If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Impact of stereotactic body radiation therapy volume on surgical patient selection, short-term survival, and long-term survival in early-stage non–small cell lung cancer
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Address for reprints: Olugbenga Okusanya, MD, Department of Surgery, Thomas Jefferson University Hospital, 1025 Walnut St, College Building Suite 607, Philadelphia, PA 19107.
Division of Thoracic Surgery, Department of Surgery, Thomas Jefferson University Hospital, Philadelphia, PaSidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, Pa
Stereotactic body radiation therapy (SBRT) is increasingly used to treat non–small cell lung cancer. The purpose of this study is to analyze relationships between facility SBRT utilization and surgical patient selection and survival after surgery.
Methods
Data on patients with TI/T2N0M0 lesions and treatment facility characteristics were abstracted from the National Cancer Database, 2008 to 2017. Facilities were stratified using an SBRT/surgery ratio previously associated with short-term survival benefit for patients treated surgically, and by a previously identified surgical volume threshold. Multiple regression analyses, Cox proportional-hazard regressions, and Kaplan–Meier log rank test were employed.
Results
In total, 182,610 patients were included. Proportion of high SBRT:surgery ratio (≥17%) facilities increased from 118 (11.5%) to 558 (48.4%) over the study period. Patients undergoing surgery at high-SBRT facilities had comparable comorbidity scores and tumor sizes to those at low-SBRT facilities, and nonclinically significant differences in age, race, and insurance status. Among low-volume surgical facilities, treatment at a high SBRT-using facility was associated with decreased 30-day mortality (1.8% vs 1.4%, P < .001) and 90-day mortality (3.3% vs 2.6%, P < .001). At high-volume surgical facilities, no difference was observed. At 5 years, a survival advantage was identified for patients undergoing resection at facilities with high surgical volumes (hazard ratio, 0.91; confidence interval, 0.90-0.93 P < .001) but not at high SBRT-utilizing facilities.
Conclusions
Differences in short-term survival following resection at facilities with high-SBRT utilization may be attributable to low surgical volume facilities. Patients treated at high volume surgical facilities do not demonstrate differences in short-term or long-term survival based on facility SBRT utilization.
Facility SBRT volume is associated with differences in 30-day, 90-day, and 5-year survival for patients undergoing surgical resection for early-stage non–small cell lung cancers (NSCLC) at low-volume surgical facilities but not for those treated at high-volume surgical facilities.
See Commentary on page 480.
Surgical resection is the standard of care for the treatment of early non–small cell lung cancer (NSCLC), the leading cause of cancer-related mortality among both men and women in the United States.
Treatment and prognosis of isolated local relapse after stereotactic body radiotherapy for clinical stage I non–small-cell lung cancer: importance of salvage surgery.
It remains unclear whether and how the expanded use of SBRT treatment for early-stage NSCLC has impacted patient selection for surgical treatment and surgical outcomes. In a recently published study, Syed and colleagues
noted a survival advantage for surgical patients based on facility SBRT volume, but this analysis did not evaluate for patient-level factors that may have impacted this trend and did not investigate long-term survival. In this study, facility-level SBRT:surgery utilization ratio greater than 17% was associated with improved 30- and 90-day survival for patients undergoing surgical resection.
The purpose of this analysis is to further explore the relationship between SBRT volume and surgical treatment with focus on patient selection, short-term survival, and long-term survival while accounting for the impact of surgical volume. Specifically, this analysis seeks to confirm the previously identified survival advantages in 30-day and 90-day surgical mortality for patients undergoing surgery at high SBRT volume facilities when segregating facilities based on a previously identified surgical volume threshold associated with improved outcomes and mortality.
Methods
Data Source
We performed a retrospective review of the National Cancer Database (NCDB) for the years 2008 to 2017. The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. This dataset captures 70% of all newly diagnosed cancers in the United States annually reported from approximately 1500 hospitals with Commission on Cancer–accredited cancer programs. The NCDB pathologic staging is classified using the current edition at the time of diagnosis of the American Joint Commission on Cancer staging guidelines, based on the tumor size and/or extension, regional lymph node metastases, and distant metastases upon clinical examination. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analysis or conclusions drawn from this study. This project was deemed exempt from the institutional review board at Thomas Jefferson University (PRC#: 2020-117A, September 7, 2021).
