Abstract
Background/Aim: Life expectancy is considered in treatment decision-making for non-metastatic prostate cancer (PCa). We explored the factors related to overall survival (OS) and investigated the association between OS and life expectancy in patients with non-metastatic PCa according to various treatment modalities. Patients and Methods: This retrospective study included 714 patients with non-metastatic PCa between 2006 and 2010 at our institute. The treatment modalities were classified as follows: radical prostatectomy (RP), androgen deprivation therapy (ADT), brachytherapy (BT) and external beam radiation therapy (EBRT). We defined life expectancy according to an abridged life table published by the Ministry of Health, Labour and Welfare in Japan. Results: The median age and initial prostate-specific antigen levels at treatment were 71 years and 8.8 ng/mL, respectively. Advanced age, Charlson comorbidity index (CCI) ≥1, cT3a stage and ADT were independent poor prognostic factors for OS. OS and life expectancy did not significantly differ in all patients (p=0.32) and in the National Comprehensive Cancer Network (NCCN) high-risk group (p=0.059). In patients with a life expectancy of <10 years, the OS was significantly shorter than life expectancy (p<0.001). Conclusion: Patients with non-metastatic PCa may live beyond their life expectancy regardless of the type of therapy and NCCN risk classification, and patients with a life expectancy of <10 years may benefit from BT rather than ADT and EBRT.
- Androgen deprivation therapy
- life expectancy
- overall survival
- prostate cancer
- radiation therapy
- radical prostatectomy
The incidence rate of prostate cancer (PCa) is estimated to be 36.0% in 2020; PCa is the second most common cancer and sixth most common cause of cancer-related deaths in men worldwide (1). The National Comprehensive Cancer Network (NCCN) guidelines recommend the treatment of non-metastatic PCa based on risk classification and life expectancy (2). It is important to consider life expectancy when deciding a treatment for PCa. Several trials have predicted life expectancy in patients with PCa (3-6) and have compared effectiveness of treatments for non-metastatic PCa (7-10). However, it is unclear whether patients who were treated for non-metastatic PCa live beyond their life expectancy.
We have previously reported the impact of radical and partial nephrectomy on life expectancy in patients diagnosed with localized renal cell carcinoma between 1980 and 2008; the findings suggested that partial nephrectomy may play a role in improving actual overall survival (OS) (11). In this study, we evaluated the prognostic factors for OS and investigated the relationship between OS and life expectancy in patients with non-metastatic PCa according to various treatment modalities. Knowledge of the impact of life expectancy in Japanese patients with PCa can aid in an appropriate treatment strategy and improved prognosis.
Patients and Methods
Patient selection and data collection. We screened patients who were diagnosed with non-metastatic PCa between 2006 and 2010 at Nara Medical University Hospital, Nara, Japan. We enrolled 714 consecutive patients with PCa treated with radical prostatectomy (RP), androgen deprivation therapy (ADT), brachytherapy (BT), or external beam radiation therapy (EBRT). Patients who were not treated for PCa were excluded. Patient characteristics and the clinicopathological data were retrospectively obtained, including age at treatment, initial prostate-specific antigen (PSA) levels, Gleason score (GS), Charlson comorbidity index (CCI), clinical T category, and NCCN risk classification. Follow-up started on the day of treatment and finished on the date of the last follow-up or the date of death. This study was approved by the Research Ethics Committee of the Nara Medical University (project identification code: 685-4). Due to the retrospective nature of this study, the requirement for informed consent was waived. The study was conducted in compliance with the study protocol and provisions of the Declaration of Helsinki (2013).
Definition of life expectancy. We defined life expectancy according to an abridged life table published by the Ministry of Health, Labour and Welfare in Japan (Table I). The date of treatment and the age at treatment for each patient was obtained and we determined their life expectancy. For example, life expectancy of a patient who was 70 years old and was treated in 2006 was determined to be 14.7 years according to the abridged life table. We compared the actual OS with the survival rate of each patient who was assumed to die at his/her indicated life expectancy.
Abridged life table for Japanese males published by the Ministry of Health, Labour and Welfare in Japan.
Statistical analysis. All data were recorded using Microsoft Excel. Continuous variables are reported as medians and interquartile ranges. Categorical variables are reported as numbers and percentages. The Mann–Whitney U-test, Fisher’s exact test, or Chi-square test was used for comparing groups, as appropriate. Cox proportional hazards regression analysis was performed to evaluate the prognostic factors for OS. Figure plotting was carried out using GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA, USA). Survival rates were examined using the Kaplan–Meier method. Each patient survival was compared using the log-rank test. All statistical tests performed were two-sided, and a p-Value <0.05 was considered statistically significant. All data were analyzed using EZR ver. 1.55 (Saitama Medical Center, Jichi Medical University, Saitama, Japan).
