Real-world treatment patterns in patients with non-metastatic non-small cell lung cancer in Greece: the ‘EVIDENCE’ study

ABSTRACT

Background

The treatment landscape of non-metastatic non-small cell lung cancer (NM-NSCLC) is rapidly evolving with recent approvals of immunotherapies and targeted therapies.

Methods

This retrospective study included 202 adults diagnosed with NM-NSCLC between 1 January 2018 and 31 December 2020 primarily aiming to capture initial management strategies.

Results

Most frequent treatment patterns among Stage I/II patients (N = 84) were surgery only (48.8%) and surgery with adjuvant chemotherapy (with/without RT; 42.9%). Among Stage III patients (N = 118), most frequent patterns were chemotherapy plus radiotherapy (44.9%) and chemotherapy only (18.6%); 58.6% of Stage IIIA patients underwent surgery (of these, 32.4% also received chemotherapy and radiotherapy).

Conclusion

Initial strategy was aligned with contemporary at that time European guidelines, setting a benchmark for understanding the future uptake of new therapies.

1. Introduction

Lung cancer is the second most commonly diagnosed cancer and the leading cause of cancer-related death worldwide, in both sexes combined, with an estimated 2.5 million incident cases and 1.8 million deaths, respectively [1]. Particularly in Greece, lung cancer had both the highest incidence (8,583 cases) and mortality (7,156 deaths) among all cancers in 2022 [2]. The main subtype of lung cancer is Non-Small Cell Lung Cancer (NSCLC), comprising about 85% of all cases, of which approximately 50% are diagnosed with Non-Metastatic (NM) disease and about one-third of them can be treated surgically, with a curative intent [3–5].

Complete surgical resection is the cornerstone of treatment for localized NSCLC [6,7]. Despite complete resection, the disease recurs in a high proportion of patients with 5-year survival rates ranging from 54% (for stage I patients) to 10–15% (for stage IIIA patients) [8–10]. For patients with early-stage disease, with contraindications for surgery, or not willing to undergo surgery, radiotherapy (RT), preferably stereotactic ablative radiotherapy (SABR), is recommended [6,7]. Prior to the immune-oncology (IO) and epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) era, platinum-based chemotherapy (PBC) was the single systemic therapy option for resected patients. Adjuvant PBC has been shown to benefit resected stage II-III patients, offering modest increases of about 4–5% in 5-year survival rates [10,11]. Though the value of adjuvant PBC in stages IA and IB is less clear, it may be considered for patients with resected stage IB disease and a primary tumor >4 cm [6,7]. Results on the use of neoadjuvant chemotherapy (ChT) in patients with resectable NSCLC have historically been conflicting [12,13]. Regarding patients with locally advanced (LA; stage III) disease, a combination of various treatment types (e.g., surgery, ChT and/or RT) is most likely to be the optimal therapeutic approach [6,7], with PBC being part of the standard-of-care. In particular, in patients with single station N2 disease, three are the main options: surgery followed by adjuvant ChT, induction ChT or induction chemoradiotherapy (CRT) followed by surgery. Noteworthy, a combination of ChT with immune checkpoint inhibitors (ICIs) is gaining ground, including the application of ICIs earlier in the treatment trajectory (neoadjuvant setting). For patients with multistation N2 or N3 resectable disease, as well as for those with unresectable stage IIIA/IIIB disease, concurrent CRT (cCRT) is the preferred treatment of choice [6,7].

Following the success of targeted agents and ICIs [mainly programmed death 1 (PD-1) and PD-ligand 1 (PD-L1) inhibitors], in metastatic NSCLC, there is a growing interest in the application of such therapies in NM-NSCLC, where clinical need remains high. The PD-L1 inhibitor durvalumab was the first immunotherapy approved by the European Medicines Agency (EMA) in September 2018 for patients with NM-NSCLC, in particular, LA, unresectable stage III NSCLC who have not progressed following CRT and with tumor cell PD-L1 expression ≥ 1% [6,7,14–16]. Until 2018, PBC was the only available systemic therapy option for unresectable NM-NSCLC, and until 2020, it was also the only available option in the pre/post-operative setting in Europe [6,7]. Since then, multiple IO agents (one anti-PD-L1 and two anti-PD1 agents) and an EGFR TKI have been approved by EMA in this setting [17–19]. As the investigation of efficacy and safety of (neo)adjuvant or perioperative IO or targeted therapies (TT) in patients with resectable NSCLC is ongoing, further therapeutic advances are anticipated [20–22].

