- Email: [email protected]
Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma
Accepted Manuscript Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma Chi-Fu Jeffrey Yang, MD, Hanghang Wang, MD, Arvind Kumar, BS, Xiaofei Wang, Ph.D., Matthew G. Hartwig, MD, MHS, Thomas A. D'Amico, MD, Mark F. Berry, MD, MHS PII:
S0012-3692(17)31330-2
DOI:
10.1016/j.chest.2017.07.032
Reference:
CHEST 1239
To appear in:
CHEST
Received Date: 2 January 2017 Revised Date:
15 July 2017
Accepted Date: 25 July 2017
Please cite this article as: Yang CFJ, Wang H, Kumar A, Wang X, Hartwig MG, D'Amico TA, Berry MF, Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma, CHEST (2017), doi: 10.1016/j.chest.2017.07.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Title: Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma
RI PT
Authors: Chi-Fu Jeffrey Yang1, MD, Hanghang Wang1, MD, Arvind Kumar1, BS, Xiaofei Wang2, Ph.D., Matthew G. Hartwig1, MD, MHS, Thomas A. D'Amico1, MD, Mark F. Berry3, MD, MHS Institutions and Affiliations: 1. Department of Surgery, Division of Thoracic Surgery, Duke University Medical Center, 3496 DUMC, Durham, North Carolina.
Department of Biostatistics and Bioinformatics, 2424 Erwin Road Suite 1102, Durham, North Carolina
3.
Department of Cardiothoracic Surgery, Stanford University Medical Center, 300 Pasteur Drive, Falk Building 2nd Floor Stanford, CA 94305-5407
M AN U
SC
2.
Corresponding Author:
Mark Berry. Falk Cardiovascular Research Center, 300 Pasteur Drive, Stanford, CA 94305. Email: [email protected]
TE D
Reprints will not be available from the authors
Sources of Support/Conflicts of Interest: The authors have no conflicts of interest to declare. Running Head: Optimal Timing of Lobectomy
EP
Word Count: 2507
Conflict of Interest: The authors have no conflicts of interest to report.
AC C
Funding: This work was not supported by any outside financial source.
ACCEPTED MANUSCRIPT 2
ABSTRACT Background: Because the relationship between timing of surgery following diagnosis of lung cancer and
RI PT
survival has not been precisely described, guidelines on what constitutes a clinically meaningful delay of resection of early-stage lung cancer do not exist. This study tested the hypothesis that increasing time between diagnosis and lobectomy for stage IA squamous cell carcinoma (SCC) was associated with worse
SC
survival.
Methods: The association between timing of lobectomy and survival for patients with clinical stage IA
M AN U
SCC in the National Cancer Data Base (2006-2011) was assessed using multivariable Cox proportional hazards analysis and restricted cubic spline (RCS) functions.
Results: The 5-year overall survival of 4,984 patients who met study inclusion criteria was 58.3% (95% CI: 56.3%-60.2%). Surgery was performed within 30 days of diagnosis in 1,811 (36%) patients while the
TE D
median time to surgery was 38 days (IQR, 23-58). In multivariable analysis, patients who had surgery 38 days or more after diagnosis had significantly worse 5-year survival than patients who had surgery earlier (Hazard Ratio (HR), 1.13 (95% CI:1.02-1.25); p=0.02). Multivariable RCS analysis demonstrated the
EP
HR associated with time to surgery increased steadily the longer resection was delayed; the threshold time
AC C
associated with statistically significant worse survival was ~90 days or greater.
Conclusion: Longer intervals between diagnosis of early-stage lung SCC and surgery are associated with worse survival. Although factors other than the timing of treatment may contribute to this finding, these results suggest that efforts to minimize delays beyond those needed to perform a complete preoperative evaluation may improve survival.
ACCEPTED MANUSCRIPT 3
INTRODUCTION Patients with early-stage non-small cell lung cancer (NSCLC) may experience delays to appropriate treatment due to difficulty in accessing primary care, wait times to see specialists, and the
RI PT
time needed to complete appropriate staging studies and evaluation.1 The impact of this delay between detection and surgical treatment on outcomes has not been well characterized.2 Studies that have
investigated the association of the time from diagnosis to surgery and survival have shown conflicting
SC
results.3-8
As such, there is no consensus on what constitutes a clinically meaningful delay of treatment. Guidelines from the National Comprehensive Cancer Network, which are based on a consensus opinion,
M AN U
state that treatment delays “should be avoided”9 without providing further specifics while the European Society for Medical Oncology10 and the American College of Chest Physicians11 do not mention any recommendations on the time to initiation of surgical intervention. The purpose of this study was to provide quantitative data on how delays to lobectomy for early-stage NSCLC impact survival and
TE D
improve the level of evidence available to support treatment guidelines and patient management. The specific goal of the study was to test the hypothesis that increasing time between diagnosis and lobectomy for clinical stage IA lung squamous cell carcinoma was associated with worse survival using the National
AC C
METHODS
EP
Cancer Data Base (NCDB).
National Cancer Database
Data from the NCDB, which is a joint project of the American College of Surgeons Commission
on Cancer (CoC) and the American Cancer Society, was used in this study. The NCDB contains data from over 30 million patients and 1,500 cancer centers in the U.S. and contains over 70% of all newly diagnosed cases of cancer in the U.S. annually.12 All clinical staging information was directly recorded in the NCDB using American Joint Committee on Cancer (AJCC) 6th edition13 and 7th edition14 TNM classifications for the years of study inclusion (2006-2011).
ACCEPTED MANUSCRIPT 4
Study design This study was approved by the Duke Medicine Institutional Review Board for Clinical
RI PT
Investigators with project approval number Pro00045337. After isolating patients from a de-identified NCDB participant user file and using the International Classification of Diseases for Oncology, 3rd edition histology and topography codes, all clinical Stage 1A (clinical T1N0M0) NSCLC patients with squamous
SC
cell carcinoma (SCC) histology in the NCDB who were treated with lobectomy as primary therapy
without either induction chemotherapy or induction radiation were selected for analysis. Patient data
M AN U
spanned from the years 2006 to 2011. Clinical diagnosis was determined by a physician. Regarding exclusion criteria, the study was restricted to patients with SCC given the relatively recently more recognized importance of histology on both NSCLC treatment and outcomes; we restricted patients to this histology so that results from the study could be directly applicable to this specific patient population. We did not include patients with adenocarcinoma because adenocarcinoma has been
TE D
categorized into different subtypes of tumors with widely ranging prognosis and include adenocarcinoma in situ and minimally invasive adenocarcinomas that have approximately 100% 5-year disease-free survival.15 We did not include bronchioloalveolar or neuroendocrine tumors because they typically have
EP
better prognoses than other lung cancers, and large cell neuroendocrine patients were excluded because of their potentially worse prognosis.15-17 In our primary analysis, patients who had surgery on the same day
AC C
of diagnosis were excluded to reduce potential confounding given that there may have been inherent but unrecorded differences between patients who had surgery on the same day of diagnosis and patients who had surgery after their diagnosis; in a sensitivity analysis we did include these patients who had surgery on the same day of diagnosis.
Statistical analysis For each patient, the number of days from diagnosis of lung cancer to surgical resection was identified. Evaluation of the impact of timing of surgery was performed using several methods. First,
ACCEPTED MANUSCRIPT 5
surgery was categorized as "early" (1-37 days from time of diagnosis to surgery) or "late" (38+ days). These cutoffs were used because our initial evaluation of the distribution of timing of surgery in the NCDB showed a median time to surgery of 38 days. Furthermore, the choice of this time was felt to be
RI PT
clinically reasonable as the timeframe of 30 days or 1 month had been a previously used threshold in the literature12,18 and by the British Thoracic Society.19 We also performed additional sensitivity analyses, described below, where we used 30 days as a cutoff and included patients who underwent surgery on the
SC
day of diagnosis. In addition, a landmark analysis was performed at 4 months to account for the
differences in survival between the early and late surgery groups that may have been due to differences in
M AN U
perioperative mortality rates. We also identified predictors of delayed surgery to 38 days or more after diagnosis using a logistic regression model that included age, sex, race, Charlson/Deyo comorbidity (CDCC) score, tumor size insurance type, education, facility location, population of area, tumor location, income, distance to facility, hospital volume, biopsy prior to lobectomy and facility type. The primary outcome was overall survival, which was defined from time of diagnosis to time of
TE D
last follow-up or death and estimated using the Kaplan-Meier product limit method. Methods of followup are described previously.20 Differences in survival between the “early” and “late” surgery groups were assessed using the log-rank test. The impact of having surgery “early” or “late” was further evaluated
EP
using a Cox proportional hazard model that included this variable as well as age, sex, race, CDCC score, insurance type, tumor size, facility type, distance from patient’s residence to hospital, median household
AC C
income, tumor location, biopsy prior to lobectomy, and hospital volume. All model covariates were chosen a priori based on clinical relevance. Next, a restricted cubic spline (RCS) analysis with 3 knots (located at the 10th, 50th, and 90th
percentiles) was used to model the relationship between time to surgery and survival.21 Spline regression fits smooth polynomial functions (i.e. first and second derivatives) between pre-defined points (knots) on a graph and joins them in a piecewise manner.22,23 Splines can be utilized to generate non-linear models between a continuous predictor variable (e.g., timing to lobectomy), and an outcome (e.g. hazard ratio), without the loss of statistical power that is frequently associated with the approach of categorization.
ACCEPTED MANUSCRIPT 6
RCS models are a specific type of spline model that smooths the curve while reducing the influence of outliers as compared to polynomial regression. As a result, RCS models minimize residuals for confounding variables, and are effective in determining threshold values of the predictor variable of
RI PT
significance.22,24-27 In this study, a multivariable RCS model was used to characterize the strength and the shape of the association between the time between diagnosis and lobectomy and survival. Primary analysis was conducted using the median value for all continuous variables as reference standard. Because sex is an
SC
important predictor of survival,28,29 we created RCS plots stratified by sex.
M AN U
Statistical analyses were performed using Stata Statistical Software: Release 13.0 (StataCorp LP, College Station, TX) and R software version 3.2.2 (Vienna, Austria). Comparisons of baseline characteristics and unadjusted outcomes were performed using the Wilcoxon Rank Sum test for continuous variables and Pearson’s chi-square test for categorical variables. A 2-sided p-value of 0.05
RESULTS
TE D
was used to define significance.
