Abstract
Background
Recent studies have suggested that more frequent postoperative surveillance imaging via computed tomography following lung cancer resection may not improve outcomes. We sought to validate these findings using a uniquely compiled dataset from the Veterans Health Administration, the largest integrated health-care system in the United States.
Methods
We performed a retrospective cohort study of veterans with pathologic stage I non-small cell lung cancer receiving surgery (2006-2016). We assessed the relationship between surveillance frequency (chest computed tomography scans within 2 years after surgery) and recurrence-free survival and overall survival.
Results
Among 6171 patients, 3047 (49.4%) and 3124 (50.6%) underwent low-frequency (<2 scans per year; every 6-12 months) and high-frequency (≥2 scans per year; every 3-6 months) surveillance, respectively. Factors associated with high-frequency surveillance included being a former smoker (vs current; adjusted odds ratio [aOR] = 1.18, 95% confidence interval [CI] = 1.05 to 1.33), receiving a wedge resection (vs lobectomy; aOR = 1.21, 95% CI = 1.05 to 1.39), and having follow-up with an oncologist (aOR = 1.58, 95% CI = 1.42 to 1.77), whereas African American race was associated with low-frequency surveillance (vs White race; aOR = 0.64, 95% CI = 0.54 to 0.75). With a median (interquartile range) follow-up of 7.3 (3.4-12.5) years, recurrence was detected in 1360 (22.0%) patients. High-frequency surveillance was not associated with longer recurrence-free survival (adjusted hazard ratio = 0.93, 95% CI = 0.83 to 1.04, P = .22) or overall survival (adjusted hazard ratio = 1.04, 95% CI = 0.96 to 1.12, P = .35).
Conclusions
We found that high-frequency surveillance does not improve outcomes in surgically treated stage I non-small cell lung cancer. Future lung cancer treatment guidelines should consider less frequent surveillance imaging in patients with stage I disease.
Surgical resection remains the standard treatment for early-stage non-small cell lung cancer (NSCLC), the leading cause of cancer-related mortality in the United States (1,2). Early-stage lung cancer is unique compared with several other malignancies in that, even after curative-intent resection, recurrence is common, occurring in 20% to 50% of patients within 5 years (3-5). Therefore, surveillance (ie, regimented follow-up imaging in asymptomatic patients) is a routine component of postoperative care, with the hope that earlier detection will allow for more aggressive treatments and better outcomes among patients who recur (6). In addition to detecting recurrence, surveillance imaging can also monitor for new primary malignancies, screen for treatment side effects, and alleviate patient anxiety about recurrence (7). However, these benefits must be weighed against the potential harms of surveilling too frequently [ie, “scanxiety” (8), false positives, unwarranted procedures, considerable health-care costs, etc] (7).
Current guidelines recommend cross-sectional, computed tomography (CT) imaging for lung cancer surveillance (7). Indeed, the long-awaited IFCT-0302 recently demonstrated that surveillance CT was superior to chest x-ray for detecting cancer recurrence and second primary malignancies, though overall survival was similar between the groups (8). However, guidelines conflict on the optimal frequency of CT surveillance, with most suggesting 6-month intervals between scans (3,9-12). This frequency is typically maintained for 2 to 3 years after resection, during which the risk of recurrence is highest; beyond that period, annual surveillance imaging is adopted given the continued risk of second primary cancers (7). However, some recent studies have suggested that more frequent surveillance imaging after cancer surgery may not improve outcomes. For example, in lung cancer, a recent analysis demonstrated that more frequent surveillance (based on the length of time between surgery and the initial CT scan after surgery) was not associated with improved outcomes (13,14). These findings, although important, were limited by the age of the dataset, nonuniform health insurance coverage (which can affect access to care), stage heterogeneity (I-III), and the absence of cause-specific survival models (13,14). In colorectal cancer, both prospective (15) and retrospective (16) studies similarly have found that more frequent surveillance (either via imaging or biochemical testing) does not improve outcomes. These findings have not been validated in a larger, modern cohort.
The Veterans Health Administration (VHA) is the largest integrated health-care system in the United States (17). As such, the VHA provides veterans with universal access to regular follow-up at little or no cost (18,19), hence mitigating several of the methodological limitations and biases that other datasets have in terms of analyzing surveillance strategies (20).
In this study, we sought to examine the association between surveillance frequency and oncologic outcomes in pathologic stage I NSCLC following surgical treatment. We hypothesized that more frequent surveillance was not associated with improved recurrence-free survival or overall survival.
Methods
Study Design and Population
We performed this retrospective cohort study using the VHA Informatics and Computing Infrastructure system, which is a repository for multiple administrative and clinical data sources within the Corporate Data Warehouse (CDW) (21). Although the VHA typically maintains highly accurate data sources, we assembled a team of dedicated researchers (including 2 data coordinators, 1 data analyst, 2 biostatisticians, and 3 physicians) to procure some additional data elements (through natural language processing techniques and chart review) and to verify the accuracy of existing data elements, which spanned a period in excess of 20 months. The St. Louis VHA Research and Development Committee and Institutional Review Board granted a waiver for consent given the deidentified nature of the analysis. Data were reported according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.
