Results of frequent CT surveillance on recurrence detection and survival after radical resection for non-small cell lung cancer

ABSTRACT

Introduction

Non-small lung cancer (NSCLC) carries a substantial risk for recurrence even after complete resection. Evidence regarding the survival impact of post-resection surveillance strategies remains limited. Danish guidelines for lung cancer recommend contrast-enhanced computed tomography (CE-CT) every 3 months for the first 2 years and every 6 months for the subsequent 3 years, a frequency twice that of major international guidelines. This study retrospectively assessed the outcomes of this high-frequency CT surveillance in Denmark, specifically focusing on recurrence detection within two years post-surgery, the potential for renewed curative-intent treatment, and post-recurrence prognosis during this initial period.

Method

A cohort of 1079 patients who underwent resection for NSCLC in the period 2019–2020 was identified from the Danish Lung Cancer Registry (DLCR). Detailed information regarding new diagnoses of lung cancer, offered treatments, and mortality was extracted from patients’ medical records.

Results

Within two years following resection, 20% of patients developed a new diagnosis of lung cancer. Of these, 28.5% presented with localized disease (stage I-II), 26% with locally advanced disease (stage III), and 45% with metastatic disease. Recurrence frequencies ranged from 13% for pathological stage I (pStage I) to 44.5% for pStage III. Forty-eight percent of patients were offered renewed curative-intent treatment, demonstrating a 2-year post-recurrence survival of 78%. In contrast, patients offered palliative care or no treatment had a 2-year post-recurrence survival of 40%.

Conclusion

The proportion of recurrences presenting with metastatic disease was lower than reported in cohort studies with less frequent surveillance. A further notable finding was the high proportion of patients offered curative-intent treatment for recurrent disease, exceeding previously reported rates. These patients demonstrated a 2 year post-recurrence survival comparable to that observed following a primary NSCLC diagnosis. The presence of symptoms at the time of recurrence was a negative prognostic indicator, even among patients receiving palliative treatment.

Introduction

Lung cancer remains one of the most prevalent malignancies globally [1]. While radical surgical interventions such as lobectomy or pneumonectomy offer the potential for cure in patients with early-stage disease, a substantial risk of recurrence persists, with recurrence rates that may approach 50% within a follow-up period exceeding 5 years [2]. The risk of recurrence has been found to increase with advanced staging from approximately 35% for stage I to 60% for stage II [3]. Notably, metastatic recurrence is observed in approximately one-third of patients following definitive surgery for non-small cell lung cancer (NSCLC) [4]. Recurrent lung cancer presents considerable therapeutic challenges and is associated with a poorer prognosis compared to primary tumours [3].

Post-treatment surveillance is recommended for early detection of potentially curable recurrences and optimizing outcomes for those with incurable disease. Surveillance-detected recurrences or second primary lung cancers (SPLC) are more amenable to radical treatment than clinically detected relapses, leading to a significantly improved prognosis [5–9].

The majority of post-resection surveillance guidelines advocate for computed tomography (CT) utilization. The European Society of Medical Oncology (ESMO) and American Society of Clinical Oncology (ASCO) guidelines recommend post-curative treatment follow-up visits every 6 months for the initial 2 years, including a chest CT scan at 12 and 24 months [10,11]. Subsequent annual follow-up visits are recommended to detect SPLC. The National Comprehensive Cancer Network (NCCN) proposes a similar surveillance strategy for resected Stage III NSCLC [12]. However, the long-term survival advantages of CT-based follow-up strategies remain a subject of ongoing debate [13,14].

Since 2010, national Danish guidelines for post-therapy lung cancer surveillance have advocated a more frequent approach than international recommendations. This involves contrast-enhanced CT (CE-CT) of the thorax and upper abdomen every 3 months for the initial 2 years post-surgery, followed by CE-CT every 6 months until 5 years post-surgery. Continued surveillance is contingent upon the patient’s suitability for further treatment. Suspicious CT findings prompt a comprehensive diagnostic work-up to establish diagnosis and clinical stage. However, contemporary data on the outcomes of the Danish surveillance program are currently lacking.