The study cohort consisted of 4,984 patients who underwent lobectomy for clinical stage IA
EP
(cT1N0) lung SCC at least 1 day after diagnosis of lung cancer (Figure 1). Median time to surgery was 38 days (IQR, 24-60) and 95% of patients underwent surgery within 4 months of diagnosis. Table 1 details
AC C
the preoperative baseline characteristics and peri-operative outcomes stratified by whether their surgery was performed before or after the median delay of 38 days. Table 2 demonstrates predictors of whether surgery occurred before or after the median delay of 38 days. Patients who underwent biopsy prior to lobectomy, older patients and black patients had increased wait times. Patients who underwent a left lower lobectomy were less likely to experience delays when compared to patients undergoing a right upper lobectomy. The 5-year overall survival was 58.3% for the entire cohort (Figure 2a). In univariable analysis, and after a median follow-up of 32.0 (range: 0.3–96.6) months for the entire cohort, patients who received
ACCEPTED MANUSCRIPT 7
surgery within 37 days of diagnosis had better survival when compared to patients who received surgery 38 days or after diagnosis (Figure 2b, 2c). After multivariable adjustment, delayed surgery beyond 37 days from time of diagnosis was found to be significantly associated with worse survival (Hazard ratio
RI PT
(HR) 1.13 (95% CI, 1.02-1.25); p=0.02). In addition, age, gender, comorbidities, tumor size, and type of insurance, facility type, and median income level were all significant predictors of survival (Table 3). Finally, when we repeated the analysis using 30 days as a cutoff between early and late, we found
SC
that the 5-year overall survival for patients who received surgery 30 days or before was significantly better than the survival for patients who received surgery after 30 days (61.5% (95% CI, 58.3%-64.6%)
M AN U
vs. 56.4% (95% CI, 53.9%-58.8%); p<0.01). Multivariable analysis, though, showed that delaying surgery beyond 30 days was not significantly associated with worse survival (HR, 1.11 (95% CI, 1.001.24); p=0.058). In our landmark analysis, patients who underwent early surgery had significantly better 5-year overall survival than patients who underwent late surgery (Figure 2d). Restricted multivariable cubic spline plots for all analyses are shown in Figure 3. Figures 3a and
TE D
3b present the relationship between delay and HR stratified by sex, Figure 3c presents the relationship between delay and HR for patients for the entire cohort. All three plots (Figures 3a-3c) demonstrate that the HR immediately begins to increase with increasing time from diagnosis to lobectomy for each cohort
EP
analyzed. For the entire cohort, the HR did not become statistically significant until approximately 90 days and beyond from time of diagnosis to time of surgery.
AC C
In a sensitivity analysis where we included patients who underwent surgery on the day of diagnosis for a total cohort of 6,723 patients, the results were consistent with those reported above. Median time to surgery was 28.5 days (IQR, 0-50). “Late” surgery, defined as surgery 29 or more days after diagnosis, was associated with worse 5-year survival (e-Figure 1) even after multivariable adjustment (e-Table 1). Furthermore, restricted cubic spline analysis showed that, as in the previous analyses, the HR increases as time between diagnosis and surgery increased (Figure 3d).
DISCUSSION
ACCEPTED MANUSCRIPT 8
This study examined the relationship between timing of lobectomy following diagnosis of lung squamous cell carcinoma and overall survival for clinical stage 1A patients. Our study found that the
RI PT
median delay between diagnosis and surgery in the NCDB was 38 days. Patients whose surgery was 38 or more days after diagnosis were found to have worse survival in both univariate and multivariate
analysis. When we characterized the non-linear relationship between timing of lobectomy and survival using restricted cubic spline analyses, we found that the hazard ratio associated with time between
SC
diagnosis and surgery steadily increased with longer delays and the threshold for when the association between time from diagnosis to surgery and worse survival became statistically significant was
M AN U
approximately 90 days.
Previous studies on delays to treatment for NSCLC differed with respect to the disease stage, the treatment characteristics included, as well as the arbitrarily selected cutoff for the definition of early versus delayed start of therapies. As a result, while some studies have demonstrated that treatment delays are associated with disease progression and worse survival,6-8,30-36 other studies have found no association
TE D
between delayed treatment and long-term outcomes,3-5,37-41 and some studies have shown an association between delayed treatment and better survival.42-45 Importantly, none of these studies specifically evaluated the impact of lobectomy on NSCLC nor have they examined stage IA alone. A strength of this
EP
study is that by using the NCDB to create a large cohort and by only examining the specific situation of lobectomy for clinical stage IA patients with SCC, we were able to limit confounders by not including a
AC C
heterogeneous group of stages, histologies or treatments. Another strength of this study was that through multivariable restricted cubic spline modeling, we characterized the nonlinear relationship between the timing of lobectomy and overall survival. Furthermore, as a population based study, these results can be generalized to the overall population of the United States, one of the strengths of using the NCDB. Our analysis showed several time points that were associated with worse survival in a statistically significant manner, including the median delay of 38 days from the analysis of the entire cohort, as well
ACCEPTED MANUSCRIPT 9
as the time of 90 days from the multivariable RCS analysis. Both analyses were consistent in showing that a longer delay led to worse outcomes. The present study also found several factors associated with delayed surgery. Patients who had a
RI PT
biopsy prior to lobectomy, who were older and black had increased times between diagnosis and surgery. Tumor location was also found to be associated with timing of lobectomy. Patients who underwent a left lower lobectomy were less likely to experience delays when compared to patients undergoing a right
SC
upper lobectomy. Potentially, a technically more straightforward operation (e.g. a left lower lobectomy) might be more quickly undertaken by some surgeons versus others. In addition, certain geographic areas
M AN U
(East south-central, and West North and South Central) were also associated with decreased wait times. This study has several limitations. First, because of the study’s retrospective design, there is a possibility for unobserved confounding and selection bias to exist. Although we used multivariable analysis to reduce bias, there are important covariates such as surgeon experience, and detailed co-morbidity and pulmonary function data that are not available in the NCDB. Lack of detailed co-morbidity and
TE D
pulmonary function data can impact findings and is an important limitation of this study. For example, patients who waited longer for surgery may have been sicker and required more pre-operative evaluation which could have delayed surgery; the “need” for delayed surgery as opposed to the actual delay of
EP
surgery may have been associated with worse outcomes. The NCDB does have co-morbidity scores and we were able to somewhat address this limitation by adjusting for co-morbidity scores in our analysis.
AC C
Second, detailed information regarding the timing of diagnosis was not available. It is unclear whether the diagnosis made for the patients was through radiographic findings, cytology or by biopsy. Given that the time of diagnosis is abstracted from patient charts, and given that the diagnosis of clinical stage IA is typically made through computed tomography (CT),9 we hypothesize that the majority of these patients had their clinical diagnosis made by CT. Third, data regarding complications does not exist in the NCDB. Fourth, our results are not necessarily generalizable to other stages of NSCLC or histology other than SCC. Lastly, cancer-specific, recurrence-free and disease-free survival are not recorded in the NCDB.
ACCEPTED MANUSCRIPT 10
In conclusion, although practical considerations may always require some time period between diagnosis of early-stage NSCLC and treatment, efforts to minimize unnecessary delays are recommended as outcomes are worse when surgical resection is delayed for clinical stage IA squamous cell carcinoma.
RI PT
Future studies should focus on determining the optimal timing of treatment for different stages of NSCLC
AC C
EP
TE D
M AN U
SC
as well as for other histologies.
ACCEPTED MANUSCRIPT 11
ACKNOWLEDGMENTS The data used in this study are derived from a de-identified NCDB file. The American College of Surgeons has executed a Business Associate Agreement that includes a data use agreement with each of
RI PT
its CoC-accredited hospitals. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methodology employed, or the conclusions drawn from these data by the investigators.
SC
C.J.Y. is the guarantor of the paper and takes responsibility for the integrity of the work as a whole. C.J.Y., H.H.W., and A.K. contributed equally in literature search, figures, study design, data
M AN U
collection, data analysis, interpretation and writing, and revisions. X.W. contributed to analyzing and interpreting the data as well as revisions. T.A.D., M.G.H., and M.F.B. contributed to design, data analysis, and interpretation and revisions.
AC C
EP
TE D
This work was not supported by any outside financial source.
ACCEPTED MANUSCRIPT 12
References
7. 8.
9. 10.
11.
12. 13. 14. 15.
16. 17. 18.
19.
20.
21.
RI PT
SC
6.
M AN U
5.
TE D
4.
EP
2. 3.
Evans WK. Prognostic implications of treatment delays in the surgical resection of lung cancer. Thoracic surgery clinics. 2013;23(2):225-232. SEER. Cancer Statistics Factsheets: Lung and Bronchus Cancer. National Cancer Institute. Bethesda, MD, . Aragoneses FG, Moreno N, Leon P, et al. Influence of delays on survival in the surgical treatment of bronchogenic carcinoma. Lung cancer. 2002;36(1):59-63. Quarterman RL, McMillan A, Ratcliffe MB, Block MI. Effect of preoperative delay on prognosis for patients with early stage non-small cell lung cancer. The Journal of thoracic and cardiovascular surgery. 2003;125(1):108-113; discussion 113-104. Shin DW, Cho J, Kim SY, et al. Delay to curative surgery greater than 12 weeks is associated with increased mortality in patients with colorectal and breast cancer but not lung or thyroid cancer. Annals of surgical oncology. 2013;20(8):2468-2476. Gomez DR, Liao KP, Swisher SG, et al. Time to treatment as a quality metric in lung cancer: Staging studies, time to treatment, and patient survival. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2015;115(2):257-263. Kanarek NF, Hooker CM, Mathieu L, et al. Survival after community diagnosis of early-stage non-small cell lung cancer. The American journal of medicine. 2014;127(5):443-449. Samson P, Patel A, Garrett T, et al. Effects of Delayed Surgical Resection on Short-Term and Long-Term Outcomes in Clinical Stage I Non-Small Cell Lung Cancer. The Annals of thoracic surgery. 2015;99(6):1906-1913. Ettinger DS, Wood DE, Akerley W, et al. Non-small cell lung cancer, version 6.2015. Journal of the National Comprehensive Cancer Network : JNCCN. 2015;13(5):515-524. Vansteenkiste J, Crino L, Dooms C, et al. 2nd ESMO Consensus Conference on Lung Cancer: early-stage non-small-cell lung cancer consensus on diagnosis, treatment and follow-up. Ann Oncol. 2014;25(8):14621474. Howington JA, Blum MG, Chang AC, Balekian AA, Murthy SC. Treatment of stage I and II non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(5 Suppl):e278S-313S. Bilimoria KY, Ko CY, Tomlinson JS, et al. Wait times for cancer surgery in the United States: trends and predictors of delays. Annals of surgery. 2011;253(4):779-785. Greene FL, American Joint Committee on Cancer., American Cancer Society. AJCC cancer staging manual. 6th ed. New York: Springer; 2002. Edge SB, American Joint Committee on Cancer. AJCC cancer staging manual. 7th ed. New York: Springer; 2010. Travis WD, Brambilla E, Noguchi M, et al. International association for the study of lung cancer/american thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2011;6(2):244-285. Rekhtman N. Neuroendocrine tumors of the lung: an update. Arch Pathol Lab Med. 2010;134(11):16281638. Bertino EM, Confer PD, Colonna JE, Ross P, Otterson GA. Pulmonary neuroendocrine/carcinoid tumors: a review article. Cancer. 2009;115(19):4434-4441. Devbhandari MP, Bittar MN, Quennell P, et al. Are we achieving the current waiting time targets in lung cancer treatment? Result of a prospective study from a large United kingdom teaching hospital. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2007;2(7):590-592. BTS recommendations to respiratory physicians for organising the care of patients with lung cancer. The Lung Cancer Working Party of the British Thoracic Society Standards of Care Committee. Thorax. 1998;53 Suppl 1:S1-8. Koike T, Koike T, Yoshiya K, Tsuchida M, Toyabe S. Risk factor analysis of locoregional recurrence after sublobar resection in patients with clinical stage IA non-small cell lung cancer. The Journal of thoracic and cardiovascular surgery. 2013;146(2):372-378. Harrell F. Regression Modelling Strategies With Applications to Linear Models, Logistic Regression, and Survival Analysis. . New York, NY: SpringerVerlag; 2001.