All adults with pathologic stage I (American Joint Committee on Cancer 7th edition) NSCLC (International Classification of Diseases for Oncology [ICD-O], 3rd ed.) who received surgical treatment in the VHA between 2006 and 2016 were included in the study. Surgical treatment was confirmed using ICD 9/10 procedure and Current Procedural Terminology codes. Exclusion criteria were patients who died within 90 days of surgery, did not have documented CT scans within the VHA during follow-up, or had symptomatic recurrences (ie, recurrence before CT confirmation imaging) (7).
Surveillance Frequency
To determine surveillance frequency, we assessed the number of chest CT scans that patients had within 2 years after surgical treatment. We excluded CT scans that occurred within 60 days of surgery because we thought that these were likely assessing postoperative complications as opposed to cancer surveillance. To better determine the true holistic frequency of surveillance scans during the first 2 years of follow-up and to reduce potential selection bias (13), we calculated the density of CT scans within that 2-year period. This was performed by dividing the number of CT scans by the length of follow-up (ie, an individual with 3 CT scans over 2 years of follow-up would have a CT density of 1.5 scans per year). For patients who recurred or died within 2 years of surgery, the number of CT scans between 60 days after surgery and the date of recurrence or death was used; this value was subsequently divided by the corresponding follow-up period before recurrence or death (ie, an individual with 3 CT scans over 1.5 years of follow-up before recurrence would have a CT density of 2.0 scans per year). We excluded scans that occurred after the date of recurrence because such imaging likely reflects symptomatic recurrence and/or additional staging scans for salvage therapy.
After calculating CT density, we grouped patients into high- and low-frequency surveillance groups based on contemporary guidelines (3,9-12). High-frequency surveillance was defined as patients with densities of 2 or more scans per year (ie, patients receiving scans every 3-6 months). Low-frequency surveillance was defined as patients with densities less than 2 scans per year (ie, patients receiving scans every 6-12 months). Patients were excluded if they had more than 4 scans per year (density > 4, likely reflecting multiple scans due to unrelated illness) or only 1 scan over the 2-year period (density = 0.5, given survival bias in the low-frequency group; Figure 1).
Figure 1.
Study consort diagram.
Covariates
We extracted several demographic, treatment, and tumor-related variables from CDW, as previously described by our group (5,17). Race was defined according to the CDW Oncology file and Facility Oncology Registry Data Standards manual. Additionally, because variations in surveillance strategies may exist between medical specialties (7,22), we assessed whether patients received follow-up with a surgeon or an oncologist.
Outcomes
Our primary outcomes were overall survival and recurrence-free survival, including time to recurrence. As described previously in VHA literature and by our group (5,23), recurrence was defined using the CDW Oncology database and supplemented with additional diagnoses suggestive of recurrence using ICD 9/10 codes. In particular, patients with any of the following were also deemed to have cancer recurrence: additional chemotherapy or radiation therapy (in the absence of another cancer diagnosis), additional lung resection (in the absence of another cancer diagnosis), malignant pleural effusion, a secondary diagnosis of metastatic lung cancer, or lung biopsy (that was followed by additional chemotherapy or radiation therapy). Overall survival was assessed using the VA Vital Status File (24). Patients were censored at the end of the study follow-up (May 1, 2020). Given potential biases in high- vs low-frequency surveillance, an additional analysis using propensity score weighting was used (Supplementary Methods, available online).
We also performed several stratified analyses to assess if potential benefits of high-frequency surveillance manifested in individuals with higher risk of recurrence. These were performed based on tumor size (<3 cm vs ≥3 cm), tumor grade (I vs II-IV), extent of resection (lobectomy vs sublobar resection), and the number of lymph nodes sampled [<10 vs ≥10 (25-27)].
Statistical Analysis
Study cohort statistics were presented comparing the high- and low-frequency groups using t tests for continuous variables and χ2 tests for categorical variables. Kruskal-Wallis tests were used for nonnormally distributed continuous variables. Factors associated with receiving high- vs low-frequency surveillance were assessed in multivariable logistic regression models. Overall survival was assessed using multivariable Cox proportional hazards models and displayed using Kaplan-Meier direct-adjusted survival curves. Noninferiority for overall survival was assessed using a 5% noninferiority margin (ie, high-frequency surveillance did not improve survival by 5% compared with low-frequency surveillance) (28,29). Recurrence-free survival was assessed using a multivariable competing risk model (Fine and Gray subdistribution hazard function) with recurrence as the outcome and death as a competing event and displayed using adjusted cumulative incidence plots. Missing data were reported in the univariate analyses, and complete case analyses were used for the multivariable models (except where missing categories were specified). P values less than .05 were considered statistically significant and were 2-tailed. All analyses were performed using SAS version 9.3 (SAS Inc, Cary, NC, USA).
Results
Study cohort characteristics
We identified 6171 veterans with pathologic stage I NSCLC undergoing surgical treatment who met inclusion criteria (Figure 1). Among these patients, the mean (SD) age was 67.5 (7.8) years, 5945 (96.3%) were male, 226 (3.7%) were female, and 3808 (61.7%) were currently smoking at the time of cancer diagnosis. Surgical delay was experienced by 1825 (29.6%) patients, with a median (interquartile range [IQR]) wait time between cancer diagnosis and surgery of 61 (40-92) days. Most veterans received a lobectomy (4348 patients [70.5%]), and the most common surgical approach was a thoracotomy (3445 patients [55.9%]). Histology most commonly showed adenocarcinomas (3373 patients [54.7%]) and higher-grade tumors (grade II, 3147 patients [54.0%]; grade III, 1749 patients [30.0%]; grade IV, 78 patients [1.3%]). Additional characteristics of the study cohort are presented in Table 1.
