Abstract
Purpose:
Accurate information regarding real-world outcomes after contemporary radiation therapy for localized prostate cancer is important for shared decision-making. We examined clinically relevant endpoints at 10 years among men treated within a national healthcare delivery system.
Methods:
We used national administrative, cancer registry, and electronic health record data for patients undergoing definitive radiation therapy with or without concurrent androgen deprivation therapy (ADT) within the Veterans Health Administration from 2005 to 2015. We used National Death Index data through 2019 for overall and prostate cancer-specific survival and identified date of incident metastatic prostate cancer using a validated natural language processing algorithm. We estimated metastasis-free, prostate cancer-specific, and overall survival using Kaplan-Meier methods.
Results:
Among 41,735 men treated with definitive radiation therapy, median age at diagnosis was 65 years and median follow up was 8.7 years. Most had intermediate (42%) and high risk (33%) disease, with 40% receiving ADT as part of initial therapy. Unadjusted 10-year metastasis-free survival was 96%, 92%, and 80% for low, intermediate, and high-risk disease. Similarly, unadjusted 10-year prostate cancer-specific survival was 98%, 97%, and 90% for low, intermediate, and high-risk disease. We found unadjusted overall survival was lower across increasing disease risk categories at 77%, 71%, and 62% for low, intermediate, and high-risk disease (p<0.001).
Conclusions:
These data provide population-based 10-year benchmarks for clinically relevant endpoints, including metastasis-free survival, among patients with localized prostate cancer undergoing radiation therapy using contemporary techniques. The survival rates for high-risk disease, in particular, suggest that outcomes have recently improved.
Keywords: Prostatic neoplasms, radiotherapy, survival, neoplasm metastasis, population
Precis:
This retrospective, population based study of 41,735 men undergoing definitive radiation therapy for localized prostate cancer in a national healthcare delivery system examined survival outcomes with contemporary treatment. In this cohort, survival outcomes, including metastasis-free survival as determined by a novel natural language processing, algorithm, are improved compared to historical studies and are comparable to those reported on contemporary clinical trials.
INTRODUCTION
Over the past two decades, radiation therapy for treatment of localized prostate cancer has evolved with respect to dose[1, 2], fractionation[3, 4], and technique (e.g., intensity-modulated radiation therapy).[5–7] In addition, the use[8–10], duration[11–13] and sequencing [14, 15] of androgen deprivation therapy (ADT) have been refined along with improved staging and risk group stratification.[16–18] For these reasons, clinical trial outcomes of radiation and hormone therapy for prostate cancer have improved over time.
While radiation therapy outcomes (i.e., metastasis, death) for prostate cancer are often derived from meta-analyses of prospective clinical trials, our contemporary understanding may be compromised due to the follow-up duration needed for longer term outcomes resulting in outdated trial data. For example, a recent meta-analysis of 12 randomized controlled trials examining ADT use and timing provided insights into metastasis-free and overall survival, but the trials spanned 1987–2010, with most patients treated in the 1990’s when techniques were actively evolving.[14] Population-based data (e.g., National Cancer Database, Surveillance, Epidemiology, and End Results (SEER)) lack key granular details that impact treatment recommendations such as development of metastatic disease, further limiting understanding of contemporary outcomes. In fact, identification of incident metastatic disease is a longstanding, critical barrier to population-based cancer outcomes assessment. Clarifying the impact of contemporary radiation techniques and hormone therapy schedules on population-based outcomes would support better shared decision-making for both clinicians and patients.
In this context, we conducted a large population-based study of contemporary outcomes in a national cohort of patients with localized prostate cancer who received definitive radiation therapy with or without ADT over the past 15 years. We used an innovative, validated, natural language processing algorithm to identify incident metastatic prostate cancer within the electronic medical records to determine important clinically relevant outcomes, including metastasis-free survival, prostate cancer-specific survival, and overall survival.[19]
METHODS
We used national electronic health record data for patients undergoing definitive radiation therapy with or without concurrent androgen deprivation therapy within the Veterans Health Administration (VHA) system from 2005 to 2015, with follow up through 2019. We queried the VHA Corporate Data Warehouse and Central Cancer Registry to identify patients with incident prostate cancer, obtaining their demographics, biopsy and treatment details. Patients who underwent definitive surgery for prostate cancer or received radiation therapy in the salvage setting for disease recurrence after primary prostatectomy were excluded from this analysis.
