, and the presence of bone, liver, and brain metastases were associated with the earlier development of synchronous lung metastases (Table 2).
Table 2. Univariable and multivariable logistic regression for analyzing the associated factors for developing lung metastases in training cohort.
| Variables | Univariable | Multivariable | ||||
|---|---|---|---|---|---|---|
| OR | 95% Cl | P-value | OR | 95% Cl | P-value | |
| Age | 1.015 | 1.006–1.025 | 0.001 | 1.010 | 0.999–1.021 | 0.086 |
| Race | 0.622 | 0.114 | ||||
| White | References | References | ||||
| Black | 1.193 | 0.758–1.878 | 0.445 | 1.174 | 0.725–1.899 | 0.515 |
| Other | 1.236 | 0.865–1.767 | 0.244 | 1.595 | 1.089–2.336 | 0.016 |
| Unknown | 0.000 | 0 .000 | 0.998 | 0.000 | 0.000 | 0.997 |
| Marital status | 0.287 | 0.144 | ||||
| Unmarried | References | References | ||||
| Married | 1.034 | 0.761–1.404 | 0.832 | 0.868 | 0.626–1.203 | 0.396 |
| Separated | 1.074 | 0.751–1.535 | 0.697 | 0.770 | 0.520–1.139 | 0.191 |
| Unknown | 0.437 | 0.174–1.096 | 0.078 | 0.352 | 0.136–0.912 | 0.032 |
| Insurance status | 0.772 | 0.418 | ||||
| Uninsured | References | References | ||||
| Insured | 0.913 | 0.496–1.683 | 0.772 | 0.730 | 0.378–1.411 | 0.349 |
| Others/Unknown | 1.284 | 0.401–4.111 | 0.674 | 1.221 | 0.357–4.174 | 0.750 |
| Household income | 1.000 | 1.000–1.000 | 0.794 | 1.000 | 1.000–1.000 | 0.622 |
| Year of diagnosis | 0.666 | 0.742 | ||||
| 2010 | References | References | ||||
| 2011 | 1.293 | 0.835–2.001 | 0.250 | 1.283 | 0.817–2.015 | 0.278 |
| 2012 | 1.367 | 0.886–2.109 | 0.158 | 1.389 | 0.889–2.172 | 0.149 |
| 2013 | 1.140 | 0.729–1.784 | 0.565 | 1.166 | 0.736–1.847 | 0.51 |
| 2014 | 1.381 | 0.902–2.117 | 0.138 | 1.348 | 0.865–2.100 | 0.187 |
| 2015 | 1.323 | 0.860–2.033 | 0.202 | 1.197 | 0.765–1.873 | 0.432 |
| Tumor size | 0.080 | 0.207 | ||||
| <2 cm | References | References | ||||
| 2–5 cm | 1.291 | 0.717–1.822 | 0.303 | 1.026 | 0.620–1.696 | 0.921 |
| >5 cm | 0.925 | 0.541–1.240 | 0.726 | 0.796 | 0.505–1.255 | 0.326 |
| Laterality | <0.001 | 0.082 | ||||
| Left | References | References | ||||
| Right | 0.704 | 0.500–0.991 | 0.044 | 0.666 | 0.467–0.951 | 0.025 |
| Bilateral | 1.556 | 1.179–2.053 | 0.002 | 0.840 | 0.620–1.138 | 0.261 |
| Grade | <0.001 | 0.000 | 0.246 | |||
| Well differentiated | References | References | ||||
| Moderately differentiated | 2.154 | 0.960–4.829 | 0.063 | 1.355 | 0.592–3.101 | 0.471 |
| Poorly differentiated | 5.288 | 2.583–10.825 | <0.001 | 1.590 | 0.736–3.437 | 0.238 |
| Undifferentiated | 6.435 | 3.139–13.195 | <0.001 | 1.890 | 0.868–4.118 | 0.109 |
| AJCC T stage | <0.001 | References | <0.001 | |||
| T1 | References | |||||
| T2 | 4.991 | 2.859–8.712 | <0.001 | 3.032 | 1.675–5.485 | <0.001 |
| T3 | 8.796 | 5.432–14.243 | <0.001 | 4.055 | 2.343–7.019 | <0.001 |
| AJCC N stage | <0.001 | <0.001 | ||||
| N0 | References | References | ||||
| N1 | 2.863 | 2.254–3.635 | <0.001 | 1.696 | 1.313–2.190 | <0.001 |
