Discerning the survival advantage among patients with prostate cancer who undergo radical prostatectomy or radiotherapy: The limitations of cancer registry data

Uncovering interpretable potential confounders in electronic medical records

Randomized clinical trials (RCT) are the gold standard for informing treatment decisions. Observational studies are often plagued by selection bias, and expert-selected covariates may insufficiently adjust for confounding. We explore how unstructured clinical text can be used to reduce selection bias and improve medical practice. We develop a framework based on natural language processing to uncover interpretable potential confounders from text. We validate our method by comparing the estimated hazard ratio (HR) with and without the confounders against established RCTs. We apply our method to four cohorts built from localized prostate and lung cancer datasets from the Stanford Cancer Institute and show that our method shifts the HR estimate towards the RCT results. The uncovered terms can also be interpreted by oncologists for clinical insights. We present this proof-of-concept study to enable more credible causal inference using observational data, uncover meaningful insights from clinical text, and inform high-stakes medical decisions.

Association of Endocrine Therapy and Dementia in Women with Breast Cancer

Purpose

Prior studies have reported differing results regarding the association between endocrine therapy (ET) in the treatment of breast cancer and dementia risk. However, existing findings may be limited by common sources of bias and confounding. Here we investigate the association of ET utilized in the definitive setting to treat non-metastatic breast cancer with dementia risk accounting for multiple potential sources of bias and confounding.

Patients and Methods

We conducted a retrospective study in SEER-Medicare of women aged ≥ 66 years with non-metastatic breast cancer. We examined the risk of all-cause dementia among ET users versus non-ET users using multivariable regression models, accounting for the competing risk of death, and using a start of the follow-up period as 12-months following breast cancer diagnosis for both groups to avoid immortal time bias.

Results

Among 25,777 individuals there were 2,869 incident dementia cases. We found a statistically significantly decreased risk of any dementia among ET users in unadjusted and adjusted models that completely attenuated when accounting for the competing risk of death (hazard ratio, 0.98; 95% confidence interval, 0.90–1.07).

Conclusion

When accounting for common sources of bias and confounding we did not find evidence to support an association between ET in the definitive treatment of non-metastatic breast cancer and dementia risk. These results suggest that ET may not be associated with dementia risk.

Comparative Survival Outcomes of High-risk Prostate Cancer Treated with Radical Prostatectomy or Definitive Radiotherapy Regimens

Background

Observational data has indicated improved survival after radical prostatectomy (RP) compared with definitive radiotherapy (RT) in men with high-risk prostate cancer (PCa).

Objective

To compare PCa-specific mortality (PCSM) and overall mortality (OM) in men with high-risk PCa treated with RP or RT, providing information on target doses and fractionations.

Design, setting, and participants

This is an observational study from the Cancer Registry of Norway. Patients were diagnosed with high-risk PCa during 2006–2015, treated with RP ≤12 mo or RT ≤15 mo after diagnosis, and stratified according to RP or RT modality; external beam radiotherapy (EBRT; 70–<74, 74–<78, or 78 Gy), hypofractionated RT or EBRT combined with brachytherapy (BT-RT).

Outcome measurements and statistical analysis

Competing risk and Kaplan-Meier methods estimated PCSM and OM, respectively. Multivariable Cox regression models evaluated hazard ratios (HRs) for PCSM and OM.

Results and limitations

In total, 9254 patients were included (RP 47%, RT 53%). RT patients were older, had poorer performance status and more unfavorable disease characteristics. With a median follow-up time of seven and eight yrs, the overall 10-yr PCSM was 7.2% (95% confidence interval [CI] 6.4–8.0) and OM was 22.9% (95% CI 21.8–24.1). Compared with RP, EBRT 70–<74 Gy was associated with increased (HR 1.88, 95% CI 1.33–2.65, p < 0.001) and BT-RT with decreased (HR 0.49, 95% CI 0.24–0.96, p = 0.039) 10-yr PCSM. Patients treated with EBRT 70–78 Gy had higher adjusted 10-yr OM than those treated with RP.

Conclusions

In men with high-risk PCa, treatment with EBRT <74 Gy was associated with increased adjusted 10-yr PCSM and OM, and BT-RT with decreased 10-yr PCSM, compared with RP.

Patient summary

In this study, we compared mortality after radical prostatectomy (RP) and radiotherapy (RT) in men with high-risk prostate cancer (PCa); the results suggest that men receiving lower-dose RT have higher, and patients receiving brachytherapy may have lower, risk of death from PCa than patients treated with prostatectomy.