We stratified patients by risk group at diagnosis using National Comprehensive Cancer Network (NCCN) definitions for low, intermediate, and high-risk groups, as well as Grade Group based on clinical staging data. Data regarding ADT use as part of the initial treatment course was derived from the VA Central Cancer Registry and using national pharmacy data similar to our prior work.[20] Because ADT is used in combination with radiotherapy in short- and longer-term regimens, we defined a cutoff to differentiate “short term” from “long term” ADT of 9 months. For example, any patient receiving <9 months of ADT as part of their initial treatment course was considered to have received short term ADT, while any patient receiving ≥9 months ADT as part of their initial treatment was considered to have received long term ADT. The Cochran-Mantel-Haenszel method was used to evaluate associations between clinical stratifications and ADT use.
Our primary outcomes were overall survival, prostate cancer-specific and metastasis-free survival. We used National Death Index data to determine cause of death and considered this as a binary variable (prostate cancer vs. any other cause). We obtained the date of incident metastatic disease utilizing a highly sensitive and specific natural language processing algorithm applied to the national electronic health records.[19] Our time-to-event for all survival endpoints was measured from date of diagnosis. We assessed survival endpoints using the Kaplan-Meier method with censoring at death or the end of the study period December 31, 2019. The Hall-Wellner method was used to generate 95% confidence bands. Log-rank testing was used for between-group comparisons, with p-value <0.05 considered significant. All analyses were conducted in SAS version 8.3. This study was approved by our institutional review board.
RESULTS
Patient Demographics and Tumor Characteristics
We included a total of 41,735 patients in this analysis with a median follow up of 8.7 years. The median age at diagnosis was 65 years. Most men had cT1 (68.0%) or cT2 (28.7%) disease, and nearly all (96.8%) were clinically node-negative according to the registry. Almost two-thirds (62%) of men had a PSA between 4 and 10 ng/mL at diagnosis. We found similar distributions of Grade Group 1 (32.0%) and 2 (31.9%) disease, with 14.1% Grade Group 3, 12.2% Grade Group 4, and 8.8% Grade Group 5 disease. In terms of NCCN risk status, 24.8% had low risk, 42.2% intermediate risk, and 32.9% had high-risk disease at diagnosis (Table 1). We found ADT was used in conjunction with definitive-intent radiation therapy in 40.3% of patients further explicated below.
TABLE 1.
Disease characteristics and use of androgen deprivation therapy in 41,735 patients undergoing definitive radiation therapy for localized prostate cancer
| Variable | All patients undergoing radiation therapy +/− ADT (N=41,735) |
|---|---|
| Age (median, years) | 65 |
| Clinical T-Stage | |
| cT1 | 28,361 (68.0%) |
| cT2 | 11,967 (28.7%) |
| cT3 | 924 (2.2%) |
| cT4 | 133 (0.3%) |
| Missing | 350 (0.8%) |
| Clinical N-Stage | |
| N0 | 40,383 (96.8%) |
| N1 | 345 (0.8%) |
| Missing | 1,007 (2.4%) |
| PSA at Diagnosis | |
| ≤ 4 | 4,665 (11.2%) |
| 4–10 | 25,887 (62.0%) |
| 10–20 | 6,778 (16.3%) |
| ≥ 20 | 3,844 (9.2%) |
| Unknown | 561 (1.3%) |
| Biopsy Grade Group | |
| 1 | 13,343 (32.0%) |
| 2 | 13,327 (31.9%) |
| 3 | 5,868 (14.1%) |
| 4 | 5,108 (12.2%) |
| 5 | 3,654 (8.8%) |
| Unknown | 435 (1.0%) |
| NCCN Risk Group | |
| Low | 10,332 (24.8%) |
| Intermediate | 17,634 (42.2%) |
| High | 13,715 (32.9%) |
| Missing | 54 (0.1%) |
| ADT Use - Initial Treatment | |
| Yes | 16,833 (40.3%) |
| No | 24,902 (59.7%) |
Abbreviations: PSA – prostate specific antigen, ADT – androgen deprivation therapy
Patterns of ADT use by Grade Group and NCCN risk groups
Radiation therapy alone was used in 59.7% of patients, and radiation with concurrent ADT in the remaining 40.3% (22.6% and 17.7% short- and long-term ADT, respectively). The use of any ADT in conjunction with definitive radiation therapy increased with increasing Grade Group, with 51.2% of patients with Grade Group 3 disease, 73.8% with Grade Group 4 disease and 81.4% of patients with Grade Group 5 receiving ADT, compared to 37.2% of patients with Grade Group 2 and 14.2% of patients with Grade Group 1 disease (p<0.001). Similarly, when examined by NCCN risk status, over two-thirds (68.4%) of men with high-risk disease received ADT as part of their initial treatment course, compared with 36.1% and 10.2% of men with intermediate and low risk, respectively (p<0.001). Most patients receiving long-term ADT had Gleason Grade 4 or 5 disease (61.4%), and 77.7% had high risk disease per NCCN. We observed higher median PSA levels to be associated with increasing treatment intensity; PSA 6.0 ng/mL for patients who received radiation therapy alone, 7.6 ng/mL for those who received radiation plus short-term ADT, and 9.7 ng/mL for patients who received radiation plus long-term ADT (p<0.001) (Table 2).