AC C
1.
ACCEPTED MANUSCRIPT 13
28.
29.
30. 31.
32.
33.
34. 35. 36.
37. 38.
39.
40.
41. 42.
RI PT
27.
SC
26.
M AN U
25.
TE D
24.
EP
23.
Adam MA, Pura J, Goffredo P, et al. Presence and Number of Lymph Node Metastases Are Associated With Compromised Survival for Patients Younger Than Age 45 Years With Papillary Thyroid Cancer. J Clin Oncol. 2015;33(21):2370-2375. Durrleman S, Simon R. Flexible regression models with cubic splines. Statistics in medicine. 1989;8(5):551-561. Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Statistics in medicine. 2010;29(9):1037-1057. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology. 1995;6(4):356-365. Molinari N, Daures JP, Durand JF. Regression splines for threshold selection in survival data analysis. Statistics in medicine. 2001;20(2):237-247. Saraiya M, Kottiri BJ, Leadbetter S, et al. Total and percent free prostate-specific antigen levels among U.S. men, 2001-2002. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2005;14(9):2178-2182. Travis WD, Brambilla E, Nicholson AG, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2015;10(9):1243-1260. Radzikowska E, Glaz P, Roszkowski K. Lung cancer in women: age, smoking, histology, performance status, stage, initial treatment and survival. Population-based study of 20 561 cases. Ann Oncol. 2002;13(7):1087-1093. Mohammed N, Kestin LL, Grills IS, et al. Rapid disease progression with delay in treatment of non-smallcell lung cancer. International journal of radiation oncology, biology, physics. 2011;79(2):466-472. Everitt S, Herschtal A, Callahan J, et al. High rates of tumor growth and disease progression detected on serial pretreatment fluorodeoxyglucose-positron emission tomography/computed tomography scans in radical radiotherapy candidates with nonsmall cell lung cancer. Cancer. 2010;116(21):5030-5037. Everitt S, Plumridge N, Herschtal A, et al. The impact of time between staging PET/CT and definitive chemo-radiation on target volumes and survival in patients with non-small cell lung cancer. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2013;106(3):288-291. Redaniel MT, Martin RM, Ridd MJ, Wade J, Jeffreys M. Diagnostic intervals and its association with breast, prostate, lung and colorectal cancer survival in England: historical cohort study using the Clinical Practice Research Datalink. PLoS One. 2015;10(5):e0126608. Buccheri G, Ferrigno D. Lung cancer: clinical presentation and specialist referral time. The European respiratory journal. 2004;24(6):898-904. Kanashiki M, Satoh H, Ishikawa H, Yamashita YT, Ohtsuka M, Sekizawa K. Time from finding abnormality on mass-screening to final diagnosis of lung cancer. Oncology reports. 2003;10(3):649-652. Kashiwabara K, Koshi S, Itonaga K, Nakahara O, Tanaka M, Toyonaga M. Outcome in patients with lung cancer found on lung cancer mass screening roentgenograms, but who did not subsequently consult a doctor. Lung cancer. 2003;40(1):67-72. Bozcuk H, Martin C. Does treatment delay affect survival in non-small cell lung cancer? A retrospective analysis from a single UK centre. Lung cancer. 2001;34(2):243-252. Kashiwabara K, Koshi S, Ota K, Tanaka M, Toyonaga M. Outcome in patients with lung cancer found retrospectively to have had evidence of disease on past lung cancer mass screening roentgenograms. Lung cancer. 2002;35(3):237-241. Loh LC, Chan LY, Tan RY, et al. Effect of time delay on survival in patients with non-small cell lung cancer--a Malaysian study. Asia-Pacific journal of public health / Asia-Pacific Academic Consortium for Public Health. 2006;18(1):69-71. Pita-Fernandez S, Montero-Martinez C, Pertega-Diaz S, Verea-Hernando H. Relationship between delayed diagnosis and the degree of invasion and survival in lung cancer. Journal of clinical epidemiology. 2003;56(9):820-825. Salomaa ER, Sallinen S, Hiekkanen H, Liippo K. Delays in the diagnosis and treatment of lung cancer. Chest. 2005;128(4):2282-2288. Annakkaya AN, Arbak P, Balbay O, Bilgin C, Erbas M, Bulut I. Effect of symptom-to-treatment interval on prognosis in lung cancer. Tumori. 2007;93(1):61-67.
AC C
22.
ACCEPTED MANUSCRIPT 14
RI PT SC M AN U TE D
45.
EP
44.
Myrdal G, Lambe M, Hillerdal G, Lamberg K, Agustsson T, Stahle E. Effect of delays on prognosis in patients with non-small cell lung cancer. Thorax. 2004;59(1):45-49. Neal RD, Allgar VL, Ali N, et al. Stage, survival and delays in lung, colorectal, prostate and ovarian cancer: comparison between diagnostic routes. Br J Gen Pract. 2007;57(536):212-219. Nadpara P, Madhavan SS, Tworek C. Guideline-concordant timely lung cancer care and prognosis among elderly patients in the United States: A population-based study. Cancer Epidemiol. 2015;39(6):1136-1144.
AC C
43.
ACCEPTED MANUSCRIPT 15
Figure Legends Figure 1
RI PT
Consort Diagram Showing Schema of Study Subject Selection
Figure 2a
Overall Survival of Clinical Stage 1A NSCLC Patients with Squamous Cell Carcinoma Undergoing
SC
Lobectomy. Represents overall survival of all patients, regardless of the number of days between
M AN U
diagnosis and surgery.
Figure 2b
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Survival is calculated from date of diagnosis. Patients who received lobectomy within 37 days after diagnosis were classified as having "early" lobectomy while those who
Figure 2c
TE D
received surgery starting on day 38 were classified as receiving "late" lobectomy.
EP
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Overall survival is calculated from date of definitive surgery. Patients who
AC C
received lobectomy within 37 days after diagnosis were classified as having "early" lobectomy while those who received surgery starting on day 38 were classified as receiving "late" lobectomy.
Figure 2d
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Overall survival is calculated from four months after date of diagnosis. Patients who received lobectomy within 37 days after diagnosis were classified as having "early"
ACCEPTED MANUSCRIPT 16
lobectomy while those who received surgery starting on day 38 were classified as receiving "late" lobectomy.
RI PT
Figure 3 Relationship between Timing to Lobectomy and Overall Survival for Patients with Clinical Stage IA NSCLC diagnosed with squamous cell carcinoma. Spline plots for adjusted Cox models excluding
SC
patients who underwent surgery on the day of diagnosis for a) men, b) women, c) the entire cohort, and d) the entire cohort including patients who underwent surgery on the day of diagnosis. Median values for all
M AN U
covariates are used as reference standard. Model was adjusted for age, race, Charlson/Deyo comorbidity (CDCC) score, insurance type, tumor size, facility type, distance from patient’s residence to hospital,
AC C
EP
TE D
median household income and hospital volume. Gray shaded area indicates 95% CIs.