| Bone metastasis | <0.001 | <0.001 | ||||
| No | References | References | ||||
| Yes | 15.403 | 7.355–32.256 | <0.001 | 5.945 | 2.373–14.894 | <0.001 |
| Brain metastasis | <0.001 | <0.001 | ||||
| No | References | References | ||||
| Yes | 17.443 | 5.340–56.981 | <0.001 | 13.375 | 3.449–51.865 | <0.001 |
| Liver metastasis | <0.001 | <0.001 | <0.001 | |||
| No | References | References | ||||
| Yes | 10.483 | 7.822–14.050 | <0.001 | 6.292 | 4.605–8.598 | <0.001 |
| Histological type | <0.001 | 0.503 | ||||
| Serous | References | References | ||||
| Endometrioid | 1.581 | 0.832–3.005 | 0.162 | 0.807 | 0.449–1.450 | 0.473 |
| Mucinous | 0.425 | 0.190–0.953 | 0.038 | 0.526 | 0.180–1.535 | 0.240 |
| Clear cell | 0.248 | 0.077–0.794 | 0.019 | 0.536 | 0.240–1.195 | 0.127 |
| Carcinosarcoma | 0.423 | 0.160–1.120 | 0.083 | 1.513 | 0.879–2.605 | 0.135 |
| Malignant Brenner | 2.166 | 0.971–4.831 | 0.059 | 0.000 | 0.000 | 0.999 |
| Carcinoma, NOS | 0.000 | 0.000 | 0.999 | 1.184 | 0.659–2.125 | 0.572 |
| Mixed | 2.160 | 0.968–4.817 | 0.060 | 1.052 | 0.632–1.751 | 0.844 |
| Other | 1.057 | 0.484–2.310 | 0.889 | 1.172 | 0.547–2.512 | 0.682 |
Bold values indicate statistical significance (P<0.05).
Nomogram development
A nomogram to predict synchronous lung metastases in patients with ovarian cancer was developed in the training cohort. The risk factors determined by multivariable logistic regression analysis, including higher T and N stage, and the development of bone, liver, and brain metastases were developed and used as the final nomogram (Figure 2).
Figure 2. Nomogram for predicting synchronous lung metastases in ovarian cancer patients.
A synchronous lung metastases nomogram was formulated on the basis of the results of multivariable logistic analysis using the rms package in R version 3.6.1. The first line shows the point assignment of each variable. Lines 2–6 indicate the variables included in the nomogram. For individual patients, each variable is assigned a point value based on tumor characteristics. The points assigned to each of the five variables are added, and the total points are displayed in the seventh line. The bottom row shows the possibility of synchronous lung metastases.
ROC curves analysis and prediction value evaluation
ROC curves were drawn to determine the predicted value of the nomogram of synchronous lung metastases in the training and validation cohorts. As shown in Figure 3A,C, ROC curves were drawn. We verified the nomogram internally and externally. The C-index was used to evaluate the prediction accuracy of the nomogram. As shown in Figure 3B, the internal verification of the nomogram was performed, and the C-index was 0.761 (0.736–0.787). As shown in Figure 3D, the external verification of the validation cohort showed that the C index was 0.757 (−0.718 to 0.795). Verification of the nomogram showed agreement with the predicted values.
Figure 3. Identification and calibration of the nomogram in the training and verification cohorts.