Table 2.
Frequency of radiation therapy alone, as well as short- and long-term ADT use analyzed by disease characteristics
| Variable | Radiation therapy alone (N=24,902) | Radiation therapy + short term ADT (N=9,436) | Radiation therapy + long term ADT (N=7,387) | p-value |
|---|---|---|---|---|
| Age (median, years) | 65.0 | 65.9 | 66.5 | <0.001 |
| T-Stage (n, %) | <0.001 | |||
| cT1–2 | 24,493 (98.4%) | 9,127 (96.7%) | 6,698 (90.7%) | |
| cT3–4 | 208 (0.8%) | 257 (2.7%) | 592 (8.0%) | |
| Missing | 201 (0.8%) | 52 (0.6%) | 97 (1.3%) | |
| Grade Group (n, %) | <0.001 | |||
| 1 | 11,428 (45.9%) | 1,460 (15.5%) | 453 (6.1%) | |
| 2 | 8,363 (33.6%) | 3,620 (38.4%) | 1,340 (18.1%) | |
| 3 | 2,862 (11.5%) | 2,069 (21.9%) | 935 (12.7%) | |
| 4 | 1,336 (5.4%) | 1,360 (14.4%) | 2,412 (32.7%) | |
| 5 | 680 (2.7%) | 853 (9.0%) | 2,121 (28.7%) | |
| Unknown | 233 (0.9%) | 74 (0.8%) | 126 (1.7%) | |
| Median PSA (ng/mL) | 6.0 | 7.6 | 9.7 | <0.001 |
| NCCN Risk Group (n, %) | <0.001 | |||
| Low | 9,278 (37.3%) | 870 (9.2%) | 184 (2.5%) | |
| Intermediate | 11,266 (45.2%) | 4,916 (52.1%) | 1,444 (19.6%) | |
| High | 4,334 (17.4%) | 3,642 (38.6%) | 5,737(77.7%) | |
| Missing | 24 (0.1%) | 8 (0.1%) | 22 (0.2%) |
Survival
Overall survival worsened as disease risk classification increased following definitive radiation therapy for localized prostate cancer. For example, 10-year overall survival was 77% for patients with low-risk, 70% for those with intermediate-risk, and 62% for those with high-risk disease at diagnosis (p<0.001). Corresponding 10-year prostate cancer-specific survival was 98% for patients with low-risk, 97% for those with intermediate-risk, and 90% for patients with high-risk disease at diagnosis (p<0.001). Similarly, 10-year metastasis-free survival for patients was 96% for low-, 92% for intermediate-, and 80% for high-risk disease (p<0.001) (Figure 1).
Figure 1.
Survival estimates for overall survival (A), prostate cancer-specific survival (B) and metastasis-free survival (C) by NCCN risk group, with 95% confidence bands.
Moreover, 10-year metastasis-free-, prostate cancer specific- and overall survival also correlated with Grade Group. Ten-year metastasis-free survival was 96% for patients with Grade Group 1, 92% for patients with Grade Group 2, 87% for those with Grade Group 3, 83% for those with Grade Group 4 and 67% for patients with Grade Group 5 disease (p<0.001). Similarly, 10-year prostate cancer specific survival for Grade 1, 2, 3, 4, and 5 disease was 98%, 97%, 94%, 93%, and 82%, respectively (p<0.001), and 10-year overall survival for Grade 1, 2, 3, 4 and 5 disease was 76%, 70%, 66%, 63%, and 53% respectively (p<0.001) (Figure 2).
Figure 2.
Survival estimates for overall survival (A), prostate cancer-specific survival (B) and metastasis-free survival (C) by Grade Group, with 95% confidence bands.