This study aimed to investigate the efficacy of the high-frequency CT surveillance protocol implemented in Denmark during the first two years following radical resection of NSCLC. Specifically, we sought to evaluate its impact on recurrence detection, the feasibility of renewed curative-intent treatment, and the prognosis within the first two years post-recurrence.

Hypotheses

Our investigation was guided by the following hypotheses:

1. Recurrence in patients with postoperative pathological Stage I (pStage I) disease primarily presents as clinical Stage I (r-cStage I). High-frequency CT surveillance will detect these recurrences before they progress into a higher stage with a less favourable prognosis.

2. High-frequency CT surveillance will across all stages lead to a higher proportion of potentially curable recurrences compared to other surveillance strategies. The earlier detection afforded by frequent CT scans may significantly impact the management and outcome of recurrent disease.

Method

Study design

The retrospective cohort study was conducted in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for cohort studies [15]. The surveillance period was from 100 to 750 days in order to capture 2-year recurrence among patients who had a clear CT scan three months post-surgery, indicating no evidence of non-radical surgery.

Throughout this manuscript, the term ‘recurrence’ refers to a subsequent diagnosis of lung cancer, irrespective of its aetiology, encompassing both true relapse of the initial tumour, metachronous lung cancer, and the development of a second primary lung cancer (SPLC).

Participating sites

The patient cohort was derived from five specialized lung cancer departments across Denmark. These departments were responsible for both the initial diagnosis of lung cancer and the subsequent diagnostic work-up for suspected recurrences in the same patient population.

Data

Data for this study were retrieved from the Danish Lung Cancer Register (DLCR) for patients treated at the five participating hospitals, supplemented by a comprehensive review of individual patient medical records. The DLCR provides accurate and comprehensive information on primary diagnostic work-up, including clinical staging, histological data, pathological stage at surgery, and dates of initial diagnosis and surgical intervention [16,17].

Patients selected from the DLCR met the following inclusion criteria: a first-time diagnosis of lung cancer between 2019 and 2020, and curative-intent surgical resection with or without adjuvant therapy. All included patients underwent a comprehensive diagnostic and staging work-up, which included CE-CT, ¹⁸Ffluorodeoxyglucose (¹⁸FDG) Positron Emission Tomography (PET)/CT, CT-guided transthoracic needle biopsy and endoscopy with endobronchial ultrasound guided transbronchial needle aspiration biopsies (EBUS-TBNA) of mediastinal lymph nodes.

The study involved a meticulous review of patient medical records to ascertain lung cancer recurrence and survival outcomes, with a minimum follow-up period of 2 years after resection. For patients diagnosed with a recurrence, the following data were recorded based on multidisciplinary team (MDT) evaluations and decisions: method of detection (symptomatic or scheduled imaging), clinical stage according to the IASLC TNM 8^th edition [18], histological diagnosis, lung cancer type, subsequent treatment, and treatment intent were recorded based on the evaluations and decisions from multidisciplinary team (MDT) meetings.

If a recurrence was diagnosed the date of the CT scan initiating the diagnostic work-up was designated as the date of diagnosis. Survival after a new diagnosis of lung cancer was evaluated with a minimum follow-up time of 2 years for all patients. The study was conducted in 2023 until November 1, ensuring that all patients would have at least 2 years of follow-up after a recurrence.

Patients who did not undergo scheduled surveillance after surgery were excluded from the final analysis. This included patients who died or were diagnosed with a recurrence prior to or at the time of their first scheduled CT scan, or within 100 days after surgery, whichever occurred first.

Calculations on resources used for surveillance

To estimate the healthcare resources allocated to surveillance during the initial two years post-resection, we calculated the number of scheduled CT scans performed quarterly from 6 to 24 months after surgery, based on the proportion of patients alive at the midpoint of this surveillance period.

Statistics

Statistical analyses were performed using STATA 18 (StataCorp, College Station, Texas, USA). Simple correlations between categorical variables were assessed using Chi-square statistics. Binary outcomes were analysed via multivariate logistic regression. Differences in median values were evaluated with the Wilcoxon rank-sum test. Survival and time-to-event data were analysed using the Cox proportional hazards regression model and the Peto – Peto – Prentice test.