ACCEPTED MANUSCRIPT
Table 1. Patient Characteristics and Perioperative Outcomes Stratified by Early (1-37 days) vs Late (38+ days) Surgery
69 (64,75)
EP
AC C
70 (64,76)
1,365 (55.9%) 1,075 (44.1%)
1,395 (54.8%) 1,149 (45.2%)
2,225 (91.2%) 153 (6.3%) 62 (2.5%)
2,266 (89.1%) 224 (8.8%) 54 (2.1%)
P-value 0.006 0.43
0.07
916 (36.0%) 1,033 (40.6%) 595 (23.4%)
M AN U
922 (37.8%) 1,013 (41.5%) 505 (20.7%)
SC
0.005
28 (1.2%) 629 (25.8%) 82 (3.4%) 1,652 (67.7%) 20 (0.8%) 29 (1.2%)
33 (1.3%) 585 (23.0%) 125 (4.9%) 1,744 (68.6%) 36 (1.4%) 21 (0.8%)
224 (9.2%) 1,529 (66.7%) 684 (28.0%) 3 (0.1%)
212 (8.3%) 1,507 (59.2%) 821 (32.3%) 4 (0.2%)
TE D
Patient Age (years), median (IQR) Sex, n (%) Male Female Race, n (%) White Black Other Charlson/Deyo Score, n (%) 0 1 2 Insurance Status, n (%) Uninsured Private Medicare Medicaid Other Government Unknown Facility Type, n (%) Community Cancer Program Comp. Comm. Cancer Program Academic/Research Program Other Diagnostic and Staging Procedure Prior to Lobectomy by Facility type Community Cancer Program No biopsy Biopsy prior to Lobectomy Other/Unknown Comp. Comm. Cancer Program No biopsy Biopsy prior to Lobectomy Other/Unknown Academic/Research Program No biopsy Biopsy prior to Lobectomy Other/Unknown Tumor Location, n (%) RLL LLL RML RUL LUL
Late Surgery (N=2,544)
RI PT
Early Surgery (N=2,440)
0.004
0.01
0.64
88 (39.3%) 133 (59.4%) 3 (1.3%)
74 (34.9%) 135 (63.7%) 3 (1.4%)
598 (39.1%) 890 (58.2%) 41 (2.7%)
553 (36.7%) 931 (61.8%) 23 (1.5%)
378 (55.3%) 302 (44.2%) 4 (0.6%)
359 (43.7%) 458 (55.8%) 4 (0.5%)
428 (17.9%) 354 (14.8%) 137 (5.7%) 838 (35.0%) 639 (26.7%)
483 (20.4%) 321 (12.9%) 121 (4.9%) 844 (33.8%) 726 (29.1%)
0.02
<0.001
0.05
ACCEPTED MANUSCRIPT
0.26
EP
AC C
3 (0.1%) 1922 (76.4%) 1456 (18.1%) 44 (1.8%) 19 (0.8%) 0 (0.0%) 73 (2.9%)
RI PT
1 (0.0%) 1,898 (78.3%) 391 (16.1%) 32 (1.3%) 27 (1.1%) 1 (0.0%) 73 (3.0%)
0.86
1,039 (43.9%) 489 (20.6%) 294 (12.4%) 143 (6.0%) 54 (2.3%) 178 (7.5%) 97 (4.1%) 36 (1.5%) 39 (1.7%)
0.04 0.01 0.62 0.13
552 (22.1%) 671 (26.9%) 701 (28.1%) 571 (22.9%)
M AN U
476 (19.8%) 650 (27.1%) 672 (28.0%) 603 (25.1%)
2,221 (91.6%) 159 (6.6%) 44 (1.8%) 22 (17,28) 10.5 (4.3, 26.8) 42 (21,76)
SC
2,134 (92.0%) 143 (6.2%) 42 (1.8%) 22 (16,27) 11.5 (4.8, 29.8) 42 (22,72)
TE D
Pathologic T stage, n (%) T0 T1 T2 T3 T4 IS Unknown Pathologic N Stage, n (%) N0 N1 N2 Tumor Size (mm), median (IQR) Distance to Facility (mi.), median (IQR) Hospital Volume (# of cases), median (IQR) Median Income, n (%) 1st Quartile 2nd Quartile 3rd Quartile 4th Quartile Population of Area, n (%) >1 million in metro area 250,000 - 1 million in metro area <250,000 in metro area >20,000 near metro area >20,000 not near metro area 2,500 – 19,999 near metro area 2,500 – 19,999 not near metro area <2,500 near metro area <2,500 not near metro area Facility Location, n (%) New England Middle Atlantic South Atlantic East North Central East South Central West North Central West South Central Mountain Pacific Education Quartile, n (%) 1st Quartile 2nd Quartile 3rd Quartile 4th Quartile 30 Day Mortality, n (%) 90 Day Mortality, n (%)
0.03
1,094 (44.6%) 570 (23.3%) 285 (11.6%) 156 (6.4%) 30 (1.2%) 174 (7.1%) 79 (3.2%) 31 (1.3%) 32 (1.3%) <0.001
113 (4.6%) 250 (10.3%) 554 (22.7%) 462 (18.9%) 324 (13.3%) 251 (10.3%) 218 (8.9%) 97 (4.0%) 171 (7.0%)
129 (5.1%) 359 (14.1%) 575 (22.6%) 583 (22.9%) 258 (10.1%) 182 (7.2%) 129 (5.1%) 101 (4.0%) 228 (9.0%)
405 (16.9%) 709 (29.5%) 807 (33.6%) 480 (20.0%) 67 (2.8%) 116 (4.8%)
452 (18.1%) 799 (32.0%) 839 (33.6%) 406 (16.3%) 89 (3.5%) 150 (5.9%)
0.005
0.30 0.15
ACCEPTED MANUSCRIPT
Table 2. Independent predictors of Early (1-37 days) vs. Late (38+ days) Surgery for patients who underwent lobectomy for cT1 N0 NSCLC P-value
1.00, 1.02 0.92, 1.16
0.006 0.58
1.31 1.07
1.04, 1.65 0.55, 2.07
1.06 1.14
0.93, 1.21 0.98, 1.34
0.76 1.21 0.79 1.64 0.60 1.00
0.44, 1.29 0.66, 2.19 0.46, 1.35 0.75, 3.61 0.27, 1.32 1.00, 1.00
1.05 1.19 1.21 1.00 1.00
0.84, 1.30 0.93, 1.52 0.26, 5.57 1.00, 1.00 1.00, 1.00
0.69 0.17 0.81 0.12 0.83
0.94 0.92 0.88
0.78, 1.14 0.75, 1.13 0.69, 1.14
0.52 0.44 0.34
0.85 1.12 1.11 0.88
0.65, 1.12 0.94, 1.32 0.96, 1.29 0.73, 1.06
0.26 0.20 0.17 0.18
1.13 0.94 0.99 0.58 0.93 0.88 0.90 0.94
0.96, 1.32 0.77, 1.15 0.76, 1.28 0.36, 0.94 0.73, 1.19 0.63, 1.23 0.54, 1.50 0.57, 1.56
0.15 0.56 0.92 0.03 0.58 0.46 0.70 0.82
1.19 0.78 0.98 0.63 0.62 0.45 0.85 1.15
0.86, 1.64 0.59, 1.06 0.73, 1.33 0.45, 0.88 0.44, 0.87 0.31, 0.64 0.57, 1.28 0.82, 1.62
0.30 0.12 0.90 0.006 0.005 <0.001 0.45 0.42
EP
AC C
RI PT
95% CI
0.03 0.85
0.37 0.10
0.31 0.54 0.39 0.22 0.21 0.69
SC
M AN U
TE D
Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref<$38,000) $38,000 to $47,999 $48,000 to $62,999 $63,000 or greater Tumor Location (Ref=RUL) RML RLL LUL LLL Population of Area (Ref= >1 million) 250,000 - 1 million in metro area <250,000 in metro area >20,000 near metro area >20,000 not near metro area 2,500 – 19,999 near metro area 2,500 – 19,999 not near metro area <2,500 near metro area <2,500 not near metro area Facility Location (Ref=New England) Middle Atlantic South Atlantic East North Central East South Central West North Central West South Central Mountain Pacific
Adjusted Odds Ratio (AOR) 1.01 1.03
ACCEPTED MANUSCRIPT
0.99 0.87 0.71
0.82, 1.20 0.70, 1.08 0.55, 0.93
0.93 0.20 0.01
1.33 0.89
1.18, 1.51 0.54, 1.47
<0.001 0.65
AC C
EP
TE D
M AN U
SC
RI PT
Education: % without HS diploma (Ref>21%) 13% to 20.9% 7% to 12.9% Less than 7% Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
ACCEPTED MANUSCRIPT
Table 3. Independent predictors of mortality following Cox proportional hazards adjustment for patients who underwent lobectomy for cT1 N0 NSCLC 95% CI 1.02, 1.25 1.03, 1.04 0.65, 0.80
EP
AC C
P-value 0.02 <0.001 <0.001
1.14 1.09
0.94, 1.38 0.70, 1.88
0.18 0.76
1.16 1.42
1.03, 1.31 1.24, 1.62
0.02 <0.001
0.46 0.84 0.55 0.35 0.71 1.00
0.28, 0.74 0.50, 1.42 0.34, 0.88 0.16, 0.76 0.36, 1.40 1.00, 1.00
0.001 0.52 0.01 0.008 0.32 0.002
0.81 0.80 0.38 1.00 1.00
0.68, 0.97 0.65, 0.99 0.05, 2.76 1.00, 1.00 1.00, 1.00
0.02 0.04 0.34 0.78 0.44
0.95 0.87 0.75
0.82, 1.10 0.75, 1.01 0.64, 0.88
0.49 0.08 <0.001
0.98 1.03 1.00 1.04
0.77, 1.25 0.89, 1.20 0.88, 1.14 0.88, 1.22
0.88 0.66 0.97 0.65
0.98 1.49
0.88, 1.09 1.06, 2.09
0.72 0.02
M AN U
SC
RI PT
Hazard Ratio (HR) 1.13 1.04 0.72
TE D
Day of Surgery (Day 38+ vs. Day 1-37) Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref=1st Quartile) 2nd Quartile 3rd Quartile 4th Quartile Tumor Location (Ref = RUL) RML RLL LUL LLL Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
SCC – Squamous Cell Carcinoma RCS – Restricted Cubic Spline NSCLC – Non-small Cell Lung Cancer NCDB – National Cancer Data Base CoC – Commission on Cancer AJCC – American Joint Committee on Cancer CDCC – Charlson/Dayo Comorbidity HR – Hazard Ratio CT – Computed Tomography
RI PT
LIST OF ABBREVIATIONS
ACCEPTED MANUSCRIPT
e-Table 1. Independent predictors of mortality following Cox proportional hazards adjustment for patients who underwent lobectomy for cT1 N0 NSCLC including patients who underwent surgery on the day of diagnosis
EP
AC C
0.92, 1.29 0.68, 1.90
0.33 0.62
RI PT
P-value 0.005 <0.001 <0.001
1.05, 1.30 1.28, 1.61
0.004 <0.001
0.57 1.08 0.63 0.55 0.83 1.00
0.38, 0.68, 0.41, 0.29, 0.46, 1.00,
0.88 1.72 0.96 1.04 1.50 1.00
0.01 0.73 0.03 0.07 0.54 <0.001
0.83 0.80 0.83 1.00 1.00
0.71, 0.67, 0.31, 1.00, 1.00,
0.96 0.96 2.24 1.00 1.00
0.02 0.02 0.71 0.68 0.78
0.94 0.88 0.74
0.83, 1.07 0.77, 1.00 0.64, 0.84
0.97 0.99 0.99 0.99
0.79, 0.87, 0.88, 0.86,
1.19 1.12 1.10 1.15
0.74 0.83 0.81 0.92
0.96 1.32
0.88, 1.06 0.97, 1.79
0.47 0.08
M AN U
1.17 1.43
95% CI 1.04, 1.26 1.03, 1.04 0.65, 0.79
SC
1.09 1.14
TE D
Day of Surgery (Day 29+ vs. Day 1-28) Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref=1st Quartile) 2nd Quartile 3rd Quartile 4th Quartile Tumor Location (Ref = RUL) RML RLL LUL LLL Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
Hazard Ratio (HR) 1.14 1.04 0.72
0.37 0.05 <0.001
Online supplements are not copyedited prior to posting and the author(s) take full responsibility for the accuracy of all data.
ACCEPTED MANUSCRIPT
EP
TE D
M AN U
SC
RI PT
e-Figure 1. Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma
AC C
Including Patients who underwent Surgery on Day of Diagnosis Stratified by Early vs Late Timing to Lobectomy. Overall survival is measured from date of diagnosis. Patients who received lobectomy within 28 days after diagnosis were classified as having "early" lobectomy while those who received surgery starting on day 29 were classified as having received "late" lobectomy.
170009
Online supplements are not copyedited prior to posting and the author(s) take full responsibility for the accuracy of all data.