In this quickly evolving landscape, there is a need to assess real-world clinical practice prior to the availability of new treatments in order to better inform future treatment decision-making. Thus, in the absence of a nationwide NSCLC patient registry, the present retrospective chart review study aimed to capture the real-world management strategies, to characterize the profile of patients with NM-NSCLC, and to determine factors guiding real-world treatment decision-making in the NM disease setting in Greece.

2. Materials & methods

2.1. Study design and population

EVIDENCE was a non-interventional, single-country, multicentre, retrospective chart review study of patients diagnosed with NM-NSCLC between 1 January 2018 and 31 December 2020 (inclusive) hereinafter referred to as the 3-year index period.

Medical charts were reviewed and assessed through a process of consecutive sampling that followed reverse chronological order based on the date of initial NSCLC diagnosis. The most recently diagnosed patients, beginning 6 months before the date of each site activation and moving back in time were assessed for fulfilling the study-specific eligibility criteria. In addition, eligible patients who attended the participating clinical sites within the context of a routine clinical visit during the patient accrual period were also considered for inclusion in the study. In order to assess potential selective enrollment, participating physicians were requested to keep a screening log of all potentially eligible patients regardless of whether they agreed or were able to participate or not; no such cases were reported in the logs.

Eligible patients should have been diagnosed with histologically or cytologically confirmed NM-NSCLC (stage IA to IIIC; pathological if available, otherwise clinical) between 1 January 2018 and 31 December 2020 (inclusive) and at least 6 months prior to inclusion in the study, should have been ≥18 years old at the time of diagnosis and should have had provided a written informed consent in case the patient was alive. All participating sites were granted a waiver of consent for deceased patients by the Institutional Review Board of the study site. Patients were excluded from the study if they had another primary malignancy that required anticancer treatment during the observation period and if they were participating in any investigational program/trial with interventions outside of routine clinical practice for the treatment of their NM-NSCLC.

The look-back period for each patient extended to the date of initial diagnostic work-up for lung cancer. The retrospective observation period for each patient was defined as the interval between initial diagnosis and end of observation defined as the earliest date of documented distant metastasis, informed consent obtainment (for patients who were alive at the time of chart abstraction onset), or death (for patients who were deceased at that time).

2.2. Study objectives

To address the study objectives, patients were divided into subgroups by NSCLC stage at initial diagnosis, i.e., Stage I/II and Stage III NSCLC patients, and whether or not they had undergone tumor resection as part of initial management strategy, i.e., resected and unresected patients, the latter also including inoperable patients.

The primary objective of this study was to capture the initial management strategies employed in the NM setting of NSCLC, overall, by disease stage, and in the resected and unresected subpopulations. Initial management strategy is defined as the treatment administered from the date of confirmation of initial NM-NSCLC diagnosis (index date) until the earliest of first disease recurrence/progression or the end of the study observation.

Secondary objectives included the analysis of the diagnostic and staging workup, rate of incidentally detected NSCLC along with incidental findings, patient profile, factors affecting the initial decision-making, surgical resection patterns, reasons for not prescribing adjuvant systemic therapy, and reasons for not performing primary site surgical resection. Additionally, the molecular/biomarker testing profile was evaluated as an exploratory objective. Lastly, the study addressed the rate of first postoperative recurrence/progression and the management strategies upon first recurrence/progression, where disease recurrence was defined as the reappearance of disease after complete surgical resection with curative intent, and disease progression as the worsening/spread of the disease among unresected patients or patients with residual disease after surgical resection (incompletely resected); these events are collectively referred to as progression events hereafter in the text.

2.3. Sample size determination

A sample size of 200 patients was deemed adequate to draw meaningful and valid conclusions for the study primary objective in relation to NM-NSCLC management pattern as it would offer a maximum margin of error of 7.1% (with 95% CI using the Clopper-Pearson exact method ranging between 42.9% and 57.1%) for the estimation of a frequency of 50% where the width of the CI is biggest. In addition, if a treatment option was not to be observed (0% frequency), the real rate would not be greater than 1.8% with a 95% confidence. Sample size determination was performed using the statistical software package SAS®, Version 9.4 (SAS Institute Inc., Cary, NC, USA).