Ethics

Patient consent for data extraction from electronic health records was waived by the hospital administration of each participating institution. This waiver was granted for the purpose of improving current medical practice, in accordance with the Danish Health Act.

Results

An unselected cohort of 1271 patients diagnosed at the five participating hospitals and subsequently resected for lung cancer was identified in the DLCR. One hundred and ninety-two patients were excluded from analysis and 1079 patients were included in the final analysis (Figure 1). The characteristics of the included patients are presented in Table 1.

Figure 1.

Patient selection for analyses.

Table 1.

Patient characteristics.

Gender  Males  Females	469 (43.5%) 610 (56.5%)
Age at surgery – median (IQR)	72 y (65–77)
Tobacco  Never-smokers  Ever-smokers  Unknown	73 (7%) 940 (87%) 66 (6%)
Pack years (N = 1014) – median (IQR)	35 [19] − 50)
Histology Groups  Adenocarcinoma  Squamous cell carcinoma  Other NSCLC1	787 (73%) 239 (22%) 53 (5%)
Pathological Stage at surgery  IA  IB  IIA  IIB  IIIA  IIIB	472 (44%) 211 (20%) 54 (5%) 205 (19%) 117 (11%) 20 (2%)
2 year recurrence rate2	214 (20%)
Type of recurrence  Recurrence of original cancer  Metachronous/new cancer  Undetermined	195 (91%) 17 (8%) 2 (1%)
Clinical Stage of recurrent cancer  IA  IB  IIA  IIB  IIIA  IIIB  IIIC  IV  Unknown	34 (16%) 9 (4%) 1 (0.5%) 17 (8%) 35 (16%) 18 (8%) 2 (1%) 88 (45%) 2 (1%)
Aim of treatment for recurrence  Curative intent  Palliative  Unknown3	102 (48%) 110 (51%) 2 (1%)
Symptoms of recurrence before confirmed on CT  No  Yes  No information in medical record	145 (68%) 56 (26%) 13 (6%)
CT scan before schedule  No  Yes  No information in medical record	153 (71.5%) 48 (22.5%) 13 (6%)

1) NSCLC = Non-Small Cell Lung Cancer. 2) Between 100–750 days after surgery. 3) Two patients did not want work-up to determine if the recurrence was potentially curable.

This cohort of 1271 patients represents 54.3% of all patients resected for lung cancer in in Denmark during 2019 and 2020 performed at the three largest of the four centres for thoracic surgery in Denmark [20].

Incidences of recurrence over time after surgery

During the surveillance period from 100 to 750 days post-surgery, following an initial clear CT scan at 3 months, 214 patients (19.8%) received a subsequent lung cancer diagnosis. Of these recurrent cancers, 61 (28.5%) were staged as localized (cStage I-II), 55 (25.7%) as locally advanced (cStage IIIA-C), and 96 (44.9%) as metastatic disease. Staging information for two patients (0.9%) was unavailable due to a lack of formal staging for their new cancer.

The proportion of patients experiencing recurrence increased with pathological stage (pStage). For patients with pStage I disease, the recurrence rate remained relatively constant throughout the 100 to 750-day surveillance period, exhibiting only a slight decrease in incidence (Figure 2(B)). Correspondingly, for increasing pStage from I to III, the hazard rates for recurrence were higher during the first year of surveillance, subsequently decreasing and tending to converge (Figure 2(B)).

Figure 2.

(A) Kaplan-Meier plot of recurrence-free survival since start of surveillance period until 2 years after surgery* for pathological stages I to III (pStage IIII). The graph represents the fraction without recurrence among patients alive and without detectable cancer at first CT 3 months after surgery, subdivided into pStage IIII. (B) Smoothed hazard estimates for daily risk of a recurrence after surgery during the surveillance period. *From 100–750 days after surgery. In the graphs, ‘0 days’ represent start of surveillance from 100 days after surgery.

The relative frequency of recurrence within the initial 2 years post-surgery approximately doubled with each increase in pStage: 12.7% for pStage I, 24.7% for pStage II, and 44.5% for pStage III (Tables 1 and 3). Patients with pStage IA disease demonstrated the lowest risk, with 52 patients (11.0%) diagnosed with recurrence within the first 2 years after surgery.