DISCUSSION
In this retrospective, population-based analysis of patients undergoing definitive radiation therapy with or without ADT for localized prostate cancer, we observed improved 10-year survival outcomes compared to those found in historical trials. Our use of an innovative natural language processing tool to identify incident metastatic disease using population-based data opens a new realm of understanding time-to-event clinically relevant outcomes for radiation therapy, extending beyond traditional use of clinical trials to characterize important disease progression outcomes. Not surprisingly, we found disease severity at diagnosis was associated with 10-year disease-specific and overall survival outcomes. Taken together, our findings suggest favorable freedom from metastasis at 10 years following radiation therapy for 8 in 10 men with high risk features at diagnosis.
Recently, much attention has been directed at defining clinically meaningful endpoints in prostate cancer research. Given the long natural history of prostate cancer and rapid advances in treatment, biochemical recurrence is often used as a primary clinical trial endpoint as recurrence occurs much sooner and more frequently than distant metastasis or prostate cancer-related death. However, multiple studies have demonstrated that biochemical recurrence is a poor surrogate for survival in patients with prostate cancer.[21, 22] For example, a recent meta-analysis of 75 trials involving 53,631 patients primarily undergoing radiation therapy for localized prostate cancer, biochemical failure-related endpoints (biochemical failure, biochemical failure-free survival, biochemical and clinical failure) all correlated poorly with overall survival. Meanwhile, metastasis-free survival correlated strongly with overall survival, with an R2 of 0.78.[21] Similarly, a recent analysis of men treated for recurrent disease after prostatectomy on NRG/RTOG 9601 found a weak correlation between first or second biochemical failure after prostatectomy and overall survival (τ = 0.25 and τ = 0.40, respectively, assessed by Kendall rank correlation), while metastasis-free survival correlated strongly with overall survival (τ = 0.86).[22] These studies, among others, have identified metastasis-free survival as a clinically relevant surrogate endpoint for overall survival, suggesting this endpoint merits further exploration and consideration in contemporary clinical trial design and population-based observational data.
On the other hand, increasing use of metastasis-free survival as an endpoint in major studies has produced heterogenous results. For example, the Meta-analysis of Randomized Trials in Cancer of the Prostate (MARCAP) consortium recently performed a meta-analysis of 12 trials involving 10,853 patients to assess the benefit of various treatment intensification strategies using radiation and ADT.[14] The 10-year metastasis-free survival in this study was 61% for patients receiving ADT in addition to radiation therapy. Nearly half (47%) of patients included had high-risk disease. Importantly, this 10-year metastasis-free survival is lower than recent trials enrolling a sizeable proportion of patients with high-risk disease.[2, 6] The 10-year metastasis-free survival we observed for patients with high-risk disease is similar to the aforementioned clinical trials that enrolled during this period, suggesting improvement in metastasis-free survival in contemporary cohorts persists in real-world settings outside of highly-regulated clinical trial environments.
This discrepancy between the metastasis-free survival observed in the MARCAP analysis of historical trials (i.e., 6 in 10 men metastasis-free at 10 years) and the metastasis-free survival observed in our study (i.e., 8 in 10 men metastasis-free at 10 years) may be due to multifactorial improvements in diagnosis and treatment over time. This is especially likely since most patients included in the MARCAP analysis were treated in the 1990s and early 2000’s and our study examined patients treated between 2005–2015. While the chronology of historical studies allows for robust follow up, advances in the delivery of radiation therapy, refinement of ADT use and timing, and improvement of therapy in the metastatic setting over time challenge historical benchmarks. In other words, these developments complicate application of historical survival outcomes to patients diagnosed and treated using contemporary techniques. Clarifying our understanding of survival outcomes in contemporary real-world cohorts is important for setting patient expectations during shared decision-making, guiding treatment intensification or de-escalation, and future clinical trial development.
Despite the advantage of identifying incident metastatic disease at the patient level within a large population, some limitations merit consideration. One is the use of NCCN risk groupings for classification of patient risk. While newer methods, such as clinical-genomic risk group classification[17] and advanced clinical prognostic stage grouping[18] provide improved patient-level prognostication of outcomes, our use of NCCN risk groupings remains common clinical practice within clinical and research communities. We also used Grade Group to better align our findings with current clinical practice. Second, we used the National Death Index to determine overall survival. While there are some limitations to cause of death attribution, overall survival observed in this study is consistent with what would be expected. Third, this analysis exclusively used data within a national health system of veterans, which may introduce subtle differences inherent to a large integrated delivery system that could limit extrapolation to the broader US population. Nonetheless, the majority of veterans with prostate cancer are cared for outside the VA system, and our comprehensive chart and national data abstraction add to the rigor and validity of our data and outcomes collection. In addition, our median age at diagnosis was 65 years, providing relevance to Medicare and younger non-Medicare eligible patients. Additionally, in this dataset lower risk patients appear to have a lower other-cause mortality risk compared to those with higher-risk disease. Age did not differ substantially between risk groups, with a median age of 64 for low risk, 65 for intermediate risk and 66 for high risk patients. The potential reasons for this discrepancy are the subject of ongoing work by our group. Finally, this work is also subject to the inherent limitations of retrospective database studies, including the potential for selection bias, data coding errors, and missing data that may have impacted our findings. While keeping these limitations in mind, our findings did reveal the expected relationships between disease severity and outcomes in real-world practice.