S0012-3692(17)31330-2
DOI:
10.1016/j.chest.2017.07.032
Reference:
CHEST 1239
To appear in:
CHEST
Received Date: 2 January 2017 Revised Date:
15 July 2017
Accepted Date: 25 July 2017
Please cite this article as: Yang CFJ, Wang H, Kumar A, Wang X, Hartwig MG, D'Amico TA, Berry MF, Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma, CHEST (2017), doi: 10.1016/j.chest.2017.07.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Title: Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma
RI PT
Authors: Chi-Fu Jeffrey Yang1, MD, Hanghang Wang1, MD, Arvind Kumar1, BS, Xiaofei Wang2, Ph.D., Matthew G. Hartwig1, MD, MHS, Thomas A. D'Amico1, MD, Mark F. Berry3, MD, MHS Institutions and Affiliations: 1. Department of Surgery, Division of Thoracic Surgery, Duke University Medical Center, 3496 DUMC, Durham, North Carolina.
Department of Biostatistics and Bioinformatics, 2424 Erwin Road Suite 1102, Durham, North Carolina
3.
Department of Cardiothoracic Surgery, Stanford University Medical Center, 300 Pasteur Drive, Falk Building 2nd Floor Stanford, CA 94305-5407
M AN U
SC
2.
Corresponding Author:
Mark Berry. Falk Cardiovascular Research Center, 300 Pasteur Drive, Stanford, CA 94305. Email: [email protected]
TE D
Reprints will not be available from the authors
Sources of Support/Conflicts of Interest: The authors have no conflicts of interest to declare. Running Head: Optimal Timing of Lobectomy
EP
Word Count: 2507
Conflict of Interest: The authors have no conflicts of interest to report.
AC C
Funding: This work was not supported by any outside financial source.
ACCEPTED MANUSCRIPT 2
ABSTRACT Background: Because the relationship between timing of surgery following diagnosis of lung cancer and
RI PT
survival has not been precisely described, guidelines on what constitutes a clinically meaningful delay of resection of early-stage lung cancer do not exist. This study tested the hypothesis that increasing time between diagnosis and lobectomy for stage IA squamous cell carcinoma (SCC) was associated with worse
SC
survival.
Methods: The association between timing of lobectomy and survival for patients with clinical stage IA
M AN U
SCC in the National Cancer Data Base (2006-2011) was assessed using multivariable Cox proportional hazards analysis and restricted cubic spline (RCS) functions.
Results: The 5-year overall survival of 4,984 patients who met study inclusion criteria was 58.3% (95% CI: 56.3%-60.2%). Surgery was performed within 30 days of diagnosis in 1,811 (36%) patients while the
TE D
median time to surgery was 38 days (IQR, 23-58). In multivariable analysis, patients who had surgery 38 days or more after diagnosis had significantly worse 5-year survival than patients who had surgery earlier (Hazard Ratio (HR), 1.13 (95% CI:1.02-1.25); p=0.02). Multivariable RCS analysis demonstrated the
EP
HR associated with time to surgery increased steadily the longer resection was delayed; the threshold time
AC C
associated with statistically significant worse survival was ~90 days or greater.
Conclusion: Longer intervals between diagnosis of early-stage lung SCC and surgery are associated with worse survival. Although factors other than the timing of treatment may contribute to this finding, these results suggest that efforts to minimize delays beyond those needed to perform a complete preoperative evaluation may improve survival.
ACCEPTED MANUSCRIPT 3
INTRODUCTION Patients with early-stage non-small cell lung cancer (NSCLC) may experience delays to appropriate treatment due to difficulty in accessing primary care, wait times to see specialists, and the
RI PT
time needed to complete appropriate staging studies and evaluation.1 The impact of this delay between detection and surgical treatment on outcomes has not been well characterized.2 Studies that have
investigated the association of the time from diagnosis to surgery and survival have shown conflicting
SC
results.3-8
As such, there is no consensus on what constitutes a clinically meaningful delay of treatment. Guidelines from the National Comprehensive Cancer Network, which are based on a consensus opinion,
M AN U
state that treatment delays “should be avoided”9 without providing further specifics while the European Society for Medical Oncology10 and the American College of Chest Physicians11 do not mention any recommendations on the time to initiation of surgical intervention. The purpose of this study was to provide quantitative data on how delays to lobectomy for early-stage NSCLC impact survival and
TE D
improve the level of evidence available to support treatment guidelines and patient management. The specific goal of the study was to test the hypothesis that increasing time between diagnosis and lobectomy for clinical stage IA lung squamous cell carcinoma was associated with worse survival using the National
AC C
METHODS
EP
Cancer Data Base (NCDB).
National Cancer Database
Data from the NCDB, which is a joint project of the American College of Surgeons Commission
on Cancer (CoC) and the American Cancer Society, was used in this study. The NCDB contains data from over 30 million patients and 1,500 cancer centers in the U.S. and contains over 70% of all newly diagnosed cases of cancer in the U.S. annually.12 All clinical staging information was directly recorded in the NCDB using American Joint Committee on Cancer (AJCC) 6th edition13 and 7th edition14 TNM classifications for the years of study inclusion (2006-2011).
ACCEPTED MANUSCRIPT 4
Study design This study was approved by the Duke Medicine Institutional Review Board for Clinical
RI PT
Investigators with project approval number Pro00045337. After isolating patients from a de-identified NCDB participant user file and using the International Classification of Diseases for Oncology, 3rd edition histology and topography codes, all clinical Stage 1A (clinical T1N0M0) NSCLC patients with squamous
SC
cell carcinoma (SCC) histology in the NCDB who were treated with lobectomy as primary therapy
without either induction chemotherapy or induction radiation were selected for analysis. Patient data
M AN U
spanned from the years 2006 to 2011. Clinical diagnosis was determined by a physician. Regarding exclusion criteria, the study was restricted to patients with SCC given the relatively recently more recognized importance of histology on both NSCLC treatment and outcomes; we restricted patients to this histology so that results from the study could be directly applicable to this specific patient population. We did not include patients with adenocarcinoma because adenocarcinoma has been
TE D
categorized into different subtypes of tumors with widely ranging prognosis and include adenocarcinoma in situ and minimally invasive adenocarcinomas that have approximately 100% 5-year disease-free survival.15 We did not include bronchioloalveolar or neuroendocrine tumors because they typically have
EP
better prognoses than other lung cancers, and large cell neuroendocrine patients were excluded because of their potentially worse prognosis.15-17 In our primary analysis, patients who had surgery on the same day
AC C
of diagnosis were excluded to reduce potential confounding given that there may have been inherent but unrecorded differences between patients who had surgery on the same day of diagnosis and patients who had surgery after their diagnosis; in a sensitivity analysis we did include these patients who had surgery on the same day of diagnosis.
Statistical analysis For each patient, the number of days from diagnosis of lung cancer to surgical resection was identified. Evaluation of the impact of timing of surgery was performed using several methods. First,
ACCEPTED MANUSCRIPT 5
surgery was categorized as "early" (1-37 days from time of diagnosis to surgery) or "late" (38+ days). These cutoffs were used because our initial evaluation of the distribution of timing of surgery in the NCDB showed a median time to surgery of 38 days. Furthermore, the choice of this time was felt to be
RI PT
clinically reasonable as the timeframe of 30 days or 1 month had been a previously used threshold in the literature12,18 and by the British Thoracic Society.19 We also performed additional sensitivity analyses, described below, where we used 30 days as a cutoff and included patients who underwent surgery on the
SC
day of diagnosis. In addition, a landmark analysis was performed at 4 months to account for the
differences in survival between the early and late surgery groups that may have been due to differences in
M AN U
perioperative mortality rates. We also identified predictors of delayed surgery to 38 days or more after diagnosis using a logistic regression model that included age, sex, race, Charlson/Deyo comorbidity (CDCC) score, tumor size insurance type, education, facility location, population of area, tumor location, income, distance to facility, hospital volume, biopsy prior to lobectomy and facility type. The primary outcome was overall survival, which was defined from time of diagnosis to time of
TE D
last follow-up or death and estimated using the Kaplan-Meier product limit method. Methods of followup are described previously.20 Differences in survival between the “early” and “late” surgery groups were assessed using the log-rank test. The impact of having surgery “early” or “late” was further evaluated
EP
using a Cox proportional hazard model that included this variable as well as age, sex, race, CDCC score, insurance type, tumor size, facility type, distance from patient’s residence to hospital, median household
AC C
income, tumor location, biopsy prior to lobectomy, and hospital volume. All model covariates were chosen a priori based on clinical relevance. Next, a restricted cubic spline (RCS) analysis with 3 knots (located at the 10th, 50th, and 90th
percentiles) was used to model the relationship between time to surgery and survival.21 Spline regression fits smooth polynomial functions (i.e. first and second derivatives) between pre-defined points (knots) on a graph and joins them in a piecewise manner.22,23 Splines can be utilized to generate non-linear models between a continuous predictor variable (e.g., timing to lobectomy), and an outcome (e.g. hazard ratio), without the loss of statistical power that is frequently associated with the approach of categorization.
ACCEPTED MANUSCRIPT 6
RCS models are a specific type of spline model that smooths the curve while reducing the influence of outliers as compared to polynomial regression. As a result, RCS models minimize residuals for confounding variables, and are effective in determining threshold values of the predictor variable of
RI PT
significance.22,24-27 In this study, a multivariable RCS model was used to characterize the strength and the shape of the association between the time between diagnosis and lobectomy and survival. Primary analysis was conducted using the median value for all continuous variables as reference standard. Because sex is an
SC
important predictor of survival,28,29 we created RCS plots stratified by sex.
M AN U
Statistical analyses were performed using Stata Statistical Software: Release 13.0 (StataCorp LP, College Station, TX) and R software version 3.2.2 (Vienna, Austria). Comparisons of baseline characteristics and unadjusted outcomes were performed using the Wilcoxon Rank Sum test for continuous variables and Pearson’s chi-square test for categorical variables. A 2-sided p-value of 0.05
RESULTS
TE D
was used to define significance.