Table 3.

Distribution of clinical stage (cStage) of recurrence at time of diagnosis of recurrence according to pStage for the 212 patients with a valid staging of their recurrence.

Pathological Stage	I	II	III	Total
Recurrence cStage, N (%)
I	28 (32.2%)	10 (15.6%)	5 (8.2%)	43 (20.3%)
II	7 (8.1%)	7 (10.9%)	4 (6.6%)	18 (8.5%)
III	16 (18.4%)	15 (23.4%)	24 (39.3%)	55 (25.9%)
IV	36 (41.4%)	32 (50.0%)	28 (45.9%)	96 (45.3%)
Total	87 (100%)	64 (100%)	61 (100%)	212 (100%)

Further potential indicators for recurrence risk within the initial 2 years post-surgery were analysed using a multivariate logistic regression model. The results of this analysis are presented in Table 2(A).

Table 2.

Logistic regressions.

A: Odds Ratios (OR) for recurrence in the first 2 years after surgery
N = 1079	OR	95% CI	P-value
Gender  Male (reference)  Female	1 1.32	0.95–1.83	0.097
Age at surgery (years)	1.00	0.98–1.02	0.948
Pathological Stage at surgery  pStage I (reference)  pStage II  pStage III	1 2.25 5.59	1.56–3.23 3.71–8.41	< 0.001  < 0.001
Histology  Adenocarcinoma (ref.)  Squamous cell carcinoma  Other NSCLC	1 1.06 2.63	0.72–1.57 1.43–4.85	0.751 0.002
B:. Odds Ratios for recurrent lung cancer in the first 2 years after surgery to be in clinical Stage I.
N = 214	OR	95% CI	P-value
Time since surgery (months)	1.11	1.04–1.18	0.001
Pathological Stage at surgery  pStage I (reference)  pStage II  pStage III	1 0.44 0.23	0.19–1.04 0.08–0.66	0.061 0.006
Histology  Adenocarcinoma (ref.)  Squamous cell carcinoma  Other NSCLC	1 1.60 4.43	0.65–3.88 1.43–13.74	0.305 0.010
C:. Odds Ratios for recurrent lung cancer in the first 2 years after surgery to be offered curative treatment.
N = 180*	OR	95% CI	P-value
Time since surgery (months)	1.07	1.01–1.14	0.016
Age at surgery (years)	0.97	0.93–1.00	0.123
Eversmoker	10.31	1.14–92.93	0.038
Symptoms of recurrence	0.46	0.22–0.97	0.041
Pathological Stage at surgery: pStage I vs. pStage II or III	2.07	1.07–3.99	0.031

A: Odds Ratios (OR) for risk of recurrence in the first 2 years after surgery with curative intent for lung cancer. OR = Odds Ratio. B:Odds Ratios for recurrent lung cancer in the first 2 years after surgery to be in clinical Stage I. C: Odds Ratios for recurrent lung cancer in the first 2 years after surgery to be curative treatment. Sex, pack-years, and histology did not have significant influence on likelihood of recurrent cancer to be curable. * Only including patients with information on smoking habits and symptoms of recurrence.

Correlation between pStage and clinical stage of recurrent cancer (r-cStage)

Among the 212 patients with fully staged recurrent cancer, a significant correlation was observed between the initial pathological stage (pStage) and the clinical stage of recurrence (r-cStage) (χ2 = 18.8, p = 0.005). Patients with pStage I disease had a significantly higher probability of being diagnosed with r-cStage I recurrence and a lower risk of r-cStage III and IV recurrence. Conversely, patients with pStage III disease were less likely to present with r-cStage I recurrence and faced a higher risk of r-cStage III and IV disease.

However, as detailed in Table 3, even among pStage I patients, only 32.2% experienced recurrence in r-cStage I. Notably, 26.5% presented with r-cStage II-III recurrence, primarily due to lymph node involvement. A further 41.4% of pStage I patients developed r-cStage IV recurrence.

