Abstract
Purpose
Utilization of neoadjuvant chemotherapy (NAC) for the management of muscle-invasive bladder cancer remains low. We sought to understand our practice of NAC use in order to design a quality improvement initiative geared towards optimizing medical oncology referral.
Materials and Methods
We identified 339 patients with ≥cT2 bladder cancer treated with radical cystectomy between 2012-2017 at our institution. We assessed the rate of referral to medical oncology, rate of NAC administration, as well as medical, patient and provider variables associated with NAC use. Bayesian logistic regression modeling identified variables associated with NAC use and chart review provided granular patient-level data.
Results
85% (n=289) of patients were referred to medical oncology and 62.5% (n=212) received NAC. Renal insufficiency, hearing loss, and treating urologist were conclusively associated with lower odds of NAC use. 46 patients were not referred to medical oncology and 50% of these had medical contraindications to cisplatin cited as the reason for no referral. 38 patients met with medical oncology but did not receive NAC. 30 (79%) had comorbidities that impacted this decision with 15 (39%) ineligible based on impaired renal function.
Conclusions
Despite the relatively high rates of medical oncology referral and NAC use in this cohort, there are still opportunities to improve the efficiency of this practice. Quality improvement initiatives could optimize the referral of patients with ≥T2 bladder cancer for consideration of cisplatin-based NAC and establish an important quality metric in the management of these patients.
Keywords: neoadjuvant chemotherapy, quality improvement, bladder cancer
Introduction
Level 1 evidence supports the use of cisplatin-based neoadjuvant chemotherapy (NAC) prior to radical cystectomy for patients with muscle-invasive bladder cancer (MIBC).1,2 NAC is a guideline-recommended treatment that offers an approximately 5% survival benefit at 5 years among patients who undergo radical cystectomy.3–6 Despite this significant oncological improvement, utilization of NAC is poor, even in academic settings. Population-based outcome studies for patients with MIBC found that only 4% of patients received NAC between 1994 – 2008; an updated analysis showed only a slight improvement in NAC use up to 19% for patients treated between 2009 – 2013.7,8 More recent series have shown rates of NAC usage ranging from 21-32% across the United States, Canada and the Netherlands. 8–11 Incorporation of NAC in the management of MIBC remains low despite its known survival benefit and recommendation for use across urologic and oncologic guidelines.
While previous studies have analyzed trends of NAC use as well as factors associated with completion of NAC, few studies have evaluated specific factors that may play a role in the under-utilization of these therapies.12,13 It is evident that NAC use improves with increasing referral rates to medical oncology.8 In addition to a multi-disciplinary approach to oncologic care, some institutions have adopted clinical practice guidelines or developed clinical risk stratification tools to improve utilization of NAC;11,14,15 yet identifying strategies to increase NAC utilization remains a challenge. To our knowledge, no previous work has focused on addressing the use of NAC through the lens of quality improvement.
Thus, we sought to evaluate and identify the factors associated with receiving cisplatin-based NAC after urologic evaluation at our institution. We hypothesized that medical, patient, and provider-specific factors all play roles in the use of NAC. In this study, we aimed to understand NAC utilization, identify modifiable factors that could improve the utilization of NAC, and inform the design of a quality improvement (QI) intervention to improve the appropriateness of medical oncology referral.
Materials and Methods
Patient Population
We retrospectively analyzed a prospectively maintained database of patients with MIBC who underwent cystectomy at our institution. Between 2012-2017, a total of 632 cases were identified. Study inclusion criteria included patients who consented to participate in our institutional database and had biopsy-proven MIBC treated with radical cystectomy for primary urothelial cancers, including those with divergent histology components. Those with “pure” non-urothelial neoplasms of the bladder (small cell, squamous, adenocarcinoma, or metastasis from non-urologic cancer) or had concurrent secondary malignancies were excluded.