(A) ROC curve for discrimination in the training cohorts. (B) Calibration plots for the actual (observed) and predicted probabilities of the nomograms in the training cohorts. (C) ROC curve for discrimination in the validation cohorts. (D) Calibration plots for the actual (observed) and predicted probabilities of the nomograms in the validation cohorts. Using the rms package in R version 3.6.1, the ROC curve and calibration diagram were drawn. (A,C) ROC curve for discrimination in the training and validation cohorts. (B,D) Calibration plots for the actual (observed) and predicted probabilities of the nomograms in the training and verification cohorts. The x-axis represents the predicted probability of the nomogram measured by logistic regression analysis, and the y-axis represents the actual probability. The vertical line represents the frequency distribution of the predicted probabilities. The dashed line represents the ideal reference line, where the predicted probability matches the observed probability. Calibration plots showed excellent calibration of the nomogram.
Survival analysis and prognostic factors of synchronous lung metastases
The 3- and 5-year overall survival rates of ovarian cancer patients were 72.2 and 58.1%, respectively. For the 411 patients with newly diagnosed lung metastases, the 3- and 5-year survival rates were 33.8 and 22.8%, respectively (Figure 4A). Kaplan–Meier analysis showed that the overall survival of married patients (Figure 4B, P=0.021), primary site surgery (Figure 4C, P<0.01), chemotherapy (Figure 4D, P<0.01), and radiation (Figure 4E, P=0.030) were higher than those of the control group. Mixed histological type (Figure 4F, P<0.001), liver metastases (Figure 4G, P=0.025), bone metastases (Figure 4H, P=0.028), and brain metastases (Figure 4I, P=0.003) correlated negatively with overall survival rate. Kaplan–Meier analysis was used to estimate the overall survival rate. The influencing factors selected by the Kaplan–Meier method were included in the multivariate Cox regression (P<0.05) to analyze the independent prognostic factors of synchronous lung metastases. Mixed histological types (P<0.001), chemotherapy (P<0.001), and primary site surgery (P<0.001) affected the overall survival of ovarian cancer patients with synchronous lung metastases (Table 3).
Figure 4. Kaplan–Meier analysis of the overall survival of ovarian cancer patients with lung metastasis.
The overall survival (OS) rate was estimated by the Kaplan–Meier method, and the log-rank test was used to compare the differences between different groups. (A) OS rate of the total population. (B) OS rates stratified by marital status. (C) OS rates stratified by primary site surgery. (D) OS rates stratified by chemotherapy. (E) OS rates stratified by radiation. (F) OS rates stratified by histological type. (G) OS rates stratified by the presence of liver metastases. (H) OS rates stratified by the presence of bone metastases. (I) OS rates stratified by the presence of brain metastases. Multivariate Cox regression results incorporating the above important factors showed that mixed histological type (hazard ratio [HR] = 2.531; 95% CI: 1.538–4.165; P<0.001) was positively correlated with overall mortality. Primary site surgery (HR = 0.315; 95% CI: 0.190–0.522; P<0.001) and chemotherapy (HR = 0.216; 95% CI: 0.139–0.335; P<0.001) were beneficial for survival (Table 3).
Table 3. Multivariable Cox regression for analyzing the associated factors for prognostic factors patients with lung metastases.
| Variables | Multivariable | ||
|---|---|---|---|
| OR | 95% Cl | P-value | |
| Histological type | 0.003 | ||
| Serous | References | ||
| Endometrioid | 1.391 | 0.764–2.532 | 0.280 |
| Mucinous | 1.081 | 0.391–2.990 | 0.880 |
| Clear cell | 1.106 | 0.515–2.374 | 0.797 |
| Carcinosarcoma | 1.457 | 0.854–2.488 | 0.167 |
| Carcinoma, NOS | 0.981 | 0.575–1.673 | 0.942 |
| Mixed | 2.531 | 1.538–4.165 | <0.001 |
| Other | 2.585 | 1.362–4.908 | 0.004 |
| Surgery (primary) | <0.001 | ||
| No | Reference | ||
| Yes | 0.315 | 0.190–0.522 | <0.001 |
| Chemotherapy | <0.001 | ||
| No | Reference | ||
| Yes | 0.216 | 0.139–0.335 | <0.001 |
Bold values indicate statistical significance (P<0.05).
