Wait times between lung cancer diagnosis and surgery: national trends, disparities, and impact on long-term survival

Introduction

Anatomic surgical resection remains the gold standard for early-stage non-small cell lung cancer (NSCLC) treatment. However, clinical practice pathways from diagnosis to surgery are highly variable (1). Long-term survival impact due to treatment delay has previously been studied with conflicting results. Some studies show increased risk and decreased survival (2-7) while others show no association (8-10). For instance, a national cohort study found that patients who underwent surgical treatment for NSCLC more than 12 weeks after diagnosis had reduced survival rates. Additionally, African American patients were more likely than White patients to experience delays in receiving care (2). Another study using data from the National Cancer Database (NCDB) [1995–2010] found that patients with delayed surgery—defined as more than 8 weeks—had higher rates of pathological upstaging at resection, shorter median survival, and increased 30-day mortality (4). As such, the impact of delay between diagnosis and definitive operative intervention for NSCLC survival remains poorly defined. Additionally, time to treatment (TTT) has not been studied as a trend or stratified by race.

Increased lung cancer screening and incidental nodule detection has resulted in an increase in early-stage lesions that require active management and surgical resection (11,12). Additionally, multiple other factors affecting decision-making for patients with early-stage disease have changed within the last decade including increased indications for tumor molecular profiling and neoadjuvant therapy, higher resolution imaging, increased utilization of positron emission tomography and computed tomography (PET-CT), and multidisciplinary tumor boards. It is unclear how these changes have impacted modern thoracic surgery workflow and wait times. Furthermore, it is not known whether these impacts are uniform across race, socioeconomic status and region.

Racial differences in survival of NSCLC are unfortunately apparent. It has been demonstrated that black patients with early-stage NSCLC have lower overall 5-year survival rates compared to white counterparts with similarly staged disease (13,14). While the disparity is partially driven by unequal surgical rates, one additional risk factor may be increased wait time between diagnosis and surgical resection. Again, there is a paucity of current studies evaluating the effect of TTT in at-risk populations.

Advances in imaging, detection, diagnosis and treatment are shaping the treatment pathway for NSCLC. Given these changes, it is important that we investigate racial disparities in access to care and potential impacts on treatment delays. In this analysis, we examine the prevalence of treatment delays and assess trends in time to surgery. We also identify factors associated with delayed treatment and evaluate the impact of these delays on overall survival. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2002/rc).

Methods

Patient population

This is a retrospective analysis of lung cancer patients in the NCDB from 2004 to 2018 (n=1,898,210). The NCDB is a joint program of the American Cancer Society and the American College of Surgeons’ Commission on Cancer. The NCDB is a hospital-based registry that includes cases from Commission on Cancer-accredited cancer programs nationwide. It encompasses 1,500 institutions, accounts for more than 72% of all cancer diagnoses in the United States and is the largest NCDB. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Exclusion criteria included patients with small cell lung cancer, clinical or pathological stage III or IV disease, induction therapy, and non-surgical management. We also excluded patients who waited for more than 124 days from the time of definitive cancer diagnosis to surgical resection, those with missing time from definitive cancer diagnosis to surgical resection, and missing follow-up data (Figure 1). This created a final cohort of 219,723 patients for analysis (Table 1). TTT was calculated by calculating the time from diagnosis to definitive surgical resection. Patients were initially stratified using the time of diagnosis to resection into groups of those who waited less than 4 weeks (n=100,363) and more than 4 weeks (n=119,360). This cutoff was chosen based on clinical significance and previous literature (2,15).

Figure 1 Total study cohort after exclusion criteria stratified based on wait times less than and greater than 4 weeks. NCDB, National Cancer Database; PUF, Participant User File.

Statistical analysis

Baseline patient characteristics were reported as either mean ± standard deviation or median with interquartile range (IQR) for continuous variables depending on overall distribution and proportions for categorical variables. Between-group comparisons were performed using Student’s t-test or Wilcoxon rank-sum test for continuous variables depending on the variable distribution. Pearson’s χ²-test was performed for categorical variables, such as demographic factors. The primary outcome assessed was 10-year overall survival. The independent variables include TTT, patient demographics, and clinical characteristics. Unadjusted survival was analyzed using the Kaplan-Meier method and compared between strata using the log-rank test. Right censoring was performed 10 years after definitive surgical resection, and patients who did not reach these follow-up times were censored on the last follow-up date.