Results
Patient characteristics. The clinicopathological characteristics of the patients are listed in Table II. The median age at treatment was 71 [interquartile range (IQR)=66-75] years. The median periods of life expectancy and actual follow-up were 168 (IQR=137-214) and 120 (IQR=92-148) months, respectively. The median initial PSA was 8.8 (IQR=5.8-15.9) ng/ml. The distribution of GS was as follows: 6, n=271 (30.0%); 7, n=302 (42.3%) and ≥8, n=141 (19.7%). With regard to the risk classification defined by the NCCN (version 4, 2019), the number of patients in each risk group was as follows: low-risk, n=176 (24.6%); intermediate-risk, n=257 (36.0%); and high-risk, n=281 (39.4%). Approximately 50% of the patients in this cohort selected BT, 10% selected primary ADT, 25% selected EBRT, and 19% selected RP as the initial treatment. Of the 714 patients, 20 (2.8%) died of PCa and 102 (14.3%) died of other causes during the follow-up.
Clinical information of patients with prostate cancer.
Prognostic factors for OS. Table III shows univariate and multivariate analysis of prognostic factors for OS. In the univariate analysis, advanced age, high PSA level, GS ≥7, CCI ≥1, cT3a, cT3b/4, intermediate- and high-risk, ADT, and EBRT were potential prognostic factors for poor OS. In the multivariate analysis, advanced age, CCI ≥1, cT3a, and ADT were independent prognostic factors for poor OS [advanced age: hazard ratio (HR)=1.07, 95% confidence interval (CI)=1.03-1.11; CCI ≥1: HR=1.69, 95% CI=1.12-2.54; cT3a: HR=2.15, 95% CI=1.03-4.52; ADT: HR=2.15, 95% CI=1.06-4.36].
Prognostic factors for overall survival.
Comparison between groups with life expectancy ≥10 and <10 years. Table IV shows a comparison of clinicopathological information between the group with life expectancy ≥10 years and that with life expectancy <10 years. The median age in the latter group was significantly higher than that in the group with life expectancy ≥10 years [79 years (IQR=78-80.3) vs. 71 years (IQR 76-75); p<0.001]. The median initial PSA levels in the group with a life expectancy <10 years were significantly higher than those in the group with life expectancy ≥10 years [12.4 ng/ml (IQR=6.9-34.8) vs. 8.9 ng/ml (IQR=5.8-16.7); p=0.0011]. There was no significant difference between the groups for patients with CCI ≥1 (p=0.22). However, there were significantly more patients with cT stage ≥3, GS ≥8, and classified as high-risk in the group with life expectancy <10 years compared to the group with life expectancy ≥10 years (p=0.0048, 0.0091 and 0.0017, respectively).
Comparison of clinicopathological information between the groups with life expectancy ≥10 and <10 years.
Prognostic factors for OS in patients with life expectancy ≥10 years. Table V shows univariate and multivariate analysis of prognostic factors for OS. In the multivariate analysis, CCI ≥1, cT3a, ADT, and EBRT were independent prognostic factors for poor OS [CCI ≥1: HR=1.86, 95% CI=1.18-2.95; cT3a: HR=2.69, 95% CI=1.14-6.38; ADT: HR=2.88, 95% CI=1.15-7.20; EBRT: HR=2.41, 95% CI=1.28-4.56].
Prognostic factors for overall survival in patients with life expectancy ≥10 years.
Comparison of actual survival with life expectancy. No significant differences were found between the actual OS and life expectancy in all patients, in the RP, ADT, BT and EBRT groups (Figure 1A, p=0.32; Figure 1B, p=0.44; Figure 1C, p=0.90; Figure 1D, p=0.83; and Figure 1E, p=0.34; respectively). In high-risk patients, the actual OS tended to be shorter than life expectancy, without statistically significant differences (Figure 2A, p=0.059). Moreover, no significant differences were found between the actual OS and life expectancy in high-risk patients when the analysis was limited to the RP, ADT, BT and EBRT groups (Figure 2B, p=0.80; Figure 2C, p=0.96; Figure 2D, p=0.21; and Figure 2E, p=0.13; respectively).
Survival curves of actual overall survival and life expectancy in all patients and treatment groups. Survival curves in all patients (A), and in the radical prostatectomy (B), androgen deprivation therapy (C), brachytherapy (D), and external beam radiation therapy group (E) are shown.
Survival curves of actual overall survival and life expectancy in patients with the National Comprehensive Cancer Network high-risk and treatment groups. Survival curves in patients with high-risk (A), and in the radical prostatectomy group with high-risk (B), in the androgen deprivation therapy group with high-risk (C), in the brachytherapy group with high-risk (D), and the external beam radiation therapy group with high-risk (E) are shown.