The study cohort consisted of 4,984 patients who underwent lobectomy for clinical stage IA
EP
(cT1N0) lung SCC at least 1 day after diagnosis of lung cancer (Figure 1). Median time to surgery was 38 days (IQR, 24-60) and 95% of patients underwent surgery within 4 months of diagnosis. Table 1 details
AC C
the preoperative baseline characteristics and peri-operative outcomes stratified by whether their surgery was performed before or after the median delay of 38 days. Table 2 demonstrates predictors of whether surgery occurred before or after the median delay of 38 days. Patients who underwent biopsy prior to lobectomy, older patients and black patients had increased wait times. Patients who underwent a left lower lobectomy were less likely to experience delays when compared to patients undergoing a right upper lobectomy. The 5-year overall survival was 58.3% for the entire cohort (Figure 2a). In univariable analysis, and after a median follow-up of 32.0 (range: 0.3–96.6) months for the entire cohort, patients who received
ACCEPTED MANUSCRIPT 7
surgery within 37 days of diagnosis had better survival when compared to patients who received surgery 38 days or after diagnosis (Figure 2b, 2c). After multivariable adjustment, delayed surgery beyond 37 days from time of diagnosis was found to be significantly associated with worse survival (Hazard ratio
RI PT
(HR) 1.13 (95% CI, 1.02-1.25); p=0.02). In addition, age, gender, comorbidities, tumor size, and type of insurance, facility type, and median income level were all significant predictors of survival (Table 3). Finally, when we repeated the analysis using 30 days as a cutoff between early and late, we found
SC
that the 5-year overall survival for patients who received surgery 30 days or before was significantly better than the survival for patients who received surgery after 30 days (61.5% (95% CI, 58.3%-64.6%)
M AN U
vs. 56.4% (95% CI, 53.9%-58.8%); p<0.01). Multivariable analysis, though, showed that delaying surgery beyond 30 days was not significantly associated with worse survival (HR, 1.11 (95% CI, 1.001.24); p=0.058). In our landmark analysis, patients who underwent early surgery had significantly better 5-year overall survival than patients who underwent late surgery (Figure 2d). Restricted multivariable cubic spline plots for all analyses are shown in Figure 3. Figures 3a and
TE D
3b present the relationship between delay and HR stratified by sex, Figure 3c presents the relationship between delay and HR for patients for the entire cohort. All three plots (Figures 3a-3c) demonstrate that the HR immediately begins to increase with increasing time from diagnosis to lobectomy for each cohort
EP
analyzed. For the entire cohort, the HR did not become statistically significant until approximately 90 days and beyond from time of diagnosis to time of surgery.
AC C
In a sensitivity analysis where we included patients who underwent surgery on the day of diagnosis for a total cohort of 6,723 patients, the results were consistent with those reported above. Median time to surgery was 28.5 days (IQR, 0-50). “Late” surgery, defined as surgery 29 or more days after diagnosis, was associated with worse 5-year survival (e-Figure 1) even after multivariable adjustment (e-Table 1). Furthermore, restricted cubic spline analysis showed that, as in the previous analyses, the HR increases as time between diagnosis and surgery increased (Figure 3d).
DISCUSSION
ACCEPTED MANUSCRIPT 8
This study examined the relationship between timing of lobectomy following diagnosis of lung squamous cell carcinoma and overall survival for clinical stage 1A patients. Our study found that the
RI PT
median delay between diagnosis and surgery in the NCDB was 38 days. Patients whose surgery was 38 or more days after diagnosis were found to have worse survival in both univariate and multivariate
analysis. When we characterized the non-linear relationship between timing of lobectomy and survival using restricted cubic spline analyses, we found that the hazard ratio associated with time between
SC
diagnosis and surgery steadily increased with longer delays and the threshold for when the association between time from diagnosis to surgery and worse survival became statistically significant was
M AN U
approximately 90 days.
Previous studies on delays to treatment for NSCLC differed with respect to the disease stage, the treatment characteristics included, as well as the arbitrarily selected cutoff for the definition of early versus delayed start of therapies. As a result, while some studies have demonstrated that treatment delays are associated with disease progression and worse survival,6-8,30-36 other studies have found no association
TE D
between delayed treatment and long-term outcomes,3-5,37-41 and some studies have shown an association between delayed treatment and better survival.42-45 Importantly, none of these studies specifically evaluated the impact of lobectomy on NSCLC nor have they examined stage IA alone. A strength of this
EP
study is that by using the NCDB to create a large cohort and by only examining the specific situation of lobectomy for clinical stage IA patients with SCC, we were able to limit confounders by not including a
AC C
heterogeneous group of stages, histologies or treatments. Another strength of this study was that through multivariable restricted cubic spline modeling, we characterized the nonlinear relationship between the timing of lobectomy and overall survival. Furthermore, as a population based study, these results can be generalized to the overall population of the United States, one of the strengths of using the NCDB. Our analysis showed several time points that were associated with worse survival in a statistically significant manner, including the median delay of 38 days from the analysis of the entire cohort, as well
ACCEPTED MANUSCRIPT 9
as the time of 90 days from the multivariable RCS analysis. Both analyses were consistent in showing that a longer delay led to worse outcomes. The present study also found several factors associated with delayed surgery. Patients who had a
RI PT
biopsy prior to lobectomy, who were older and black had increased times between diagnosis and surgery. Tumor location was also found to be associated with timing of lobectomy. Patients who underwent a left lower lobectomy were less likely to experience delays when compared to patients undergoing a right
SC
upper lobectomy. Potentially, a technically more straightforward operation (e.g. a left lower lobectomy) might be more quickly undertaken by some surgeons versus others. In addition, certain geographic areas
M AN U
(East south-central, and West North and South Central) were also associated with decreased wait times. This study has several limitations. First, because of the study’s retrospective design, there is a possibility for unobserved confounding and selection bias to exist. Although we used multivariable analysis to reduce bias, there are important covariates such as surgeon experience, and detailed co-morbidity and pulmonary function data that are not available in the NCDB. Lack of detailed co-morbidity and
TE D
pulmonary function data can impact findings and is an important limitation of this study. For example, patients who waited longer for surgery may have been sicker and required more pre-operative evaluation which could have delayed surgery; the “need” for delayed surgery as opposed to the actual delay of
EP
surgery may have been associated with worse outcomes. The NCDB does have co-morbidity scores and we were able to somewhat address this limitation by adjusting for co-morbidity scores in our analysis.
AC C
Second, detailed information regarding the timing of diagnosis was not available. It is unclear whether the diagnosis made for the patients was through radiographic findings, cytology or by biopsy. Given that the time of diagnosis is abstracted from patient charts, and given that the diagnosis of clinical stage IA is typically made through computed tomography (CT),9 we hypothesize that the majority of these patients had their clinical diagnosis made by CT. Third, data regarding complications does not exist in the NCDB. Fourth, our results are not necessarily generalizable to other stages of NSCLC or histology other than SCC. Lastly, cancer-specific, recurrence-free and disease-free survival are not recorded in the NCDB.
ACCEPTED MANUSCRIPT 10
In conclusion, although practical considerations may always require some time period between diagnosis of early-stage NSCLC and treatment, efforts to minimize unnecessary delays are recommended as outcomes are worse when surgical resection is delayed for clinical stage IA squamous cell carcinoma.
RI PT
Future studies should focus on determining the optimal timing of treatment for different stages of NSCLC
AC C
EP
TE D
M AN U
SC
as well as for other histologies.
ACCEPTED MANUSCRIPT 11
ACKNOWLEDGMENTS The data used in this study are derived from a de-identified NCDB file. The American College of Surgeons has executed a Business Associate Agreement that includes a data use agreement with each of
RI PT
its CoC-accredited hospitals. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methodology employed, or the conclusions drawn from these data by the investigators.
SC
C.J.Y. is the guarantor of the paper and takes responsibility for the integrity of the work as a whole. C.J.Y., H.H.W., and A.K. contributed equally in literature search, figures, study design, data
M AN U
collection, data analysis, interpretation and writing, and revisions. X.W. contributed to analyzing and interpreting the data as well as revisions. T.A.D., M.G.H., and M.F.B. contributed to design, data analysis, and interpretation and revisions.
AC C
EP
TE D
This work was not supported by any outside financial source.
ACCEPTED MANUSCRIPT 12
References
7. 8.
9. 10.
11.
12. 13. 14. 15.
16. 17. 18.
19.
20.
21.
RI PT
SC
6.
M AN U
5.
TE D
4.
EP
2. 3.
Evans WK. Prognostic implications of treatment delays in the surgical resection of lung cancer. Thoracic surgery clinics. 2013;23(2):225-232. SEER. Cancer Statistics Factsheets: Lung and Bronchus Cancer. National Cancer Institute. Bethesda, MD, . Aragoneses FG, Moreno N, Leon P, et al. Influence of delays on survival in the surgical treatment of bronchogenic carcinoma. Lung cancer. 2002;36(1):59-63. Quarterman RL, McMillan A, Ratcliffe MB, Block MI. Effect of preoperative delay on prognosis for patients with early stage non-small cell lung cancer. The Journal of thoracic and cardiovascular surgery. 2003;125(1):108-113; discussion 113-104. Shin DW, Cho J, Kim SY, et al. Delay to curative surgery greater than 12 weeks is associated with increased mortality in patients with colorectal and breast cancer but not lung or thyroid cancer. Annals of surgical oncology. 2013;20(8):2468-2476. Gomez DR, Liao KP, Swisher SG, et al. Time to treatment as a quality metric in lung cancer: Staging studies, time to treatment, and patient survival. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2015;115(2):257-263. Kanarek NF, Hooker CM, Mathieu L, et al. Survival after community diagnosis of early-stage non-small cell lung cancer. The American journal of medicine. 2014;127(5):443-449. Samson P, Patel A, Garrett T, et al. Effects of Delayed Surgical Resection on Short-Term and Long-Term Outcomes in Clinical Stage I Non-Small Cell Lung Cancer. The Annals of thoracic surgery. 2015;99(6):1906-1913. Ettinger DS, Wood DE, Akerley W, et al. Non-small cell lung cancer, version 6.2015. Journal of the National Comprehensive Cancer Network : JNCCN. 2015;13(5):515-524. Vansteenkiste J, Crino L, Dooms C, et al. 2nd ESMO Consensus Conference on Lung Cancer: early-stage non-small-cell lung cancer consensus on diagnosis, treatment and follow-up. Ann Oncol. 2014;25(8):14621474. Howington JA, Blum MG, Chang AC, Balekian AA, Murthy SC. Treatment of stage I and II non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(5 Suppl):e278S-313S. Bilimoria KY, Ko CY, Tomlinson JS, et al. Wait times for cancer surgery in the United States: trends and predictors of delays. Annals of surgery. 2011;253(4):779-785. Greene FL, American Joint Committee on Cancer., American Cancer Society. AJCC cancer staging manual. 6th ed. New York: Springer; 2002. Edge SB, American Joint Committee on Cancer. AJCC cancer staging manual. 7th ed. New York: Springer; 2010. Travis WD, Brambilla E, Noguchi M, et al. International association for the study of lung cancer/american thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2011;6(2):244-285. Rekhtman N. Neuroendocrine tumors of the lung: an update. Arch Pathol Lab Med. 2010;134(11):16281638. Bertino EM, Confer PD, Colonna JE, Ross P, Otterson GA. Pulmonary neuroendocrine/carcinoid tumors: a review article. Cancer. 2009;115(19):4434-4441. Devbhandari MP, Bittar MN, Quennell P, et al. Are we achieving the current waiting time targets in lung cancer treatment? Result of a prospective study from a large United kingdom teaching hospital. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2007;2(7):590-592. BTS recommendations to respiratory physicians for organising the care of patients with lung cancer. The Lung Cancer Working Party of the British Thoracic Society Standards of Care Committee. Thorax. 1998;53 Suppl 1:S1-8. Koike T, Koike T, Yoshiya K, Tsuchida M, Toyabe S. Risk factor analysis of locoregional recurrence after sublobar resection in patients with clinical stage IA non-small cell lung cancer. The Journal of thoracic and cardiovascular surgery. 2013;146(2):372-378. Harrell F. Regression Modelling Strategies With Applications to Linear Models, Logistic Regression, and Survival Analysis. . New York, NY: SpringerVerlag; 2001.