We used propensity score matching to establish balanced treatment groups for survival analysis comparisons. A logistic regression model was fitted with waiting for more than 4 weeks from the time of cancer diagnosis to definitive surgical resection to calculate the propensity score. Covariates included in the model were age sex, race, insurance type, Charlson Deyo score, facility type, multi-facility treatment, clinical and pathological staging, year of diagnosis, and distance between a patient’s residence and the treating hospital. Next, a 1:1 greedy nearest neighbor match was performed using a caliper of 0.01 of the logit of the propensity score computed by this model. Standardized mean differences (instead of P values) were reported after matching to assess covariate balance. A standardized mean difference of 10% or less was deemed to be the ideal balance, while a standardized mean difference of 20% or less was deemed to be an acceptable balance.

To better assess the effect of wait times between cancer diagnosis and definitive surgical resection on long-term mortality, we also constructed a multivariable Cox regression model (Figure 2), in which wait times were treated as a categorical variable with 11 groups (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and >10 weeks). Variables included for adjustment were similar to those used for propensity score matching. The clustering of patients within each center was accounted for using a robust sandwich variance estimator. The proportional hazards assumption was assessed by plotting martingale residuals, which displayed an appropriate degree of scatter, confirming that the assumption was met. Furthermore, the C-index was 0.997, reflecting an excellent model fit. This enabled us to identify that significant increase in mortality was seen starting from 4 weeks and beyond.

Figure 2 A Cox regression model was used to determine the likelihood of mortality based on time between diagnosis and surgical resection. Hazard ratios with 95% confidence intervals are reported in each week.

We also constructed a multivariable logistic regression model to evaluate factors associated with waiting for more than 4 weeks between diagnosis and definitive surgery similar to prior literature (2,16,17). Variables included in this model for were age, race, sex, insurance type, Charlson-Deyo score, year of diagnosis, facility type, multi-facility treatment, staging, and distance between a patient’s residence and the treating hospital. No formal adjustments for multiple hypothesis testing were made, and the analyses should be interpreted as exploratory. All tests were two-tailed with an alpha level of 0.05. All tests were two-tailed with an alpha level of 0.05. All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, North Carolina) and R version 4.3.0 (Vienna, Austria).

Results

Between 2004–2018, 1,898,210 patients were diagnosed with lung cancer while 219,723 met the inclusion criteria (Figure 1). In the unmatched cohort, the median age was 68.5 (IQR, 61–75) years and 100,757 (46.07%) were men (P<0.001). There were 188,717 (85.89%) non-Hispanic White patients, 17,065 (7.77%) non-Hispanic Black patients, and 6,064 (2.76%) Hispanic/Latino patients (P<0.001). There were 10,029 (4.57%) Medicaid patients, 136,144 (61.97%) Medicare patients, 65,666 (29.89%) privately insured patients, and 3,109 (1.41%) uninsured patients (P<0.001). The Charlson-Deyo index was used as a surrogate for co-morbidities with 115,384 (51.51%) patients having a Charlson-Deyo score of 0, 69,9961 (31.84%) with a score of 1, 24,237 (11.03%) with a score of 2, and 10,141 (4.62%) with a score of 3 or greater (P<0.001; Table 1).

The median time between diagnosis and surgery was 30.5 (IQR, 5–50) days and it increased from 25 (IQR, 0–47) days in 2004 to 38 (IQR, 17–59) days in 2018 (P<0.001; Figure 3). Most patients were treated at comprehensive community cancer centers (87,046; 40.50%), while 44,662 (20.32%) were performed at cancer centers, 10,609 (4.84%) at community programs, and 75,829 (44.34%) at academic centers (P<0.001).

Figure 3 Trends in wait times between diagnosis and surgery by year.

After propensity matching, each of the groups were well matched (Table 1). In the matched study cohort, 83,839 patients waited greater than 4 weeks and 83,839 patients waited less than or equal to 4 weeks for surgery. The remaining patients were not included in the final propensity analysis. The median age was 68 (IQR, 61–75) years and 76,647 (45.7%) were men (P=0.60). The number of individuals undergoing diagnosis and resection in the same day decreased over the study period from 28.7% to 19.0% (P<0.001).

A Cox regression model adjusted for age, sex, insurance status, Charlson-Deyo, year, facility, and stage was used to determine the likelihood of mortality based on time between diagnosis and surgical resection. At week 3, the adjusted hazard ratio was 1.03 [95% confidence interval (CI): 0.99–1.07; P<0.05] and at week 4 adjusted hazard ratio was 1.05 (95% CI: 1.01–1.08; P<0.001) showing a transition to a statistically significant increase in risk of mortality after waiting 4 or more weeks between diagnosis and surgical treatment. As such, 4 weeks was selected as the inflection point for further analysis (Figure 2).

