Computational risk model for predicting 2-year malignancy of pulmonary nodules using demographic and radiographic characteristics

OBJECTIVES

To determine whether discriminatory performance of a computational risk model in classifying pulmonary lesion malignancy using demographic, radiographic, and clinical characteristics is superior to the opinion of experienced providers. We hypothesized that computational risk models would outperform providers.

METHODS

Outcome of malignancy was obtained from selected patients enrolled in the NAVIGATE trial (NCT02410837). Five predictive risk models were developed using an 80:20 train-test split: univariable logistic regression model based solely on provider opinion, multivariable logistic regression model, random forest classifier, extreme gradient boosting model, and artificial neural network. Area under the receiver operating characteristic curve achieved during testing of the predictive models was compared to that of prebiopsy provider opinion baseline using the DeLong test with 10,000 bootstrapped iterations.

RESULTS

The cohort included 984 patients, 735 (74.7%) of which were diagnosed with malignancy. Factors associated with malignancy from multivariable logistic regression included age, history of cancer, largest lesion size, lung zone, and positron-emission tomography positivity. Testing area under the receiver operating characteristic curve were 0.830 for provider opinion baseline, 0.770 for provider opinion univariable logistic regression, 0.659 for multivariable logistic regression model, 0.743 for random forest classifier, 0.740 for extreme gradient boosting, and 0.679 for artificial neural network. Provider opinion baseline was determined to be the best predictive classification system.

CONCLUSIONS

Computational models predicting malignancy of pulmonary lesions using clinical, demographic, and radiographic characteristics are inferior to provider opinion. This study questions the ability of these models to provide additional insight into patient care. Expert clinician evaluation of pulmonary lesion malignancy is paramount.

Is Minimally Invasive Resection of Large Thymoma Appropriate?

BACKGROUND

Current practice patterns suggest open rather than minimally invasive (MIS) approaches for thymomas >4 cm. We hypothesized there would be similar perioperative outcomes and overall survival between open and MIS approaches for large (>4 cm) thymoma resection.

METHODS

The National Cancer Database was queried for patients who underwent thymectomy from 2010 to 2020. Surgical approach was characterized as either open or MIS. The primary outcome was overall survival and secondary outcomes were margin status, and length of stay (LOS). Differences between approach cohorts were compared after a 1:1 propensity match.

RESULTS

Among 4121 thymectomies, 2474 (60%) were open and 1647 (40%) were MIS. Patients undergoing MIS were older, had fewer comorbidities, and had smaller tumors (median; 4.6 vs 6 cm, P < .001). In the unmatched cohort, MIS and open had similar 90-day mortality (1.1% vs 1.8%, P = .158) and rate of positive margin (25.1% vs 27.9%, P = .109). MIS thymectomy was associated with shorter LOS (2 (1-4) vs 4 (3-6) days, P < .001). Propensity matching reduced the bias between the groups. In this cohort, overall survival was similar between the groups by log-rank test (P = .462) and multivariate cox hazard analysis (HR .882, P = .472). Multivariable regression showed shorter LOS with MIS approach (Coef -1.139, P < .001), and similar odds of positive margin (OR 1.130, P = .150).

DISCUSSION

MIS has equivalent oncologic benefit to open resection for large thymomas, but is associated with shorter LOS. When clinically appropriate, MIS thymectomy may be considered a safe alternative to open resection for large thymomas.

The Society of Thoracic Surgeons Database Analysis: Comparing Sublobar Techniques in Stage IA Lung Cancer

BACKGROUND

This study compares sublobar resections-wedge resection and segmentectomy-in clinical stage IA lung cancers. It tests the hypothesis that overall survival after wedge resection is similar to segmentectomy.

METHODS

Adults undergoing wedge resection or segmentectomy for clinical stage IA lung cancer were identified from The Society of Thoracic Surgeons General Thoracic Surgery Database. Eligible patients were linked to the Centers for Medicare and Medicaid Services database using a matching algorithm. The primary outcome was long-term overall survival. Propensity scores overlap weighting (PSOW) adjustment of wedge resection using validated covariates was used for group difference mitigation. Kaplan-Meier and Cox regression models analyzed survival. All-cause first readmission, and morbidity and mortality were examined using PSOW regression models.

RESULTS

Of 9756 patients, 6141 met inclusion criteria, comprising 2154 segmentectomies and 3987 wedge resections. PSOW reduced differences between the groups. Unadjusted perioperative mortality was comparable, but wedge resection showed lower major morbidity rates. Weighted regression analysis indicated reduced mortality and major morbidity risks in wedge resection. Kaplan-Meier analysis revealed no mortality difference between groups, which was confirmed by PSOW Cox regression models. The cumulative risk of readmission was also comparable for both groups, with Cox Fine-Gray models showing no difference in rehospitalization risks.

CONCLUSIONS

In clinical stage IA lung cancer, relative to segmentectomy, wedge resection has comparable overall survival and lower perioperative morbidity, suggesting it is an equally effective option for the broader population of patients with clinical stage IA lung cancer, not only those at highest risk of complications.

Risk Factors for Chronic Atrial Fibrillation Following Lung Lobectomy

INTRODUCTION

Postoperative atrial fibrillation (POAF) is a common complication following lung lobectomy and is associated with increased risk of stroke, mortality, and prolonged hospital length of stay. The purpose of this study was to define the risk factors for POAF after lobectomy, hypothesizing that operative approach would be associated with risk of chronic POAF.

METHODS

The TriNetX database was used to identify adult patients with no history of arrythmia receiving elective lung lobectomy for cancer from 7/6/2003-7/6/2023. Patients were categorized by approach: video-assisted thoracoscopic surgery (VATS) or open. The outcome of interest was the presence of POAF occurring at 1-3 months ("early") and 12-24 months postop ("chronic"). Propensity matching was performed to reduce bias between cohorts.

RESULTS

We identified 22,998 patients: 8472 (36.8%) who received open and 14,526 (63.2%) VATS lobectomy. The rate of early POAF was 3.7% of VATS and 5.3% of open patients. The rate of chronic POAF was 5.5 % of VATS patients and 6.2% of open lobectomy patients. Propensity matching decreased bias between the approach groups, creating 7942 pairs for analysis. After matching, the risk of early POAF was greater in the open approach (5.5% open vs 3.4% VATS, risk ratio 1.607 (95% confidence interval 1.385-1.865), P < 0.001). Chronic POAF was (also) higher in the open approach (6.3% open vs 5.2% VATS, Risk Ratio 1.211 (95%CI 1.067-1.374), P = 0.003).

CONCLUSIONS

Postoperative atrial fibrillation (POAF) occurs more commonly after open lobectomy, both acutely and chronically. Providers should counsel patients about the risk of chronic arrythmia after lung resection.

Frailty is associated with 90-day mortality in urgent thoracic surgery conditions

Objective

In patients undergoing elective thoracic surgery, frailty is associated with worse outcomes. However, the magnitude by which frailty influences outcomes of urgent thoracic surgery (UTS) is unknown.

Methods

We identified patients admitted with a UTS condition from January to September 2017 in the National Readmissions Database. UTS conditions were classified as esophageal perforation, hemo/pneumothorax, rib fracture, and obstructed hiatal hernia. Outcome of interest was mortality within 90 days of index admission. Frailty score was calculated using a deficit accumulation method. Cox proportional hazard modeling was used to calculate a hazard ratio for each combination of UTS disease type and frailty score, adjusted for sex, insurance payor, hospital size, and hospital and patient location, and was compared with the effect of frailty on elective lung lobectomy.

Results

We identified 107,487 patients with a UTS condition. Among UTS conditions overall, increasing frailty elements were associated with increased mortality (hazard ratio, 2270; 95% CI, 1463-3523; P < .001). Compared with patients without frailty undergoing elective lobectomy, increasing frailty demonstrated trending toward increased mortality in all diagnoses. The magnitude of the effect of frailty on 90-day mortality differed depending on the disease and level of frailty.

Conclusions

The effect of frailty on 90-day mortality after admission for urgent thoracic surgery conditions varies by disease type and level of frailty. Among UTS disease types, increasing frailty was associated with a higher 90-day risk of mortality. These findings suggest a valuable role for frailty evaluation in both clinical settings and administrative data for risk assessment.