Data Collection
After institutional IRB approval, we gathered information on referral to medical oncology and administration of cisplatin-based NAC. We collected demographic information and data on overall mortality, disease-specific mortality, and recurrence rates. The location of nearest oncologist and travel distances were determined and calculated through Google search and Google Maps™ based on the patient’s home address. We identified medical, patient, and provider-related variables through review of documentation in the electronic medical record (EMR).
Statistical Analysis
Rates of NAC utilization were determined using simple proportions. We endeavored to identify variables that were associated with the administration of NAC. Continuous variables included age, body mass index (BMI), and Eastern Cooperative Oncology Group (ECOG) score while all others were categorical variables. A multivariate Bayesian logistic regression model was used to estimate the association of variables with log-odds of NAC use.16 Compared to classical statistical modeling, this approach allows for estimation of the probability of association of different variables and helps differentiate true effect from noise.17 Random intercept modeling was used to account for physician behavior, using a theoretical bell curve to estimate whether physicians are making similar decisions or if variability exists across providers. Analysis was completed in R version 3.5.0 utilizing the ‘brms’ package.
The results were interpreted in the context of 95% uncertainty intervals instead of p-values. If the 95% uncertainty interval for the odds ratio includes 0, the association is said to be “inconclusive.” In cases where results are “inconclusive,” probability of association >80% is defined as “potentially associated” while those variables with association <80% were not considered for additional evaluation or future intervention. This will inform our decisions concerning the design of a QI initiative moving forward.
Qualitative Analysis of Electronic Medical Record
We identified two sub-cohorts of interest: patients who were not referred to medical oncology and those who were referred but did not receive NAC. Through review of the EMR, we identified medical, patient and provider factors documented as reasons for omission of medical oncology referral or lack of NAC usage after medical oncology consultation.
Results
Neoadjuvant Chemotherapy Use and Factors Associated with NAC
Between 2012 and 2017, we identified 339 patients who met inclusion criteria (Table 1). 289 (85%) of patients with MIBC were referred to medical oncology for consideration of NAC and 212 patients (62.5%) received NAC prior to cystectomy. For patients who received NAC, 2-year mortality after diagnosis was 22.4%, compared to 33.6% for those who did not (OR 0.53, 95% UI 0.30-0.92). Similarly, 2-year overall mortality after cystectomy was 29.4% for patients who received NAC, compared to 42.4% for those who did not (OR 0.57, 95% UI 0.33 – 0.97). 2-year rate of recurrence after cystectomy for patients who received NAC was 15.2%, compared to 17.3% for those who did not (OR 1.16, 95% UI 0.61 – 2.27).
Study cohort demographic and clinical characteristics
| Demographics | Clinical Characteristics | ||
|---|---|---|---|
| Age, Median (IQR) | Clinical AJCC stage, No. (%) | ||
| Age | 66 (59 – 74) | II (T2NxMx) | 221 (65.2%) |
| Sex, No. (%) | III (T3NxMx) | 99 (29.2%) | |
| Female | 78 (23.0%) | IV (T4NxMx) | 19 (5.6%) |
| Male | 261 (77.0%) | Medical Co-morbidities, No. (%) | |
| Race/Ethnicity, No. (%) | Estimated Glomerular Filtration Rate (eGFR) < 60 | 92 (27.1%) | |
| Caucasian | 320 (94.4%) | Hearing Loss | 31 (9.1%) |
| Black | 11 (3.2%) | Peripheral Neuropathy | 17 (5.0%) |
| Other | 8 (2.4%) | Congestive Heart Failure | 8 (2.4%) |
| Marital Status, No. (%) | Solitary Kidney | 9 (2.7%) | |