A multivariate logistic regression model (Table 2(B)) was used to estimate the likelihood of an r-cStage I recurrence based on pStage, histology, and time from surgery. The probability of an r-cStage I recurrence significantly increased with time since surgery and for non-adenocarcinoma or non-squamous cell carcinoma NSCLC histology, while it decreased with increasing pStage.

Significance of symptoms and CT scan before schedule

A strong correlation was observed between patients reporting symptoms suggestive of recurrence and undergoing an unscheduled CT scan among the 201 out of 214 patients for whom data on both conditions were available (χ2 = 25.3, p < 0.001). An unscheduled CT scan that was diagnostic for recurrence was associated with 5.5-fold increased odds (95% CI: 2.7–11.1, p < 0.001) of the patient also reporting symptoms of recurrence. Furthermore, patients who reported symptoms at the time of diagnosis were significantly more likely to have r-cStage IV disease and less likely to have r-cStage I disease (χ2 = 11.7, p = 0.008). The odds ratio for patients to be in r-cStage III or IV was 3.2 (95% CI: 1.4–7.2, p = 0.006) if they had reported symptoms at the time of the CT scan.

Time from surgery to diagnosis of curative versus palliative recurrent lung cancer

A median difference of 74.5 days was observed in the time from surgery to diagnosis between recurrences offered palliative versus curative-intent treatment (p = 0.005). Notably, diagnoses leading to palliative treatment occurred significantly earlier.

Relapse of original cancer versus metachronous new lung cancer

Among the 214 patients diagnosed with recurrent lung cancer within two years of surgery, 17 (7.9%) were identified as having a second primary lung cancer (SPLC) according to the evaluation on a multidisciplinary team (MDT) meeting based on the criterions proposed by Martini & Melamed [19] or molecular testing, while 195 (91.1%) were considered to have a relapse of their original cancer. The remaining two cases were undetermined. Thirteen of the 17 cases of SPLC were in r-cStage I and 10 in rcStage IA. Sixteen of the 17 cases of SPLC were found in smokers or former smokers.

Among the 52 patients who experienced recurrent cancer despite having pStage IA at the time of initial surgery, 7 (13.5%) were categorized as SPLC rather than a relapse. For these patients, time since surgery was not an indicator for the new cancer to be assessed as an SPLC (OR = 0.99, 95% CI: 0.90–1.08).

The 17 patients classified as having an SPLC had a significantly higher probability of being offered renewed curative treatment (χ2 = 8.56, p = 0.003). In a Cox regression analysis, their hazard ratio (HR) for death within two years from the diagnosis of the new lung cancer, compared to patients with a relapse of their original cancer, was 0.31 (95% CI: 0.12–0.86, p = 0.024).

Table 2(C) presents the results of a multivariate logistic regression analysis for other indicators of a recurrence being deemed curable. Positive predictors for a curable recurrence included pStage I at surgery, a history of smoking, and increased time since surgery, whereas the presence of symptoms was a negative predictor.

Survival after recurrence

The 2-year survival curves for patients experiencing recurrence, stratified by the clinical stage of recurrent cancer, are presented in Figure 3(A). For context, the 2-year survival rate for patients who had no detectable recurrence on their first post-operative CT scan and remained free of recurrence during the 100 to 750-day surveillance period was 94%.

Figure 3.

(A) Kaplan-meier survival plots by clinical stage of recurrent lung cancer. This graph shows survival for patients diagnosed with recurrent NSCLC within 2 years of surgery, stratified by the clinical stage of their recurrence. Time is measured from the date of the CT scan that diagnosed the recurrent cancer. (B) Kaplan-meier survival plots for patients diagnosed with recurrent lung cancer. This graph illustrates overall survival for patients diagnosed with recurrent NSCLC within 2 years of surgery. Time is measured from the date of the CT scan that diagnosed the recurrent cancer. (C) Kaplan-meier plots of 2-year survival for patients with palliative recurrence, stratified by symptom presence. This graph displays survival for patients diagnosed with recurrent lung cancer during the 100 to 750-day surveillance period who were not candidates for renewed curative-intent treatment. Patients are segregated based on the presence or absence of symptoms at the time of their recurrence diagnosis.