Early Repair of Rib Fractures Is Associated with Superior Length of Stay and Total Hospital Cost: A Propensity Matched Analysis of the National Inpatient Sample

Background and Objectives: Previous studies have suggested that early scheduling of the surgical stabilization of rib fractures (SSRF) is associated with superior outcomes. It is unclear if these data are reproducible at other institutions. We hypothesized that early SSRF would be associated with decreased morbidity, length of stay, and total charges. Materials and Methods: Adult patients who underwent SSRF for multiple rib fractures or flail chest were identified in the National Inpatient Sample (NIS) by ICD-10 code from the fourth quarter of 2015 to 2016. Patients were excluded for traumatic brain injury and missing study variables. Procedures occurring after hospital day 10 were excluded to remove possible confounding. Early fixation was defined as procedures which occurred on hospital day 0 or 1, and late fixation was defined as procedures which occurred on hospital days 2 through 10. The primary outcome was a composite outcome of death, pneumonia, tracheostomy, or discharge to a short-term hospital, as determined by NIS coding. Secondary outcomes were length of hospitalization (LOS) and total cost. Chi-square and Wilcoxon rank-sum testing were performed to determine differences in outcomes between the groups. One-to-one propensity matching was performed using covariates known to affect the outcome of rib fractures. Stuart-Maxwell marginal homogeneity and Wilcoxon signed rank matched pair testing was performed on the propensity-matched cohort. Results: Of the 474 patients who met the inclusion criteria, 148 (31.2%) received early repair and 326 (68.8%) received late repair. In unmatched analysis, the composite adverse outcome was lower among early fixation (16.2% vs. 40.2%, p < 0.001), total hospital cost was less (USD114k vs. USD215k, p < 0.001), and length of stay was shorter (6 days vs. 12 days) among early SSRF patients. Propensity matching identified 131 matched pairs of early and late SSRF. Composite adverse outcomes were less common among early SSRF (18.3% vs. 32.8%, p = 0.011). The LOS was shorter among early SSRF (6 days vs. 10 days, p < 0.001), and total hospital cost was also lower among early SSRF patients (USD118k vs. USD183k late, p = 0.001). Conclusion: In a large administrative database, early SSRF was associated with reduced adverse outcomes, as well as improved hospital length of stay and total cost. These data corroborate other research and suggest that early SSRF is preferred. Studies of outcomes after SSRF should stratify analyses by timing of procedure.

Impact of Diabetes on Pathologic Response to Multimodality Therapy for Esophageal Cancer

BACKGROUND

The incidence of esophageal cancer has increased faster than that of most cancers. Evidence from other malignant neoplasms suggests that diabetic patients have a worse response to multimodality therapy. We hypothesized that diabetic patients with esophageal cancer will have a decreased response to neoadjuvant chemotherapy and radiation therapy compared with nondiabetic patients.

METHODS

A retrospective study of The Society of Thoracic Surgeons General Thoracic Surgery Database identified all patients who had an esophagectomy after neoadjuvant therapy for esophageal cancer between 2012 and 2019. Patients were compared on the basis of the presence of diabetes. A pathologic complete response (pCR) was defined as ypT0 N0. The χ2 and Wilcoxon rank sum tests were used to compare patients' demographic and clinical characteristics between those with and those without diabetes. Multivariable logistic regression was used to evaluate the predictors of response to neoadjuvant therapy.

RESULTS

Of the 9171 patients who met inclusion criteria, 2011 (22%) patients were diabetic and 7160 (78%) patients were nondiabetic. Patients with diabetes were older, more likely to be male, and more likely to have all comorbidities. Univariate analysis revealed that diabetic patients were less likely to have pCR (16% vs 18%; P = .026). Although multivariable analysis showed a trend toward diabetic patients' having lower odds of achieving pCR, diabetes was not independently associated with pCR (odds ratio, 0.89; 95% CI, 0.78-1.01; P = .075).

CONCLUSIONS

Diabetic patients may be less likely than nondiabetic patients to achieve pCR after neoadjuvant treatment of esophageal cancer. This suggests the need for further exploration as diabetic patients with esophageal cancer can potentially benefit from different treatment paradigms compared with their nondiabetic counterparts.

Lymph Node Yield in Lung Cancer Resection is Associated With Demographic and Institutional Factors

INTRODUCTION

Lymphadenectomy is routinely performed during surgical resection of nonsmall cell lung cancer (NSCLC). Lymph node yield and number of nodal stations sampled are important prognostic markers viewed as surrogates of surgical quality. The purpose of this study was to identify factors associated with these quality metrics after resection of NSCLC.

MATERIALS AND METHODS

We identified NSCLC patients undergoing resection at a single institution from 2010 to 2021. Cases were matched to detailed pathologist reports, which included lymph node yield and number of stations sampled. Demographic and clinical characteristics were analyzed individually using unadjusted linear regression to identify factors associated with lymph node yield and number of stations sampled. Multivariable linear regression analyses were performed to evaluate the same end points, using covariates determined through stepwise-backwards selection.

RESULTS

The study cohort included 836 patients. Multivariable regression demonstrated that male sex, history of cardiothoracic surgery, and individual pathologist were independently associated with lymph node yield. Among 18 pathologists, interpathologist coefficients with respect to lymph node yield varied from -5.61 to 11.25. Multivariable regression demonstrated White race and history of cardiothoracic surgery to be independently associated with number of nodal stations sampled, as well as individual surgeon and pathologist.

CONCLUSIONS

Lymph node yield and number of nodal stations sampled after NSCLC resection may vary based on patient demographic and clinical characteristics, as well as institutional factors. These factors should be accounted for when using these metrics as markers of surgical quality and prognosis of NSCLC.

Uncovering Health-Care Disparities Through Patient Decisions in Lung Cancer Surgery

INTRODUCTION

Declining cancer surgery represents a conflict between patients' rights to autonomy and providers' perspectives of best practice. We hypothesize that, among patients with nonmetastatic lung cancer, patient demographics would be associated with different rates of declination of lung cancer surgery.

METHODS

Patients with nonmetastatic lung cancer from 2004 to 2018 in the National Cancer Database were identified. Patients were categorized into two groups based on surgical treatment: surgical resection and declined surgery. Patient characteristics were compared using bivariate and multivariate models to identify factors associated with surgical declination. Additionally, we performed subgroup analyses of cT1N0M0 patients with no comorbidities. Survival analysis done using multivariate cox analysis and Kaplan-Meier survival analysis.

RESULTS

478,757 patients were identified. In a multivariate model, declining surgery was associated with increased age (odds ratio 1.09, 1.09-1.10), non-Hispanic Black race (odds ratio 1.95, 1.73-2.21), nonprivate insurance, and lower Socioeconomic Status. In a subgroup of cT1N0M0 patients with no comorbidities, declining surgery was associated with increasing age, non-Hispanic Black race, nonprivate insurance, and socioeconomic status. Patient's that declined surgery demonstrated lower overall survival when compared to patients that underwent surgical resection (5 y overall survival: declined surgery 40% versus underwent resection 72%, P < 0.001).

CONCLUSIONS

Although early-stage lung cancer is potentially curable, many patients decline guideline-based surgery, and have worse overall survival. There are social and economic factors associated with patients declining lung cancer surgery. Providers have an ethical responsibility to understand the basis of patient's decision to decline recommended surgery and address endemic disparities related to race and access to care.

National enrollment of lung cancer clinical trials is disproportionate based on race and health care access

OBJECTIVE

Despite declining lung cancer mortality in the United States, survival differences remain among racial and ethnic minorities in addition to those with limited health care access. Improvements in lung cancer treatment can be obtained through clinical trials, yet there are disparities in clinical trial enrollment of other cancer types. This study aims to evaluate disparities in lung cancer clinical trial enrollment to inform future enrollment initiatives.

METHODS

We analyzed patients with non-small cell lung cancer from the National Cancer Database (2004-2018), categorizing them as enrolled or not enrolled in clinical trials based on "rx_summ_other" data element. Clinical, demographic, and institutional factors associated with trial enrollment were assessed using bivariate and multivariate analysis, adjusting for institutional-level clustering.

RESULTS

A total of 1924 (0.12%) patients with lung cancer were enrolled in clinical trials. Enrolled patients were predominantly non-Hispanic White (82%), with greater socioeconomic status, treated at academic programs (67%), and had private insurance (42%) or Medicare (44%). They also traveled further for treatment compared with unenrolled patients (56 vs 27 miles, P < .001). After adjusting for demographic and clinical factors, lung cancer trial enrollment was significantly less likely among Black (odds ratio, 0.55; 95% confidence interval, 0.5-0.7, P < .001) and Hispanic (0.66; 95% confidence interval, 0.5-0.9, P = .01) patients. Patients with Medicaid or uninsured, in the lowest socioeconomic status group, and those treated at community-based cancer programs were the least likely to enroll.

CONCLUSIONS

Enrollment in lung cancer trials disproportionally excludes minority patients, those in the lowest socioeconomic status, community cancer programs, and the underinsured. These disparities in demographic and access for trial participation show a need for improved enrollment strategies.

Evaluating the Optimal Time Between Diagnosis and Surgical Intervention for Early-Stage Lung Cancer

INTRODUCTION

There is no consensus on the optimal timing for lung cancer surgery. We aim to evaluate the impact of timing of surgical intervention. We hypothesize delay in intervention is associated with worse overall survival and higher pathologic upstaging in early-stage lung cancer.

METHODS

We identified patients with cT1/2N0M0 nonsmall cell lung cancer in the National Cancer Database from 2004 to 2018. Patients were categorized by time to surgery groups: early (<26 d), average (26-60 d), and delayed (61-365 d). Primary outcome was overall survival and secondary outcome was pathologic upstaging. Multivariate models and survival analyses were used to determine factors associated with time from diagnosis to surgery, pathologic upstaging, and overall survival.