In conclusion, our study found favorable 10-year metastasis-free survival using contemporary diagnosis and radiation therapy treatment approaches among men with localized prostate cancer, higher than observed in some historical trial data. This information is valuable for setting patient expectations for treatment during shared decision-making, and as benchmarks in future trial design. Furthermore, these data may serve as a basis for future studies identifying patients at high risk of metastatic disease for whom treatment escalation is indicated, as well as those at lower risk of metastatic disease for whom ADT may be de-escalated or omitted.
ACKNOWLEDGEMENTS
We thank Steven Kronenberg for his excellent technical assistance.
Funding:
This work is supported by NIH NCI R01 CA242559 (AT, TAS).
ABBREVIATIONS
- ADT
Androgen Deprivation Therapy
- SEER
Surveillance, Epidemiology and End Results
- VHA
Veterans Health Administration
- NCCN
National Comprehensive Cancer Network
- PSA
Prostate-Specific Antigen
Footnotes
CrediT Author Statement: D.J. Herr: Conceptualization, Methodology, Writing – Original Draft. D.A. Elliott - Conceptualization, Methodology Writing – Review & Editing. G. Duchesne – Writing – Review & Editing. K.D. Stensland – Investigation, Writing – Review & Editing. M.E.V. Caram – Conceptualization, Writing – Review & Editing. C. Chapman – Investigation, Writing – Review & Editing. J.A. Burns - Investigation, Formal Analysis, Writing - Review & Editing. B.K. Hollenbeck – Writing, Review & Editing. J.B. Sparks – Writing – Review & Editing. C. Shin – Methodology, Writing – Review & Editing. A. Zaslavsky – Conceptualization, Writing - Review & Editing. A. Tsodikov – Conceptualization, Methodology, Writing – Review & Editing. T.A. Skolarus - Conceptualization, Methodology, Writing – Review & Editing.
COI Statement: We have no conflicts of interest to disclose.
REFERENCES
- 1.Dearnaley DP, et al. , Escalated-dose versus control-dose conformal radiotherapy for prostate cancer: long-term results from the MRC RT01 randomised controlled trial. Lancet Oncol, 2014. 15(4): p. 464–73. [DOI] [PubMed] [Google Scholar]
- 2.Kerkmeijer LGW, et al. , Focal Boost to the Intraprostatic Tumor in External Beam Radiotherapy for Patients With Localized Prostate Cancer: Results From the FLAME Randomized Phase III Trial. J Clin Oncol, 2021. 39(7): p. 787–796. [DOI] [PubMed] [Google Scholar]
- 3.Widmark A, et al. , Ultra-hypofractionated versus conventionally fractionated radiotherapy for prostate cancer: 5-year outcomes of the HYPO-RT-PC randomised, non-inferiority, phase 3 trial. Lancet, 2019. 394(10196): p. 385–395. [DOI] [PubMed] [Google Scholar]
- 4.Dearnaley D, et al. , Conventional versus hypofractionated high-dose intensity-modulated radiotherapy for prostate cancer: 5-year outcomes of the randomised, non-inferiority, phase 3 CHHiP trial. Lancet Oncol, 2016. 17(8): p. 1047–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Brand DH, et al. , Intensity-modulated fractionated radiotherapy versus stereotactic body radiotherapy for prostate cancer (PACE-B): acute toxicity findings from an international, randomised, open-label, phase 3, non-inferiority trial. Lancet Oncol, 2019. 20(11): p. 1531–1543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Murthy V, et al. , Prostate-Only Versus Whole-Pelvic Radiation Therapy in High-Risk and Very High-Risk Prostate Cancer (POP-RT): Outcomes From Phase III Randomized Controlled Trial. J Clin Oncol, 2021. 39(11): p. 1234–1242. [DOI] [PubMed] [Google Scholar]
- 7.Morris WJ, et al. , Androgen Suppression Combined with Elective Nodal and Dose Escalated Radiation Therapy (the ASCENDE-RT Trial): An Analysis of Survival Endpoints for a Randomized Trial Comparing a Low-Dose-Rate Brachytherapy Boost to a Dose-Escalated External Beam Boost for High- and Intermediate-risk Prostate Cancer. Int J Radiat Oncol Biol Phys, 2017. 98(2): p. 275–285. [DOI] [PubMed] [Google Scholar]