Among the 214 patients diagnosed with recurrent lung cancer within two years of initial resection, 102 (47.7%) were offered treatment with curative intent, while 110 (51.4%) received palliative treatment. Two patients declined further work-up for staging and opted against active treatment. Of the 102 patients offered curative-intent treatment, the majority received chemo-radiotherapy (54, 52.9%). Other curative modalities included lung surgery (27, 26.5%), various types of surgery for extra-thoracic metastases (7, 6.9%), stereotactic body radiation therapy (9, 8.8%), and microwave ablation (1, 1%). Information was missing for 4 patients (3.9%).

The survival curves from the moment of recurrent cancer diagnosis, comparing those offered curative versus palliative treatment, are shown in Figure 3(B). The 2-year survival rates for patients diagnosed with recurrence and offered renewed curative treatment and palliative treatment were 78% and 40%, respectively.

In a Cox regression analysis from the time of recurrence, the hazard ratio (HR) for death within two years was 0.28 (95% CI: 0.18–0.43, p < 0.001) favouring curative-intent treatment. For the 73 patients with cStage II-III recurrence, where survival curves based on r-cStage did not visibly separate in Figure 3(A), Cox regression analysis revealed an HR for death within two years from the renewed lung cancer diagnosis of 0.35 (95% CI: 0.18–0.67, p = 0.002), again favouring curative-intent treatment.

Significance of symptoms when a recurrence is diagnosed

Among the 214 patients who experienced recurrence within the surveillance period (100 to 750 days post-surgery), 110 patients (51.4%) were offered palliative treatment. Of these, 65 (59.1%) were asymptomatic at the time of recurrence diagnosis, 37 (33.6%) reported symptoms, and symptom information was missing for 8 patients (7.3%).

As depicted in Figure 3(C), the survival of palliative patients presenting with symptoms was significantly inferior to that of asymptomatic patients (Peto – Peto – Prentice test: p = 0.002).

Of the 102 patients offered new curative-intent treatment, 18 (17.6%) presented with symptoms, while 80 (78.4%) were asymptomatic. Symptom information was missing for 4 patients (3.9%). Although the number of events/deaths within two years following recurrence diagnosis was substantially lower than in the palliative group, the presence of symptoms at the time of recurrence diagnosis was similarly associated with inferior survival even among patients receiving curative treatment (Peto – Peto – Prentice test: p = 0.043).

Resources used for surveillance

With a surveillance CT performed every 3 months during the initial two years post-resection, 644 CT scans were required to detect 13 recurrences among 100 resected pStage I patients. This translates to approximately 50 CT scans per surveillance-detected recurrence. This figure is less than half the 111 CT scans needed to detect one lung cancer in the NELSON lung cancer screening trial [21] and is comparable to the number of scans used for detecting lung cancer during surveillance of incidental lung nodules [22]. For patients in pStage III, about 7 CT scans were needed for each surveillance-detected recurrence in the first year.

Discussion

The 2-year recurrence rates in the current study with 12.7% for pStage I, 24.7% for pStage II, and 44.5% for pStage III was for pStage I and III within the 95% confidence intervals (CI) found in the meta-analyses by Rajaram et al. [23] while for pStage II it was somewhat less than their lower 95% CI limit of 27%. In the National Lung Screening Trial (NLST) the 2-year recurrence rate for pStage I was 10.2% (95% CI: 7.4–13.6%) and for pStage II 20.2% (95% CI: 13.1–28.4%) [24].