Definitive radiotherapy for prostate cancer in Norway 2006-2015: Temporal trends, performance and survival

BACKGROUND

More studies are needed to document nation-wide use and effectiveness of curative definitive radiotherapy (Def-RT) in the treatment of prostate cancer (PCa).

PATIENTS AND METHODS

For 38,960 men diagnosed with PCa without distant metastases from 2006 to 2015 data from the Norwegian Prostate Cancer Registry and a national radiotherapy database (NoRadBase) was analyzed. Overall survival and PCa-specific mortality were described comparing EQD-2 < 74 Gy ("low-dose") with EQD-2 ≥ 74 Gy ("escalated dose").

RESULTS

Use of Def-RT decreased (27-24%) whereas the proportion of radical prostatectomies (RPs) increased (31-38%). In high-risk patients the use of RP doubled (18-36%), while the proportion of Def-RT remained stable (about 35%). Before 2010, almost a quarter of patients received low-dose Def-RT with gradual increase of escalated Def-RT thereafter. Escalated Def-RT was associated with significantly more favorable 10-year PCa-specific mortality (4.4% [95% CI: 2.7-10.7%]) than observed after low-dose Def- RT (8.8% [95% CI: 6.2-9.8%), with the most beneficial effects in high-risk patients. Our analyses indicated the need to expand the NoRadBase by consensus-based quality measures.

CONCLUSION

In this nationwide cohort, the overall use of Def-RT decreased slightly. In high-risk patients the provision of Def-RT remained stable and was accompanied by doubling of patients with RP and reduction of a "no curative treatment" strategy. Escalated dose Def-RT significantly reduced 10-year PCa-specific mortality compared to low-dose Def-RT. Aiming for cancer care equity national radiotherapy registries for PCa should regularly monitor data based on consensus-based quality measures enabling feedback to the responsible hospitals.

Quantifying treatment selection bias effect on survival in comparative effectiveness research: findings from low-risk prostate cancer patients

BACKGROUND

Comparative effectiveness research (CER) using national registries influences cancer clinical trial design, treatment guidelines, and patient management. However, the extent to which treatment selection bias (TSB) affects overall survival (OS) in cancer CER remains poorly defined. We sought to quantify the TSB effect on OS in the setting of low-risk prostate cancer, where 10-year prostate cancer-specific survival (PCSS) approaches 100% regardless of treatment modality.

METHODS

The Surveillance, Epidemiology, and End Results database was queried for patients with low-risk prostate cancer (cT1-T2a, PSA < 10, and Gleason 6) who received radical prostatectomy (RP), brachytherapy (BT), or external beam radiotherapy (EBRT) from 2005 to 2015. The TSB effect was defined as the unadjusted 10-year OS difference between modalities that was not due to differences in PCSS. Propensity score matching was used to estimate the TSB effect on OS due to measured confounders (variables present in the database and associated with OS) and unmeasured confounders.

RESULTS

A total of 50,804 patients were included (8845 RP; 18,252 BT; 23,707 EBRT) with a median follow-up of 7.4 years. The 10-year PCSS for the entire cohort was 99%. The 10-year OS was 92.9% for RP, 83.6% for BT, and 76.9% for EBRT (p < 0.001). OS differences persisted after propensity score matching of RP vs. EBRT (7.4%), RP vs. BT (4.6%), and BT vs. EBRT (3.7%) (all p < 0.001). The TSB effect on 10-year OS was estimated to be 15.0% for RP vs. EBRT (8.6% measured, 6.4% unmeasured), 8.5% for RP vs. BT (4.8% measured, 3.7% unmeasured), and 6.5% for BT vs. EBRT (3.1% measured, 3.4% unmeasured).

CONCLUSIONS

Patients with low-risk prostate cancer selected for RP exhibited large OS differences despite similar PCSS compared to radiotherapy, suggesting OS differences are almost entirely driven by TSB. The quantities of these effects are important to consider when interpreting prostate cancer CER using national registries.

Prostate cancer dose-response, fractionation sensitivity and repopulation parameters evaluation from 25 international radiotherapy outcome data sets

OBJECTIVE

This study was undertaken to model the biochemical free survival at 5 years and to evaluate the parameters defining dose-response curve, dose-fractionation radiosensitivity and repopulation.