AC C
1.
ACCEPTED MANUSCRIPT 13
28.
29.
30. 31.
32.
33.
34. 35. 36.
37. 38.
39.
40.
41. 42.
RI PT
27.
SC
26.
M AN U
25.
TE D
24.
EP
23.
Adam MA, Pura J, Goffredo P, et al. Presence and Number of Lymph Node Metastases Are Associated With Compromised Survival for Patients Younger Than Age 45 Years With Papillary Thyroid Cancer. J Clin Oncol. 2015;33(21):2370-2375. Durrleman S, Simon R. Flexible regression models with cubic splines. Statistics in medicine. 1989;8(5):551-561. Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Statistics in medicine. 2010;29(9):1037-1057. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology. 1995;6(4):356-365. Molinari N, Daures JP, Durand JF. Regression splines for threshold selection in survival data analysis. Statistics in medicine. 2001;20(2):237-247. Saraiya M, Kottiri BJ, Leadbetter S, et al. Total and percent free prostate-specific antigen levels among U.S. men, 2001-2002. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2005;14(9):2178-2182. Travis WD, Brambilla E, Nicholson AG, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2015;10(9):1243-1260. Radzikowska E, Glaz P, Roszkowski K. Lung cancer in women: age, smoking, histology, performance status, stage, initial treatment and survival. Population-based study of 20 561 cases. Ann Oncol. 2002;13(7):1087-1093. Mohammed N, Kestin LL, Grills IS, et al. Rapid disease progression with delay in treatment of non-smallcell lung cancer. International journal of radiation oncology, biology, physics. 2011;79(2):466-472. Everitt S, Herschtal A, Callahan J, et al. High rates of tumor growth and disease progression detected on serial pretreatment fluorodeoxyglucose-positron emission tomography/computed tomography scans in radical radiotherapy candidates with nonsmall cell lung cancer. Cancer. 2010;116(21):5030-5037. Everitt S, Plumridge N, Herschtal A, et al. The impact of time between staging PET/CT and definitive chemo-radiation on target volumes and survival in patients with non-small cell lung cancer. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2013;106(3):288-291. Redaniel MT, Martin RM, Ridd MJ, Wade J, Jeffreys M. Diagnostic intervals and its association with breast, prostate, lung and colorectal cancer survival in England: historical cohort study using the Clinical Practice Research Datalink. PLoS One. 2015;10(5):e0126608. Buccheri G, Ferrigno D. Lung cancer: clinical presentation and specialist referral time. The European respiratory journal. 2004;24(6):898-904. Kanashiki M, Satoh H, Ishikawa H, Yamashita YT, Ohtsuka M, Sekizawa K. Time from finding abnormality on mass-screening to final diagnosis of lung cancer. Oncology reports. 2003;10(3):649-652. Kashiwabara K, Koshi S, Itonaga K, Nakahara O, Tanaka M, Toyonaga M. Outcome in patients with lung cancer found on lung cancer mass screening roentgenograms, but who did not subsequently consult a doctor. Lung cancer. 2003;40(1):67-72. Bozcuk H, Martin C. Does treatment delay affect survival in non-small cell lung cancer? A retrospective analysis from a single UK centre. Lung cancer. 2001;34(2):243-252. Kashiwabara K, Koshi S, Ota K, Tanaka M, Toyonaga M. Outcome in patients with lung cancer found retrospectively to have had evidence of disease on past lung cancer mass screening roentgenograms. Lung cancer. 2002;35(3):237-241. Loh LC, Chan LY, Tan RY, et al. Effect of time delay on survival in patients with non-small cell lung cancer--a Malaysian study. Asia-Pacific journal of public health / Asia-Pacific Academic Consortium for Public Health. 2006;18(1):69-71. Pita-Fernandez S, Montero-Martinez C, Pertega-Diaz S, Verea-Hernando H. Relationship between delayed diagnosis and the degree of invasion and survival in lung cancer. Journal of clinical epidemiology. 2003;56(9):820-825. Salomaa ER, Sallinen S, Hiekkanen H, Liippo K. Delays in the diagnosis and treatment of lung cancer. Chest. 2005;128(4):2282-2288. Annakkaya AN, Arbak P, Balbay O, Bilgin C, Erbas M, Bulut I. Effect of symptom-to-treatment interval on prognosis in lung cancer. Tumori. 2007;93(1):61-67.
AC C
22.
ACCEPTED MANUSCRIPT 14
RI PT SC M AN U TE D
45.
EP
44.
Myrdal G, Lambe M, Hillerdal G, Lamberg K, Agustsson T, Stahle E. Effect of delays on prognosis in patients with non-small cell lung cancer. Thorax. 2004;59(1):45-49. Neal RD, Allgar VL, Ali N, et al. Stage, survival and delays in lung, colorectal, prostate and ovarian cancer: comparison between diagnostic routes. Br J Gen Pract. 2007;57(536):212-219. Nadpara P, Madhavan SS, Tworek C. Guideline-concordant timely lung cancer care and prognosis among elderly patients in the United States: A population-based study. Cancer Epidemiol. 2015;39(6):1136-1144.
AC C
43.
ACCEPTED MANUSCRIPT 15
Figure Legends Figure 1
RI PT
Consort Diagram Showing Schema of Study Subject Selection
Figure 2a
Overall Survival of Clinical Stage 1A NSCLC Patients with Squamous Cell Carcinoma Undergoing
SC
Lobectomy. Represents overall survival of all patients, regardless of the number of days between
M AN U
diagnosis and surgery.
Figure 2b
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Survival is calculated from date of diagnosis. Patients who received lobectomy within 37 days after diagnosis were classified as having "early" lobectomy while those who
Figure 2c
TE D
received surgery starting on day 38 were classified as receiving "late" lobectomy.
EP
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Overall survival is calculated from date of definitive surgery. Patients who
AC C
received lobectomy within 37 days after diagnosis were classified as having "early" lobectomy while those who received surgery starting on day 38 were classified as receiving "late" lobectomy.
Figure 2d
Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma Stratified by Early vs Late Timing to Lobectomy. Overall survival is calculated from four months after date of diagnosis. Patients who received lobectomy within 37 days after diagnosis were classified as having "early"
ACCEPTED MANUSCRIPT 16
lobectomy while those who received surgery starting on day 38 were classified as receiving "late" lobectomy.
RI PT
Figure 3 Relationship between Timing to Lobectomy and Overall Survival for Patients with Clinical Stage IA NSCLC diagnosed with squamous cell carcinoma. Spline plots for adjusted Cox models excluding
SC
patients who underwent surgery on the day of diagnosis for a) men, b) women, c) the entire cohort, and d) the entire cohort including patients who underwent surgery on the day of diagnosis. Median values for all
M AN U
covariates are used as reference standard. Model was adjusted for age, race, Charlson/Deyo comorbidity (CDCC) score, insurance type, tumor size, facility type, distance from patient’s residence to hospital,
AC C
EP
TE D
median household income and hospital volume. Gray shaded area indicates 95% CIs.