| Married | 246 (72.6%) | BMI, median (IQR) | 28.0 (24.3 – 30.7) |
| Single | 56 (16.5%) | ECOG Scale of Performance Status, No. (%) | |
| Widowed | 19 (5.6%) | 0 | 174 (51.3%) |
| Divorced | 17 (5.0%) | 1 | 138 (40.7%) |
| Unknown | 1 (0.3%) | 2 | 22 (6.5%) |
| Primary Insurance, No. (%) | 3 | 4 (1.2%) | |
| Private | 144 (42.5%) | Not documented | 1 (0.3%) |
| Public | 187 (55.2%) | Pathology, No. (%) | |
| Uninsured | 8 (2.4%) | Divergent Histology | 115 (33.9%) |
| Geographic variables; median (IQR) | |||
| Distance to our institution | 110.5 miles (40.3 – 119.7) | ||
| Distance to nearest medical oncologist | 13.5 miles (3.6 – 17.0) | ||
The Bayesian regression model with a horseshoe prior on population level effects resulted in two variables with conclusively lower odds of receiving NAC: the presence of renal insufficiency (95% uncertainty interval (UI) −2.73 to −0.94) and hearing loss (95% UI −3.23 to −0.27; Figure 1). A diagnosis of CHF was inconclusive (95% UI −6.57 – 0.22) but potentially associated with decreased odds of patients receiving NAC. Likewise, increasing age (95% UI −0.64 – 0.04), poor functional status (ECOG score >2; 95% UI −0.61 – 0.06), and greater distance to our institution (95% UI −0.56 – 0.09) yielded inconclusive results, but are potentially associated with decreased odds of receiving NAC. Conversely, higher BMI (95% UI −0.08 – 0.58) and greater distance to nearest oncologist (95% UI −0.06 – 0.71) yielded inconclusive results but are potentially associated with increased odds of receiving NAC (Table 2).
Bayesian regression models to identify variables of interest
A. Normal plot: This figure shows the results of Bayesian logistic regression modeling assuming a normal distribution on the population level parameters for variables associated with NAC. Median effect size highlighted with dark bar.
B. Horseshoe prior plot: Relative to use of a normal prior, the use of a horseshoe prior reduces small associations in the data that arise most likely from noise and emphasizes particularly strong associations. A horseshoe prior reduces variables that are near zero effect toward zero, while maintaining those variables that appear to have a true association with use of NAC. Median effect size highlighted with dark bar.
Results of Bayesian Regression Modeling on variable that may impact NAC use
| Variable | Effect Size, Lower 95% – Upper 95% UI | Interpretation |
|---|---|---|
| Renal Insufficiency (eGFR < 60) | −1.78 −2.73 - −0.94 |
Conclusively (>95% probability) associated with lower log-odds of receiving NAC. |
| Hearing Loss | −1.78 −3.23 - −0.27 |
Conclusively (>95% probability) associated with lower log-odds of receiving NAC. |
| CHF | −1.62 −6.57 – 0.22 |
Probably (>80% probability) associated with lower log-odds of receiving NAC. |
| Increasing BMI | 0.18 −0.08 – 0.58 |
Probably (>80% probability) associated with higher log-odds of receiving NAC. |
| Increasing Age | 0.25 −0.64 – 0.04 |
Probably (>80% probability) associated with lower log-odds of receiving NAC. |
| Increasing ECOG | −0.21 −0.61 – 0.06 |
Probably (>80% probability) associated with lower log-odds of receiving NAC. |
| Distance to nearest medical oncologist | 0.23 −0.06 – 0.71 |
Probably (>80% probability) associated with higher log-odds of receiving NAC. |
| Distance to our institution | −0.16 −0.56 – 0.09 |
Probably (>80% probability) associated with lower log-odds of receiving NAC. |
Random intercept modeling, as plotted in the horseshoe prior plot, demonstrates significant variability in treatment decisions among urologists and oncologists. Though inconclusive, the individual urologist and oncologist potentially influenced the decision to administer NAC, with some physicians being associated with higher log-odds of NAC administration and others being associated with lower log-odds of NAC administration (Figure 1).