All patients 18 years of age or older with clinical T1 or T2 N0M0 lesions were considered for inclusion. Patients treated with neoadjuvant therapy were excluded from the analysis; patients with preoperative cytology or histology consistent with carcinoid tumor were further excluded. Patients with additional malignancies and unknown survival were also excluded from this study. Patient specific data on age, sex, race, median income, Charlson–Deyo Comorbidity Index (CCI) score, and median tumor size was abstracted. Abstracted facility-level data included hospital type and geographic setting. Finally, clinical outcomes including hospital length of stay, readmission, tumor histology, 30-day surgical mortality, 90-day surgical mortality, and 5-year survival status were abstracted.
Facility SBRT and Surgical Volume Designations
Data were analyzed with respect to facility volume. We utilized the SBRT:surgery volume ratio previously identified by Syed and colleagues.
Facilities performing greater than or equal to an SBRT:surgery ratio of ≥17% in a given year were designated as high SBRT facilities based on the previously identified survival advantage. Hospitals performing ≥40 lobectomies per year were designated high-volume facilities based on previous analyses,
and those performing fewer were designated as low-volume operative facilities. For these calculations, per annuum facility volume was calculated using data from the full NCDB.
Statistical Analysis
Pearson χ2 tests and Student t-tests were utilized to compare patient, facility, and treatment characteristics. Univariable regression analyses were performed comparing patient, hospital, and treatment characteristics. Variables that were of high clinical importance or that were statistically significant on univariable analysis were considered for inclusion in multiple logistic regression modeling. Covariates in the multivariable model included SBRT:surgery ratio, year of diagnosis, patient age, sex, race, CCI score, insurance status, facility type, and type of surgery. Kaplan–Meier log rank test and Cox proportional-hazards ratios were used to evaluate 5-years survival. All analyses utilized STATA/SE 15.1 statistical software (StataCorp LLC).
Results
Treatment Volumes
In total, 182,610 patients were included in the analysis. The number of patients undergoing surgical resection increased over the study period: 13,702 to 23,003 (P < .001), as did SBRT procedure volume: 1089 to 6252, (P < .001) (Figure 1, Video Abstract). The proportion of high SBRT:surgery ratio (H-SBRT; ≥17%) facilities increased from 118 (11.5%) to 558 (48.4%) over the study period (Figure 2).
Figure 1Surgical and stereotactic body radiation therapy volume and ratio by year. SBRT, Stereotactic body radiation therapy.
There were statistically significant differences in patient characteristics between cohorts treated at low SBRT:surgery ratio (L-SBRT) facilities and H-SBRT facilities (Figure 3). Patients differed in median age (L-SBRT: 69 years vs H-SBRT: 68 years, P < .001), sex (L-SBRT: 45.3% males vs H-SBRT: 44.3% male, P < .001), and race (L-SBRT: 84.9% White vs H-SBRT: 87% White, P < .001). Other statistically significant demographic differences between groups including sex, insurance status, median income, and facility type are detailed in Table 1. Groups did not differ in rate of CCI >0 (L-SBRT: 49.1% vs H-SBRT: 48.7%, P < .072).
Figure 3Impact of SBRT volume on surgical patient selection and outcomes. SBRT, Stereotactic body radiation therapy; NSCLC, non–small cell lung cancer; NCDB, National Cancer Database.
There were statistically significant differences in patient characteristics between cohorts treated and L-SBRT facilities and H-SBRT facilities. Patients differed in median age (L-SBRT: 69 years vs H-SBRT: 68 years, P < .001), sex (L-SBRT: 45.3% males vs H-SBRT: 44.3% male, P < .001), and race (L-SBRT: 84.9% White vs H-SBRT: 87% White, P < .001). Other statistically significant demographic differences between groups including sex, insurance status, median income, and facility type are detailed in Table 1. Groups did not differ in rate of CCI >0 (L-SBRT: 49.1% vs H-SBRT: 48.7%, P < .072).
When we compared outcomes, inpatient length of stay did not differ between groups. Rates of readmission differed between groups (L-SBRT: 5.4% readmission vs H-SBRT: 5.1% readmission, P < .0.28), as did 30-day mortality (L-SBRT: 1.6% mortality vs H-SBRT: 1.3% mortality, P < .001) and rates of 90-day mortality (L-SBRT: 2.9% vs H-SBRT: 2.4%, P < .001).