Discussion
Ovarian cancer is the seventh most common cancer among women and the eighth most common cause of cancer death worldwide, with a 5-year overall survival rate of <50% [8]. Two-thirds of the patients are already at advanced stages at the time of diagnosis (Stage III/IV) [9]. When the lungs are affected, the main route of metastasis is through the pleura. Lung metastases usually represent as visceral pleura involvement and continuous infiltration. Occasionally, isolated lesions are observed. Invasion of lymphatic and blood vessels also occurs [10]. The incubation period from the diagnosis of ovarian cancer to the development of lung metastases can be as long as 108 months [11]. Compared with standard chemotherapy treatment alone, early detection of lung metastases can increase the chances of timely, more aggressive treatments, which may lead to prolonged survival [4]. Active chemotherapy can significantly reduce the tumor load and metastasis of ovarian cancer [12]. Surgical removal of isolated lung metastatic lesions is reasonable [13]. Targeted therapy is also a promising treatment for metastatic ovarian cancer [14]. Routine imaging studies, such as computed tomography or magnetic resonance imaging, have not shown high sensitivity and specificity when diagnosing micrometastases <1 cm [15]. Therefore, there is a need for a non-invasive method to predict the likelihood of synchronous lung metastases in ovarian cancer patients. We used data from the SEER database to develop and validate the predicted nomogram, which demonstrated significant discernment and calibration capabilities and can provide a personalized estimation of the likelihood of synchronous lung metastases in ovarian cancer patients.
To the best of our knowledge, the present study is the first to generate a risk model based on clinical and tumor characteristics through population-based surveillance, epidemiology, and final result databases to predict the risk of synchronous lung metastases in newly diagnosed ovarian cancer patients. We found that the higher the AJCC T and N stages, the higher the likelihood of metastases which is similar to likelihood of bone metastasis of ovarian cancer and the findings of other types of tumor metastases research [16–18]. Previous studies have shown that poor differentiation and lymph node involvement are risk factors for distant metastasis [4]. We found that liver metastases, brain metastases, and bone metastases are risk factors for synchronous lung metastases. If distant metastases are found in other parts of the body, it means that the cancer has metastasized [19], and the probability of lung metastases is higher.
We verified the nomogram internally and externally. The nomogram of synchronous lung metastases includes five factors: AJCC T stage, AJCC N stage, bone metastases, liver metastases, and brain metastases. The nomogram showed agreement between the predicted results and the observed results in the verification. In addition, the C-indices of internal verification and external verification of the nomogram were 0.761 (0.736–0.787) and 0.757 (0.718–0.795), respectively, indicating consistency with the predicted values. For patients with a higher risk of synchronous metastases predicted by this model, imaging examination should be performed on time to diagnose the occurrence of lung metastases in the initial period, so as to better guide clinical procedures.
The determination of prognostic factors related to synchronous lung metastases in these patients may help doctors to provide targeted treatment strategies for patients at different risk levels and improve patient survival and quality of life. Previous studies have shown that lung metastases can significantly worsen the prognosis of patients [20]. The median survival time for the diagnosis of distant disease is 12 months [5]. In this study, the 3- and 5-year survival rates for 411 patients with synchronous lung metastases were 33.8 and 22.8%, respectively, similar to other studies [21,22]. Primary site surgical treatment and chemotherapy can improve overall survival. Therefore, for patients with ovarian cancer with synchronous lung metastases, active surgery, and chemotherapy are encouraged. At the same time, the mixed histological type is a high-risk factor for mortality, and physicians should attach great importance to it. The present study has several limitations that should be noted. The main limitation is that the variables used to construct the nomogram only used clinico-pathological features because there were no important tumor biomarkers in the SEER database. Another limitation is that although the established nomogram shows good discrimination and verification capabilities, it still requires further verification based on large-scale external queues. Third, only patients with synchronous lung metastases were analyzed. Since they may not be recorded in the SEER databases, metachronous lung metastases that occurred later in the disease were not analyzed. This was a retrospective study. The patients were selected from the hospital, so there was a selection bias.
Conclusion
Lung metastasis is an independent risk factor affecting the prognosis of patients with ovarian cancer. In the first diagnosis of ovarian cancer, early detection of synchronous lung metastases through routine screening is beneficial for high-risk patients.
