The NCDB Public Benchmark Report was queried for patients with NSCLC between 2004 and 2020. Patients included are described in Supplemental Table 1. Exclusion criteria included histology suggestive of small cell carcinoma or non-NSCLC pulmonary metastases, and diagnosis after 2020. Data did not include individual or identifying information and the study was approved by the NCDB. This study was approved by the university Institutional Review Board.

Patient and tumor information including age at diagnosis, histology, stage, grade, treatment, and months elapsed from diagnosis to date of death/last contact were extracted. Histology was stratified into adenocarcinoma, adenosquamous, neuroendocrine, and other histologies.

Staging in the NCDB was assigned according to the 7^th edition of the American Joint Commission on Cancer (AJCC), acknowledging the change from AJCC 7^th edition to 8^th edition in 2018. Stage was assigned the value of the pathologic stage, or clinical stage if pathologic stage was unavailable. Stage values used were from time of diagnosis, excluding patients with occult or stage 0 disease.

Grade was assigned according to the 7^th edition of the AJCC, acknowledging the change to the 8^th edition staging system in 2018. Grade was defined according to pathologic diagnosis of cancer resemblance to normal tissue. If multiple grades were assigned, the highest grade was used.

Surgery performed was categorized into: wedge resection, segmental resection, lobectomy/bilobectomy, and pneumonectomy. Surgery included only definitive surgery to the primary tumor.

The use of systemic chemotherapy was assessed for use as first-line treatment. Patients were grouped by whether they received systemic chemotherapy or not. Immunotherapy and targeted therapy, described by the NCDB and henceforth as “immunotherapy”, was assessed for use as first-line treatment. Patients were grouped by whether immunotherapy or targeted therapy was administered or not, beginning in 2011 when immunotherapy was FDA-approved.

Overall survival (OS) was assessed using months elapsed from the date of diagnosis to the date of last contact or death, whichever occurred first. Patients diagnosed after 2017 were excluded from analysis to ensure 5-years of follow-up between diagnosis and date of reporting (2023).

Analyses were performed using R software (version 4.4.0). Patients were analyzed in cohorts (Supp.Table 1), excluding patients with missing/unknown data. Descriptive statistics are represented in percentages. Changes in proportions are reported as percentage-point changes. Logistic regression was used to estimate change in percentages by year. A sensitivity analysis fit separate logistic regressions for Stage I diagnosis and grade before and after January 1, 2018 to account for the change in AJCC guidelines. A linear spline regression with one knot at 2014 was used to analyze immunotherapy due to nonlinearity at that point. OS stratified by year was examined using Kaplan-Meier and Cox regressions. P-values <0.05 were considered significant.

2,028,553 patients were identified with NSCLC (Supp.Table 2). Over the study period from 2004-2020, there was a 15% increase in patients with Stage I disease at diagnosis. Over time, patients were more likely to be diagnosed with Stage I disease (OR1.05, 95%CI 1.05-1.05, p<0.001) (Figure 1a) (Supp.Table 3).

The NSCLC cohort was grouped by adenocarcinoma, adenosquamous, neuroendocrine, squamous, and other histologies. Of patients with adenocarcinoma, there was a 14% increase in Stage I disease, and over time patients were more likely to be diagnosed with Stage I disease (OR1.03, 95%CI 1.03-1.04, p<0.001) (Figure 1b).

Of adenosquamous patients, there was a 2% increase in patients with Stage I disease at diagnosis. Over the study period, there was no significant change in stage (OR1.00, 95%CI 0.99-1.00, p=0.52) (Figure 1c). Incorporating the change in AJCC staging system in 2018, patients with adenosquamous histology were less likely to be diagnosed with Stage I disease prior to 2018 (OR0.99, 95%CI 0.98-1.00, p=0.018), and there was no significant change in Stage I after 2018 (OR0.97, 95%CI 0.90-1.05, p=0.41).

Of patients with squamous histology, there was a 8% increase in patients with Stage I disease at diagnosis. Over the study period, patients were more likely to be diagnosed with Stage I (OR1.02, 95%CI 1.02-1.02, p<0.001) (Figure 1d).

In patients with neuroendocrine and other histologies, there was a 37% and 16% increase in patients with Stage I disease, respectively. Over time, patients in both cohorts were more likely to be diagnosed with Stage I (Neuroendocrine: OR1.12, 95%CI 1.11-1.12, p<0.001; Other: OR1.12, 95%CI 1.11-1.12, p<0.001) (Figure 1e-f).