METHODS

It was carried out a literature search on Pubmed to retrieve data sets of patients treated with external beam radiation therapy of 1.8-4.0 Gy per fraction and overall treatment time of 3 to 10 weeks. 10 groups were identified, based on risk class and androgen deprivation therapy (ADT). Dose-response curve D₅₀ (dose at 50% probability of control) and g₅₀ (steepness), α/β (dose-fractionation radiosensitivity), and repopulation parameters, d_prolif and T_prolif , were calculated. Bootstrap-based cross-validation was performed and median and 95% CI (confidence interval) were evaluated.

RESULTS

25 data sets, including 20,310 patients, were considered. The median (95% CI) D₅₀ and g₅₀ values were 62 (CI 53 - 66) Gy and 1.6 (0.8 - 2.4). ADT patients showed lower values of D₅₀ and g₅₀ (57 ± 5 Gy and 1.1 ± 0.4) compared to no-ADT patients (65 ± 2 Gy and 2.3 ± 0.6), with p < 0.0001 and p = 0.002. If we did not consider any dependence on overall treatment time, the median (95% CI) value of α/β was 1.4 (1.0 - 1.9) Gy with p < 0.0001 for all patients. The median values of d_{proli f} and T_prolif were 0.0 to 0.3 Gy/d and 18-40 days.

CONCLUSION

Dose-response curve resulted dependent on risk class and ADT, with higher steepness for no-ADT patients. Low values of dose-fractionation radiosensitivity were found, supporting the use of moderate hypofractionated radiotherapy in each risk class. A limited dependence on repopulation was observed.

ADVANCES IN KNOWLEDGE

Prostate cancer response to moderate hypofractionated radiotherapy was reliably quantified considering risk class and androgen deprivation therapy.

Stroke and thromboembolic events in men with prostate cancer treated with definitive radiation therapy with or without androgen deprivation therapy

BACKGROUND

There is conflicting evidence regarding the association between androgen deprivation therapy (ADT) for prostate cancer (PC) and the risk of developing stroke and thromboembolic events. Our study evaluated the association between ADT use and development of stroke, transient ischemic attack (TIA), deep vein thrombosis (DVT), and pulmonary embolism (PE) in a homogenous group of men with PC treated with definitive radiation therapy (RT) after controlling for multiple sources of confounding.

METHODS

Observational cohort study of patients diagnosed with PC at the US Department of Veterans Affairs between 1 January 2001 and October 31, 2015 and treated with definitive RT. Exposure was initiation of ADT within 1 year of PC diagnosis. Primary outcomes were development of stroke, TIA, DVT, or PE.

RESULTS

44,246 men with median follow-up of 6.8 years. The overall cumulative incidences of stroke, TIA, DVT, and PE at 10 years were 6.0, 3.0, 3.4, and 1.9%, respectively. In the multivariable competing risks model, there was a significant association between ADT and stroke (subdistribution hazard ratio (SHR) = 1.19, 95% CI = 1.09-1.30, p < 0.01), TIA (SHR = 1.24, 95% CI = 1.08-1.41, p < 0.01), and DVT (SHR = 1.18, 95% CI = 1.04-1.34, p < 0.01). ADT was only associated with PE in men receiving ADT for > 1 year (SHR = 1.34, 95% CI = 1.06-1.69, p-value = 0.03).

CONCLUSION

We observed an increase in the risk of stroke, TIA, and DVT in men receiving ADT and an increased risk of PE in men receiving long-term ADT. These results highlight concerns regarding long-term risks of ADT on stroke and thromboembolic events in the treatment of PC.

Comparative effectiveness of radical prostatectomy with adjuvant radiotherapy versus radiotherapy plus androgen deprivation therapy for men with advanced prostate cancer

BACKGROUND

Men with locally advanced prostate cancer (LAPCa) or regionally advanced prostate cancer (RAPCa) are at high risk for death from their disease. Clinical guidelines support multimodal approaches, which include radical prostatectomy (RP) followed by radiotherapy (XRT) and XRT plus androgen deprivation therapy (ADT). However, there are limited data comparing these substantially different treatment approaches. Using Surveillance, Epidemiology, and End Results (SEER)-Medicare data, this study compared survival outcomes and adverse effects associated with RP plus XRT versus XRT plus ADT in these men.

METHODS

SEER-Medicare data were queried for men with cT3-T4N0M0 (LAPCa) or cT3-T4N1M0 (RAPCa) prostate cancer. Propensity score methods were used to balance cohort characteristics between the treatment arms. Survival analyses were analyzed with the Kaplan-Meier method and Cox proportional hazards models.