RESULTS

In multivariate model, advanced age, non-Hispanic Black patients, nonprivate insurance, low median income and education, and treatment at low-volume facilities were less likely to undergo early intervention and compared to the average group were more likely to receive delayed intervention. Pathologic upstaging was more likely in the delayed group (odds ratio 1.11, 1.07-1.14) compared to early group (odds ratio 0.96, 0.93-0.99). Early intervention was associated with improved overall survival (hazard ratio 0.93, 0.91-0.95), while delayed intervention was associated with inferior survival (hazard ratio 1.11, 1.09-1.14).

CONCLUSIONS

Expeditious surgical intervention is associated with lower rates of pathologic upstaging and improved overall survival in early-stage lung cancer. Delays in surgery are associated with social and economic factors, suggesting disparities in access to surgery. Lung cancer surgery should be performed as quickly as possible to maximize oncologic outcomes.

Affordable Care Act Medicaid Expansion is Associated With Increased Utilization of Minimally Invasive Lung Resection for Early Stage Lung Cancer

OBJECTIVE

Minimally invasive lung resection (MILR) is underutilized in the United States. Under the Affordable Care Act (ACA), 39 states adopted Medicaid expansion, while 12 did not. Although Medicaid expansion has been associated with improved access to cancer care, its effect on utilization of MILR is unclear. We hypothesize that MILR would increase in Medicaid expansion states.

METHODS

The National Cancer Database was queried for adult patients from 2010 to 2018 with cT1/2N0M0 non-small cell lung cancer who received surgical resection by wedge, segmentectomy, or lobectomy. Patients were grouped by whether they received care in a state without Medicaid expansion vs expansion in January 2014. The outcome of interest was MILR (defined as video-assisted or robotic-assisted thoracoscopy) relative to open. Multivariable difference in differences (DID) cross-sectional analysis was used to estimate the average treatment effect (ATE) of Medicaid expansion.

RESULTS

There were 41,439 patients who met inclusion criteria: 20,446 (49.3%) in expansion states and 20,993 (50.7%) in non-expansion states. Multivariable DID analysis showed that Medicaid expansion was associated with an increase in Medicaid insurance type with an ATE of 7.4% (95% CI 7.1-7.7%, P = .002). Medicaid expansion was also associated with increased MILR utilization in unadjusted analysis (10,278/20,446 (50.3%) vs 9,953/20,993 (47.4%), p < .001) and in multivariable DID analysis (ATE 0.6%, 95% CI 0.3-0.8%, P = .008).

CONCLUSIONS

Although Medicaid expansion was associated with increased utilization of MILR for early stage lung cancer, the treatment effect was modest. This suggests that barriers in access to MILR are larger than simply access to care.

Comparing Thoracoscopic and Robotic Lobectomy Using a Nationally Representative Database

BACKGROUND

Studies of robotic lobectomy (Robot-L) have been performed using data from high-volume, specialty centers which may not be generalizable. The purpose of this study was to compare mortality, length of stay (LOS), and cost between Robot-L and thoracoscopic lobectomy (VATS-L) using a nationally representative database hypothesizing they would be similar.

METHODS

The Premier Healthcare Database was used to identify patients receiving elective lobectomy for lung cancer from 2009 to 2019. Patients were categorized as receiving Robot-L or VATS-L using ICD-9/10 codes. Survey methodology and patient level weighting were used to correct for sampling error and estimation of a nationally representative sample. A propensity match analysis was performed to reduce bias between the groups. Primary outcome of interest was in-hospital mortality. Secondary outcomes were LOS and patient charges.

RESULTS

Among 62 698 patients, 19 506 (31.1%) underwent Robot-L and 43 192 (68.9%) underwent VATS-L. Differences between the groups included age, race, comorbidities, and insurance type. A propensity matched cohort demonstrated similar in-hospital mortality for Robot-L and VATS-L (.9% vs .9%, respectively, P = .91). Patients who underwent Robot-L had a shorter LOS (4 vs 5d, respectively, P < .001) but higher patient charges (90 593.0 vs 72 733.3 USD, respectively, P < .001).

CONCLUSIONS

In a nationally representative database, Robot-L and VATS-L had similar mortality. Although Robot-L was associated with shorter hospitalization, it was also associated with excess charges of almost $20,000. As Robot-L is now the most common approach for lobectomy in the U.S., further study into the cost and benefit of robotic surgery is warranted.

Motor Vehicle Protective Device Usage Associated with Decreased Rate of Flail Chest: A Retrospective Database Analysis

Background and Objectives: Protective equipment, including seatbelts and airbags, have dramatically reduced the morbidity and mortality rates associated with motor vehicle collisions (MVCs). While generally associated with a reduced rate of injury, the effect of motor vehicle protective equipment on patterns of chest wall trauma is unknown. We hypothesized that protective equipment would affect the rate of flail chest after an MVC. Materials and Methods: This study was a retrospective analysis of the 2019 iteration of the American College of Surgeons Trauma Quality Program (ACS-TQIP) database. Rib fracture types were categorized as non-flail chest rib fractures and flail chest using ICD-10 diagnosis coding. The primary outcome was the occurrence of flail chests after motor vehicle collisions. The protective equipment evaluated were seatbelts and airbags. We performed bivariate and multivariate logistic regression to determine the association of flail chest with the utilization of vehicle protective equipment. Results: We identified 25,101 patients with rib fractures after motor vehicle collisions. In bivariate analysis, the severity of the rib fractures was associated with seatbelt type, airbag status, smoking history, and history of cerebrovascular accident (CVA). In multivariate analysis, seatbelt use and airbag deployment (OR 0.76 CI 0.65-0.89) were independently associated with a decreased rate of flail chest. In an interaction analysis, flail chest was only reduced when a lap belt was used in combination with the deployed airbag (OR 0.59 CI 0.43-0.80) when a shoulder belt was used without airbag deployment (0.69 CI 0.49-0.97), or when a shoulder belt was used with airbag deployment (0.57 CI 0.46-0.70). Conclusions: Although motor vehicle protective equipment is associated with a decreased rate of flail chest after a motor vehicle collision, the benefit is only observed when lap belts and airbags are used simultaneously or when a shoulder belt is used. These data highlight the importance of occupant seatbelt compliance and suggest the effect of motor vehicle restraint systems in reducing severe chest wall injuries.

Is Sarcomatoid Lung Cancer Associated With Inferior Overall Survival? A National Cancer Database Analysis

INTRODUCTION

Sarcomatoid lung cancer has mainly been described in case series and single institution reviews. Although often associated with a poor prognosis, the overall survival compared to other forms of nonsmall cell lung cancer (NSCLC) is unknown. We hypothesize that sarcomatoid lung cancers have worse overall survival relative to other forms of NSCLC.

MATERIALS AND METHODS

In this retrospective cohort study, we identified adult patients with nonmetastatic NSCLC from 2004 to 2018 in the National Cancer Database. Patients were categorized by histology as sarcomatoid, adenocarcinoma, or squamous cell carcinoma. We compared clinical and demographic characteristics between the groups. The primary outcome of overall survival was analyzed using Kaplan-Meier analysis. Multivariable Cox analysis was used to analyze factors associated with overall survival in sarcomatoid patients undergoing surgery.

RESULTS

Among 1,259,109 patients with lung cancer, there were 5223 (0.4%) sarcomatoid cancers. Sarcomatoid patients were more likely to be male, of Hispanic ethnicity, have fewer comorbidities, and receive treatment at an academic program. Despite higher cT- and M-stages, patients with sarcomatoid cancer were more likely to undergo surgical resection in multivariate analysis (odds ratio = 1.8 [confidence interval 1.60-2.11]; P < 0.001). Among nonmetastatic patients, overall survival was lower for sarcomatoid cancer relative to other histologies in Kaplan-Meier analysis (median survival sarcomatoid 17.6 mo versus nonsarcomatoid 31.5 mo, P < 0.001).

CONCLUSIONS

This National Cancer Database study confirms the findings of smaller studies that sarcomatoid cancer is associated with inferior overall survival compared to other NSCLCs. Given the inferior prognosis, further studies regarding optimal staging practices are appropriate.

Disparities and access to thoracic surgeons among esophagectomy patients in the United States

Esophagectomy is a complex operation with significant morbidity and mortality. Previous studies have shown that sub-specialization is associated with improved esophagectomy outcomes. We hypothesized that disparities would exist among esophagectomy patients regarding access to thoracic surgeons based on demographic, geographic, and hospital factors. The Premier Healthcare Database was used to identify adult inpatients receiving esophagectomy for esophageal and gastric cardia cancer, Barrett's esophagus, and achalasia from 2015 to 2019 using ICD-10 codes. Patients were categorized as receiving their esophagectomy from a thoracic versus non-thoracic provider. Survey methodology was used to correct for sampling error. Backwards selection from bivariable analysis was used in a survey-weighted multivariable logistic regression to determine predictors of esophagectomy provider specialization. During the study period, 960 patients met inclusion criteria representing an estimated population size of 3894 patients. Among them, 1696 (43.5%) were performed by a thoracic surgeon and 2199 (56.5%) were performed by non-thoracic providers. On multivariable analysis, factors associated with decreased likelihood of receiving care from a thoracic provider included Black (OR 0.41, p < 0.001), Other (OR 0.21, p < 0.001), and Unknown race (OR 0.22, p = 0.04), and uninsured patients (OR 0.53, p = 0.03). Urban hospital setting was associated with an increased likelihood of care by a thoracic provider (OR 4.43, p = 0.001). In this nationally representative study, Nonwhite race, rural hospital setting, and lower socioeconomic status were factors associated with decreased likelihood of esophagectomy patients receiving care from a thoracic surgeon. Efforts to address these disparities and provide appropriate access to thoracic surgeons is warranted.