Table 1.
Demographic, treatment-, and tumor-related characteristics of Veterans with stage I NSCLC undergoing surgical treatment
| Characteristics | Full population (N = 6171) | Surveillance frequency |
||
|---|---|---|---|---|
| Low (n = 3047) | High (n = 3124) | P | ||
| Demographics | ||||
| Mean age (SD), y | 67.5 (7.8) | 67.3 (8.0) | 67.7 (7.5) | .03 |
| Sex, No. (%) | .32 | |||
| Male | 5945 (96.3) | 2928 (96.1) | 3017 (96.6) | |
| Female | 226 (3.7) | 119 (3.9) | 107 (3.4) | |
| Race, No. (%)a | <.001 | |||
| White | 5116 (82.9) | 2445 (80.2) | 2671 (85.5) | |
| Black | 1046 (14.7) | 521 (17.1) | 386 (12.4) | |
| Other | 80 (1.3) | 42 (1.4) | 38 (1.2) | |
| Unknown | 68 (1.1) | 39 (1.3) | 29 (0.9) | |
| Smoking status at time of diagnosis, No. (%) | .001 | |||
| Current | 3808 (61.7) | 1949 (64.0) | 1859 (59.5) | |
| Former | 2276 (36.9) | 1056 (34.7) | 1220 (39.1) | |
| Never | 87 (1.4) | 42 (1.4) | 45 (1.4) | |
| Mean BMI, kg/m2 (SD)b | 27.3 (5.3) | 27.2 (5.3) | 27.3 (5.3) | .29 |
| Area Deprivation Index, Mean (SD)b | .42 | |||
| Q1 | 1537 (25.0) | 779 (25.7) | 758 (24.3) | |
| Q2 | 1536 (25.0) | 758 (25.0) | 778 (25.0) | |
| Q3 | 1539 (25.0) | 734 (24.2) | 805 (25.9) | |
| Q4 | 1535 (25.0) | 762 (25.1) | 773 (24.8) | |
| Charlson/Deyo score, median (IQR) | 6.9 (2.2) | 6.8 (2.2) | 6.9 (2.2) | .02 |
| Treatment characteristics | ||||
| Wait time to surgery | ||||
| Median (IQR), d | 61 (40,92) | 63 (41,95) | 59 (39,90) | .003 |
| >12 wk, No. (%) | 1825 (29.6) | 958 (31.4) | 867 (27.8) | .002 |
| Tumor size, No. (%) | .51 | |||
| ≤10 mm | 603 (9.8) | 299 (9.8) | 304 (9.7) | |
| 11-20 mm | 2591 (42.0) | 1307 (42.9) | 1284 (41.1) | |
| 21-30 mm | 1606 (26.0) | 764 (25.1) | 842 (27.0) | |
| 31-40 mm | 861 (14.0) | 434 (14.2) | 427 (13.7) | |
| 40+ mm | 358 (5.8) | 171 (5.6) | 187 (6.0) | |
| Unknown | 152 (2.5) | 72 (2.4) | 80 (2.6) | |
| Resection, No. (%)b | .03 | |||
| Lobectomy | 4348 (70.5) | 2173 (71.4) | 2175 (69.6) | |
| Wedge | 1405 (22.8) | 652 (21.4) | 753 (24.1) | |
| Segment | 368 (6.0) | 190 (6.2) | 178 (5.7) | |
| Pneumonectomy | 47 (0.8) | 29 (1.0) | 18 (0.6) | |
| Surgical approach, No. (%)b | .03 | |||
| VATS | 2716 (44.1) | 1298 (42.7) | 1418 (45.5) | |
| Thoracotomy | 3445 (55.9) | 1744 (57.3) | 1701 (54.5) | |
| Histology, No. (%)b | .04 | |||
| Adenocarcinoma | 3373 (54.7) | 1712 (56.2) | 1661 (53.2) | |
| Squamous cell | 2044 (33.1) | 986 (32.4) | 1058 (33.9) | |
| Other | 752 (12.2) | 348 (11.4) | 404 (12.9) | |
| Grade, No. (%)b | .64 | |||
| I | 859 (14.7) | 431 (14.9) | 428 (14.5) | |
| II | 3147 (54.0) | 1566 (54.3) | 1581 (53.6) | |
| III | 1749 (30.0) | 853 (29.6) | 896 (30.4) | |
| IV | 78 (1.3) | 34 (1.2) | 44 (1.5) | |
| Postoperative outcomes | ||||
| Major complications, No. (%) | 711 (11.5) | 340 (11.2) | 371 (11.8) | .38 |
| Pneumonia | 302 (4.9) | 153 (5.0) | 149 (4.8) | .65 |
| MI | 59 (1.0) | 21 (0.7) | 38 (1.2) | .03 |
| Empyema | 44 (0.7) | 21 (0.7) | 23 (0.7) | .83 |
| Renal failure | 65 (1.1) | 28 (0.9) | 37 (1.2) | .31 |
| Respiratory/cardiac failure | 401 (6.5) | 197 (6.5) | 204 (6.5) | .92 |
| Stroke | 17 (0.3) | 9 (0.3) | 8 (0.3) | .77 |
| 30-day readmission, No. (%) | 393 (6.4) | 195 (6.4) | 198 (6.3) | .92 |
Defined according to the American College of Surgery Facility Oncology Registry Data Standards manual. This manual documents the “other” category, but it is not defined further. ADI = Area Deprivation Index; BMI = body mass index; IQR = interquartile range; MI = myocardial infarction; NSCLC = non-small cell lung cancer; VATS = video-assisted thoracoscopic surgery.