Patients not referred to Medical Oncology
46 patients were not referred to medical oncology prior to cystectomy (Table 3). We divided the reasons patients were not referred into medical, physician, and patient-related factors. 23 (50%) of patients were not referred to medical oncology due to medical comorbidities including poor renal function (n=6, 13%), presence of divergent histology (n=8, 17%), cardiac history (n=2, 4%), age (n=2, 4%), hearing loss (n=2, 4%), poor functional status (ECOG score >2 or subjective assessment in clinic; n=1, 2%), and other health issues (n=3, 7%), including anemia, pulmonary embolism (PE) with previous chemotherapy, and Crohn’s disease. 12 (26%) of patients did not have referral due to physician-related factors, including no documented NAC discussion (n=11, 24%) or logistical issues with the consultation being completed locally (n=1, 2%). 11 (24%) did not receive NAC due to patient factors; reasons noted included: local symptoms (n=5, 11%), personal preference after discussing options with their urologist (n=5,11%), or fear of delaying surgery (n=1, 2%; Table 4).
Characteristics of patients not referred to medical oncology
| Demographics | Clinical Characteristics | ||
|---|---|---|---|
| Age, Median (IQR) | Clinical AJCC stage, No. (%) | ||
| Age | 71 (65 – 78) | II (T2NxMx) | 34 (73.9%) |
| Sex, No. (%) | III (T3NxMx) | 10 (21.7%) | |
| Female | 11 (23.9%) | IV (T4NxMx) | 2 (4.3%) |
| Male | 35 (76.1%) | Medical Co-morbidities, No. (%) | |
| Race/Ethnicity, No. (%) | Estimated Glomerular Filtration Rate (eGFR) < 60 | 19 (41.3%) | |
| Caucasian | 45 (97.8%) | Hearing Loss | 5 (10.9%) |
| Black | 1 (2.2%) | Peripheral Neuropathy | 1 (2.2%) |
| Other | 0 (0%) | Congestive Heart Failure | 2 (4.3%) |
| Marital Status, No. (%) | Solitary Kidney | 1 (2.2%) | |
| Married | 35 (76.1%) | BMI, median (IQR) | 26.9 (22.9 – 29.8) |
| Single | 5 (10.9%) | ECOG Scale of Performance Status, No. (%) | |
| Widowed | 5 (10.9%) | 0 | 17 (37.0%) |
| Divorced | 1 (2.2%) | 1 | 22 (47.8%) |
| Primary Insurance, No. (%) | 2 | 5 (10.9%) | |
| Private | 21 (45.7%) | 3 | 1 (2.2%) |
| Public | 24 (52.2%) | Pathology, No. (%) | |
| Uninsured | 1 (2.2%) | Divergent Histology | 16 (34.8%) |
| Geographic variables; median (IQR) | |||
| Distance to our institution | 133.3 (54.3 – 177.5) | ||
| Distance to nearest medical oncologist | 12.2 (3.9 – 15.7) | ||
Documented reasons for no referral to medical oncology
| Medical | Patient | Physician | |||
|---|---|---|---|---|---|
| Divergent Pathology | 8 | Preference | 5 | No discussion documented | 11 |
| Renal function (eGFR <60) | 6 | Local symptoms | 5 | Logistics/Transfer of care | 1 |
| Cardiac history | 2 | Delay surgery | 1 | ||
| Age | 2 | ||||
| Hearing Loss | 2 | ||||
| ECOG score | 1 | ||||
| Other | 3 | ||||
Patients referred to Medical Oncology who did not receive NAC
38 patients had a medical oncology consultation but did not proceed with NAC prior to radical cystectomy (Table 5). For this group, there was often a combination of medical, patient, or physician-related factors documented in their discussion regarding chemotherapy. 30 (79%) patients had medical factors cited, including decreased renal function (eGFR < 60; n=15, 39%), cardiac history (n=7, 18%), age (n=5, 13%), neuropathy (n=4, 11%), hearing loss (n=3, 8%), presence of divergent histology (n=2, 5%) or other co-morbidities (n=6, 16%) (e.g. pulmonary embolism on lifelong anti-coagulation, chronic nausea and vomiting complicated by Mallory-Weiss tear, severe anemia, recurrent septic episodes, myelodysplastic syndrome, and severe chronic obstructive pulmonary disease). Patient-related factors influenced the decision for NAC due to personal preference (n=5, 13%), past experience with chemotherapy (n=2, 5%), concern of delaying surgery (n=2, 5%) and local symptoms (n=1, 3%). In two cases, there were physician factors impacting the decision and both were due to logistics/coordination of care - one patient lived too far to travel for NAC and another had issues scheduling an appointment with a local oncologist (Table 6).