Subgroup Analysis, Surgical Volume
Data were further analyzed regarding facility surgical volume (Table 2). Among facilities with low surgical volume, those with low SBRT:surgery ratios (<17%) and high SBRT:surgery ratios (≥17%) differed in age, sex, race, insurance status, median income, frequency of Charlson-Deyo Score ≥1 (L-Surg/L-SBRT: 49.7% vs L-Surg/H-SBRT: 48.4%, P < .001), year of diagnosis, surgery type, open surgical approach (L-Surg/L-SBRT: 57.6% open vs L-Surg/H-SBRT: 53.8% open, P < .001), academic hospital type (L-Surg/L-SBRT: 23.9% academic vs L-Surg/H-SBRT: 29.9% academic, P < .001), and geographic setting. Groups did not differ in mean tumor size. Regarding outcomes, groups differed in length of state, rate of readmission, 30-day mortality (L-Surg/L-SBRT: 1.9% mortality vs L-Surg/H-SBRT: 1.3% mortality, P < .001) and 90-day mortality (L-Surg/L-SBRT: 3.2% mortality vs L-Surg/H-SBRT: 2.5% mortality P < .001).
Table 2Low and high surgical volume facilities compared by SBRT:surgery ratio
Among facilities with high surgical volumes, there were differences between facilities based on SBRT:surgery ratio in terms of patient age, race, insurance status, income, Charlson-Deyo score ≥1 (H-Surg/L-SBRT: 47.1% vs H-Surg/H-SBRT: 49.4%, P < .001), tumor size, year of diagnosis, surgery type, surgical approach (H-Surg/L-SBRT: 39.8% open vs H-Surg/H-SBRT: 37.4%, P < .001), facility type (H-Surg/L-SBRT: 68.3% academic vs H-Surg/H-SBRT: 61.6% academic, P < .001), and geographic setting. Groups did not differ in readmission rate, 30-day mortality (H-Surg/L-SBRT: 1.1% mortality vs H-Surg/H-SBRT: 1.2% mortality, P = .32), or 90-day mortality (H-Surg/L-SBRT: 2.1% mortality vs H-Surg/H-SBRT: 2.1%, P < .001).
Regression Analysis
On multivariable regression for 30-day survival (Table 3), SBRT:surgery ratio ≥17% was significant: odds ratio (OR), 0.91 (confidence interval [CI], 0.83-0.99), P = .027. Other significant associations with 30-day survival on multivariable analysis included high surgical volume: OR, 0.81 (CI, 0.74-0.91), P < .001, year of diagnosis, age, female sex, Black race: OR, 1.16 (CI, 1.01-1.34), P = .039, insurance status, facility type: OR, community 1.36 (CI, 1.22-1.73), P < 001, CCI score, and extent of resection.
Table 3Univariable and multivariable modeling for 30-day and 90-day mortality
On multivariable analysis for 90-day survival, SBRT:surgery ratio ≥17% was significant: OR, 0.93 (CI, 0.87-0.99), P = .03. Other variables significantly associated with 90-day survival on multivariable regression included surgical volume: OR, 0.84 (CI, 0.78-0.91), P < .001, year of diagnosis, age, female sex, private insurance status, facility type: OR, community 1.50 (CI, 1.32-1.70), P < 001, CCI score, and extent of resection.
Long-Term Survival
Five-year survival data were analyzed with Kaplan–Meier log rank test (Figure 4) and Cox proportional-hazards ratio (Table 4). Kaplan–Meier demonstrated inferior survival for patients treated at low-volume surgical facilities, and a statistically significant difference between patients treated at high SBRT-utilizing (H-SBRT) low surgical volume facilities and low-SBRT-utilizing (L-SBRT) low surgical volume facilities (H-SBRT: 67.61% survival vs L-SBRT: 63.93% survival, P < .01). Survival for patients undergoing resection at high-volume surgical facilities did not differ based on SBRT utilization (H-SBRT: 70.09% vs L-SBRT: 70.71%, P = .23). On Cox multivariable regression, hazard ratio (HR) for operation at high-volume surgical facility was significant: HR, 0.84 (CI, 0.86-0.90), P < .001 but treatment at high SBRT-utilizing facility was not: HR, 0.99 (CI, 0.97-1.00), P = .097.
This analysis has demonstrated the rapid expansion of SBRT utilization since 2008 for the treatment of early-stage NSCLC. It has identified an association between facility SBRT- utilization and short-term and long-term survival following surgical resection. These differences appear isolated to patients undergoing resection at low-volume surgical facilities, as neither an association between SBRT-utilization and short-term nor long-term survival was identified for those undergoing operation at high-volume surgical facilities.