Overall survival in the propensity-matched cohort was significantly worse among patients who waited more than 4 weeks for definitive surgery compared to those treated within 4 weeks, when analyzing combined stage I and II NSCLC patients (P<0.001; Figure 4). This trend remained consistent when examining stage I and stage II patients separately (P<0.001; Figures 5,6). Patients treated within 4 weeks had a median survival of 8.09 months (95% CI: 8.02–8.16), whereas those treated after 4 weeks had a median survival of 7.51 months (95% CI: 7.43–7.60). Among stage I patients, median survival was 9.27 months (95% CI: 9.18–9.37) for early treatment compared to 8.03 months (95% CI: 7.94–8.13) for delayed treatment. A similar pattern was observed in stage II patients, where treatment within 4 weeks was associated with a median survival of 5.95 months (95% CI: 5.82–6.13) versus 5.34 months (95% CI: 5.12–5.50) for those treated later.

Figure 4 The overall survival probability comparing propensity matched pairs with >4 or ≤4 weeks wait times. Patients treated ≤4 weeks had a median survival of 8.09 months (95% CI: 8.02–8.16), whereas those treated >4 weeks had a median survival of 7.51 months (95% CI: 7.43–7.60). CI, confidence interval.

Figure 5 The survival probability in stage I NSCLC comparing propensity matched pairs with >4 or ≤4 weeks wait times. Among stage I patients, median survival was 9.27 months (95% CI: 9.18–9.37) for ≤4 weeks compared to 8.03 months (95% CI: 7.94–8.13) for >4 weeks. CI, confidence interval; NSCLC, non-small cell lung cancer.

Figure 6 The survival probability in stage II NSCLC comparing propensity matched pairs with >4 or ≤4 weeks wait times. Patients had a median survival of 5.95 months (95% CI: 5.82–6.13) for ≤4 weeks vs. 5.34 months (95% CI: 5.12–5.50) for >4 weeks. CI, confidence interval; NSCLC, non-small cell lung cancer.

Additionally, significant racial disparities in the interval between diagnosis and surgery were observed (Figure 7). The median wait time for non-Hispanic White patients was 31.0 (IQR, 8–52) days. Non-Hispanic Black patients waited 37.0 (IQR, 9–62) days and had 20% increased odds of longer wait times compared to non-Hispanic White patients [odds ratio (OR) 1.2, 95% CI: 1.23–1.32, P<0.001]. Hispanic/Latino patients waited 32.0 (IQR, 2–55) days and had statistically similar wait times compared to White patients (OR 0.95, 95% CI: 0.90–1.01, P=0.08). These odds ratios were adjusted for age, sex, insurance status, Charlson-Deyo index, year, facility, and stage.

Figure 7 Logistic regression analysis of racial disparities in the interval between diagnosis and surgery adjusted for age, sex, insurance status, Charlson-Deyo, year, facility, and stage. Odds ratio reported with 95% confidence intervals.

Based on the NCDB facility classifications, when compared to academic or research centers, patients treated at community centers have 12% higher odds of waiting greater than 4 weeks from diagnosis to surgery (OR 1.12, 95% CI: 1.07–1.17; P<0.001; Figure 8). Conversely, those treated at cancer centers have decreased odds of waiting greater than 4 weeks (OR 0.96, 95% CI: 0.94–0.99; P<0.001; Figure 8) when compared to academic facilities.

Figure 8 Logistic regression analysis of disparities based on facility in the interval between diagnosis and surgery adjusted for age, sex, insurance status, Charlson-Deyo, year, facility, and stage. Odds ratio reported with 95% confidence intervals. *, P value for all odds ratios was statistically significant.

Discussion

This study is the largest cohort studied to date looking at wait times from diagnosis to definitive surgical treatment for early-stage NSCLC. Wait time increased significantly over the study period from 25 to 38 days. This change is likely multifactorial. More patients are undergoing diagnostic procedures prior to invasive surgical intervention in the form of CT guided biopsy, navigational bronchoscopy with biopsy and/or mediastinal staging. This is supported by the significant decrease in the proportion of patients undergoing diagnosis and treatment in the same day. The benefit of the staggered approach is the reduction of patients with benign lesions undergoing surgical resection, while the drawback is increasing the risk for TTT delay. It has been shown that additional risk stratification and multidisciplinary evaluation increases wait time, and this varies greatly between individual providers (18).