Characteristics of patients referred to medical oncology who did not receive NAC
| Demographics | Clinical Characteristics | ||
|---|---|---|---|
| Age, Median (IQR) | Clinical AJCC stage, No. (%) | ||
| Age | 70.9 (65.3 – 79.0) | II (T2NxMx) | 29 (76.3%) |
| Sex, No. (%) | III (T3NxMx) | 7 (18.4%) | |
| Female | 10 (26.3%) | IV (T4NxMx) | 2 (5.3%) |
| Male | 28 (73.7%) | Medical Co-morbidities, No. (%) | |
| Race/Ethnicity, No. (%) | Estimated Glomerular Filtration Rate (eGFR) < 60 | 24 (63.2%) | |
| Caucasian | 38 (100%) | Hearing Loss | 6 (15.8%) |
| Black | 0 (0%) | Peripheral Neuropathy | 6 (15.8%) |
| Other | 0 (0%) | Congestive Heart Failure | 4 (10.5%) |
| Marital Status, No. (%) | Solitary Kidney | 2 (5.3%) | |
| Married | 26 (68.4%) | BMI, median (IQR) | 26.4 (23.6 – 29.0) |
| Single | 8 (21.1%) | ECOG Scale of Performance Status, No. (%) | |
| Widowed | 2 (5.3%) | 0 | 20 (52.6%) |
| Divorced | 2 (5.3%) | 1 | 13 (34.2%) |
| Primary Insurance, No. (%) | 2 | 3 (7.9%) | |
| Private | 15 (39.5%) | 3 | 2 (5.3%) |
| Public | 23 (60.5%) | Pathology, No. (%) | |
| Geographic variables; median (IQR) | Divergent Histology | 10 (26.3%) | |
| Distance to our institution | 101.3 (38.0 – 158.5) | ||
| Distance to nearest medical oncologist | 10.6 (3.6 – 16.3) | ||
Documented reasons for not receiving NAC after consultation with medical oncology
| Medical | Patient | Physician | |||
|---|---|---|---|---|---|
| Renal function (eGFR <60) | 15 | Preference | 5 | Logistics/Transfer of care | 2 |
| Cardiac history | 7 | Previous experience chemotherapy | 2 | ||
| Age | 5 | Delaying surgery | 2 | ||
| Neuropathy | 4 | Local symptoms | 1 | ||
| Hearing Loss | 3 | ||||
| Divergent pathology | 2 | ||||
| Other | 6 | ||||
Discussion
Despite changes in surgical and medical therapy, mortality resulting from urothelial bladder cancer has not improved in the last 30 years.18 Within this cohort, renal insufficiency (eGFR < 60ml/min) and hearing loss were conclusively associated with lower odds of receiving NAC. Congestive heart failure, older age and higher ECOG also seem to decrease the odds of receiving NAC. These findings are consistent with established contraindications to platinum-based chemotherapy, namely ECOG performance status > 2, GFR <60mL/min, as well as hearing loss and neuropathy.19 We hypothesize that increased distance to local medical oncologists is potentially associated with increased odds of NAC use because those patients may be more likely to travel to our institution for medical oncology evaluation where rates of NAC use are high. While known medical contraindications were associated with decreased NAC use, the variability of medical oncology referral across urologists provides an avenue for targeted intervention. Urologists are the gatekeepers for the management of MIBC and referral to medical oncology is a modifiable component of this care delivery pathway.