The proportion of recurrences in r-cStage IV of 44.9% in the current study is well below the proportion of 82.8% with distant metastatic or both distant metastatic and local/regional found by Boyd et al. [25] in their retrospective study of recurrences over a 5-year follow-up period on 975 surgically treated patients in pStage I-II (84.1% in pStage I) between 1995–2005. In that study there was no strict follow-up regime, but the patients were generally seen every 3 to 6 months, typically with a chest X-ray for the first 5 years. Further imaging, such as CT or PET was generally obtained if the patient presented with symptoms. Among participants in the NLST study who underwent complete resection for pStage III NSCLC the 5-year cumulative incidence of recurrence was 20.1% of which 81.9% had metastatic disease [24]. Details regarding the type and frequency of post-resection follow-up were unavailable, as patients were managed across multiple, distinct surgical centres. Similar outcome was reported by Lou et al. for a cohort of 1,294 patients that underwent complete resection for pStage III NSCLC between 2004–2009 where 19.9% had a recurrence and 73.2% of these had distant metastases [26]. The patients in this cohort had been followed with a CT scan at least every 6 to 12 months for the first 2 years after resection. For Asian patients Choi et al. [27] found a somewhat lower proportion of 66.2% with metastatic disease among patients with recurrence in their retrospective study of 242 patients resected for pStage I NSCLC between 1995–2012 and followed with CT or PET/CT every 6 months thereafter. In a more recent retrospective study of 949 resected Asian patients with NSCLC in pStage I-II (73.6% in pStage I) between 2010–2015 Jeong et al. [28] found 64.4% of the recurrences to be metastatic. The postoperative surveillance for these patients had been with a chest CT every 3–6 months for the first 2 years and every 6–12 months until 5 years.

Thus a markedly lower proportion of recurrences at clinical Stage IV was observed in this study compared to previous cohort studies that utilized less systematic and frequent postoperative CT surveillance. Notably, the two Asian studies, utilizing a somewhat more systematic and frequent CT surveillance approach, reported proportions closer to those observed in the current study. This suggests that intensive follow-up with contrast-enhanced CT may facilitate earlier detection of recurrence.

The proportion of patients in the current cohort to be offered curative-intent treatment at recurrence (47.7%) was considerable higher than reported in the randomized prospective trial by Westeel et al. [14], even when compared only with the cohort of patients followed with CT every 6 month where just 29.4% of detected recurrences were potentially curable. In a retrospective study by Moore et al. [29] 379 of 923 patients (41.1%) treated with curative intent for stage III NSCLC experienced a recurrence and only 42 of these (11.1%) were treated with curative intent for their recurrence. Clinical follow-up and imaging investigations of these patients had been performed at the discretion of the treating physician.

The observed 2-year survival rate of 78% for patients who received renewed curative treatment following a diagnosis of locoregional recurrence aligns with the 73–82% range previously reported by Bowes et al. [30] in their comprehensive review of treatment patterns and survival outcomes for early-stage non-small cell lung cancer (NSCLC) with locoregional recurrence. The post-recurrence survival for the patients in the current study for the first 2 years was for those with a recurrence in cStage III equal to that of unselected newly diagnosed lung cancer patients in Denmark in the same clinical stage and time period while those with a recurrence in cStage IIIIV in that comparison even had a better survival [31].

The positive predictive value of a smoking history for post-recurrence survival is likely attributable to the high incidence of recurrences presenting as early-stage SPLC among individuals with a history of smoking.

The prognostic significance of symptoms at the time of cancer recurrence, as identified in this study, has been consistently corroborated by numerous prior investigations. A meta-analysis by Calman et al. [32] encompassing four separate studies, revealed a Hazard Rate (HR) of 0.61 (95% CI: 0.50–0.74) for death in the follow-up period relative to symptomatic recurrences. Similar survival benefit of asymptomatic recurrences has been observed in other cohorts. Lou et al. [21] reported a nearly twofold increase in median survival for patients with asymptomatic recurrence in a cohort of 346 patients who underwent surgery for Stage IIIA Non-Small Cell Lung Cancer (NSCLC). Further supporting these findings, a retrospective cohort study by Banerji et al. [22] of 249 patients who underwent complete surgical resection for pStage I-II NSCLC found that among the 68% who experienced recurrence (69.2% of whom were symptomatic at recurrence diagnosis), patients with symptomatic recurrence had a median survival of 0.7 years from the time of recurrence, compared to 1.6 years for asymptomatic patients (p = 0.004). These consistent findings highlight the critical prognostic implications of symptom presentation at the time of cancer recurrence. Nevertheless, the potential for lead-time bias in these findings cannot be excluded.

Boyd et al. [25] compared the timeline from primary treatment to recurrence diagnosis for cases with local versus distant failure and identified no significant difference. Conversely, the present investigation found that cases classified as incurable were diagnosed approximately 2.5 months earlier than cases presenting with potentially curable recurrence.