Data were missing for BMI (287 patients), ADI (24 patients), extent of resection (3 patients), surgical approach (10 patients), histology (2 patients), and grade (338 patients).
Surveillance frequency
The median number of CT scans during follow-up was 3 (IQR = 2-4) (see Supplementary Figure 1, available online), corresponding to a median (IQR) CT density of 2 (1.5-2.4) scans per year. Based on these densities, 3047 (49.4%) patients were designated as receiving low-frequency surveillance (ie, <2 scans per year) and 3124 (50.6%) were designated as receiving high-frequency surveillance (ie, ≥2 scans per year). Factors associated with higher likelihood of receiving high-frequency surveillance included being a former smoker (vs current, adjusted odds ratio [aOR] = 1.18, 95% CI = 1.05 to 1.33; Supplementary Table 1, available online), receiving a wedge resection (vs lobectomy, aOR = 1.21, 95% CI = 1.05 to 1.39), and having follow-up with an oncologist (aOR = 1.43, 95% CI = 1.28 to 1.60); factors associated with lower likelihood of high-frequency surveillance included African American race (vs White race, aOR = 0.64, 95% CI = 0.55 to 0.75). Of note, several tumor- and treatment-related variables were unassociated with surveillance frequency, including tumor size, tumor grade, histology (ie, adenocarcinoma vs squamous cell carcinoma), positive margin, and adequacy of lymph node collection.
Recurrence-free survival
With a median (IQR) follow-up of 7.4 (3.4-12.5) years, recurrence was observed in 1360 (22.0%) patients. The median (IQR) time to recurrence was 21.3 (14.0-37.8) months in the low-frequency groups and 23.7 (13.4-41.1) months in the high-frequency group (P = .046; Table 2; Figure 2). Using multivariable competing risk models, high-frequency surveillance was not associated with improved recurrence-free survival after controlling for tumor- and treatment-related factors (adjusted hazard ratio = 0.93, 95% CI = 0.83 to 1.04, P = .22; Table 2; Supplementary Table 2, available online). In subanalyses, high-frequency surveillance remained unassociated with recurrence-free survival across different tumor sizes, tumor grades, extents of resection, and lymph node sampling thresholds (Supplementary Table 3, available online).
Table 2.
Recurrence-free and overall survival in high- vs low-frequency surveillance groups
| Outcomes | Low frequency | High frequency | P |
|---|---|---|---|
| (n = 3047) | (n = 3124) | ||
| Recurrence-free survival | |||
| Adjusted hazard ratio (95% CI)a | 0.93 (0.83 to 1.04) | [1 reference] | |
| Median time to recurrence (IQR), mo | 21.3 (14.0 to 37.8) | 23.7 (13.4 to 41.1) | .046 |
| Cumulative incidence (95% CI) | |||
| 1 y | 4.1 (3.4 to 4.8) | 4.2 (3.5 to 4.9) | .22 |
| 3 y | 15.7 (14.4 to 17.0) | 15.4 (14.2 to 16.7) | |
| 5 y | 19.6 (18.2 to 21.1) | 19.6 (18.2 to 21.1) | |
| Overall survival | |||
| Adjusted hazard ratio (95% CI)a | 1.04 (0.96 to 1.12) | [1 reference] | |
| Percent survival (95% CI) | |||
| 1 y | 98.1 (97.5 to 98.5) | 96.6 (95.9 to 97.2) | .35 |
| 3 y | 79.2 (77.7 to 80.6) | 79.8 (78.4 to 81.2) | |
| 5 y | 65.7 (64.0 to 67.4) | 64.4 (62.6 to 66.1) |
Models control for age, sex, race, smoking status, BMI, Charlson/Deyo score, Area Deprivation Index, distance from hospital, annual hospital volume, surgical year, endobronchial ultrasound or mediastinoscropy, type of operation, incision type, histology, grade, tumor size, lymph node collection, major complication, positive margin, readmission, follow-up with oncologist. BMI = body mass index; CI = confidence interval; IQR = interquartile range.
Figure 2.
Cumulative incidence of recurrence-free survival for high- vs low-frequency surveillance groups.
Overall survival
The rates of 1-year, 3-year, and 5-year survival were 97.3% (95% CI = 96.9% to 97.7%), 79.5% (95% CI = 78.5% to 80.5%), and 65.1% (95% CI = 63.8% to 66.3%), respectively. The median overall survival was 65.7 (64.0-67.4) months in the low-frequency groups and 64.4 (62.6-66.1) months in the high-frequency group (Table 2; Figure 3). Using multivariable Cox regression, high-frequency surveillance was not associated with improved overall survival after controlling for tumor- and treatment-related factors (adjusted hazard ratio = 1.04, 95% CI = 0.96 to 1.12, P = .35; Table 2; Supplementary Table 4 (available online); P = .02 for prespecified noninferiority margin). Similarly, when stratifying by tumor size, tumor grade, extent of resection, and adequacy of lymph node sampling, high-frequency surveillance remained unassociated with overall survival (Supplementary Table 3, available online).