Our institutional rate of NAC administration was 62.5% between 2012 and 2017 - higher than previously reported rates in the literature (21 to 32% across multiple series).8–11 This high rate of NAC use is attributed to the formal collaboration between the departments of urology and medical oncology through weekly multi-disciplinary clinics and tumor boards, which facilitate referral to medical oncology. For providers and health systems who aim to increase institutional NAC use, initial steps include understanding the current state of NAC use and medical oncology referral rates. Local initiatives should identify opportunities to increase communication and collaboration with medical oncologists.
Despite our high rate of medical oncology referral and NAC use, we identified two groups of patients who could benefit from QI interventions to optimize referral to medical oncology. First, there are patients who are candidates for chemotherapy and should have the opportunity to weigh risks and benefits through discussion with medical oncologists. The most common reasons (50%) for not referring patients to medical oncology were medical comorbidities. These included impaired renal function, poor functional status, presence of divergent histology, hearing loss and cardiac comorbidities. Of these factors, only renal function and functional status have identifiable cut offs for cisplatin-based chemotherapy regimens in the literature, namely GFR < 60ml/min and ECOG > 2. Proper assessment of hearing, neurologic and cardiac impairments are more nuanced and are unlikely to be performed in a urology clinic. Similarly, patients with components of divergent histology could still benefit from a balanced discussion about NAC. Furthermore, 23 (50%) patients not referred to medical oncology lacked medical contraindications to therapy, suggesting potential missed opportunities for medical oncology consultation. Therefore, medical oncology referral should be offered to most patients unless, of course, a patient or their family declines evaluation.
The second group of patients who stand to benefit from improvements in care pathways, are those with contra-indications to cisplatin-based NAC and who should have an expedited path to surgery. Renal function and performance status can be assessed by urologists. Patients with ECOG performance status >2 and decreased renal function, as agreed upon by local medical oncologists, could bypass an oncologic evaluation saving time and, possibly, money. For instance in our cohort, 15 (39%) of the patients referred to medical oncology who did not receive NAC were ineligible for cisplatin-based chemotherapy based on renal function alone.19 To address these issues, we have designed a QI intervention focused on physician education and prospectively measuring a new quality metric - rate of appropriate referral (see Figure 2). This approach will optimize the efficiency of the care pathway for MIBC patients at our institution. It will continue to promote a high rate of medical oncology referral with the aim of reducing the variability of referral across providers. Incorporating this new quality indicator could ultimately shed light on target rates of medical oncology referral.
Proposed quality metric: rate of appropriate NAC referral
In this QI intervention, we will hold an educational session for our urologists who focus on bladder cancer to emphasize the contraindications to NAC and highlight scenarios in which medical oncology referral may still be indicated (e.g. hearing loss, neuropathy and divergent histology). We will also design a patient NAC education pamphlet to discuss benefits and risks of NAC. We will use an EMR prompt to encourage urologists to discuss NAC with patients with ≥T2 bladder cancer. It will list absolute contra-indications to NAC agreed upon by our medical oncology colleagues and help document reasons if medical oncology referral was not pursued.19 In this manner, we aim to maximize referrals for those patients who do not have medical contra-indications for NAC, limit inappropriate referrals and expedite cystectomy for appropriate patients, and facilitate a comparison of pre and post-intervention appropriate referral rates and institutional NAC use. These findings could help define ideal rates of medical oncology referral, NAC use, and establish potential quality benchmarks for MIBC management.