Surgical subspecialization is associated with higher rate of rib fracture stabilization: a retrospective database analysis

Background

Surgical stabilization of rib fractures (SSRF) is performed on only a small subset of patients who meet guideline-recommended indications for surgery. Although previous studies show that provider specialization was associated with SSRF procedural competency, little is known about the impact of provider specialization on SSRF performance frequency. We hypothesize that provider specialization would impact performance of SSRF.

Methods

The Premier Hospital Database was used to identify adult patients with rib fractures from 2015 and 2019. The outcome of interest was performance of SSRF, defined using International Classification of Diseases—10th Revision Procedure Coding System coding. Patients were categorized as receiving their procedures from a thoracic, general surgeon, or orthopedic surgeon. Patients with missing or other provider types were excluded. Multivariate modeling was performed to evaluate the effect of surgical specialization on outcomes of SSRF. Given a priori assumptions that trauma centers may have different practice patterns, a subgroup analysis was performed excluding patients with ‘trauma center’ admissions.

Results

Among 39 733 patients admitted with rib fractures, 2865 (7.2%) received SSRF. Trauma center admission represented a minority (1034, 36%) of SSRF procedures relative to other admission types (1831, 64%, p=0.15). In a multivariable analysis, thoracic (OR 6.94, 95% CI 5.94–8.11) and orthopedic provider (OR 2.60, 95% CI 2.16–3.14) types were significantly more likely to perform SSRF. In further analyses of trauma center admissions versus non-trauma center admissions, this pattern of SSRF performance was found at non-trauma centers.

Conclusion

The majority of SSRF procedures in the USA are being performed by general surgeons and at non-trauma centers. ‘Subspecialty’ providers in orthopedics and thoracic surgery are performing fewer total SSRF interventions, but are more likely to perform SSRF, especially at non-trauma centers. Provider specialization as a barrier to SSRF may be related to competence in the SSRF procedures and requires further study.

Type

Therapeutic/care management.

Level of evidence

IV

Age as a Barrier to Surgical Stabilization of Rib Fractures in Patients with Flail Chest

INTRODUCTION

Although randomized trials demonstrate a benefit to surgical stabilization of rib fractures (SSRF), SSRF is rarely performed. We hypothesized older patients were less likely to receive SSRF nationally.

METHODS

The 2016 National Inpatient Sample was used to identify adults with flail chest. Comorbidities and receipt of SSRF were categorized by ICD-10 code. Univariable testing and Multivariable regression were performed to determine the association of demographic characteristics and comorbidities to receipt of SSRF.

RESULTS

1021 patients with flail chest were identified, including 244 (23.9%) who received SSRF. Patients ≥70 years were less likely to receive SSRF. (<70 yrs 201/774 [26.0%] vs ≥70 43/247 [17.4%], P = .006) and had higher risk of death (<70 yrs 39/774 [5.0%] vs ≥70 33/247 [13.4%], P < .001) In multivariable modeling, only age ≥70 years was associated with SSRF (OR .591, P = .005).

CONCLUSION

Despite guideline-based support of SSRF in flail chest, SSRF is performed in <25% of patients. Age ≥70 years is associated with lower rate of SSRF and higher risk of death. Future study should examine barriers to SSRF in older patients.

Modeling the COVID Pandemic: Do Delays in Surgery Justify Using Stereotactic Radiation to Treat Low-Risk Early Stage Non-Small Cell Lung Cancer?

INTRODUCTION

It was suggested that stereotactic radiation (SBRT) is an "alternative if no surgical capacity is available" for non-small cell lung cancer (NSCLC) care during the COVID-19 pandemic. The purpose of this study was to compare the oncologic outcomes of delayed surgical resection and early SBRT among operable patients with early stage lung cancer.

METHODS

The National Cancer Database was queried for patients with cT1aN0M0 NSCLC who underwent surgery or SBRT (2010-2016) with no comorbidity. Patients with any comorbidities or age >80 were excluded. The outcome of interest was overall survival. Delays in surgical care were modeled using different times from diagnosis to surgery. A 1:1 propensity match was performed and survival was analyzed using multivariable Cox regression.

RESULTS

Of 6720 healthy cT1aN0M0 NSCLC patients, 6008 (89.4%) received surgery and 712 (10.6%) received SBRT. Among surgery patients, time to surgery >30 d was associated with inferior survival (HR > 1.4, P ≤ 0.013) compared with patients receiving surgery ≤14 d. Relative to SBRT, surgery demonstrated superior survival at all time points evaluated: 0-30 d, 31-60 d, 61-90 d, and >90 d (all P < 0.001). Among a propensity-matched cohort of 256 pairs of patients, delayed surgery (>90 d) remained association with better overall survival relative to early SBRT (5-year survival 76.9% versus 32.3%, HR = 0.266, P < 0.001).

CONCLUSIONS

Although longer time to surgery is associated with inferior survival among surgery patients, delayed surgery is superior to early SBRT. Surgical resection should remain the standard of care to treat operable early stage lung cancer despite delays imposed by the COVID-19 pandemic.

Morbidity After Esophagectomy Is Higher for Benign Than Malignant Disease

BACKGROUND

Esophagectomy carries a high risk of morbidity and mortality. The most common indication for esophagectomy is esophageal cancer, with fewer than 5% of esophagectomies performed for benign disease. We hypothesized that esophagectomy for benign disease is associated with a higher risk of operative and postoperative complications.

METHODS

A retrospective study of The Society of Thoracic Surgeons database was performed to identify all patients who had an esophagectomy from 2010 to 2018. Patients who had an emergent or palliative esophagectomy were excluded. Patients were compared based on the indication for operation, malignant vs benign disease. A 1:1 propensity score matching of The Society of Thoracic Surgeons risk factors was performed and outcomes compared between the matched cohorts.

RESULTS

Of 16,392 patients, 14,871 (91%) had malignant disease and 1521 (9%) had benign disease that met inclusion criteria. Patients with malignant disease were older (P < .001), more likely to be male (83% vs 56%, P < .001), and had more comorbidities (P < .001). There were 1362 propensity-matched pairs. Malignant esophagectomies were more likely to be performed with a minimally invasive vs an open approach (P < .001). Benign operations had more intraoperative blood transfusions (P < .001). Patients undergoing esophagectomy for benign disease had more prolonged intubations (P = .02) and postoperative blood transfusions (P = .001). Benign disease had more major morbidities (P = .001) but similar postoperative mortality (P = .62).

CONCLUSIONS

Esophagectomy for benign disease is associated with worse perioperative morbidity compared with esophagectomy for malignant disease. Given these findings patients should be counseled on expected outcomes, and this variable should be considered for inclusion in the composite score for risk assessment.

Contemporary outcomes of esophageal and gastroesophageal junction neuroendocrine tumors

Adenocarcinoma and squamous cell esophageal cancers have been extensively studied in the literature. Esophageal neuroendocrine (NET)/carcinoid tumors are less commonly studied and have only been described in small series. The purpose of this study was to describe the demographics and natural history of esophageal NETs, as well as optimal treatments. We hypothesized that surgical resection would be the best treatment of esophageal NETs. The National Cancer Database was used to identify adult patients with esophageal or gastroesophageal junction (GEJ) cancer from 2004 to 2018. Patients were characterized as carcinoid/NET, adenocarcinoma, or squamous cell cancer. Clinical and demographic characteristics were compared between the histology groups. The primary outcome was overall survival, which was assessed by multivariable Cox analysis. Multivariable Cox analysis was also used to analyze factors associated with survival among NET patients who underwent surgery. Among 206,321 patients with esophageal cancer, 1,563 were NETs (<0.01%). Relative to the other two histologies, NETs were associated with younger age, female sex, and advanced clinical stage at diagnosis. Multivariate analysis suggested that NETs were less likely to be treated with surgical resection (OR 0.51, P < 0.001). Nonetheless, surgical resection was associated with improved survival (HR 0.64, P = 0.003). Among patients with NETs who received surgery, neoadjuvant therapy was associated with improved overall survival (HR 0.38, P = 0.013). NET of the esophagus presents with more advanced disease than other common histologies. Among patients with nonmetastatic cancer, surgical resection appears to be the best treatment. Neoadjuvant systemic therapy may offer survival benefit, but future studies are necessary.

Contemporary outcomes of surgical resection for chest wall chondrosarcoma

Objective

Chondrosarcoma is the most common primary malignant chest wall tumor and is historically associated with poor prognosis. Recommendations regarding surgical excision are on the basis of small, single-institution studies. We used a large national database to assess outcomes of surgery for chest wall chondrosarcoma (CWC) hypothesizing that surgical excision remains standard of care.