- 8.Bolla M, et al. , Short Androgen Suppression and Radiation Dose Escalation for Intermediate- and High-Risk Localized Prostate Cancer: Results of EORTC Trial 22991. J Clin Oncol, 2016. 34(15): p. 1748–56. [DOI] [PubMed] [Google Scholar]
- 9.Jones CU, et al. , Radiotherapy and short-term androgen deprivation for localized prostate cancer. N Engl J Med, 2011. 365(2): p. 107–18. [DOI] [PubMed] [Google Scholar]
- 10.Roach M 3rd, et al. , Short-term neoadjuvant androgen deprivation therapy and external-beam radiotherapy for locally advanced prostate cancer: long-term results of RTOG 8610. J Clin Oncol, 2008. 26(4): p. 585–91. [DOI] [PubMed] [Google Scholar]
- 11.Nabid A, et al. , Duration of Androgen Deprivation Therapy in High-risk Prostate Cancer: A Randomized Phase III Trial. Eur Urol, 2018. 74(4): p. 432–441. [DOI] [PubMed] [Google Scholar]
- 12.Denham JW, et al. , Short-term androgen suppression and radiotherapy versus intermediate-term androgen suppression and radiotherapy, with or without zoledronic acid, in men with locally advanced prostate cancer (TROG 03.04 RADAR): 10-year results from a randomised, phase 3, factorial trial. Lancet Oncol, 2019. 20(2): p. 267–281. [DOI] [PubMed] [Google Scholar]
- 13.Bolla M, et al. , Duration of androgen suppression in the treatment of prostate cancer. N Engl J Med, 2009. 360(24): p. 2516–27. [DOI] [PubMed] [Google Scholar]
- 14.Kishan AU, et al. , Androgen deprivation therapy use and duration with definitive radiotherapy for localised prostate cancer: an individual patient data meta-analysis. Lancet Oncol, 2022. 23(2): p. 304–316. [DOI] [PubMed] [Google Scholar]
- 15.Roach M, et al. , Sequence of hormonal therapy and radiotherapy field size in unfavourable, localised prostate cancer (NRG/RTOG 9413): long-term results of a randomised, phase 3 trial. Lancet Oncol, 2018. 19(11): p. 1504–1515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zumsteg ZS, et al. , A new risk classification system for therapeutic decision making with intermediate-risk prostate cancer patients undergoing dose-escalated external- beam radiation therapy. Eur Urol, 2013. 64(6): p. 895–902. [DOI] [PubMed] [Google Scholar]
- 17.Spratt DE, et al. , Development and Validation of a Novel Integrated Clinical-Genomic Risk Group Classification for Localized Prostate Cancer. J Clin Oncol, 2018. 36(6): p. 581–590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Dess RT, et al. , Development and Validation of a Clinical Prognostic Stage Group System for Nonmetastatic Prostate Cancer Using Disease-Specific Mortality Results From the International Staging Collaboration for Cancer of the Prostate. JAMA Oncol, 2020. 6(12): p. 1912–1920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Alba PR, et al. , Ascertainment of Veterans With Metastatic Prostate Cancer in Electronic Health Records: Demonstrating the Case for Natural Language Processing. JCO Clin Cancer Inform, 2021. 5: p. 1005–1014. [DOI] [PubMed] [Google Scholar]
- 20.Caram MEV, et al. , Factors influencing treatment of veterans with advanced prostate cancer. Cancer, 2021. 127(13): p. 2311–2318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gharzai LA, et al. , Intermediate clinical endpoints for surrogacy in localised prostate cancer: an aggregate meta-analysis. Lancet Oncol, 2021. 22(3): p. 402–410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Jackson WC, et al. , Biochemical Failure Is Not a Surrogate End Point for Overall Survival in Recurrent Prostate Cancer: Analysis of NRG Oncology/RTOG 9601. J Clin Oncol, 2022. 40(27): p. 3172–3179. [DOI] [PMC free article] [PubMed] [Google Scholar]