The present study is the first to use population-based SEER database to generate a risk model based on clinical and tumor characteristics to predict the risk of synchronous lung metastases in newly diagnosed ovarian cancer patients with high accuracy. The present study preliminarily determined the prognostic factors related to synchronous lung metastases in patients with ovarian cancer, which will help doctors to provide targeted treatment strategies for patients at different risk levels and improve the survival rate and quality of life of patients.
Abbreviations
- AJCC
American Joint Committee on Cancer
- CI
confidence interval
- C-index
consistency index/Harrell’s concordance index
- OR
odds ratio
- ROC
receiver operating characteristic
- SEER
Surveillance, Epidemiology, and Final Results
Data Availability
https://seer.cancer.gov/data/ is available for the Surveillance, Epidemiology, and End Results Program database.
Competing Interests
The authors declare that there are no competing interests associated with the manuscript.
Funding
This work was supported by the National Natural Science Foundation Project [grant number 81771615].
Author Contribution
(I) Conception and design: Yufei Yuan. (II) Administrative support: Guiqin Bai. (III) Provision of study materials or patients: Fanfan Guo. (IV) Collection and assembly of data: Yufei Yuan and Yidan Zhang. (V) Data analysis and interpretation: Ruoran Wang. (VI) Manuscript writing: Yufei Yuan. (VII) Final approval of manuscript: all authors.
Ethics Approval
Since the data collected from the Surveillance, Epidemiology, and End Results: data were anonymized and de-identified prior to release, informed patient consent was not required in our study.
References
- 1.Torre L.A., Trabert B., DeSantis C.E., Miller K.D., Samimi G., Runowicz C.D. et al. (2018) Ovarian cancer statistics, 2018. CA Cancer J. Clin. 68, 284–296 10.3322/caac.21456 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Weidle U.H., Birzele F., Kollmorgen G. and Rueger R. (2016) Mechanisms and targets involved in dissemination of ovarian cancer. Cancer Genomics Proteomics 13, 407–423 10.21873/cgp.20004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jelovac D. and Armstrong D.K. (2011) Recent progress in the diagnosis and treatment of ovarian cancer. CA Cancer J. Clin. 61, 183–203 10.3322/caac.20113 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Deng K., Yang C., Tan Q., Song W., Lu M., Zhao W. et al. (2018) Sites of distant metastases and overall survival in ovarian cancer: a study of 1481 patients. Gynecol. Oncol. 150, 460–465 10.1016/j.ygyno.2018.06.022 [DOI] [PubMed] [Google Scholar]
- 5.Cormio G., Rossi C., Cazzolla A., Resta L., Loverro G., Greco P. et al. (2003) Distant metastases in ovarian carcinoma. Int. J. Gynecol. Cancer 13, 125–129 10.1136/ijgc-00009577-200303000-00004 [DOI] [PubMed] [Google Scholar]
- 6.Ayub I.I., Thangaswamy D., Joseph L.D. and Meenakshisundaram M. (2018) Lung parenchymal and endobronchial metastases from ovarian carcinoma. J. Bronchol. Interv. Pulmonol. 25, 235–238 10.1097/LBR.0000000000000463 [DOI] [PubMed] [Google Scholar]
- 7.Cheng J., Ou X. and Ye Y. (2018) Prominent pulmonary metastases without concurrent intra-abdominal metastasis in a case of ovarian carcinoma. Clin. Nucl. Med. 43, 349–351 10.1097/RLU.0000000000002045 [DOI] [PubMed] [Google Scholar]
- 8.Webb P.M. and Jordan S.J. (2017) Epidemiology of epithelial ovarian cancer. Best Pract. Res. Clin. Obstet. Gynaecol. 41, 3–14 10.1016/j.bpobgyn.2016.08.006 [DOI] [PubMed] [Google Scholar]
- 9.Johnson C. and Jazaeri A.A. (2018) Diagnosis and management of immune checkpoint inhibitor-related toxicities in ovarian cancer: a series of case vignettes. Clin. Ther. 40, 389–394 10.1016/j.clinthera.2018.02.011 [DOI] [PubMed] [Google Scholar]