Methods

The National Cancer Databases for bone and soft tissue were merged to identify patients with chondrosarcoma from 2004 to 2018. Clinical and demographic characteristics of CWC were compared with chondrosarcoma from other sites. The primary outcome was overall survival described using Kaplan-Meier estimate. Univariable and multivariable Cox analysis was used to determine risk factors for poor survival among CWC patients who underwent surgery. Multivariable analysis of predictors of margin status was performed because of worse prognosis associated with positive margins.

Results

Among 11,925 patients with chondrosarcoma, 1934 (16.2%) had a CWC. Relative to other sites, CWC was associated with older age, male sex, White race, surgical resection, and care at a nonacademic institution. CWC was associated with 1-, 3-, 5-, and 10-year survival of 91.5%, 82.0%, 75.5%, and 62.7%, respectively. In univariable analysis, survival was associated with surgery (hazard ratio, 0.02; P < .001) and adversely affected by positive margins (hazard ratio, 2.66; P < .001). Multivariable analysis showed larger tumor size was independently associated with increased risk for positive margins (odds ratio, 1.04; 95% CI, 1.011-1.075).

Conclusions

CWC represents a different cohort of patients relative to chondrosarcoma from other sites. Surgical excision remains the optimal treatment, and positive margins are associated with poor prognosis.

Monitoring for Recurrence After Esophagectomy

BACKGROUND

Current guidelines for follow-up after esophagectomy suggest only history and physical examination (HPE). With recent advances in chemotherapy and immunotherapy for patients with recurrent esophageal cancer, we hypothesized that surveillance imaging (SI) would identify patients with cancer recurrence earlier and improve long-term survival.

METHODS

A retrospective review of all patients undergoing esophagectomy for esophageal cancer at a single institution between 2007 and 2018 was conducted. Patients were categorized as recurrence detected through SI or recurrence detected through HPE alone. Patients were excluded if recurrence occurred within 3 months of esophagectomy.

RESULTS

During the study period, 225 esophageal cancer patients underwent an esophagectomy. Among these, 101 (44.9%) had SI and 124 (55.1%) had routine follow-up with HPE. There were 88 recurrences (39.1%) with median follow-up of 12 months. Rate of recurrence was similar based on screening method: 41 of 101 (40.6%) by SI and 47 of 124 (37.9%) by HPE (P = .68). Among patients with recurrence, recipients of additional treatment were also similar between groups, 36 of 41 (87.8%) by SI and 34 of 47 (72.3%) by HPE (P = .468). Among those who had a recurrence, the median overall survival was significantly longer in those undergoing SI at 23 months compared with those who received HPE at 16 months (P = .047).

CONCLUSIONS

SI after esophagectomy is not associated with improved detection of recurrence, but is associated with improved overall survival once recurrence is detected. These data suggest that earlier identification of esophageal cancer recurrence may have survival benefit. Standardizing SI may prove beneficial for patients after esophagectomy.

Combined Radiomic and Visual Assessment for Improved Detection of Lung Adenocarcinoma Invasiveness on Computed Tomography Scans: A Multi-Institutional Study

Objective

The timing and nature of surgical intervention for semisolid abnormalities are dependent upon distinguishing between adenocarcinoma-in-situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (INV). We sought to develop and evaluate a quantitative imaging method to determine invasiveness of small, ground-glass lesions on computed tomography (CT) chest scans.

Methods

The study comprised 268 patients from 4 institutions with resected (<=3 cm) semisolid lesions with confirmed histopathological diagnosis of MIA/AIS or INV. A total of 248 radiomic texture features from within the tumor nodule (intratumoral) and adjacent to the nodule (peritumoral) were extracted from manually annotated lung nodules of chest CT scans. The datasets were randomly divided, with 40% of patients used for training and 60% used for testing the machine classifier (Training D_Train, N=106; Testing, D_Test, N=162).

Results

The top five radiomic stable features included four intratumoral (Laws and Haralick feature families) and one peritumoral feature within 3 to 6 mm of the nodule (CoLlAGe feature family), which successfully differentiated INV from MIA/AIS nodules with an AUC of 0.917 [0.867-0.967] on D_Train and 0.863 [0.79-0.931] on D_Test. The radiomics model successfully differentiated INV from MIA cases (<1 cm AUC: 0.76 [0.53-0.98], 1-2 cm AUC: 0.92 [0.85-0.98], 2-3 cm AUC: 0.95 [0.88-1]). The final integrated model combining the classifier with the radiologists’ score gave the best AUC on D_Test (AUC=0.909, p<0.001).

Conclusions

Addition of advanced image analysis via radiomics to the routine visual assessment of CT scans help better differentiate adenocarcinoma subtypes and can aid in clinical decision making. Further prospective validation in this direction is warranted.

Risk of Each of the Five Lung Lobectomies: A Society of Thoracic Surgery Database Analysis

BACKGROUND

The perioperative risk of pulmonary lobectomy as a solitary procedure has been extensively studied, yet the differences in outcomes between each lobe, which have unique anatomy and a different amount of lung parenchyma, are entirely unknown. The purpose of this study is to define the risk of each of the five lobectomies.

METHODS

The Society of Thoracic Surgery Database was queried for patients undergoing lobectomy between 2008 and 2018. Patient and disease characteristics, operative variables, major morbidity and 30-day mortality were examined. A multivariable logistic regression model (using the same variables in the current STS lobectomy risk model) was developed to assess for the contribution to lobectomy site to adverse outcomes.

RESULTS

65,006 patients were analyzed. Adjusted perioperative mortality rate is lowest for RML (0.63%) intermediate for RUL, LUL and LLL (1.08-1.24%), and highest for RLL (1.63%). The adjusted major morbidity rate is lowest for RML (5.36%) intermediate for LLL and LUL (7.82-8.33%), and highest for RUL and RLL (8.94-9.32%). Adjusted intraoperative transfusion rate is lowest for RML (1.37%) intermediate for RLL and LLL (1.81-1.94%) and highest for RUL and LUL (2.47-2.72%).

CONCLUSIONS

There are clear differences in postoperative outcomes by lobectomy location. Mortality, major morbidity, and transfusion rate are lowest for RML, but vary across other lobectomies. These differences should be appreciated when evaluating risk of operation, deciding upon best therapy, counseling patients, and comparing outcomes.

Iatrogenic rib fractures and the associated risks of mortality

PURPOSE

Rib fractures, though typically associated with blunt trauma, can also result from complications of medical or surgical care, including cardiopulmonary resuscitation. The purpose of this study is to describe the demographics and outcomes of iatrogenic rib fractures.

METHODS

Patients with rib fractures were identified in the 2016 National Inpatient Sample. Mechanism of injury was defined as blunt traumatic rib fracture (BTRF) or iatrogenic rib fracture (IRF). IRF was identified as fractures from the following mechanisms: complications of care, drowning, suffocation, and poisoning. Differences between BTRF and IRF were compared using rank-sum test, Chi-square test, and multivariable regression.

RESULTS

34,644 patients were identified: 33,464 BTRF and 1180 IRF. IRF patients were older and had higher rates of many comorbid medical disorders. IRF patients were more likely to have flail chest (6.1% versus 3.1%, p < 0.001). IRF patients were more likely to have in-hospital death (20.7% versus 4.2%, p < 0.001) and longer length of hospitalization (11.8 versus 6.9 days, p < 0.001). IRF patients had higher rates of tracheostomy (30.2% versus 9.1%, p < 0.001). In a multivariable logistic regression of all rib fractures, IRF was independently associated with death (OR 3.13, p < 0.001). A propensity matched analysis of IRF and BTRF groups corroborated these findings.

CONCLUSION

IRF injuries are sustained in a subset of extremely ill patients. Relative to BTRF, IRF is associated with greater mortality and other adverse outcomes. This population is understudied. The etiology of worse outcomes in IRF compared to BTRF is unclear. Further study of this population could address this disparity.

Drain Amylase: A Simple and Versatile Method of Detecting Esophageal Anastomotic Leaks

BACKGROUND

Anastomotic leak following esophagectomy is a significant cause of morbidity. Perianastomotic drain amylase is accurate in detecting leaks, but it is unclear whether its accuracy is affected by comorbid conditions, anastomotic method, or anastomotic location. We hypothesized that drain amylase would accurately discriminate leak in a variety of settings.

METHODS

We reviewed 290 consecutive patients undergoing esophagectomy with gastric conduit reconstruction. Patient comorbidities, operative variables, and drain amylase were collected. The diagnosis of a leak was based on the level of intervention required, and was characterized as "clinically significant" if it required wound opening, endoscopic or surgical intervention. Receiver operating characteristic curves analysis was performed to determine the accuracy of amylase to detect leak for each patient variable.