Figure 3.
Overall survival for high- vs low-frequency surveillance groups.
Propensity-matched analysis
A total of 5346 matched pairs were included in the propensity weighted analysis (2673 low-frequency surveillance, 2673 high-frequency surveillance). The groups were comparable in terms of demographic-, tumor-, and treatment-related characteristics (Supplementary Table 5, available online). High-frequency surveillance was not associated with improved overall survival (HR = 1.00, 95% CI = 0.93 to 1.07, P = .93) or recurrence-free survival (HR = 1.01, 95% CI = 0.91 to 1.14, P = .81; Supplementary Table 6, available online).
Discussion
In this study, we examined postresection imaging surveillance frequencies in veterans with pathologic stage I NSCLC. In our cohort of over 6000 patients, we found that more frequent surveillance was not associated with improved recurrence-free or overall survival. These findings were maintained across several different subanalyses, including by tumor size, tumor grade, type of resection (ie, lobectomy vs sublobar resection), and lymph node sampling adequacy. These findings, which are in agreement with a growing body of evidence (7,13,14), suggest that postresection surveillance after stage I NSCLC resection should occur at least annually but that no clear benefit exists for more than semiannual surveillance within 2 years after surgery.
With the emergence of robust lung cancer screening programs, especially in the VHA (30,31), cancer survivorship—the experience of living for extended cancer-free periods after curative-intent treatment—is increasingly common in early-stage lung cancer (32). Unfortunately, even in early-stage disease, recurrence remains relatively frequent, ranging from 20% to 50% (4,33). An important challenge of this cancer survivorship period, therefore, is to determine the optimal frequency of surveillance imaging. Although the potential benefits of more frequent surveillance may be intuitive (earlier detection, more aggressive treatment, etc), these must be weighed against the potential harms of overly aggressive surveillance, including ballooning health-care costs, patient anxiety, and unwarranted medical procedures (34).
This study joins an expanding body of evidence showing that more aggressive surveillance is not associated with improved cancer-specific outcomes. McMurry and colleagues (13) performed a Commission on Cancer special study of 4463 patients with stage I-III NSCLC who underwent chest CT (2006-2007). They divided patients into screening intensity intervals (3-, 6-, and 12-month) based on the time between surgery and the first postoperative CT scan and found that more frequent surveillance was not associated with improved survival. Subramanian and colleagues (14) reached similar conclusions in a subset of patients with stage I NSCLC. Analogous findings have been noted in colorectal cancer studies evaluating both imaging-based and lab-based (ie, carcinoembryonic antigen) surveillance (16). Our study expands on these findings by using modern data from the VHA, the largest integrated health-care system in the United States, to examine surveillance density. This method not only better reflects the holistic surveillance strategy of the postoperative period (as opposed to the first scan) but also removes selection bias associated with variable follow-up periods among patients who recur or die (13). Regardless of the distinct methodological approaches, the conclusions concur in suggesting that more aggressive surveillance is currently of little to no benefit in patients with early-stage NSCLC following resection.
An important finding of our study is that despite substantial variability in surveillance frequencies among veterans, important tumor- and treatment-related factors did not seem to drive this variability. It is plausible that disease- and patient-related factors should influence surveillance regimens (7). For example, inadequate lymph node sampling—which has been associated with worse outcomes in patients with NSCLC (27,35)—could logically drive clinicians toward a more aggressive surveillance strategy. Other factors such as sublobar resection, close margins, tumor size, and patient smoking status could similarly influence provider practices. However, we did not observe this in our data. Importantly, when stratifying our analysis by several of these “high-risk” features, we did not observe any benefit to high-frequency surveillance. It remains to be seen if other high-risk groups—such as patients with genetically aggressive tumors (36)—may benefit from more frequent surveillance. Similarly, advances in blood-based recurrence testing (eg, cell-free DNA) could further guide the frequency of surveillance imaging. These findings further highlight the complexities of designing patient-centered surveillance regimens.
It is also notable that oncology follow-up was associated with higher frequency surveillance. Oncologists and surgeons may have fundamentally different surveillance approaches based on unique training and practice experiences. More dedicated studies are needed to properly understand this finding.
A critical challenge of modern medicine is the dissemination and implementation of research findings into evidence-based practice (37). Despite growing evidence from this study and others (13,14) that more frequent surveillance imaging does not improve lung cancer outcomes, current guidelines from the National Comprehensive Cancer Network, the American Society for Clinical Oncology, the American College of Chest Physicians, and the American Society for Thoracic Surgery recommend chest CT every 6 months for the first 2-3 years following lung cancer resection (3,7,9,10,12). We believe, given our study and others, that these guidelines should be liberalized, at least, to recommending a chest CT scan every 6-12 months for patients with pathologic stage I disease. Such regimens should be sufficient for detecting recurrence as well as new primary cancers during extended follow-up (because annual surveillance is the current standard for lung cancer screening programs) (38). Surveillance and screening in patients with resected stage I lung cancer have many similarities; however, there is considerable debate as to when surveillance ends and screening resumes. Recommending annual imaging throughout may provide an opportunity to develop new, unambiguous, unified clinical practice guidelines. Liberalizing these guidelines will help to streamline care and reduce unnecessary costs and anxiety associated with more frequent surveillance.