RESULTS

From 1992 to 2009, 13,856 men (≥65 years old) were diagnosed with LAPCa or RAPCa: 6.1% received RP plus XRT, and 23.6% received XRT plus ADT. At a median follow-up of 14.6 years, there were 2189 deaths in the cohort, of which 702 were secondary to prostate cancer. Regardless of the tumor stage or the Gleason score, the adjusted 10-year prostate cancer-specific survival and 10-year overall survival favored men who underwent RP plus XRT over men who underwent XRT plus ADT. However, RP plus XRT versus XRT plus ADT was associated with higher rates of erectile dysfunction (28% vs 20%; P = .0212) and urinary incontinence (49% vs 19%; P < .001).

CONCLUSIONS

Men with LAPCa or RAPCa treated initially with RP plus XRT had a lower risk of prostate cancer-specific death and improved overall survival in comparison with those men treated with XRT plus ADT, but they experienced higher rates of erectile dysfunction and urinary incontinence.

Impact of prior local therapy on overall survival in men with metastatic castration-resistant prostate cancer: Results from Shared Equal Access Regional Cancer Hospital

OBJECTIVES

To evaluate the impact of previous local treatment on survival in men with newly diagnosed metastatic castration-resistant prostate cancer.

METHODS

We carried out a retrospective study of patients newly diagnosed with metastatic castration-resistant prostate cancer in the year 2000 or later from eight Veterans Affairs Medical Centers. Patients were categorized based on prior local therapy (none, prostatectomy ± radiation or radiation alone). Overall and cancer-specific survival was estimated by the Kaplan-Meier method. Cox proportional hazards regression models were used to test the association between prior local treatment and survival.

RESULTS

Of 729 patients, 284 (39%) underwent no local treatment, 176 (24%) underwent radical prostatectomy ± radiation and 269 (37%) underwent radiation alone. On multivariable analysis, men with prior prostatectomy had improved overall (hazard ratio 0.71, P = 0.005) and cancer-specific survival (hazard ratio 0.55, P < 0.001) compared with men with no prior local therapy. This improvement in overall (hazard ratio 0.89, P = 0.219) and cancer-specific survival (hazard ratio 0.87, P = 0.170) was not seen in men with prior radiation alone. After further adjusting for comorbidity with the Charlson Comorbidity Index, patients with prior prostatectomy still had improved overall survival (hazard ratio 0.70, P = 0.003), whereas this was not seen in patients who received prior radiation alone (hazard ratio 0.88, P = 0.185).

CONCLUSIONS

Independent of patient- and disease-related factors, men with metastatic castration-resistant prostate cancer who had undergone prior radical prostatectomy have improved overall and cancer-specific survival compared with those with no prior local therapy.

Comparative Effectiveness of Local and Systemic Therapy for T4 Prostate Cancer

OBJECTIVE

To evaluate the comparative effectiveness of local vs systemic therapy among patients diagnosed with nonmetastatic clinical T4 prostate cancer.

METHODS

Using the National Cancer Database men with clinical T4N0-1M0 prostate cancer from 2004 to 2013 were identified. Local therapy was defined as radiation (RT with androgen deprivation therapy [ADT]), surgery (radical prostatectomy with ADT), or combined radiation plus surgery (radical prostatectomy plus RT with ADT). Systemic therapy was defined as ADT or chemotherapy alone. The primary outcome of overall survival was estimated using the Kaplan-Meier method. Factors associated with overall survival were determined by Cox proportional hazards models.

RESULTS

A total of 1914 patients were included in our analysis, 1559 received local therapy and 355 received systemic therapy. Median 5-year survival for local vs systemic therapy was 41.5 and 28.2 months, respectively. On multivariable analysis, local therapy was associated with increased overall survival compared to systemic therapy (hazard ratio [HR] = 0.52; 95% confidence interval [CI] 0.44-0.62, P < .001). Comparing local therapy treatment modalities, both radiation (HR = 0.44; 95% CI 0.36-0.53, P < .001) and surgery (HR = 0.67; 95% CI 0.55-0.82, P < .001) were associated with increased overall survival compared to systemic therapy. Among those receiving local therapy, more patients were treated with radiation (n = 709/1559 or 45.5%) compared to surgery (n = 560/1559 or 35.9%) or combined radiation plus surgery (n = 290/1559 or 18.6%) with 5-year overall survival by treatment type being 61%, 51.4%, and 62.2%, respectively.

CONCLUSION

Local therapy for clinical T4 prostate cancer is associated with improved overall survival. Due to the retrospective, nonrandomized nature of the study design, a clinical trial is needed to better define the efficacy of local therapy in this high-risk patient population.