We next assessed the grade at diagnosis over time. Of 1,152,308 patients with NSCLC, there was a 17% increase in well-differentiated or moderately-differentiated histology at diagnosis. Over time, patients were more likely to have well-differentiated histology (OR1.04, 95%CI 1.04-1.04, p<0.001) (Figure 2a) (Supp.Table 4).

Of patients with adenocarcinoma, there was a 6% increase in patients with well- or moderately-differentiated histology at diagnosis, and over time patients were more likely to have well-differentiated histology (OR1.01, 95%CI 1.01-1.01, p<0.001) (Figure 2b).

Of patients with adenosquamous histology, there was a 4% increase in patients with well- or moderately-differentiated histology at diagnosis, but there was no significant change in grade (OR1.00, 95%CI 0.98-1.03, p=0.77) (Figure 2c). When accounting for the change in AJCC staging in 2018, there was still no significant change in grade before 2018 (OR1.01, 95%CI 0.98-1.04, p=0.57), but after 2018 patients were significantly less likely to be diagnosed with well-differentiated cancer (OR0.77, 95%CI 0.59-0.99, p=0.044).

Of patients with squamous histology, there was a 6% increase with well- or moderately-differentiated histology at diagnosis, but over time patients were less likely to be found to have well-differentiated histology (OR0.99, 95%CI 0.99-0.99, p<0.001) (Figure 2d). When accounting for the change in AJCC staging, there was a decrease in the likelihood of well-differentiated histology prior to 2018 (OR0.99, 95%CI 0.98-0.99, p<0.001), but no significant change after 2018 (OR1.02, 95%CI 0.96-1.08, p=0.50).

Of patients with neuroendocrine histology, there was a 57% increase in patients with well- or moderately-differentiated histology, and over time patients were more likely to be found to have well-differentiated histology (OR1.18, 95%CI 1.18-1.19, p<0.001) (Figure 2e).

Of patients with other histologies, there was a 3% decrease in patients with well- or moderately-differentiated histology, but patients were more likely to be found to have well-differentiated histology (OR1.09, 95%CI 1.08-1.10, p<0.001) (Figure 2e). When accounting for the change in AJCC staging, patients were more likely to have well-differentiated histology prior to 2018 (OR1.14, 95%CI 1.13-1.14, p<0.001), and patients were less likely to have well-differentiated grade after 2018 (OR0.80, 95%CI 0.66-0.96, p=0.015).

The cohort was evaluated for systemic chemotherapy use as first-line treatment. Of 1,980,499 NSCLC patients, there was a 6% decrease in systemic treatment, and patients were less likely to receive systemic treatment over time (OR0.98, 95%CI 0.98-0.98, p<0.001) (Figure 3a) (Supp.Table 5).

Of patients with adenocarcinoma, there was a 4% decrease in use of systemic treatment, and patients were less likely to receive systemic treatment over time (OR0.98, 95%CI 0.98-0.99, p<0.001) (Figure 3b).

Of patients with adenosquamous histology, there was a 3% increase in systemic treatment; however, there was no change in likelihood of use of systemic treatment (OR1.00, 95%CI 1.00-1.01, p=0.78) (Figure 3c).

Of patients with squamous histology, there was a 1% decrease in use of systemic treatment, and patients were significantly less likely to receive systemic treatment (OR0.99, 95%CI 0.99-0.99, p<0.001) (Figure 3d).

Of patients with neuroendocrine and other histologies, there was a 16% and 13% decrease in use of systemic treatment, respectively. Patients were less likely to receive systemic treatment over time (Neuroendocrine: OR0.95, 95%CI 0.95-0.95, p<0.001; Other: OR0.96, 95%CI 0.96-0.96, p<0.001) (Figure 3e-f).

The cohort was evaluated to assess use of immunotherapy. Of 1,265,149 NSCLC patients, there was a 24.8% increase in immunotherapy treatment from 2011-2020. Patients were more likely to receive immunotherapy over time (OR1.55, 95%CI 1.55-1.56, p<0.001) (Figure 4a) (Supp.Table 6). Although usage increased rapidly from 2011 to 2014 (OR 5.20), it increased but more slowly from 2014 to 2020 (OR1.50) (Supp.Fig.1a).

Of patients with adenocarcinoma, there was a 24.8% increase in use of immunotherapy. Patients were more likely to receive immunotherapy over time (OR1.47, 95%CI 1.46-1.47, p<0.001) (Figure 4b). Although usage increased rapidly from 2011 to 2014 (OR 5.69), it increased but more slowly from 2014 to 2020 (OR 1.40) (Supp.Fig.1b).

