Going the distance in acoustic neuroma resection: microsurgical outcomes at high-volume centers of excellence

High-volume hospitals are associated with improved surgical outcomes for acoustic neuromas (ANs). Due to the benign and slow-growing nature of ANs, many patients travel to geographically distant cities, states, or countries for their treatment. However, the impact of travel burden to high-volume centers, as well as its relative benefit are poorly understood. We compared post-operative outcomes between AN patients that underwent treatment at local, low-volume hospitals with those that traveled long distances to high-volume hospitals. The National Cancer Database was used to analyze AN patients that underwent surgery (2004–2015). Patients in the lowest quartile of travel distance and volume (Short-travel/Low-Volume: STLV) were compared to patients in the highest quartile of travel distance and volume (Long-travel/High-Volume: LTHV). Only STLV and LTHV cases were included for analysis. Of 13,370 cases, 2,408 met inclusion criteria. STLV patients (n = 1,305) traveled a median of 6 miles (Interquartile range [IQR] 3–9) to low-volume centers (median 2, IQR 1–3 annual cases) and LTHV patients (n = 1,103) traveled a median of 143 miles [IQR 103–230, maximum 4,797] to high-volume centers (median 34, IQR 28–42 annual cases). LTHV patients had lower Charlson/Deyo scores (p = 0.001), mostly received care at academic centers (81.7% vs. 39.4%, p < 0.001), and were less likely to be minorities (7.0% vs. 24.2%, p < 0.001) or underinsured (4.2% vs. 13.8%, p < 0.001). There was no difference in average tumor size. On multivariable analysis, LTHV predicted increased likelihood of gross total resection (odds ratio [OR] 5.6, 95% confidence interval [CI] 3.8–8.4, p < 0.001), longer duration between diagnosis and surgery (OR 1.3, 95% CI 1.0-1.6, p = 0.040), decreased length of hospital stay (OR 0.5, 95% CI 0.4–0.7, p < 0.001), and greater overall survival (Hazard Ratio [HR] 0.6, 95% CI 0.4–0.95, p = 0.029). There was no significant difference in 30-day readmission on adjusted analysis. Although traveling farther to high-volume centers was associated with greater time between diagnosis and treatment for AN patients, they experienced superior postoperative outcomes compared to patients who received treatment locally at low-volume centers. Enabling access and travel to high-volume centers may improve AN patient outcomes.


Introduction
Acoustic neuromas (ANs) are benign nerve sheath tumors of the vestibular nerve, occurring in approximately 1 in 100,000 people. Treatment options include observation, radiation, and microsurgical resection. Improved outcomes have been observed after surgery for ANs at high-volume surgical centers, with lower morbidity and mortality and improved facial nerve and hearing preservation [1][2][3][4][5][6][7][8][9]. These findings are consistent with volume-outcome relationships that have been identified for numerous surgical procedures.
With the increasing nationwide focus on patient-centered and value-based care, initiatives such as the Leapfrog Group have instituted minimum surgeon and hospital standards for certain common high-risk surgical procedures such as bariatric surgery, carotid endarterectomy, and total hip replacement. Currently, there are no skull base neurosurgery or otolaryngology procedures on this list. The need for highvolume, multidisciplinary dedicated teams for complex medical conditions such as ANs have led to the concept of Centers of Excellence (CoE), [10]. although there is no formal definition of CoE in regard to AN treatment. There is a growing body of research on this topic, in addition to the developing role of social media and patient-centered organizations such as the Acoustic Neuroma Association (ANA) in patient decision-making. Therefore, patients diagnosed with ANs may have the choice to pursue treatment at a local, low-volume hospital versus traveling larger distances to reach a high-volume center. The objective of this study was to determine the impact of travel burden and hospital volume on post-operative outcomes of patients with ANs.

Patient population
The National Cancer Database (NCDB) is a joint program of the Commission on Cancer and the American Cancer Society. Nearly 70% of new cancer cases among patients ≥ 18 years of age in the United States from over 1500 hospital clinics are included in the database. In our analysis, we included patients diagnosed with an acoustic neuroma / vestibular schwannoma (histology codes 9560 or 9570, and site code C72.4) between 2004 and 2015. Patients were included if they underwent surgery. Patients who did not receive surgery or had an unknown surgical history were excluded.