Comparison of actual survival with life expectancy according to the life expectancy. In patients with a life expectancy of <10 years, no significant difference was found between the actual OS and life expectancy (Figure 3A, p<0.001). In patients with a life expectancy of <10 years and low- or intermediate-risk, no significant differences were found between the actual OS and life expectancy (Figure 3B, p=0.20 and Figure 3C, p=0.12, respectively). In contrast, the actual OS was significantly shorter than the life expectancy in patients with a life expectancy of <10 years and high-risk (Figure 3D, p<0.001). In patients with a life expectancy of <10 years and CCI=0, the actual OS was significantly shorter than the life expectancy (Figure 3E, p<0.001); while, in patients with a life expectancy of <10 years and CCI ≥1, no significant difference was found between actual OS and life expectancy (Figure 3F, p=0.11). In the ADT and EBRT groups with life expectancy <10 years, the actual OS was significantly shorter than life expectancy (Figure 4A, p=0.027 and Figure 4B, p=0.024, respectively). In contrast, no significant difference was found between the actual OS and life expectancy in the BT group with a life expectancy of <10 years (Figure 4C; p=0.076). We could not compare the actual OS with life expectancy in the RP group because there was only 1 patient with a life expectancy of <10 years in the RP group. In patients with a life expectancy of <10 years, no significant differences were found in the OS between the ADT, BT and EBRT groups (Figure 4D, ADT vs. BT, p=0.37 and ADT vs. EBRT, p=0.53, respectively). In patients with a life expectancy of ≥10 years, no significant difference was found between the actual OS and life expectancy (Figure 5A, p=0.24). In the RP, ADT, BT and EBRT groups with a life expectancy of ≥10 years, no significant differences were found between the actual OS and life expectancy (Figure 5B, p=0.23; Figure 5C, p=0.15; Figure 5D, p=0.28; and Figure 5E, p=0.80; respectively). In patients with a life expectancy ≥10 years, the OS in the RP group was significantly longer than that in the ADT and EBRT groups (Figure 5F; RP vs. ADT, p<0.001; and RP vs. EBRT, p=0.0074). Additionally, the OS in the BT group was significantly longer than that in ADT or EBRT group in patients with a life expectancy ≥10 years (Figure 5F; BT vs. ADT, p<0.001; and BT vs. EBRT, p=0.0011). However, there was no significant difference in the OS between the RP and BT groups in patients with a life expectancy of ≥10 years (Figure 5F, RP vs. BT, p=0.66).
Survival curves of the actual overall survival and life expectancy in patients with the National Comprehensive Cancer Network risk classification and Charlson comorbidity index (CCI) according to the life expectancy. Survival curves in patients with a life expectancy of <10 years (A), in patients with life expectancy of <10 years and low-, intermediate-, or high-risk (B, C, and D), and in patients with life expectancy of <10 years and CCI=0 or CCI ≥1 (E and F), are shown.
Survival curves of the actual overall survival (OS) and life expectancy in different treatment groups and analysis of actual OS between the treatment groups with life expectancy <10 years. Survival curves in the androgen deprivation therapy (ADT), external beam radiation therapy (EBRT), and brachytherapy (BT) groups with life expectancy <10 years (A, B and C) are shown. OS between the ADT, BT, and EBRT groups with life expectancy <10 years (D) are shown.
Survival curves of the actual overall survival (OS) and life expectancy in different treatment groups and analysis of actual OS between the groups with life expectancy ≥10 years. Survival curves in patients with life expectancy ≥10 years (A), and the radical prostatectomy (RP), androgen deprivation therapy (ADT), brachytherapy (BT) and external beam radiation therapy (EBRT) groups with life expectancy ≥10 years (B, C, D and E) are shown. OS between the RP, ADT, BT and EBRT groups with life expectancy ≥10 years groups (F) are shown.
Discussion
The present study showed that advanced age, CCI ≥1, cT3a and ADT were independent predictive factors for OS in patients with non-metastatic PCa. Boorjian et al. (12) compared the long-term survival in patients with high-risk PCa after RP and EBRT, with or without ADT. They reported that patients who received EBRT with ADT had a significantly increased risk of all-cause mortality (ACM), and differences in age and comorbidity status between patients who received EBRT and those who received RP may have affected ACM. In the present study, the median age in the EBRT and RP groups was 73 and 71 years, respectively; and the proportion of patients who underwent EBRT with CCI ≥1 and that of patients who underwent RP with CCI ≥1 was 54.3% and 31.9%, respectively. Patients in the EBRT group were significantly older and tended to have more comorbidities than those in the RP group. However, EBRT may not be an independent predictive factor for OS. Sun et al. (13) compared the efficacy of RP, radiotherapy, and surveillance in terms of OS in patients with non-metastatic PCa. They reported that in patients with a life expectancy of <10 years, RP and radiotherapy failed to differ in terms of OS, whereas in patients with a life expectancy of ≥10 years, RP was associated with an improved OS than that with surveillance and radiotherapy. Berg et al. (14) compared OS of EBRT+BT versus RP in comparatively young (≤65yr) and healthy men (CCI=0) with high-risk localized PCa. They reported that RP was associated with benefit of OS compared with EBRT+BT. In the present study, RP and BT were associated with improved OS than that with EBRT in patients with a life expectancy of ≥10 years. The proportions of patients with high-risk and life expectancy of ≥10 years who underwent RP, BT, or EBRT were 32.1%, 8.2%, or 80.0%, respectively. EBRT tended to be preferred in high-risk patients, but it was an independent predictive factor for the OS in patients with a life expectancy of ≥10 years. Therefore, RP and BT may improve OS more than EBRT. This finding may help patients with a life expectancy of ≥10 years to decide on the type of therapy.