Our study has several limitations. First, this is a retrospective analysis of data from a single institution which may not be generalizable to other health systems and practice settings. Second, the Bayesian regression horseshoe prior plots require subjective interpretation; however, this interpretation is useful for identifying variables of interest for future QI interventions. Our sample size is relatively small, but this is balanced by the granularity of the data provided in order to design our local intervention. Our identification of physician, patient or medical factors associated with NAC use are based upon review of the medical record and this qualitative analysis is limited to what is explicitly documented in the chart. It is important to note that while GFR<60ml/min was used as cutoff of decreased renal function precluding cisplatin-based chemotherapy there may be patients in which renal function could be optimized or where chemotherapy may remain an option depending on local oncologist comfort and expertise with alternative regimens (for example split-dose cisplatin).20,21 Finally, as advances in immunotherapy and molecular profiling of bladder cancer provide additional options for patients there will be a growing role for multi-disciplinary care.
Conclusions
Our use of NAC use is consistent with our high referral rate to medical oncology. Despite this, we identified patients not referred who may have benefited from medical oncology consultation and others with absolute contraindications to NAC who could have bypassed the referral process. A Quality improvement initiative that will follow endeavors to increase the rate of appropriate medical oncology referral, maintain our high rate of NAC use and establish a possible quality indicator for this practice.
Footnotes
Publisher's Disclaimer: DISCLAIMER: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our subscribers we are providing this early version of the article. The paper will be copy edited and typeset, and proof will be reviewed before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to The Journal pertain.
References
- 1.International collaboration of trialists on behalf of the Medical Research Council Advanced Bladder Cancer Working Party, EORTC Genito-Urinary Group, Australian Bladder Cancer Study Group NCI of CCT, Group, Finnbladder, Norwegian Bladder Cancer Study Group and CUE de TO, Group (CUETO). Neoadjuvant cisplatin,methotrexate,and vinblastine chemotherapy for muscle-invasive bladder cancer: a randomised controlled trial. Lancet. 1999;354(9178):533–540. doi: 10.1016/S0140-6736(99)02292-8 [DOI] [PubMed] [Google Scholar]
- 2.Grossman B, Natale R, Tangen C, Crawford D. Neoadjuvant Chemotherapy plus Cystectomy Compared with Cystectomy Alone for Locally Advanced Bladder Cancer. N Engl J Med. 2003;349:859–866. doi: 10.1056/NEJMoa022148 [DOI] [PubMed] [Google Scholar]
- 3.Clark PE, Agarwal N, Biagioli MC, et al. Bladder cancer. J Natl Compr Canc Netw. 2013;11(4):446–475. doi: 10.6004/jnccn.2013.0059 [DOI] [PubMed] [Google Scholar]
- 4.Alfred Witjes J, Lebret T, Comperat EM, et al. Updated 2016 EAU Guidelines on Muscle-invasive and Metastatic Bladder Cancer. Eur Urol. 2017;71(3):462–475. doi: 10.1016/j.eururo.2016.06.020 [DOI] [PubMed] [Google Scholar]
- 5.Chang SS, Bochner BH, Chou R, et al. Treatment of Non-Metastatic Muscle-Invasive Bladder Cancer: AUA/ASCO/ASTRO/SUO Guideline. J Urol. 2017;198(3):552–559. doi: 10.1016/j.juro.2017.04.086 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vale CL. Neoadjuvant chemotherapy in invasive bladder cancer: Update of a systematic review and meta-analysis of individual patient data. Eur Urol. 2005;48(2):202–206. doi: 10.1016/j.eururo.2005.04.006 [DOI] [PubMed] [Google Scholar]