Regarding the differentiation between relapse of the original cancer versus a SPLC Westeel et al. [14] found 11% of diagnosed cancers to be a SPLC. According to the medical records, 7.9% of the patients in the current study with a new diagnosis of lung cancer were judged to have a SPLC. For those who had been in pStage IA it was nearly twice as high, at 13.5%. In a retrospective study of 342 patients with curatively resected NSCLC from Graz, Austria, SPLC was diagnosed in 7.3% over a median follow-up time of 83 months [2] while an earlier review has reported an incidence rate of 1–2% per year for SPLC [33]. The median follow-up time in the study by Westeel et al. was 7.2 years corresponding to an average annual incidence rate of about 1.5%. In the current study, 7.9% of cancers diagnosed within the initial 2 years were considered to be SPLC, corresponding to an annual incidence rate of nearly 4%.

We were unable to identify a group of patients with a risk of recurrence that could be considered negligible. Despite patients in pStage I had a significantly lower risk of recurrence within the first 2 years, the risk was still 13% and only 28 of 87 (32%) of those diagnosed again with lung cancer were in rcStage I while 23 of 87 (26%) were in rcStage II-III, which for the vast majority was because of involvement of hilar (N1) or mediastinal lymph nodes (N2), and 36 of 87 (41%) were in rcStage IV.

Thus, we could not confirm our hypothesis 1.

With reference to hypothesis 2, the results support the assumption that frequent CT surveillance increases the proportion of potentially curable recurrences, which may influence both the management and outcomes of recurrent disease.

The main limitation of this study is the retrospective design, without randomization and a non-CT or low-frequency CT control arm for comparison, which severely limits our ability to determine the possible benefits of the current high-frequency surveillance program in Denmark versus less frequent surveillance. Furthermore, the study was based on data from the DLCR and information from the patients’ medical records with conclusions from MDT evaluations of the diagnosis, staging, and treatment possibilities for the detected recurrencies. The conclusions from the MDT meetings may be less than perfect. It is possible that a proportion of the curable recurrences in fact represents slow growing synchronous lung cancers as part of multifocal adenocarcinoma rather than relapse of the original lung cancer.

Another limitation is the duration of follow-up in the current study, which was limited to a two-year period, thereby potentially ignoring long-term outcomes beyond this time limit.

The retrospective nature of the analysis could introduce a risk of patient selection bias and uncontrolled for confounding variables, potentially influencing the study outcomes. Having said that, it is worth mentioning that the size of the primary cohort was relatively large, including 54% of all patients in Denmark undergoing resection for lung cancer on 3 of the 4 centres in Denmark for thoracic surgery in the two years in question, and the cohorts from the participating diagnostic departments were complete and unselected samples of all resected patients in those years with near to no patients lost to follow-up. Both circumstances enhance the likelihood that the current results apply to the entire population of resected patients in Denmark.

Conclusion

We observed a persistent risk of recurrence across all pathological stages. Even among pStage I patients, a substantial proportion of recurrences presented at advanced stages.

A markedly lower proportion of recurrences at clinical Stage IV was observed in this study compared to previous cohort studies that utilized less systematic and frequent post-operative CT surveillance. This could suggest that intensive follow-up with contrast-enhanced CT facilitates detection of recurrences at an earlier stage.

Furthermore, the proportion of patients diagnosed with recurrence received renewed curative-intent treatment was considerably higher in our study than reported in prior retrospective and prospective investigations and the patients demonstrated a post-recurrence survival consistent with results from other studies and equal to or better than that of unselected newly diagnosed lung cancer patients in Denmark in the same clinical stage and time period.

Collectively, these findings suggest a clinically significant benefit from systematic and frequent postoperative follow-up with contrast-enhanced CT in optimizing recurrence detection and improving access to curative-intent therapies, which could ultimately lead to enhanced survival. Conversely, the presence of symptoms at the time of recurrence negatively impacted prognosis, even within the palliative cohort.

Funding Statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Disclosure statement

No potential conflict of interest was reported by the author(s).