This study has some limitations. First, veterans represent a unique cohort of patients, which can limit generalizability. Despite this, lung cancer treatment patterns and outcomes have been demonstrated to be very similar between veterans and the general population (17). Second, we were unable to ascertain the reason for each chest CT scan in the study (ie, scans ordered for nonsurveillance purposes). It is important, however, to use such incidental scans in lieu of scheduled surveillance imaging, when appropriate, to reduce excessive health-care use. Third, we were unable to assess surveillance noncompliance among patients. Further, our cohort excluded patients without documented CT scans in the VA system after surgery. We presume such patients received follow-up scans outside of the VA; however, the study results should be interpreted accordingly. Finally, one of the implicit conclusions of this study is that our current arsenal of salvage therapies is largely ineffective at treating recurrent and metastatic disease given the universally poor outcomes among patients who recur (7,39). Therefore, some may assume that as treatment options improve, earlier detection of recurrence will lead to better outcomes given these better therapeutic options (40,41). Although this is possible, our study showed no improvement in time to recurrence detection between high- and low-frequency groups. Therefore, even with advances in treatment options, it is unlikely that more frequent scanning will demonstrably improve outcomes.
In conclusion, more frequent surveillance is not associated with improved oncologic outcomes in pathologic stage I NSCLC following surgical treatment. Future lung cancer treatment guidelines should consider less frequent surveillance, especially in low-risk groups such as patients with stage I disease.
Supplementary Material
Contributor Information
Brendan T Heiden, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA.
Daniel B Eaton, Jr, VA St. Louis Health Care System, St. Louis, MO, USA.
Su-Hsin Chang, VA St. Louis Health Care System, St. Louis, MO, USA; Division of Public Health Sciences, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA.
Yan Yan, VA St. Louis Health Care System, St. Louis, MO, USA; Division of Public Health Sciences, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA.
Martin W Schoen, VA St. Louis Health Care System, St. Louis, MO, USA; Division of Hematology and Medical Oncology, Department of Internal Medicine, Saint Louis University School of Medicine, St. Louis, MO, USA.
Theodore S Thomas, VA St. Louis Health Care System, St. Louis, MO, USA; Divisions of Hematology and Oncology, Department of Medicine, Washington University School of Medicine, St. Louis, MO, USA.
Mayank R Patel, VA St. Louis Health Care System, St. Louis, MO, USA.
Daniel Kreisel, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA; VA St. Louis Health Care System, St. Louis, MO, USA.
Ruben G Nava, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA; VA St. Louis Health Care System, St. Louis, MO, USA.
Bryan F Meyers, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA.
Benjamin D Kozower, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA.
Varun Puri, Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA; VA St. Louis Health Care System, St. Louis, MO, USA.
Funding
This work was supported by Merit Award # 1I01HX002475-01A2 from the United States (U.S.) Department of Veterans Affairs (VP, SHC, YY, DBE) and 5T32HL007776-25 from the National Institutes of Health (BTH).
Notes
Role of the funder: The funders had no role in the study design; data collection, analysis, or interpretation; or writing of the manuscript or decision to submit it for publication.
Disclosures: All authors report no conflicts of interest.
Author contributions: Conceptualization: BTH, VP. Data curation: DBE, VP. Formal Analysis: DBE, VP. Methodology: BTH, BDE, S-HC, YY, VP. Supervision: VP. Writing—original draft: BTH, VP. Writing—review and editing: all authors.
Disclaimers: The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.
Prior presentations: Society of Thoracic Surgery 58th Annual Meeting, January 29-31, 2022, Miami Beach, FL.
Data availability
The patient-level data that were used in this study are maintained by the US Department of Veterans Affairs. VA data are freely available to VA-affiliated researchers with VA-secured computing access and an appropriate VA study protocol approval. For more information, visit https://www.virec.research.va.gov or contact the VA Information Resource Center at VIReC@va.gov. Additional inquiries can be directed to the corresponding author (BTH).