- 7.Booth CM, Robert Siemens D, Peng Y, Mackillop WJ. Patterns of referral for perioperative chemotherapy among patients with muscle-invasive bladder cancer: A population-based study. Urol Oncol Semin Orig Investig. 2014;32(8):1200–1208. doi: 10.1016/j.urolonc.2014.05.012 [DOI] [PubMed] [Google Scholar]
- 8.Booth CM, Karim S, Brennan K, Siemens DR, Peng Y, Mackillop WJ. Perioperative chemotherapy for bladder cancer in the general population: Are practice patterns finally changing? Urol Oncol Semin Orig Investig. 2018;36(3):89.e13–89.e20. doi: 10.1016/j.urolonc.2017.11.015 [DOI] [PubMed] [Google Scholar]
- 9.Reardon ZD, Patel SG, Zaid HB, et al. Trends in the Use of perioperative chemotherapy for localized and locally advanced muscle-invasive bladder cancer: A sign of changing tides. Eur Urol. 2015;67(1):165–170. doi: 10.1016/j.eururo.2014.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hermans TJN, Fransen Van De Putte EE, Horenblas S, et al. Perioperative treatment and radical cystectomy for bladder cancer - A population based trend analysis of 10,338 patients in the Netherlands. Eur J Cancer. 2016;54:18–26. doi: 10.1016/j.ejca.2015.11.006 [DOI] [PubMed] [Google Scholar]
- 11.Duplisea JJ, Mason RJ, Reichard CA, et al. Trends and disparities in the use of neoadjuvant chemotherapy for muscle-invasive urothelial carcinoma. Can Urol Assoc J. 2019;13(2):24–28. doi: 10.5489/cuaj.5405 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hugar LA, Yabes JG, Turner RM, et al. Rate and Determinants of Completing Neoadjuvant Chemotherapy in Medicare Beneficiaries With Bladder Cancer: A SEER-Medicare Analysis. Urology. November 2018. doi: 10.1016/J.UROLOGY.2018.11.001 [DOI] [PubMed] [Google Scholar]
- 13.Huo J, Ray-Zack MD, Shan Y, et al. Discerning Patterns and Quality of Neoadjuvant Chemotherapy Use Among Patients with Muscle-invasive Bladder Cancer. Eur Urol Oncol. August 2018. doi: 10.1016/j.euo.2018.07.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Miles BJW, Fairey AS, Eliasziw M, et al. Referral and treatment rates of neoadjuvant chemotherapy in muscle-invasive bladder cancer before and after publication of a clinical practice guideline. J Can Urol Assoc. 2010;4(4):263–267. doi: 10.5489/cuaj09134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Culp SH, Dickstein RJ, Grossman HB, et al. Refining Patient Selection for Neoadjuvant Chemotherapy before Radical Cystectomy. J Urol. 2014;191(l):40–47. doi: 10.1016/j.juro.2013.07.061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bürkner P-C. Advanced Bayesian Multilevel Modeling with the R Package brms. R J 2018;10(1):395–411. doi: 10.32614/RJ-2018-017 [DOI] [Google Scholar]
- 17.Carvalho CM, Poison NG, Scott JG. The horseshoe estimator for sparse signals. Biometrika. 2010;97(2):465–480. doi: 10.1093/biomet/asq017 [DOI] [Google Scholar]
- 18.de Vere White R, Lara PN, Black PC, Evans CP, Dall’Era M. Framing Pragmatic Strategies to Reduce Mortality From Bladder Cancer: An Endorsement From the Society of Urologic Oncology. J Clin Oncol. 2020:JC01901731. doi: 10.1200/JC0.19.01731 [DOI] [PubMed] [Google Scholar]
- 19.Gaisky MD. Panel Proposes Consensus Definition of Cisplatin Ineligibility Criteria; 2011. https://www.healio.com/hematology-oncology/genitourinary-cancer/news/print/hemonc-today/%7Bbaba0de7-bdef-489e-8aec-3f3037091af6%7D/panel-proposes-consensus-definition-of-cisplatin-ineligibility-criteria. Accessed January 12, 2020.
- 20.Hussain S A study of split-dose cisplatin-based neo-adjuvant chemotherapy in muscle-invasive bladder cancer. Oncol Lett. January 2012. doi: 10.3892/ol.2012.563 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Osterman CK, Babu DS, Geynisman DM, et al. Efficacy of Split Schedule Versus Conventional Schedule Neoadjuvant Cisplatin - Based Chemotherapy for Muscle - Invasive Bladder Cancer. Oncologist. 2019;24(5):688–690. doi: 10.1634/theoncologist.2018-0561 [DOI] [PMC free article] [PubMed] [Google Scholar]