RESULTS

53/290 (18.3%) esophagectomies had an anastomotic leak, of which 33/290 (11.4%) were clinically significant. Drain amylase was a strong predictor of anastomotic leak on postoperative days 3-7, regardless of patient comorbidities, location of anastomosis, or technique of anastomosis, but was less accurate in the diagnosis of leak in current smokers (AUC 0.530 vs 0.752, p= 0.006). A maximum drain amylase value no higher than 35 on postoperative 3, 4 or 5 was 88% sensitive in detecting leak at any point postoperatively. A value >=150 was 88% specific in diagnosing leak CONCLUSIONS: Drain amylase is a versatile method for early detection of anastomotic leaks. Its accuracy is unaffected by neoadjuvant treatment, location or type of anastomosis or patient comorbidities, but may be less accurate in active smokers.

Surgery following neoadjuvant chemoradiation therapy in clinical N3 esophageal cancer results in improved survival: a propensity-matched analysis

Esophageal cancer patients with extensive nodal metastases have poor survival, and benefit of surgery in this population is unclear. The aim of this study is to determine if surgery after neoadjuvant chemoradiotherapy (nCRT) improves overall survival (OS) in patients with clinical N3 (cN3) esophageal cancer relative to chemoradiation therapy (CRT) alone. The National Cancer Database was queried for all patients with cN3 esophageal cancer between 2010 and 2016. Patients who met inclusion criteria (received multiagent chemotherapy and radiation dose ≥30 Gy) were divided into two cohorts: CRT alone and nCRT + surgery. 769 patients met inclusion criteria, including 560 patients who received CRT alone, and 209 patients who received nCRT + surgery. The overall 5-year survival was significantly lower in the CRT alone group compared to the nCRT + surgery group (11.8% vs 18.0%, P < 0.001). A 1:1 propensity matched cohort of CRT alone and nCRT + surgery patients also demonstrated improved survival associated with surgery (13.11 mo vs 23.1 mo, P < 0.001). Predictors of survival were analyzed in the surgery cohort, and demonstrated that lymphovascular invasion was associated with worse survival (HR 2.07, P = 0.004). Despite poor outcomes of patients with advanced nodal metastases, nCRT + surgery is associated with improved OS. Of those with cN3 disease, only 27% underwent esophagectomy. Given the improved OS, patients with advanced nodal disease should be considered for surgery. Further investigation is warranted to determine which patients with cN3 disease would benefit most from esophagectomy, as 5-year survival remains low (18.0%).

Traffic stops do not prevent traffic deaths

OBJECTIVES

Amid growing calls for police reform, it is imperative to reassess whether police actions designed to improve public safety are associated with injury prevention. This study aims to examine the relationship between the police traffic stops (PTSs) and motor vehicle crash (MVC) deaths at the state level. We hypothesize that increased PTSs would be associated with reduced MVC deaths.

METHODS

We retrospectively analyzed PTSs and MVC deaths at the state level from 2004 to 2016. Police traffic stops data were from 33 state patrols from the Stanford Open Policing Project. The MVC deaths data were collected from the National Highway Traffic Safety Administration. The vehicle miles traveled data were from the Federal Highway Administration Office of Highway Policy Information. All data were adjusted per 100 million vehicle miles traveled (100MVMT) and were analyzed as state-level time series cross-sectional data. The dependent variable was MVC deaths per 100MVMT, and the independent variable was number of PTSs per 100MVMT. We performed panel data analysis accounting for random and fixed state effects and changes over time.

RESULTS

Thirty-three state patrols with 235 combined years were analyzed, with a total of 161,153,248 PTSs. The PTS rate varied by state and year. Nebraska had the highest PTS rate (3,637/100MVMT in 2004), while Arizona had the lowest (0.17/100MVMT in 2009). Motor vehicle crash deaths varied by state and year, with the highest death rate occurring in South Carolina in 2005 (2.2/100MVMT) and the lowest in Rhode Island in 2015 (0.57/100MVMT). After accounting for year and state-level variability, no association was found between PTS and the MVC death rates.

CONCLUSION

State patrol traffic stops are not associated with reduced MVC deaths. Strategies to reduce death from MVC should consider alternative strategies, such as motor vehicle modifications, community-based safety initiatives, improved access to health care, or prioritizing trauma system.

LEVEL OF EVIDENCE

Retrospective epidemiological study, level IV.

The Impact of Biopsy Tool Choice and Rapid On-Site Evaluation on Diagnostic Accuracy for Malignant Lesions in the Prospective: Multicenter NAVIGATE Study

BACKGROUND

The diagnostic yield of electromagnetic navigation bronchoscopy (ENB) is impacted by biopsy tool strategy and rapid on-site evaluation (ROSE) use. This analysis evaluates usage patterns, accuracy, and safety of tool strategy and ROSE in a multicenter study.

METHODS

NAVIGATE (NCT02410837) evaluates ENB using the superDimension navigation system (versions 6.3 to 7.1). The 1-year analysis included 1215 prospectively enrolled subjects at 29 United States sites. Included herein are 416 subjects who underwent ENB-aided biopsy of a single lung lesion positive for malignancy at 1 year. Use of a restricted number of tools (only biopsy forceps, standard cytology brush, and/or bronchoalveolar lavage) was compared with an extensive multimodal strategy (biopsy forceps, cytology brush, aspirating needle, triple needle cytology brush, needle-tipped cytology brush, core biopsy system, and bronchoalveolar lavage).

RESULTS

Of malignant cases, 86.8% (361/416) of true positive diagnoses were obtained using extensive multimodal strategies. ROSE was used in 300/416 cases. The finding of malignancy by ROSE reduced the total number of tools used. A malignant ROSE call was obtained in 71% (212/300), most (88.7%; 188/212) by the first tool used (49.5% with aspirating needle, 20.2% with cytology brush, 17.0% with forceps). True positive rates were highest for the biopsy forceps (86.9%) and aspirating needle (86.6%). Use of extensive tool strategies did not increase the rates of pneumothorax (5.5% restricted, 2.8% extensive) or bronchopulmonary hemorrhage (3.6% restricted, 1.1% extensive).

CONCLUSION

These results suggest that extensive biopsy tool strategies, including the aspirating needle, may provide higher true positive rates for detecting lung cancer without increasing complications.

Robotic approach has improved outcomes for minimally invasive resection of mediastinal tumors

BACKGROUND

The optimal minimally invasive surgical (MIS) approach to mediastinal tumors is unknown. There are limited reports comparing the outcomes of resection with robotic-assisted and video-assisted thoracoscopy (RATS vs VATS). We hypothesized that patients who underwent RATS would have improved outcomes.

METHODS

The National Cancer Database was queried for all patients who underwent MIS for any mediastinal tumor from 2010-2016. Patients were determined to have an adverse composite outcome if they had any of the adverse perioperative outcomes; conversion to open procedure, 90-day mortality, 30-day readmission, and positive pathologic margins. Secondary outcomes of interest were length-of-stay (LOS) and overall survival. Multivariable logistic regression was used to assess likelihood of having a composite adverse outcome based on surgical approach.

RESULTS

856 patients were included: 402 (47%) underwent VATS and 454 (53%) underwent RATS. RATS resections were associated with fewer conversions (4.9% vs 14.7%, p<0.001), fewer positive margins (24.3% vs 31.6%, p=0.02), shorter LOS (3.8d vs 4.3d, p=0.01) and less composite adverse events (36.7% vs 51.3%, p<0.001). Multivariate analysis showed RATS (OR 0.44, p<0.001) was independently associated with decreased likelihood of composite adverse outcome, even among tumors >4 cm (OR 0.45, p=0.001). Overall survival was similar between the two groups.

CONCLUSIONS

Among patients who underwent MIS for a mediastinal tumor, RATS had fewer adverse outcomes than VATS, even for tumors ≥4 cm. These data suggests that RATS may be the preferred technique for patients who are candidates for minimally invasive resection of mediastinal tumors.

Palliative Surgery Outcomes for Patients with Esophageal Cancer: An NCDB Analysis

INTRODUCTION

Many patients with esophageal cancer are not candidates for surgical resection with curative intent, given the advanced stage of disease at presentation. Palliative surgery is one treatment option, but relative survival of palliative surgery has not been described. This study aims to describe the outcomes of palliative surgery in patients with esophageal cancer.

METHODS

We used the National Cancer Database to identify patients with esophageal cancer who received palliative surgery or non-surgical palliation-which consisted of palliative radiation and palliative chemotherapy without any surgery. The outcome of interest was overall survival. Characteristics of patients were compared between the palliative surgery group and the non-surgical group using rank sum test or chi square test. Survival differences between groups were compared using Kaplan Meier estimate and log rank test, and Cox proportional hazards model.

RESULTS

A total of 14,589 patients were included in the analysis, including 2,812 (19.2%) receiving palliative surgery and 11,777 (80.7%) receiving non-surgical palliation (6,512 palliative radiation and 5,265 palliative chemotherapy). Median overall survival in palliative surgery patients was 5.5 mo, shorter than non-surgical palliation (6.4 mo, P = 0.004). However, when correcting for age, sex, nodal status, metastases, Charlson score, histology, academic center, and private insurance, there was no difference in survival between palliative surgery and non-surgical palliation in Cox proportional hazard modeling (HR 1.03 (0.975-1.090), P = 0.281).