Within the study period, Non-Hispanic Black patients had significantly longer wait times when compared to White or Hispanic patients. The reason for this disparity is not elucidated in this analysis but has been investigated extensively elsewhere (16,19,20). Explanation for disparities in delivery of care for lung cancer include but are not limited to implicit bias, health literacy, access to specialist providers, and lack of trust in the health system (17). Prior literature has demonstrated that Black Americans are 15% less likely to be diagnosed early, 19% less likely to receive surgical treatment and 12% less likely to survive five years compared with White Americans (21). Further study is urgently needed to understand what interventions may be effective in addressing this health care disparity, particularly as the harm of increased wait times from diagnosis to surgery are further understood.

This analysis also revealed that patients treated at community centers had higher odds of experiencing treatment delays compared to those treated at academic centers (OR 1.12, 95% CI: 1.07–1.17; P<0.001). While this finding may be influenced by several factors, it’s important to note that a significant proportion of Non-Hispanic Black and Hispanic/Latino patients were treated at community centers. Specifically, 30.3% of Non-Hispanic Black patients and 31.7% of Hispanic patients in this cohort received care at community centers. This finding is similar to reported literature which has shown that patients in rural areas were less likely to be treated in academic centers and also had overall lower odds of survival than patients (22,23).

Overall survival decreased significantly when wait times were longer than 4 weeks between diagnosis and surgery. This was evident for both stage I and stage II NSCLC together and independently. Retrospective database and single institution studies have already demonstrated the negative effects of long wait times on mortality after lung cancer diagnosis. Furthermore, while single institution studies have more clinical granularity, large database studies can have more statistical power. Despite some conflicting findings, well powered and designed studies uniformly report that survival is affected with 4–8 weeks of delay with early-stage lung cancer (2-7), similar to the findings of this large national investigation. In this context, the increasing TTT over the period demonstrated by this study should be particularly alarming to the thoracic surgical community. Our analysis showed that patients treated greater than 4 weeks from diagnosis was significantly more associated with lower survival. This is different from recently published literature, with TTT ranging from 6 to 12 weeks (2,4,7). This discrepancy emphasizes the need for further research to explore the critical thresholds of delay that truly impact survival outcomes. It also highlights the importance of optimizing care pathways to minimize delays, particularly for early-stage lung cancer patients.

This study has several important limitations. It is a retrospective study and carries an inherent risk of bias. It employs a large database without granular preoperative clinical data and as such propensity matching is limited by the data provided. Notable datapoints that are not collected include preoperative pulmonary function tests, specific comorbid conditions, functional and nutritional status. The NCDB does not collect information about staging method used which may significantly alter time to surgery. Overall survival is used as a surrogate for cancer specific survival as this degree of specificity is not provided by the database. Time to surgical resection is also provided by NCDB; and as such, it is difficult to determine patients that may have had long-term surveillance for suspicious pulmonary nodules, which may affect the findings of this study. Finally, preoperative workup details are not available, and this limits our ability to determine causality for the delays.

Despite these limitations, this analysis adds key evidence to the growing literature on the impact of delays from diagnosis to treatment on patients with early-stage NSCLC. With the number of pulmonary nodules discovered incidentally and via screening continuing to increase nationally (11), there may likely be increased stress on already saturated health care system with at-risk populations being disproportionately affected. Every effort should be made to mitigate the potential deleterious effects of an overburdened system. Delivery of care should be streamlined and TTT must be regarded as a quality metric. A team-based approach reduces wait times and should be adopted if possible.

Conclusions

In conclusion, this study demonstrates that longer wait times from diagnosis to surgery significantly decrease overall survival for early-stage NSCLC patients, particularly when wait times exceed 4 weeks. Racial disparities were observed, with non-Hispanic Black patients experiencing longer wait times, which may reflect broader healthcare access issues. Despite the study’s limitations, these findings emphasize the urgent need to streamline treatment pathways and reduce delays, especially in vulnerable populations. Further research is required to better understand the impact of these delays and to inform strategies to optimize care delivery.