Statistical analysis
The primary exposures were travel distance and hospital volume. Within the NCDB, travel distance is defined as the number of miles as the crow flies between a patient's residential zip code centroid and the zip code of the treating hospital. Annual surgical volume was calculated using unique facility identifiers. Travel distance and surgical volume were each divided into quartiles. Patients in the lowest quartile of travel distance and the lowest quartile of hospital volume were designated as Short-Travel/Low-Volume: STLV. Patients in the highest quartile of travel distance and the upper quartile of hospital volume were categorized as Long-Travel/High-Volume: LTHV. Patients not meeting criteria for either group were excluded from the final analysis. Outcomes included extent of resection (coded in the NCDB starting in 2010, and categorized as biopsy, subtotal resection, or gross total resection), receipt of postoperative radiation, post-treatment 30-day readmission, overall survival, time between diagnosis and surgery, and length of stay (LOS). Long LOS was defined as the top quartile of LOS.
Binary outcomes were investigated using Pearson's chisquare tests for univariable analysis and logistic regression for multivariable analysis. Continuous and ordinal variables were analyzed using t-tests and Wilcoxon rank-sum tests, respectively. Univariable and multivariable analyses of overall survival (OS) were completed with Cox proportional hazards models with results presented as hazard ratios (HRs). Multivariable analyses were adjusted a priori for the following patient, clinical, tumor, and hospital characteristics: age at diagnosis, race/ethnicity, Charlson-Deyo comorbidity score, insurance status, median household income of zip code, tumor laterality, tumor size, hospital type, hospital region, and hospital safety-net burden. The Charlson-Deyo score is a weighted age-independent comorbidity score calculated from the sum of scores for specified conditions to predict long-term survival [11]. Safety-net burden of the treatment facility was defined as the proportion of Medicaid and uninsured patients treated at each hospital included in the NCDB [12]. Survival analysis was also adjusted for the receipt of postoperative radiation. Race/ethnicity included white, black, Hispanic, Asian/Pacific Islander, or other race. Patients were categorized as underinsured if they were uninsured or covered by Medicaid. Safety-net burden was defined as the proportion of Medicaid and uninsured patients treated at each hospital. Statistical analysis was completed with Stata MP Version 14.1 (Stata Corp LP). All tests were two-sided with statistical significance defined as p < 0.05.  Multivariable regression analyses were completed with both hospital volume and residential distance from the hospital treated as continuous variables, and then as categorical variables in quartiles. Hospital volume was significantly associated with increased likelihood of GTR and longer duration between diagnosis and surgery, reduced odds of 30-day readmission and prolonged length of stay, and lower HR of death (Table 2). No consistent relationship was identified between distance and any of the outcomes studied.

Results
Overall, 13,370 AN patients underwent surgery in the NCDB between 2004 and 2015. Of these, 1,305 patients were classified as STLV and traveled a median of 6 miles (Interquartile range [IQR] 3-9) to 439 low-volume centers, which performed a median of 2 (IQR 1-3) annual cases. In comparison, the LTHV cohort (n = 1,103) traveled a median of 143 miles (IQR 103-230) to 31 high-volume centers, which performed a median of 34 annual cases (IQR 28-42). Patient and hospital characteristics are summarized in Table 1. Regional variation in surgical volume and travel distance is shown in Fig. 1.
Geographic variations included more LTHV patients treated in centers located in the South Atlantic ( ANs. As many patients with ANs do not live near a highvolume center, we sought to determine if overcoming the travel burden to a high-volume center instead of receiving care locally at a low-volume center resulted in a meaningful clinical benefit. Given that ANs are typically slow-growing

Discussion
In this study, we examined the impact of travel distance and hospital volume on treatment characteristics and post-surgical outcomes for a nationwide cohort of newly diagnosed rates, length of stay, costs, need for postoperative intervention for hydrocephalus, and nonroutine discharge [1, 2, 4-6, 8, 9]. For AN surgery, which is highly specialized and is frequently performed by a specific neurosurgery-otolaryngology team (rather than a variety of on-call surgeons) at a given institution, surgeon volume and hospital volume are often strongly correlated.
In the current work, we found that annual hospital operative volume was independently associated with an increased likelihood of GTR and longer OS, as well as a reduced likelihood of long hospital stay or 30-day readmission. Operative mortality has been previously studied in the NCDB [13]. However, a survival benefit associated with high-volume hospitals has not been previously reported and is among the important findings of this study. Due to the predominantly benign pathology of ANs, this likely reflects a difference in the quality of surgical management rather than the natural history of the disease.
GTR is the goal of AN surgery when it can be done safely, as GTR is associated with reduced tumor recurrence that could necessitate reoperation or radiation [14]. However, the goal of GTR for benign schwannomas is superseded by preservation of facial nerve function, and in some cases, hearing preservation or protection of the brainstem. Higher rates of GTR, lower complication rates, and increased likelihood of routine discharge have previously been associated with greater surgeon/hospital volume and experience [1,5,8,9,15]. However, some studies have identified greater morbidity associated with GTR due to adherence between the facial nerve and tumor, particularly just medial to the porous acousticus [16]. Our finding that both surgical morbidity and mortality are lower for high-volume hospitals suggests that highly experienced AN surgeons can achieve GTR while also minimizing the risk of complications. and treated in an elective fashion, patients may have time to arrange for longer travel to a high-volume center without the risk of disease progression or clinical deterioration. Our study demonstrates that despite an expected delay in treatment associated with long travel distance (as well as likely delays in establishing care with a second provider), patients treated at high-volume hospitals are more likely to receive maximal tumor resection and have improved short-and long-term post-operative outcomes (including overall survival) as compared to those treated locally at low-volume centers. This is the first study to characterize the distancevolume relationship on outcomes for ANs.