CONCLUSIONS

Palliative surgery in advanced esophageal cancer is associated with poor overall survival but is similar to other palliative modalities. Palliative Surgery for esophageal cancer patients should be used sparingly given these poor outcomes.

Concurrent diagnosis of anxiety increases postoperative length of stay among patients receiving esophagectomy for esophageal cancer

OBJECTIVE

Psychiatric comorbidities disproportionately affect patients with cancer. While identified risk factors for prolonged length of stay (LOS) after esophagectomy are primarily medical comorbidities, the impact of psychiatric comorbidities on perioperative outcomes is unclear. We hypothesized that psychiatric comorbidities would prolong LOS in patients with esophageal cancer.

METHODS

The 2016 National Inpatient Sample (NIS) was used to identify patients with esophageal cancer receiving esophagectomy. Concurrent psychiatric illness was categorized using Clinical Classifications Software Refined for ICD-10, creating 34 psychiatric diagnosis groups (PDGs). Only PDGs with >1% prevalence in the cohort were included in the analysis. The outcome of interest was hospital LOS. Bivariable testing was performed to determine the association of PDGs and demographic factors on LOS using rank sum test. Multivariable regression analysis was performed using backward selection from bivariable testing (α ≤ 0.05).

RESULTS

We identified 1,730 patients who underwent esophagectomy for esophageal cancer in the 2016 NIS. The median LOS was 8 days (IQR 5-12). In bivariable testing, a concurrent diagnosis of anxiety was the only PDG associated with LOS (9 days (IQR 6-14) with anxiety diagnosis versus 8 days (IQR 5-12) with no anxiety diagnosis, p = 0.022). Multivariable modeling showed an independent association between anxiety diagnosis and increased LOS (OR 4.82 (1.25-25.23), p = 0.022). Anxiety was not associated with increased hospital cost or in-hospital mortality.

CONCLUSIONS

This analysis demonstrates an independent effect of anxiety prolonging postoperative LOS after esophagectomy in the United States. These findings may influence perioperative care, patient expectations, and resource allocation.

Wedge Resection Offers Similar Survival to Segmentectomy for Typical Carcinoid Tumors

Current guidelines recommend anatomic lung resection of typical bronchopulmonary carcinoids. Typical carcinoid tumors have excellent prognosis and sublobar resection has been associated with noninferior long-term survival. It's unclear whether wedge resection is acceptable for small typical carcinoid tumors. We hypothesize there is no difference in survival between wedge resection and segmentectomy for Stage I typical bronchopulmonary carcinoid tumors. Using the National Cancer Database from 2010 to 2016, we identified clinical T1N0M0 typical bronchopulmonary carcinoid tumors by wedge resection or segmentectomy. Short-term outcomes included length of stay, lymph nodes evaluated, pathologic node status, positive margin status, and 90-day mortality. Primary outcome was overall survival and estimated using Kaplan-Meier survival analysis. 821 patients were identified: 677 receiving wedge resection, 144 receiving segmentectomy. Segmentectomy was more commonly performed in an academic setting (70.0% vs 57.3%, P = 0.005). The mean tumor size for segmentectomy was 1.7 cm versus 1.4 cm for wedge resection (P < 0.001). There was no difference in LOS, positive margin status, and 90-day mortality between groups. There were significantly more lymph nodes evaluated in segmentectomy patients (median 4 vs 0, P < 0.001), but there was no difference in positive lymph node status (5.3% vs 2.6%, P = 0.165). The OS was similar between wedge and segmental resection (P = 0.613): 3-year survival (93.5% vs 92.8%) and 5-year survival (83.8% vs 84.9%). Wedge resection and segmentectomy have similar survival for Stage I typical bronchopulmonary carcinoids in a large national database. This analysis suggests nonanatomic, parenchymal-sparing resection should be considered an appropriate alternative for Stage I typical bronchopulmonary carcinoids.

Does Timing of Robotic Esophagectomy Adoption Impact Short-Term Postoperative Outcomes?

BACKGROUND

Robotic esophagectomies are increasingly common and are reported to have superior outcomes compared with an open approach; however, it is unclear if all institutions can achieve such outcomes. We hypothesize that early adopters of robotic technique would have improved short-term outcomes.

METHODS

The National Cancer Database (2010-2016) was used to identify robotic esophagectomies. Early adopters were defined as programs which performed robotic esophagectomies in 2010-2011, late adopters in 2012-2013. Outcomes of esophagectomies performed between 2014 and 2016 were compared and included length of stay, number of lymph nodes evaluated, readmission, conversion rate, and 90-day mortality. Multivariable regressions, accounting for robotic esophagectomy volume, were used to control for confounding factors.

RESULTS

There were 37 early adopters and 35 late adopters. Between 2014 and 2016, 683 robotic esophagectomies were performed: 446 (65.3%) by early adopters and 237 (34.7%) by late adopters. Early adopters were more likely to be academic programs (96.2 versus 72.8%, P < 0.01). Other clinical and demographic variables were similar. Late adopters were found to have decreased a number of lymph nodes evaluated (coefficient -2.407, P = 0.004) compared with early adopters. There were no significant differences in length of stay, readmissions, rate of positive margins, conversion from robotic to open, or 90-day mortality.

CONCLUSIONS

When accounting for robotic esophagectomy volume, late adoption of robotic esophagectomy was associated with a reduced lymph node harvest, but other postoperative outcomes were similar. These data suggest that programs can safely start new robotic esophagectomy programs, but must ensure an adequate case load.

Robotic minimally invasive esophagectomy provides superior surgical resection

INTRODUCTION

Robotic minimally invasive esophagectomy (RMIE) and "traditional" minimally invasive esophagectomy techniques (tMIE) have reported superior outcomes relative to open techniques. Differences in the outcomes of these two approaches have not been examined. We hypothesized that short-term outcomes of RMIE would be superior to tMIE.

METHODS AND PROCEDURES

The National Cancer Database was used to analyze outcomes of patients undergoing RMIE and tMIE from 2010 to 2016. Patients with clinical metastatic disease were excluded. Trends in the number of procedures performed with each approach were described using linear regression testing. Primary outcome of interest was 90-day mortality rate. Secondary outcomes of interest were positive surgical margin rate, number of lymph nodes (LN) removed, adequate lymphadenectomy (> 15 LNs), length of hospitalization (LOS), readmission rate, and conversion to open rate. Outcomes of RMIE and tMIE were compared using Wilcoxon rank sum test and chi square test as appropriate. Multivariable regression was also performed to reduce the impact of differences in the cohorts of patients receiving RMIE and tMIE.

RESULTS

6661 minimally invasive esophagectomies were performed from 2010 to 2016 (1543/6661 (23.2%) RMIE and 5118/6661 (76.8%) tMIE). Over the study period, the proportion of RMIE increased from 10.4% (64/618) in 2010 to 27.2% (331/1216) in 2016 (p < 0.001) (Fig. 1). The primary outcome of 90-day mortality was similar between RMIE and tMIE (92/1170 (7.4%) vs 305/4148 (7.9%), p = 0.558) (Table 2). RMIE and tMIE also had similar readmission rate (6.3 vs 7%, p = 0.380). There was no difference between the cohorts based on sex, age, race, insurance, and tumor size. The cohorts of patients receiving RMIE and tMIE differed in that RMIE patients had lower rates of elevated Charlson scores, were more likely to be treated at an academic institution, had a higher rate of advanced clinical T-stage and clinical nodal involvement, and had received neoadjuvant therapy. In a univariate analysis, RMIE had a lower rate of positive margin (3.9 vs 6.1%, p = 0.001), more mean lymph nodes evaluated (16.6 ± 9.74 vs 16.1 ± 10.08 p = 0.018), lower conversion to open rate (5.4 vs 11.4%, p < 0.001), and a shorter mean length of stay (12.1 ± 10.39 vs 12.8 ± 11.18 days, p < 0.001). In multivariable analysis, RMIE was associated with lower risk of conversion to open (OR 0.51, 95% CI: 0.37-0.70, p < 0.001) and lower rate of positive margin (OR 0.62, 95% CI: 0.41-0.93, p = 0.021).). Additionally, in a multivariable logistic regression, RMIE demonstrated superior adequate lymphadenectomy (> 15 LNs) (OR 1.18, 95% CI 1.02-1.37, p < 0.032).

CONCLUSION

In the National Cancer Database, robotic esophagectomy is associated with superior rate of conversion to open and positive surgical margin status. We speculate enhanced dexterity and visualization of RMIE facilitates intraoperative performance leading to improvement in these outcomes.

Retrograde signaling by a mtDNA-encoded non-coding RNA preserves mitochondrial bioenergetics

Alveolar epithelial type II (AETII) cells are important for lung epithelium maintenance and function. We demonstrate that AETII cells from mouse lungs exposed to cigarette smoke (CS) increase the levels of the mitochondria-encoded non-coding RNA, mito-RNA-805, generated by the control region of the mitochondrial genome. The protective effects of mito-ncR-805 are associated with positive regulation of mitochondrial energy metabolism, and respiration. Levels of mito-ncR-805 do not relate to steady-state transcription or replication of the mitochondrial genome. Instead, CS-exposure causes the redistribution of mito-ncR-805 from mitochondria to the nucleus, which correlated with the increased expression of nuclear-encoded genes involved in mitochondrial function. These studies reveal an unrecognized mitochondria stress associated retrograde signaling, and put forward the idea that mito-ncRNA-805 represents a subtype of small non coding RNAs that are regulated in a tissue- or cell-type specific manner to protect cells under physiological stress.