Acknowledgments

This manuscript was recently presented at the AATS conference in Los Angeles, CA on May 8, 2023.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2002/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2002/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2002/coif). H.J.S. and A.R.B. are consultants for Intuitive and receive consulting fees. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. West H, Hu X, Zhang S, et al. Treatment Patterns and Outcomes in Resected Early-stage Non-small Cell Lung Cancer: An Analysis of the SEER-Medicare Data. Clin Lung Cancer 2023;24:260-8. [Crossref] [PubMed]
2. Heiden BT, Eaton DB Jr, Engelhardt KE, et al. Analysis of Delayed Surgical Treatment and Oncologic Outcomes in Clinical Stage I Non-Small Cell Lung Cancer. JAMA Netw Open 2021;4:e2111613. [Crossref] [PubMed]
3. Goldstraw P, Chansky K, Crowley J, et al. The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2016;11:39-51. [Crossref] [PubMed]
4. 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. Ann Thorac Surg 2015;99:1906-12; discussion 1913. [Crossref] [PubMed]
5. Huang CS, Hsu PK, Chen CK, et al. Delayed surgery after histologic or radiologic-diagnosed clinical stage I lung adenocarcinoma. J Thorac Dis 2020;12:615-25. [Crossref] [PubMed]
6. Yang CJ, Wang H, Kumar A, et al. Impact of Timing of Lobectomy on Survival for Clinical Stage IA Lung Squamous Cell Carcinoma. Chest 2017;152:1239-50. [Crossref] [PubMed]
7. Myrdal G, Lambe M, Hillerdal G, et al. Effect of delays on prognosis in patients with non-small cell lung cancer. Thorax 2004;59:45-9.
8. Visser MR, van Lanschot JJ, van der Velden J, et al. Quality of life in newly diagnosed cancer patients waiting for surgery is seriously impaired. J Surg Oncol 2006;93:571-7. [Crossref] [PubMed]
9. Aragoneses FG, Moreno N, Leon P, et al. Influence of delays on survival in the surgical treatment of bronchogenic carcinoma. Lung Cancer 2002;36:59-63. [Crossref] [PubMed]
10. Quarterman RL, McMillan A, Ratcliffe MB, et al. Effect of preoperative delay on prognosis for patients with early stage non-small cell lung cancer. J Thorac Cardiovasc Surg 2003;125:108-13; discussion 113-4. [Crossref] [PubMed]
11. Gould MK, Tang T, Liu IL, et al. Recent Trends in the Identification of Incidental Pulmonary Nodules. Am J Respir Crit Care Med 2015;192:1208-14. [Crossref] [PubMed]
12. Lung Cancer Incidence and Mortality with Extended Follow-up in the National Lung Screening Trial. J Thorac Oncol 2019;14:1732-42. [Crossref] [PubMed]
13. Bach PB, Cramer LD, Warren JL, et al. Racial differences in the treatment of early-stage lung cancer. N Engl J Med 1999;341:1198-205. [Crossref] [PubMed]
14. Wolf A, Alpert N, Tran BV, et al. Persistence of racial disparities in early-stage lung cancer treatment. J Thorac Cardiovasc Surg 2019;157:1670-1679.e4. [Crossref] [PubMed]
15. Mayne NR, Bajaj SS, Powell J, et al. Extended Delay to Treatment for Stage III-IV Non-Small-Cell Lung Cancer and Survival: Balancing Risks During the COVID-19 Pandemic. Clin Lung Cancer 2022;23:e362-e376. [Crossref] [PubMed]
16. Shugarman LR, Mack K, Sorbero ME, et al. Race and sex differences in the receipt of timely and appropriate lung cancer treatment. Med Care 2009;47:774-81. [Crossref] [PubMed]
17. Nadeem MF, Kaiser LR. Disparities in Health Care Delivery Systems. Thorac Surg Clin 2022;32:13-21. [Crossref] [PubMed]
18. Tang A, Ahmad U, Raja S, et al. How Much Delay Matters? How Time to Treatment Impacts Overall Survival in Early Stage Lung Cancer. Ann Surg 2023;277:e941-7. [Crossref] [PubMed]
19. Holmes JA, Chen RC. Racial Disparities in Time From Diagnosis to Treatment for Stage I Non-Small Cell Lung Cancer. JNCI Cancer Spectr 2018;2:pky007. [Crossref] [PubMed]
20. Navuluri N, Morrison S, Green CL, et al. Racial Disparities in Lung Cancer Screening Among Veterans, 2013 to 2021. JAMA Netw Open 2023;6:e2318795. [Crossref] [PubMed]
21. American Lung Association: State of Lung Cancer, 2022. Available online: https://www.lung.org/research/state-of-lung-cancer
22. Logan CD, Feinglass J, Halverson AL, et al. Rural-urban survival disparities for patients with surgically treated lung cancer. J Surg Oncol 2022;126:1341-9. [Crossref] [PubMed]
23. Tesema GA, Stirling RG, Wah W, et al. Geographic variation in delay to surgical treatment among non-small cell lung cancer patients. Lung Cancer 2025;199:108077. [Crossref] [PubMed]