ACCEPTED MANUSCRIPT
Table 1. Patient Characteristics and Perioperative Outcomes Stratified by Early (1-37 days) vs Late (38+ days) Surgery
69 (64,75)
EP
AC C
70 (64,76)
1,365 (55.9%) 1,075 (44.1%)
1,395 (54.8%) 1,149 (45.2%)
2,225 (91.2%) 153 (6.3%) 62 (2.5%)
2,266 (89.1%) 224 (8.8%) 54 (2.1%)
P-value 0.006 0.43
0.07
916 (36.0%) 1,033 (40.6%) 595 (23.4%)
M AN U
922 (37.8%) 1,013 (41.5%) 505 (20.7%)
SC
0.005
28 (1.2%) 629 (25.8%) 82 (3.4%) 1,652 (67.7%) 20 (0.8%) 29 (1.2%)
33 (1.3%) 585 (23.0%) 125 (4.9%) 1,744 (68.6%) 36 (1.4%) 21 (0.8%)
224 (9.2%) 1,529 (66.7%) 684 (28.0%) 3 (0.1%)
212 (8.3%) 1,507 (59.2%) 821 (32.3%) 4 (0.2%)
TE D
Patient Age (years), median (IQR) Sex, n (%) Male Female Race, n (%) White Black Other Charlson/Deyo Score, n (%) 0 1 2 Insurance Status, n (%) Uninsured Private Medicare Medicaid Other Government Unknown Facility Type, n (%) Community Cancer Program Comp. Comm. Cancer Program Academic/Research Program Other Diagnostic and Staging Procedure Prior to Lobectomy by Facility type Community Cancer Program No biopsy Biopsy prior to Lobectomy Other/Unknown Comp. Comm. Cancer Program No biopsy Biopsy prior to Lobectomy Other/Unknown Academic/Research Program No biopsy Biopsy prior to Lobectomy Other/Unknown Tumor Location, n (%) RLL LLL RML RUL LUL
Late Surgery (N=2,544)
RI PT
Early Surgery (N=2,440)
0.004
0.01
0.64
88 (39.3%) 133 (59.4%) 3 (1.3%)
74 (34.9%) 135 (63.7%) 3 (1.4%)
598 (39.1%) 890 (58.2%) 41 (2.7%)
553 (36.7%) 931 (61.8%) 23 (1.5%)
378 (55.3%) 302 (44.2%) 4 (0.6%)
359 (43.7%) 458 (55.8%) 4 (0.5%)
428 (17.9%) 354 (14.8%) 137 (5.7%) 838 (35.0%) 639 (26.7%)
483 (20.4%) 321 (12.9%) 121 (4.9%) 844 (33.8%) 726 (29.1%)
0.02
<0.001
0.05
ACCEPTED MANUSCRIPT
0.26
EP
AC C
3 (0.1%) 1922 (76.4%) 1456 (18.1%) 44 (1.8%) 19 (0.8%) 0 (0.0%) 73 (2.9%)
RI PT
1 (0.0%) 1,898 (78.3%) 391 (16.1%) 32 (1.3%) 27 (1.1%) 1 (0.0%) 73 (3.0%)
0.86
1,039 (43.9%) 489 (20.6%) 294 (12.4%) 143 (6.0%) 54 (2.3%) 178 (7.5%) 97 (4.1%) 36 (1.5%) 39 (1.7%)
0.04 0.01 0.62 0.13
552 (22.1%) 671 (26.9%) 701 (28.1%) 571 (22.9%)
M AN U
476 (19.8%) 650 (27.1%) 672 (28.0%) 603 (25.1%)
2,221 (91.6%) 159 (6.6%) 44 (1.8%) 22 (17,28) 10.5 (4.3, 26.8) 42 (21,76)
SC
2,134 (92.0%) 143 (6.2%) 42 (1.8%) 22 (16,27) 11.5 (4.8, 29.8) 42 (22,72)
TE D
Pathologic T stage, n (%) T0 T1 T2 T3 T4 IS Unknown Pathologic N Stage, n (%) N0 N1 N2 Tumor Size (mm), median (IQR) Distance to Facility (mi.), median (IQR) Hospital Volume (# of cases), median (IQR) Median Income, n (%) 1st Quartile 2nd Quartile 3rd Quartile 4th Quartile Population of Area, n (%) >1 million in metro area 250,000 - 1 million in metro area <250,000 in metro area >20,000 near metro area >20,000 not near metro area 2,500 – 19,999 near metro area 2,500 – 19,999 not near metro area <2,500 near metro area <2,500 not near metro area Facility Location, n (%) New England Middle Atlantic South Atlantic East North Central East South Central West North Central West South Central Mountain Pacific Education Quartile, n (%) 1st Quartile 2nd Quartile 3rd Quartile 4th Quartile 30 Day Mortality, n (%) 90 Day Mortality, n (%)
0.03
1,094 (44.6%) 570 (23.3%) 285 (11.6%) 156 (6.4%) 30 (1.2%) 174 (7.1%) 79 (3.2%) 31 (1.3%) 32 (1.3%) <0.001
113 (4.6%) 250 (10.3%) 554 (22.7%) 462 (18.9%) 324 (13.3%) 251 (10.3%) 218 (8.9%) 97 (4.0%) 171 (7.0%)
129 (5.1%) 359 (14.1%) 575 (22.6%) 583 (22.9%) 258 (10.1%) 182 (7.2%) 129 (5.1%) 101 (4.0%) 228 (9.0%)
405 (16.9%) 709 (29.5%) 807 (33.6%) 480 (20.0%) 67 (2.8%) 116 (4.8%)
452 (18.1%) 799 (32.0%) 839 (33.6%) 406 (16.3%) 89 (3.5%) 150 (5.9%)
0.005
0.30 0.15
ACCEPTED MANUSCRIPT
Table 2. Independent predictors of Early (1-37 days) vs. Late (38+ days) Surgery for patients who underwent lobectomy for cT1 N0 NSCLC P-value
1.00, 1.02 0.92, 1.16
0.006 0.58
1.31 1.07
1.04, 1.65 0.55, 2.07
1.06 1.14
0.93, 1.21 0.98, 1.34
0.76 1.21 0.79 1.64 0.60 1.00
0.44, 1.29 0.66, 2.19 0.46, 1.35 0.75, 3.61 0.27, 1.32 1.00, 1.00
1.05 1.19 1.21 1.00 1.00
0.84, 1.30 0.93, 1.52 0.26, 5.57 1.00, 1.00 1.00, 1.00
0.69 0.17 0.81 0.12 0.83
0.94 0.92 0.88
0.78, 1.14 0.75, 1.13 0.69, 1.14
0.52 0.44 0.34
0.85 1.12 1.11 0.88
0.65, 1.12 0.94, 1.32 0.96, 1.29 0.73, 1.06
0.26 0.20 0.17 0.18
1.13 0.94 0.99 0.58 0.93 0.88 0.90 0.94
0.96, 1.32 0.77, 1.15 0.76, 1.28 0.36, 0.94 0.73, 1.19 0.63, 1.23 0.54, 1.50 0.57, 1.56
0.15 0.56 0.92 0.03 0.58 0.46 0.70 0.82
1.19 0.78 0.98 0.63 0.62 0.45 0.85 1.15
0.86, 1.64 0.59, 1.06 0.73, 1.33 0.45, 0.88 0.44, 0.87 0.31, 0.64 0.57, 1.28 0.82, 1.62
0.30 0.12 0.90 0.006 0.005 <0.001 0.45 0.42
EP
AC C
RI PT
95% CI
0.03 0.85
0.37 0.10
0.31 0.54 0.39 0.22 0.21 0.69
SC
M AN U
TE D
Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref<$38,000) $38,000 to $47,999 $48,000 to $62,999 $63,000 or greater Tumor Location (Ref=RUL) RML RLL LUL LLL Population of Area (Ref= >1 million) 250,000 - 1 million in metro area <250,000 in metro area >20,000 near metro area >20,000 not near metro area 2,500 – 19,999 near metro area 2,500 – 19,999 not near metro area <2,500 near metro area <2,500 not near metro area Facility Location (Ref=New England) Middle Atlantic South Atlantic East North Central East South Central West North Central West South Central Mountain Pacific
Adjusted Odds Ratio (AOR) 1.01 1.03
ACCEPTED MANUSCRIPT
0.99 0.87 0.71
0.82, 1.20 0.70, 1.08 0.55, 0.93
0.93 0.20 0.01
1.33 0.89
1.18, 1.51 0.54, 1.47
<0.001 0.65
AC C
EP
TE D
M AN U
SC
RI PT
Education: % without HS diploma (Ref>21%) 13% to 20.9% 7% to 12.9% Less than 7% Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
ACCEPTED MANUSCRIPT
Table 3. Independent predictors of mortality following Cox proportional hazards adjustment for patients who underwent lobectomy for cT1 N0 NSCLC 95% CI 1.02, 1.25 1.03, 1.04 0.65, 0.80
EP
AC C
P-value 0.02 <0.001 <0.001
1.14 1.09
0.94, 1.38 0.70, 1.88
0.18 0.76
1.16 1.42
1.03, 1.31 1.24, 1.62
0.02 <0.001
0.46 0.84 0.55 0.35 0.71 1.00
0.28, 0.74 0.50, 1.42 0.34, 0.88 0.16, 0.76 0.36, 1.40 1.00, 1.00
0.001 0.52 0.01 0.008 0.32 0.002
0.81 0.80 0.38 1.00 1.00
0.68, 0.97 0.65, 0.99 0.05, 2.76 1.00, 1.00 1.00, 1.00
0.02 0.04 0.34 0.78 0.44
0.95 0.87 0.75
0.82, 1.10 0.75, 1.01 0.64, 0.88
0.49 0.08 <0.001
0.98 1.03 1.00 1.04
0.77, 1.25 0.89, 1.20 0.88, 1.14 0.88, 1.22
0.88 0.66 0.97 0.65
0.98 1.49
0.88, 1.09 1.06, 2.09
0.72 0.02
M AN U
SC
RI PT
Hazard Ratio (HR) 1.13 1.04 0.72
TE D
Day of Surgery (Day 38+ vs. Day 1-37) Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref=1st Quartile) 2nd Quartile 3rd Quartile 4th Quartile Tumor Location (Ref = RUL) RML RLL LUL LLL Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
SCC – Squamous Cell Carcinoma RCS – Restricted Cubic Spline NSCLC – Non-small Cell Lung Cancer NCDB – National Cancer Data Base CoC – Commission on Cancer AJCC – American Joint Committee on Cancer CDCC – Charlson/Dayo Comorbidity HR – Hazard Ratio CT – Computed Tomography
RI PT
LIST OF ABBREVIATIONS
ACCEPTED MANUSCRIPT
e-Table 1. Independent predictors of mortality following Cox proportional hazards adjustment for patients who underwent lobectomy for cT1 N0 NSCLC including patients who underwent surgery on the day of diagnosis
EP
AC C
0.92, 1.29 0.68, 1.90
0.33 0.62
RI PT
P-value 0.005 <0.001 <0.001
1.05, 1.30 1.28, 1.61
0.004 <0.001
0.57 1.08 0.63 0.55 0.83 1.00
0.38, 0.68, 0.41, 0.29, 0.46, 1.00,
0.88 1.72 0.96 1.04 1.50 1.00
0.01 0.73 0.03 0.07 0.54 <0.001
0.83 0.80 0.83 1.00 1.00
0.71, 0.67, 0.31, 1.00, 1.00,
0.96 0.96 2.24 1.00 1.00
0.02 0.02 0.71 0.68 0.78
0.94 0.88 0.74
0.83, 1.07 0.77, 1.00 0.64, 0.84
0.97 0.99 0.99 0.99
0.79, 0.87, 0.88, 0.86,
1.19 1.12 1.10 1.15
0.74 0.83 0.81 0.92
0.96 1.32
0.88, 1.06 0.97, 1.79
0.47 0.08
M AN U
1.17 1.43
95% CI 1.04, 1.26 1.03, 1.04 0.65, 0.79
SC
1.09 1.14
TE D
Day of Surgery (Day 29+ vs. Day 1-28) Age (year) Female vs. Male Race (Ref=White) Black Other CDCC Score (Ref=0) 1 2+ Insurance Status (Ref=Uninsured) Private Medicare Medicaid Other Government Unknown Tumor Size Facility Type (Ref=Community) Comprehensive Academic Other Distance to Facility Hospital Volume Median Income (Ref=1st Quartile) 2nd Quartile 3rd Quartile 4th Quartile Tumor Location (Ref = RUL) RML RLL LUL LLL Primary Site Biopsy (Ref = No biopsy) Biopsy prior to Lobectomy Other/Unknown
Hazard Ratio (HR) 1.14 1.04 0.72
0.37 0.05 <0.001
Online supplements are not copyedited prior to posting and the author(s) take full responsibility for the accuracy of all data.
ACCEPTED MANUSCRIPT
EP
TE D
M AN U
SC
RI PT
e-Figure 1. Overall Survival of Clinical Stage IA NSCLC Patients with Squamous Cell Carcinoma
AC C
Including Patients who underwent Surgery on Day of Diagnosis Stratified by Early vs Late Timing to Lobectomy. Overall survival is measured from date of diagnosis. Patients who received lobectomy within 28 days after diagnosis were classified as having "early" lobectomy while those who received surgery starting on day 29 were classified as having received "late" lobectomy.
170009
Online supplements are not copyedited prior to posting and the author(s) take full responsibility for the accuracy of all data.