Radiation therapy to the primary in metastatic prostate cancer: palliation only or altering tumor biology?

PURPOSE OF REVIEW

Despite improvement in systemic treatment, the prognosis of men with de novo metastatic prostate cancer remains poor. Treating the local disease may not only reduce the occurrence of local urologic symptoms, but also slow the metastatic process, either by reducing the seeding from the primary tumor or by altering the microenvironment and thus minimizing the formation of new metastatic sites.

RECENT FINDINGS

Retrospective and population-based studies have suggested that the addition of local treatment to systemic therapy may improve survival in this patient group. The aim of this review is to discuss the biologic rationale of such an approach, present and discuss the current available evidence, with a focus on radiation-based treatments. It is key to also address the issue of patient selection as not all patients with metastatic prostate cancer will benefit from the treatment of the primary tumor.

SUMMARY

Retrospective and population-based research suggests a survival benefit of prostatectomy or radiotherapy in metastatic prostate cancer patients. Clinical trials evaluating the role of prostate radiotherapy in the metastatic setting are ongoing.

Review of the comparative effectiveness of radical prostatectomy, radiation therapy, or expectant management of localized prostate cancer in registry data

Evidence regarding the effectiveness of treatment for prostate cancer is primarily based on randomized controlled trials. Long-term outcomes are generally difficult to evaluate within experimental studies and may benefit from large pools of observational data. We conducted a systematic review of administrative and registry studies to evaluate the comparative effectiveness of treatment for clinically localized prostate cancer on overall and prostate-cancer specific mortality.

MATERIALS AND METHODS

In accordance with the preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P, 2015), we conducted a systematic search of Ovid Medline and Embase (1946-February 2017) and identified studies that evaluated the relationship between types of treatment for localized prostate cancer and mortality. Additional articles were identified through manual search. Randomized, prospective, and single institution studies were excluded. The risk of bias for each study was evaluated with the Newcastle Ottawa scale. Multivariable adjusted hazard ratios were reported to evaluate overall and cancer-specific mortality.

RESULTS

We screened 4,721 studies and included for review, 19 that were published between 2001 and 2015. The pooled population included 228,444 patients. Countries of origin included the United States, Canada, China, Switzerland, the Netherlands, and Sweden, and the sources included administrative (n = 6) and cancer registry or prostate databases (n = 11). Overall and cancer-specific mortality were lowest among definitive treatment arms as compared to conservative therapy with no treatment, observation, or active surveillance. Radiotherapy was associated with worse overall and cancer-specific mortality than radical prostatectomy.

CONCLUSION

Although observational studies using large, population-based cohorts have the potential for bias, we found consistent evidence that high-quality observational studies may be used to evaluate the comparative effectiveness of prostate cancer treatment. Methodologic limitations of observational data should be considered.

Association between type 2 diabetes, curative treatment and survival in men with intermediate- and high-risk localized prostate cancer

OBJECTIVE

To investigate whether curative prostate cancer (PCa) treatment was received less often by men with both PCa and Type 2 diabetes mellitus (T2DM) as little is known about the influence of T2DM diagnosis on the receipt of such treatment in men with localized PCa.

SUBJECTS AND METHODS

The Prostate Cancer database Sweden (PCBaSe) was used to obtain data on men with T2DM and PCa (n = 2210) for comparison with data on men with PCa only (n = 23 071). All men had intermediate- (T1-2, Gleason score 7 and/or prostate-specific antigen [PSA] 10-20 ng/mL) or high-risk (T3 and/or Gleason score 8-10 and/or PSA 20-50 ng/mL) localized PCa diagnosed between 1 January 2006 and 31 December 2014. Multivariate logistic regression was used to calculate the odds ratios (ORs) for receipt of curative treatment in men with and without T2DM. Overall survival, for up to 8 years of follow-up, was calculated both for men with T2DM only and for men with T2DM and PCa.

RESULTS

Men with T2DM were less likely to receive curative treatment for PCa than men without T2DM (OR 0.78, 95% confidence interval 0.69-0.87). The 8-year overall survival rates were 79% and 33% for men with T2DM and high-risk PCa who did and did not receive curative treatment, respectively.

CONCLUSIONS

Men with T2DM were less likely to receive curative treatment for localized intermediate- and high-risk PCa. Men with T2DM and high-risk PCa who received curative treatment had substantially higher survival times than those who did not. Some of the survival differences represent a selection bias, whereby the healthiest patients received curative treatment. Clinicians should interpret this data carefully and ensure that individual patients with T2DM and PCa are not under- nor overtreated.