In terms of life expectancy, patients with non-metastatic PCa may live beyond their life expectancy regardless of the type of therapy and NCCN risk classification. Additionally, patients with a life expectancy of ≥10 years may live beyond their life expectancy regardless of the type of therapy in the present study. However, patients with a life expectancy of <10 years may not live beyond their life expectancy. Patients with a life expectancy of <10 years had more risk factors than those with a life expectancy ≥10 years; therefore, the actual OS may be shorter than life expectancy. Patients with low- and intermediate-risk may benefit from therapy even if their life expectancy is <10 years. In patients with CCI=0 and life expectancy <10 years, the actual OS was significantly shorter than life expectancy. However, the actual OS was similar to life expectancy in patients with CCI ≥1 and life expectancy <10 years. Matthes et al. (15) assessed the associations between comorbidities with primary treatment of PCa and those with prostate cancer-specific mortality (PCSM) as opposed to other-cause mortality (OCM) in Switzerland. The probability of OCM was higher in patients with comorbidities than in those without, but this was not observed for PCSM. In this study, the proportion of patients with life expectancy of <10 years was only 10.6%. The small sample size may have caused the difference between the findings of the present study and those of previous studies. Boehm et al. (3) examined age- and CCI-adjusted 10-year OCM rates by treatment type (surveillance, RP, BT, EBRT and ADT), and also assessed OCM rates for different treatment types relative to a life expectancy of 10 years. They found that OCM rates varied significantly according to primary treatment, despite adjusting for age and CCI. Patients who received RP and BT had better OCM and life expectancy rates than patients who received ADT or EBRT. In this study, ADT was an independent predictive factor for OS, but the actual OS was similar to life expectancy, regardless of the type of therapy. However, the actual OS was significantly shorter than the life expectancy in the ADT and EBRT groups even if their life expectancy was <10 years, and the actual OS was similar to life expectancy in the BT group when life expectancy was <10 years. The proportions of patients with life expectancy of <10 years and CCI ≥1 who underwent BT, ADT and EBRT were 22.7%, 56.8% and 56.3%, respectively. Patients in the BT group tended to have fewer comorbidities than those in the ADT and EBRT groups. Therefore, patients treated with BT may benefit from the therapy even if their life expectancy is <10 years. We could not examine the actual OS and life expectancy in the RP group stratified according to age category due to the small number of patients with a life expectancy of <10 years.
Our study has several limitations. Firstly, this was a retrospective study from a single institution, which may lead to selection bias. Secondly, life expectancy was calculated from the abridged life table for Japanese; therefore, our findings may not be applicable in other countries. Thirdly, the number of patients between the treatment groups was significantly different. Moreover, there was a difference between the median follow-up period and median life expectancy. Therefore, the number of patients should be increased, and further follow-up is needed.
Conclusion
In conclusion, the study suggests that patients with non-metastatic PCa may live beyond their life expectancy regardless of the type of therapy and NCCN risk classification, and patients with a life expectancy of <10 years may benefit from BT rather than ADT and EBRT. This information may help in treatment decision-making for elderly men with non-metastatic PCa.
Acknowledgements
The Authors would like to thank all patients who participated in this study for their important contributions. We also wish to thank Mariko Yoshimura (Department of Urology, Nara Medical University, Nara, Japan) for invaluable help with obtaining and summarizing the data used in this study.
Footnotes
Authors’ Contributions
SH, MM, NT and KF contributed to conception and design, acquisition of patients’ data and analysis and interpretation of data. YN, MM, KT, and NT performed the treatment. SH and NT contributed to acquisition and interpretation of data. SH and MM contributed to the analysis and interpretation of data. All Authors have been involved in drafting the manuscript and revising it critically for important intellectual content and approved the version to be published. All Authors have participated sufficiently in this work to take public responsibility for appropriate portions of the content.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received October 4, 2022.
- Revision received October 16, 2022.
- Accepted October 18, 2022.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.