References
- 1. Siegel RL, Miller KD, Jemal A.. Cancer statistics, 2020. CA A Cancer J Clin. 2020;70(1):7-30. doi: 10.3322/caac.21590. [DOI] [PubMed] [Google Scholar]
- 2. Puri V, Crabtree TD, Bell JM, et al. Treatment outcomes in stage I lung cancer: a comparison of surgery and stereotactic body radiation therapy. J Thorac Oncol. 2015;10(12):1776-1784. doi: 10.1097/JTO.0000000000000680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Colt HG, Murgu SD, Korst RJ, Slatore CG, Unger M, Quadrelli S.. Follow-up and surveillance of the patient with lung cancer after curative-intent therapy: diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(suppl 5):e437S-e454S. doi: 10.1378/chest.12-2365. [DOI] [PubMed] [Google Scholar]
- 4. Taylor MD, Nagji AS, Bhamidipati CM, et al. Tumor recurrence after complete resection for non-small cell lung cancer. Ann Thorac Surg. 2012;93(6):1813-1820. discussion 1820-1. doi: 10.1016/j.athoracsur.2012.03.031. [DOI] [PubMed] [Google Scholar]
- 5. 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(5):e2111613. doi: 10.1001/jamanetworkopen.2021.11613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wong ML, McMurry TL, Stukenborg GJ, et al. Impact of age and comorbidity on treatment of non-small cell lung cancer recurrence following complete resection: a nationally representative cohort study. Lung Cancer. 2016;102:108-117. doi: 10.1016/j.lungcan.2016.11.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Heiden BT, Subramanian MP, Puri V, Kozower BD.. Striking a balance: surveillance of non-small cell lung cancer after resection. J Thorac Cardiovasc Surg. 2021;162(3):680-684. doi: 10.1016/j.jtcvs.2020.10.166. [DOI] [PubMed] [Google Scholar]
- 8. Westeel V, Foucher P, Scherpereel A, et al. Chest CT scan plus x-ray versus chest x-ray for the follow-up of completely resected non-small-cell lung cancer (IFCT-0302): a multicentre, open-label, randomised, phase 3 trial. Lancet Oncol. 2022;23(9):1180-1188. doi: 10.1016/S1470-2045(22)00451-X. [DOI] [PubMed] [Google Scholar]
- 9. Jaklitsch MT, Jacobson FL, Austin JHM, et al. The American Association for Thoracic Surgery guidelines for lung cancer screening using low-dose computed tomography scans for lung cancer survivors and other high-risk groups. J Thorac Cardiovasc Surg. 2012;144(1):33-38. doi: 10.1016/j.jtcvs.2012.05.060. [DOI] [PubMed] [Google Scholar]
- 10. National Comprehensive Cancer Network. Non-Small Cell Lung Cancer, Version 6.2020. NCCN guidelines in oncology. https://www.nccn.org/professionals/physician_gls/pdf/nscl_blocks.pdf. Published 2020. Accessed January 9, 2021.
- 11. Postmus PE, Kerr KM, Oudkerk M, et al. ; ESMO Guidelines Committee. Early and locally advanced non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2017;28(suppl 4):iv1-iv21. doi: 10.1093/annonc/mdx222. [DOI] [PubMed] [Google Scholar]
- 12. Schneider BJ, Ismaila N, Aerts J, et al. Lung cancer surveillance after definitive curative-intent therapy: ASCO guideline. J Clin Oncol. 2020;38(7):753-766. doi: 10.1200/JClinOncol.19.02748. [DOI] [PubMed] [Google Scholar]
- 13. McMurry TL, Stukenborg GJ, Kessler LG, et al. More frequent surveillance following lung cancer resection is not associated with improved survival: a nationally representative cohort study. Ann Surg. 2018;268(4):632-639. doi: 10.1097/SLA.0000000000002955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Subramanian M, Liu J, Greenberg C, et al. Imaging surveillance for surgically resected stage I non-small cell lung cancer: is more always better? J Thorac Cardiovasc Surg. 2019;157(3):1205-1217.e2. doi: 10.1016/j.jtcvs.2018.09.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Primrose JN, Perera R, Gray A, et al. ; FACS Trial Investigators. Effect of 3 to 5 years of scheduled CEA and CT follow-up to detect recurrence of colorectal cancer: the FACS randomized clinical trial. JAMA. 2014;311(3):263-270. doi: 10.1001/JAMA.2013.285718. [DOI] [PubMed] [Google Scholar]
- 16. Snyder RA, Hu C-Y, Cuddy A, et al. ; Alliance for Clinical Trials in Oncology Network Cancer Surveillance Optimization Working Group. Association between intensity of posttreatment surveillance testing and detection of recurrence in patients with colorectal cancer. JAMA. 2018;319(20):2104-2115. doi: 10.1001/JAMA.2018.5816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Heiden BT, Eaton DBJ, Chang S-H, et al. Comparison between veteran and non-veteran populations with clinical stage I non-small cell lung cancer undergoing surgery [published online ahead of print]. Ann Surg. 2021. doi: 10.1097/SLA.0000000000004928. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mulligan CR, Meram AD, Proctor CD, Wu H, Zhu K, Marrogi AJ.. Unlimited access to care: effect on racial disparity and prognostic factors in lung cancer. Cancer Epidemiol Biomarkers Prev. 2006;15(1):25-31. doi: 10.1158/1055-9965.EPI-05-0537. [DOI] [PubMed] [Google Scholar]
- 19. Sherer EA, Fisher DA, Barnd J, Jackson GL, Provenzale D, Haggstrom DA.. The accuracy and completeness for receipt of colorectal cancer care using Veterans Health Administration administrative data. BMC Heal Serv Res. 2016;16(1):1-10. doi: 10.1186/S12913-016-1294-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Nattinger AB, Schapira MM, Warren JL, Earle CC.. Methodological issues in the use of administrative claims data to study surveillance after cancer treatment. Med Care. 2002;40(suppl 8):IV- IV74. [DOI] [PubMed] [Google Scholar]
- 21. Zullig LL, Smith VA, Jackson GL, et al. Colorectal cancer statistics from the Veterans Affairs Central Cancer Registry. Clin Colorectal Cancer. 2016;15(4):e199-e204. doi: 10.1016/j.clcc.2016.04.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Korst RJ, Kansler AL, Port JL, Lee PC, Altorki NK.. Accuracy of surveillance computed tomography in detecting recurrent or new primary lung cancer in patients with completely resected lung cancer. Ann Thorac Surg. 2006;82(3):1009-1015; discussion 1015. doi: 10.1016/j.athoracsur.2006.03.062. [DOI] [PubMed] [Google Scholar]
- 23. Tarlov E, Lee TA, Weichle TW, et al. Reduced overall and event-free survival among colon cancer patients using dual system care. Cancer Epidemiol Biomarkers Prev. 2012;21(12):2231-2241. doi: 10.1158/1055-9965.EPI-12-0548. [DOI] [PubMed] [Google Scholar]
- 24. Sohn M-W, Arnold N, Maynard C, Hynes DM.. Accuracy and completeness of mortality data in the Department of Veterans Affairs. Popul Health Metr. 2006;4:2. doi: 10.1186/1478-7954-4-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Samson P, Crabtree T, Broderick S, et al. Quality measures in clinical stage I non-small cell lung cancer: improved performance is associated with improved survival. Ann Thorac Surg. 2017;103(1):303-311. doi: 10.1016/j.athoracsur.2016.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. American College of Surgeons. CoC Quality of Care Measures 2020 surveys. https://www.facs.org/Quality-Programs/Cancer/NCDB/qualitymeasurescocweb. Accessed December 21, 2020.