Minimally invasive esophagectomy is associated with superior survival, lymphadenectomy and surgical margins: propensity matched analysis of the National Cancer Database

Despite excellent short-term outcomes of minimally invasive esophagectomy (MIE), there is minimal data on long-term outcomes compared to open esophagectomy. MIE's superior visualization may have improved lymphadenectomy and complete resection rate and therefore improved long-term outcomes. We hypothesized that MIE would have superior long-term survival. Patients undergoing an esophagectomy for cancer between 2010 and 2016 were identified in the National Cancer Database. MIE included laparoscopic/robotic approach, and conversions were categorized as open. A 1:1 propensity match was performed. Lymphadenectomy and margin status were compared between MIE and open using Stuart Maxwell marginal homogeneity and Wilcoxon matched-pair signed-rank test. Survival was compared using log-rank test. 13,083 patients were identified: 8,906 (68%) open and 4,177 (32%) MIE. Propensity matching identified 3,659 'pairs' of MIE and open esophagectomy patients. Among them, MIE was associated with higher number lymph nodes examined (16 vs. 14, P < 0.001) and similar number of positive lymph nodes (0 vs. 0, P = 0.33). MIE had higher rate of negative pathologic margin (95 vs. 93.5%, P < 0.001). MIE was also associated with shorter hospitalization (9 vs. 10 days, P < 0.001). Survival was improved among MIE patients (46.6 vs. 41.4 months for open, P = 0.003) and among pathologic node-negative patients (71.4 vs. 61.5 months, P = 0.005). These data suggest that MIE has improved short-term outcomes (improved lymphadenectomy, pathologic margins, and length of stay) and also associated improved overall survival. The etiology of superior overall survival is likely secondary to many factors related and unrelated to surgical approach.

Mortality After Esophagectomy: Analysis of Individual Complications and Their Association with Mortality

BACKGROUND

The relationship between individual complications and esophagectomy mortality is unclear. The influence of comorbidities on the impact of complications on operative mortality is also unknown. We sought to assess the impact of individual complications and the effect of coexisting comorbidities on operative mortality following esophagectomy.

METHODS

All gastric conduit esophagectomies performed for cancer from 2008 to 2017 in the Society of Thoracic Surgery database were identified. Chi square was utilized to identify postoperative events associated with operative mortality. Multivariable logistic regression analysis was performed, utilizing postoperative events, to determine the risk-adjusted effect on operative mortality for each postoperative event. To assess the effect of preoperative comorbidities, a second logistic regression analysis was performed, incorporating preoperative characteristics.

RESULTS

Of 11,943 esophagectomy patients, 63.9% had a postoperative event and 3.3% died, which did not change over the study period. The postoperative events with the highest impact on operative mortality were respiratory distress syndrome (OR 7.48 (95% CI 5.23-10.7)), reintubation (OR 6.55 (4.61-9.30)), and renal failure (OR 5.97 (4.08-8.75)). Anastomotic leak requiring reoperation was associated with increased operative mortality (OR 1.48 (1.03-2.14)), but medically managed leak was not. Incorporating preoperative characteristics into the operative mortality model had little effect on odds ratio for death for individual postoperative events.

CONCLUSIONS

In the Society of Thoracic Surgery database, 64% of patients suffer postoperative events and 3.3% die following esophagectomy. The independent association of certain postoperative events with mortality is an objective method of terming a complication "major" and may aid efforts to reduce mortality.

Antibiotic Resistance Is Associated With Morbidity and Mortality After Decortication for Empyema

BACKGROUND

Previous studies of decortication for empyema demonstrated that patient characteristics are associated with mortality, but the relationship of infectious pathogen to outcome has not been described. Our objective was to analyze the association of microbiology and antibiotic resistance with postoperative mortality after decortication for empyema. We hypothesized that bacterial pathogens, antibiotic resistance, and patient characteristics would all contribute to perioperative morbidity and mortality.

METHODS

Patients undergoing pulmonary decortication for empyema from January 1, 2010 to October 1, 2017 were reviewed retrospectively. Cases were matched with microbiology cultures. Outcomes of interest were a composite of death, tracheostomy, initial ventilator support greater than 48 hours, or unexpected intensive care unit readmission. Antibiotic resistance was categorized as present or absent, and the number of antibiotics with resistance was counted for each patient. We describe the relationship of patient characteristics, antibiotic resistance, and microbiology to mortality.

RESULTS

During the study period, 185 patients underwent decortication, 118 of whom had a diagnosis of primary empyema (63.8%). Positive culture results were present in 79 of 185 patients (43%). The most common isolate was Streptococcus, which was present in 29 of 79 (37%), followed by Staphylococcus in 19 of 79 (24%). Of 79 patients, 11 had fungal infections (13.9%). In addition, 16 of 79 patients had polymicrobial empyema (20%). Of 185 patients, 30 experienced the composite adverse outcome (16.2%). In multivariable regression, the composite adverse outcome was associated with emphysema, Candida in pleural culture, and antibiotic resistance count.

CONCLUSIONS

Perioperative mortality and morbidity after decortication for empyema is considerable. In this cohort, infections with increasing antibiotic resistance are associated with morbidity and mortality among patients with empyema.

Outcomes of robotic esophagectomies for esophageal cancer by hospital volume: an analysis of the national cancer database

BACKGROUND

Robotic minimally invasive esophagectomies (RMIE) have been associated with superior outcomes; however, it is unclear if these are specific to robotic technique or are present only at high-volume institutions. We hypothesize that low-volume RMIE centers would have inferior outcomes.

METHODS

The National Cancer Database (NCDB) identified patients receiving RMIE from 2010 to 2016. Based on the total number of RMIE performed by each hospital system, the lowest quartile performed ≤ 9 RMIE over the study period. Ninety-day mortality, number of lymph nodes evaluated, margins status, unplanned readmissions, length of stay (LOS), and overall survival were compared. Regression models were used to account for confounding.

RESULTS

1565 robotic esophagectomies were performed by 212 institutions. 173 hospitals performed ≤ 9 RMIE (totaling 478 operations over the study period, 30.5% of RMIE) and 39 hospitals performed > 9 RMIE (1087 operations, 69.5%). Hospitals performing > 9 RMIE were more likely to be academic centers (90.4% vs 66.2%, p < 0.001), have patients with advanced tumor stage (65.3% vs 59.8%, p = 0.049), andadministered preoperative radiation (72.8% vs 66.3%, p = 0.010). There were no differences based on demographics, nodal stage, or usage of preoperative chemotherapy. On multivariable regressions, hospitals performing ≤ 9 RMIE were associated with a greater likelihood of experiencing a 90-day mortality, a reduced number of lymph nodes evaluated, and a longer LOS; however, there was no association with rates of positive margins or unplanned readmissions. Median overall survival was decreased at institutions performing ≤ 9 RMIE (37.3 vs 51.5 months, p < 0.001). Multivariable Cox regression demonstrated an association with poor survival comparing hospitals performing ≤ 9 to > 9 RMIE (HR 1.327, p = 0.018).

CONCLUSION

Many robotic esophagectomies occur at institutions which performed relatively few RMIE and were associated with inferior short- and long-term outcomes. These data argue for regionalization of robotic esophagectomies or enhanced training in lower volume hospitals.

Conversion from thoracoscopic to open pneumonectomy is not associated with short- or long-term mortality

BACKGROUND

Thoracoscopic pneumonectomy commonly requires conversion to thoracotomy. We hypothesize that conversion would lead to worse short- and long-term outcomes compared with operations completed thoracoscopically.

METHODS

The National Cancer Database identified patients who underwent a pneumonectomy from 2010 to 2016. Trends in conversion were described using linear regression. Multivariable regression of factors associated with conversion was performed. Short-term outcomes included duration of stay, number of lymph nodes harvested, margin status, readmission, and 90-day mortality. Long-term outcome was evaluated as overall survival.

RESULTS

A total of 8,037 patients underwent a pneumonectomy. The rate of attempted thoracoscopic pneumonectomies increased from 11% to 22% (P < .001) and the rate of conversion decreased from 39% to 33% (P = .011). On multivariable analysis, a greater patient comorbidity index and pathologic nodal-stage 2 disease were associated with an increased rate of conversion. The mean number of lymph nodes evaluated was greater as was the duration of stay in the conversion group, but conversion to open thoracotomy was not associated with positive surgical margins, readmission rate, 90-day mortality, or survival.

CONCLUSION

Thoracoscopic pneumonectomy is performed with increasing frequency and decreasing conversion rate. Conversion is associated with a greater duration of stay but other short- and long-term outcomes are similar. This observation suggests minimal harm in conversion from minimally invasive to open pneumonectomy.