This analysis was intended to redemonstrate the previously identified survival benefit presented by Syed and colleagues with consideration for patient-level factors that may influence survival, and while utilizing of a previously established facility-level surgical volume threshold associated with mortality benefit. Second, we sought to investigate long-term survival after resection. Regarding patient selection, there are several ways in which the presence or absence of SBRT as a treatment option may impact patient and provider decision-making regarding surgical care, thereby impacting patient demographics. First, there is the possibility of a purification effect, whereby poor operative candidates are consciously or subconsciously directed toward SBRT therapy over surgical resection. In theory, the effect of such a trend would be most pronounced at those facilities that perform a low volume of surgical resection and have a proportionally higher utilization of SBRT. In this analysis, we have not found clinically significant differences between this cohort and others in terms of patient age, comorbidity profile, tumor size, or income (Table 2). Second, there is a possibility that at low-volume surgical facilities, the addition of SBRT programs has driven providers to offer surgery to poorer surgical candidates than may have otherwise been considered for operative management. Comparing comorbidity scores, this effect would most likely be seen at low surgical volume facilities with proportionally greater utilization of SBRT. We have seen no evidence of this trend in our analysis, either. More data are needed, particularly regarding patient functional status, pulmonary function testing, and frailty, to better understand the relationship between the availability of SBRT treatment and implications for changes in the demographics of those patients undergoing surgical resection.
Regarding surgical volume, our results differ from the previous analysis. We find that the effect size of surgical volume on both measures of short-term surgical mortality is larger than that observed for SBRT-volume: H-SBRT: OR, 0.91 (CI, 0.83-0.99), P = .027 versus H-Surg: OR, 0.82 (CI, 0.74-0.91), P < .001 (30-day mortality). This differs from the previous analysis, which found only top quartile surgical volume (>85 procedures per year) significantly impacted 30-day surgical mortality: OR, 0.78 (CI, 0.69-0.88), P < .001, and 90-day surgical mortality: OR, 0.82 (CI, 0.75-0.89), P < .001, and found comparable associations in survival benefit observed for H-SBRT: OR, 0.80 (CI, 0.73-0.88), P < .001 and H-Surg: OR, 0.78 (CI, 0.69-0.88), P < .001. The reason for these different findings is unclear, and further study is needed. Of note, our analysis did not include facilities with no SBRT program, as we believe incorporating these hospitals into the analysis confounds the central question of how SBRT programs impact patient selection and outcomes. In our analysis, the effect size of facility type, insurance status, and patient comorbidity profile were each larger than that observed for facility SBRT-utilization on short-term mortality.
Together, the observed differences in short-term and long-term mortality, coupled with nonclinically significant differences observed in patient demographics, suggest that high SBRT-utilization may serve as a proxy for other important facility-level characteristics impacting short-term and long-term mortality for patients rather than influencing outcomes by altering patient selection for surgery. More granular detail on baseline patient demographics and functional status is needed to further test this hypothesis.
There are several notable limitations to this study. First, the data used in the analysis are derived from a deidentified NCDB files. Because of the retrospective nature of the data utilized in this analysis, selection bias cannot be discounted. Another limitation is the lack of granularity surrounding physician and provider decision-making surrounding choices of type of treatment. Further study is needed surrounding the issue of shared clinician and patient decision-making regarding the use of SBRT as opposed to surgical treatment. Finally, stratifying facility volume by the 17% SBRT:surgery ratio cutoff may miss other important inflection points in the interaction between these treatment modalities, and further analysis is needed to fully establish the relationship between these services and surgical outcomes.
In conclusion, we present an analysis of the relationship between SBRT utilization and surgical volume on surgical survival in the setting of early NSCLC. This study has demonstrated that for surgical care at facilities with low surgical volume, SBRT utilization is associated with differences in short-term and long-term morality. Further study is needed to evaluate the relationship between expanding SBRT programs, surgical patient selection, and mortality.
Conflict of Interest Statement
Tyler Grenda reports a one-time consulting fee from Astra Zeneca for “Fireside Chat for Thoracic Surgeons.” Nathaniel R. Evans, III, reports consulting fees from BMS Foundation, Intuitive Surgical, BMS, and Genentech. All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Treatment and prognosis of isolated local relapse after stereotactic body radiotherapy for clinical stage I non–small-cell lung cancer: importance of salvage surgery.
The volume–outcome relationship for lung cancer resection is complex, and the strength of the association depends on the details of the statistical modeling approach, the manner in which volume is measured, and the outcome of interest.1 Institutional stereotactic body radiation therapy (SBRT) practice patterns were recently suggested as an important component of the lung cancer resection volume–outcome relationship.2 In this study published in the Journal, Till and colleagues3 have expounded upon the volume–outcome relationship for lung cancer resection by accounting for institutional SBRT experience.
00816-9&id=fx1.jpg){kind=link}
00816-9&id=fx2.jpg){kind=link}
00816-9&id=gr1.jpg){kind=link}
00816-9&id=gr2.jpg){kind=link}
00816-9&id=gr3.jpg){kind=link}
00816-9&id=gr4.jpg){kind=link}