Volume-outcome relationship in AN
Prior studies have characterized the volume-outcome relationship in ANs. High surgeon volume has been associated with reduced risk for postoperative complications, shorter length of stay, and lower hospital charges [1]. Similarly, high volume centers have been shown to have lower complication         for all AN patients, it remains unclear how to either increase access to high-volume centers or optimize outcomes in lower volume centers for patients with limited resources.
To date, there are limited data regarding the impact of travel burden to high volume centers on outcomes for skull base tumors such as ANs. Patients' decisions to travel substantial distances to high-volume centers are complex and multifactorial. While financial resources and healthcare literacy likely play a significant role, the blossoming of online patient education, social media, and social networking such as the Acoustic Neuroma Association in which patients can directly exchange information regarding the facets of their care undoubtedly have impacted the way patients choose their surgeons. Recognizing the value of traveling to highvolume AN hospitals is increasingly important, as certain employers and insurance carriers may consider providing financial incentives to seeking care at CoE [22].

Limitations
In this study, we calculated hospital volume based on unique hospital identifiers provided by the NCDB. Unfortunately, unique surgeon identifiers were not available, so we were unable to determine surgeon volume. Limitations of this study include the fact that surgeon experience may not equate to surgeon volume (i.e., a high-volume practice in a given year may not necessarily reflect many years of high volume). The same is true for hospitals, which may have a referral pattern that persists even when experienced surgeons retire or leave. High-volume surgeons can move to new institutions and create new high-volume centers; in these situations, the centers have a relatively little experience whereas the surgeon may have decades of it.
Furthermore, retrospective analysis of the NCDB has its own limitations. Data are derived only from Commission on Cancer (CoC)-accredited cancer programs. Additionally, important anatomic characteristics (such as the position of the facial and vestibular nerves relative to the tumor), surgical details (including the approach used), and outcomes that may be important to quality of life (such as postoperative facial nerve and hearing status), were not available. Although the NCDB includes geographic regions which each encompass several states, future analyses would benefit from specific location data that specify radius to a major metropolitan center. Finally, because NCDB data include only the "first course of treatment" (all methods of treatment administered before disease progression or recurrence), the need for additional surgeries or radiation for residual or recurrent tumor cannot be assessed.

Long travel high volume versus short travel low volume
We found that LTHV patients had an increased likelihood of GTR, shorter hospital length of stay, and greater overall survival, suggesting that overcoming the travel burden in order to receive care at a high-volume center is worth the wait for those with the socioeconomic means to do so.
The impact of residential distance on CNS tumor outcome was studied by Johnson et al. using the NCDB [17]. The study population was adolescents and young adults with primary brain and intraspinal tumors. For patients with low-grade astrocytic tumors and ependymomas, increased distance to the hospital was associated with a reduced risk of death and mediated by facility volume. These findings mirror our results for AN, which suggests that this effect may specifically pertain to tumor types that are relatively slow growing and have very high survival.
Interestingly, a study on GBM by Lopez Ramos et al. used methods similar to the current work and also identified a significant survival advantage for GBM patients traveling long distances for high volume care [18]. While rates of GTR were the same between STLV and LTHV hospitals, the utilization of standard-of-care adjuvant chemotherapy was significantly higher at LTHV hospitals, suggesting that aspects of care other than surgical technique may play a critical role in defining outcomes in patients with more aggressive CNS tumors.
In the current work, LTHV patients had fewer medical comorbidities than STLV patients and had lower proportion of Black, Hispanic, or Asian/Pacific Islander patients. They were also more likely to have private insurance and less likely to be underinsured. Unlike STLV hospitals, LTHV hospitals were mostly academic centers. An effect of race on AN management has been previously analyzed, with one prior study of the SEER database reporting that Black patients with ANs were more likely to receive recommendations against surgical resection compared with White patients [19]. Regarding surgical outcomes, studies have identified a higher rate of mortality of African American patients undergoing resection by low-caseload surgeons [20]. This has led prior authors to suggest that perhaps underresourced patient populations should be preferentially managed with minimally invasive options such as observation, fractionated stereotactic radiotherapy, or stereotactic radiosurgery [20]. However, treatment disparities in ANs may also extend to radiation therapy, as an NCDB study found that patients treated with Gamma Knife had greater proportions of private insurance, higher incomes, treatment in academic centers and high-volume facilities compared to those treated with a Linear Accelerator [21]. Although the ideal solution to these findings is an equitable distribution of care