Of patients with adenosquamous histology, there was a 24% increase in use of immunotherapy over the study period. Patients were more likely to receive immunotherapy over time (OR1.63, 95%CI 1.59-1.68 p<0.001) (Figure 4c). Although usage increased rapidly from 2011 to 2014 (OR 28.0), it increased but at a slower rate from 2014 to 2020 (OR1.55) (Supp.Fig.1c).

Patients with squamous histology had a 25.9% increase in use of immunotherapy over time. Patients were more likely to receive immunotherapy over time (OR1.81, 95%CI 1.80-1.83, p<0.001) (Figure 4d). Although usage increased rapidly from 2011 to 2014 (OR 3.03), it increased but more slowly from 2014 to 2020 (OR1.76) (Supp.Fig.1d).

Of patients with neuroendocrine histology and other histologies, there was a 9.4% and 24.8% increase in immunotherapy over the study period, respectively. Patients were more likely to receive immunotherapy (Neuroendocrine: OR1.55, 95%CI 1.51-1.59, p<0.001; Other histologies: OR1.62, 95%CI 1.60-1.63, p<0.001) (Figure 4e-f). Although usage increased rapidly for these patients from 2011 to 2014 (Neuroendocrine: OR2.31; Other: OR4.94), it increased but more slowly from 2014 to 2020 (Neuroendocrine: OR1.55; Other: OR1.56) (Supp.Fig.1e-f).

We assessed surgery performed over the study period. Of 596,409 NSCLC patients, there was a decrease in likelihood of receiving a pneumonectomy (OR0.90, 95%CI 0.90-0.91, p<0.001), but segmentectomies and wedge resections were only shown to increase by 4.4%, although the study was not powered to asses the statistical significance (Figure 5a) (Supp.Table 7).

Of adenocarcinoma patients, there was a decrease in surgery performed overall by 6%. There was a decrease in likelihood of receiving a pneumonectomy (OR0.90, 95%CI 0.90-0.90, p<0.001), but segmentectomies and wedge resections only increased by 5.1% over the study period (Figure 5b).

Of patients with adenosquamous histology, there was a decrease in likelihood of receiving a pneumonectomy (OR0.91, 95%CI 0.90-0.92, p<0.001) over the study period (Figure 5c), and the proportion of segmentectomies and wedge resections increased by 3.4%.

Of patients with squamous histology, there was a decrease in likelihood of receiving a pneumonectomy (OR0.91, 95%CI 0.91-0.91, p<0.001) (Figure 5d). This group had a decrease in segmentectomies and wedge resections by 4.5%.

Of patients with neuroendocrine and other histologies, there was a decrease in likelihood of receiving a pneumonectomy (Neuroendocrine: OR0.91, 95%CI 0.91-0.92, p<0.001; Other: OR0.91, 95%CI 0.90-0.91, p<0.001). Segmentectomies and wedge resections increased by 2.4% for neuroendocrine histology and 2.9% for other histologies (Figure 5e-f).

A 5-year survival analysis was performed to evaluate observed survival as an unadjusted temporal association. Cox regression was performed from 2011 through 2017. For the overall cohort, each subsequent year was associated with decreased risk of mortality (HR0.97/year, 95%CI 0.97-0.97, p<0.001) (Figure 6a) (Supp.Table 8). The estimated probability of survival for year of diagnosis increased over time (Figure 7a).

Of adenocarcinoma patients, each year was associated with decreased risk of mortality (HR0.98/year, 95%CI 0.98-0.98, p<0.001) (Figure 6b). Additionally, the estimated probability of survival for year of diagnosis increased over time (Figure 7b).

Of patients with adenosquamous histology, survival appeared consistent; however, later years of diagnosis were associated with slightly increased risk of mortality (HR1.01/year, 95%CI 1.00-1.01, p<0.001) (Figure 6c). The estimated probability of survival for year of diagnosis did not show a significant trend (Figure 7c).

Of patients with squamous histology, each subsequent year was associated with decreased risk of mortality (HR0.99/year, 95%CI 0.99-0.99, p<0.001) (Figure 6d). The probability of survival for year of diagnosis increased over time (Figure 7d).

Of patients with neuroendocrine and other histologies, each year was associated with decreased risk of mortality (Neuroendocrine: HR0.95/year, 95%CI 0.95-0.95, p<0.001; Other: HR0.95/year, 95%CI 0.95-0.95, p<0.001) (Figure 6e-f). In both groups the probability of survival for year of diagnosis increased over time (Figure 7e-f).