- 27. Heiden BT, Eaton DB, Chang S-H, et al. Assessment of updated commission on cancer guidelines for intraoperative lymph node sampling in early-stage non-small cell lung cancer [published online ahead of print]. J Thorac Oncol. 2022. doi: 10.1016/j.jtho.2022.08.009. [DOI] [PubMed] [Google Scholar]
- 28. Althunian TA, de Boer A, Groenwold RHH, Klungel OH.. Defining the noninferiority margin and analysing noninferiority: an overview. Br J Clin Pharmacol. 2017;83(8):1636-1642. doi: 10.1111/bcp.13280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Burdett S, Pignon JP, Tierney J, et al. Adjuvant chemotherapy for resected early-stage non-small cell lung cancer. Cochrane Database Syst Rev. 2015;(3):CD011430. doi: 10.1002/14651858.CD011430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Kinsinger LS, Anderson C, Kim J, et al. Implementation of lung cancer screening in the Veterans Health Administration. JAMA Intern Med. 2017;177(3):399-406. doi: 10.1001/jamainternmed.2016.9022 [DOI] [PubMed] [Google Scholar]
- 31. Ganti AK, Klein AB, Cotarla I, Seal B, Chou E.. Update of incidence, prevalence, survival, and initial treatment in patients with non-small cell lung cancer in the US. JAMA Oncol. 2021;7(12):1824. doi: 10.1001/jamaoncol.2021.4932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Jacobsen PB, DeRosa AP, Henderson TO, et al. Systematic review of the impact of cancer survivorship care plans on health outcomes and health care delivery. J Clin Oncol. 2018;36(20):2088-2100. doi: 10.1200/JClinOncol.2018.77.7482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Uramoto H, Tanaka F.. Recurrence after surgery in patients with NSCLC. Transl Lung Cancer Res. 2014;3(4):242-249. doi: 10.3978/j.issn.2218-6751.2013.12.05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Smith TJ, Hillner BE.. Bending the cost curve in cancer care. N Engl J Med. 2011;364(21):2060-2065. doi: 10.1056/NEJMSB1013826 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Gajra A, Newman N, Gamble GP, Kohman LJ, Graziano SL.. Effect of number of lymph nodes sampled on outcome in patients with stage I non-small-cell lung cancer. J Clin Oncol. 2003;21(6):1029-1034. doi: 10.1200/JClinOncol.2003.07.010 [DOI] [PubMed] [Google Scholar]
- 36. Kratz J, Van den Eeden S, He J, Jablons D, Mann M.. A prognostic assay to identify patients at high risk of mortality despite small, node-negative lung tumors. JAMA. 2012;308(16):1629-1631. doi: 10.1001/JAMA.2012.13551 [DOI] [PubMed] [Google Scholar]
- 37. Heiden BT, Tetteh E, Robbins KJ, et al. Dissemination and implementation science in cardiothoracic surgery: a review and case study. Ann Thorac Surg. 2022;114(2):373-382. doi: 10.1016/J.ATHORACSUR.2021.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Krist AH, Davidson KW, Mangione CM, et al. US Preventive Services Task Force. Screening for lung cancer. JAMA. 2021;325(10):962-970. doi: 10.1001/jama.2021.1117 [DOI] [PubMed] [Google Scholar]
- 39. Subotic D, Van SP, Grigoriu B.. Optimising treatment for post-operative lung cancer recurrence. Eur Respir J. 2016;47(2):374-378. doi: 10.1183/13993003.01490-2015 [DOI] [PubMed] [Google Scholar]
- 40. Howlader N, Forjaz G, Mooradian MJ, et al. The effect of advances in lung-cancer treatment on population mortality. N Engl J Med. 2020;383(7):640-649. doi: 10.1056/NEJMoa1916623 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Tong BC. Commentary: channeling goldilocks. J Thorac Cardiovasc Surg. 2021;162(3):685-686. doi: 10.1016/J.JTCVS.2020.11.064. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The patient-level data that were used in this study are maintained by the US Department of Veterans Affairs. VA data are freely available to VA-affiliated researchers with VA-secured computing access and an appropriate VA study protocol approval. For more information, visit https://www.virec.research.va.gov or contact the VA Information Resource Center at VIReC@va.gov. Additional inquiries can be directed to the corresponding author (BTH).