- 10.Julian C.G., Goss J., Blanchard K. and Woodruff J.D. (1974) Biologic behavior of primary ovarian malignancy. Obstet. Gynecol. 44, 873–884 [PubMed] [Google Scholar]
- 11.Fuller A.F. Jr, Scannell J.G. and Wilkins E.W. Jr (1985) Pulmonary resection for metastases from gynecologic cancers: Massachusetts General Hospital experience, 1943-1982. Gynecol. Oncol. 22, 174–180 10.1016/0090-8258(85)90024-1 [DOI] [PubMed] [Google Scholar]
- 12.Anderson K., Lawson K.A., Simmons-Menchaca M., Sun L., Sanders B.G. and Kline K. (2004) Alpha-TEA plus cisplatin reduces human cisplatin-resistant ovarian cancer cell tumor burden and metastasis. Exp. Biol. Med. (Maywood) 229, 1169–1176 10.1177/153537020422901112 [DOI] [PubMed] [Google Scholar]
- 13.Kerr V.E. and Cadman E. (1985) Pulmonary metastases in ovarian cancer. Analysis of 357 patients. Cancer 56, 1209–1213 10.1002/1097-0142(19850901)56:5<1209::AID-CNCR2820560542>3.0.CO;2-Y [DOI] [PubMed] [Google Scholar]
- 14.Satpathy M., Wang L., Zielinski R.J., Qian W., Wang Y.A., Mohs A.M. et al. (2019) Targeted drug delivery and image-guided therapy of heterogeneous ovarian cancer using HER2-targeted theranostic nanoparticles. Theranostics 9, 778–795 10.7150/thno.29964 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fischerova D. and Burgetova A. (2014) Imaging techniques for the evaluation of ovarian cancer. Best Pract. Res. Clin. Obstet. Gynaecol. 28, 697–720 10.1016/j.bpobgyn.2014.04.006 [DOI] [PubMed] [Google Scholar]
- 16.Zhang C., Guo X., Peltzer K., Ma W., Qi L., Zhang Y. et al. (2019) The prevalence, associated factors for bone metastases development and prognosis in newly diagnosed ovarian cancer: a large population based real-world study. J. Cancer 10, 3133–3139 10.7150/jca.30335 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hou G., Zheng Y., Wei D., Li X., Wang F., Tian J. et al. (2019) Development and validation of a SEER-based prognostic nomogram for patients with bone metastatic prostate cancer. Medicine (Baltimore) 98, e17197 10.1097/MD.0000000000017197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gardner A.B., Charo L.M., Mann A.K., Kapp D.S., Eskander R.N. and Chan J.K. (2020) Ovarian, uterine, and cervical cancer patients with distant metastases at diagnosis: most common locations and outcomes. Clin. Exp. Metastasis 37, 107–113 10.1007/s10585-019-10007-0 [DOI] [PubMed] [Google Scholar]
- 19.Barbolina M.V. (2018) Molecular mechanisms regulating organ-specific metastases in epithelial ovarian carcinoma. Cancers (Basel) 10, 444 10.3390/cancers10110444 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhang Y., Guo X., Wang G., Ma W., Liu R., Han X. et al. (2018) Real-world study of the incidence, risk factors, and prognostic factors associated with bone metastases in women with uterine cervical cancer using Surveillance, Epidemiology, and End Results (SEER) Data Analysis. Med. Sci. Monit. 24, 6387–6397 10.12659/MSM.912071 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang X., Li X., Su S. and Liu M. (2017) Marital status and survival in epithelial ovarian cancer patients: a SEER-based study. Oncotarget 8, 89040–89054 10.18632/oncotarget.21648 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Peres L.C., Cushing-Haugen K.L., Kobel M., Harris H.R., Berchuck A., Rossing M.A. et al. (2019) Invasive epithelial ovarian cancer survival by histotype and disease stage. J. Natl. Cancer Inst. 111, 60–68 10.1093/jnci/djy071 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
https://seer.cancer.gov/data/ is available for the Surveillance, Epidemiology, and End Results Program database.