2.4. Statistical analysis

Statistical analysis was mainly of a descriptive nature. Categorical variables were summarized using frequencies and percentages. The Shapiro–Wilk test was used to test the normality of the distribution of continuous variables, which are presented as mean (standard deviation; SD) in the case of a normal distribution of data; otherwise, the median (interquartile range; IQR) is presented. For variables which do not follow a normal distribution in at least one of the study (sub)populations, a uniform presentation of median (IQR) was applied.

The association of patient, disease and institution characteristics with the initial management strategies utilized in the subpopulations with Stage I/II and Stage III disease were examined using univariable and multivariable logistic regression models. Next, the potential influence of the confounding factors of type of institution and geographic region on the associations was examined through multivariable logistic regression analysis. The factors were considered confounders if, after being added in the model, the crude beta coefficient would alter by at least 10% in either direction and also if the observed relationship between the examined factor and the outcome was statistically significant. Finally, a stepwise method based on minimizing Akaike Information Criterion (AIC) was used to determine the factors to be included in the multivariable models and the results of only the best fitted models are presented. In case any of the two aforementioned variables was considered as a confounder, it was not included in the stepwise procedure but it was added directly in the final multivariable model.

All statistical tests were two-sided and were performed at a 0.05 significance level. In terms of handling missing data, only partial missing dates were imputed. In the context of the logistic analysis, covariates with at least 20% missing data rate were excluded from the analysis.

Statistical analysis was performed using SAS® software, Version 9.4 (SAS Institute Inc., Cary, NC, USA).

3. Results

3.1. Patient disposition

Over a 6-month accrual period from 28 May 2021 (First Patient In) to 30 November 2021 (Last Patient In), a total of 203 NM-NSCLC patients were consecutively enrolled by pulmonologists or oncologists practicing in 7 public or private sector hospital centers/clinics located across 4 geographic regions of Greece (“Attica,” “Central Macedonia,” “Crete,” and “Eastern Macedonia and Thrace”). One patient did not meet all eligibility criteria; thus, the analysis was performed in 202 patients. Of the eligible patients, 55.0% (111/202) and 45.0% (91/202) were enrolled by physicians practicing at a publicly owned academic (university) institution and a privately owned hospital clinic, respectively. Furthermore, 69.3% of patients were enrolled by sites located in Attica which is the most populous region of Greece.

Among the 202 NM-NSCLC patients included in the study, a higher proportion was newly diagnosed with Stage III disease (58.4%) compared with Stage I/II disease (41.6%) (Figure 1(a)). Of the total study population, 58.9% had undergone primary tumor resection as part of the initial management strategy, with a relatively balanced distribution of Stage I, II and III patients (Figure 1(a)). Complete resection (surgical margin status of R0) was achieved in 91.7% (99/108) of the patients with available data. The proportion of resected patients among Stage I/II and Stage III NSCLC patients was 94.0% (79/84) and 33.9% (40/118), respectively, the latter mostly comprising Stage IIIA NSCLC patients. Among Stage IIIA NSCLC patients, 58.6% (34/58) had undergone primary tumor resection, while this proportion was 6.4% (3/47) and 25.0% (3/12) among Stage IIIB and Stage IIIC NSCLC patients, respectively.

Figure 1.

Patient disposition in the study and length of retrospective observation period.

(a) Distribution of NM-NSCLC patients by disease stage/substage at initial diagnosis, resection status and disease stage at the end of observation. (b) Length of retrospective observation period per disease stage at initial diagnosis and resection status as initial management strategy.

Box-plots depict median with interquartile range (Q1-Q3), including whiskers that extend from minimum to maximum values.

^aDisease staging was based on the AJCC/UICC TNM 8th Edition criteria in 93.1% (188/202) and the 7^th Edition in 5.9% (12/202), while for 2 patients the criteria were unknown/not available. Staging method was pathological (± clinical) for 75.2% (152/202) and clinical (only) for 24.8% (50/202). ^b Excluding one patient with Stage III NSCLC for whom disease substage was unknown. End of observation is defined as the earliest date of documented distant metastasis, informed consent obtainment (for patients who were alive at the time of chart abstraction onset), or death (for patients who were deceased at that time). Imputation of the missing date of death has been implemented for eight patients. N: number of patients; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; Q1: 25% Percentile; Q3: 75% Percentile.

In the overall NM-NSCLC population, median (IQR) time from initial NSCLC diagnosis until the end of observation was 15.8 (10.7–23.6) months. At the time of inclusion in the study, 32.2% of the patients were deceased; information on the cause of death is presented in Supplementary Figure S1A. During the retrospective observation period, 39.1% (79/202) of the NM-NSCLC patients (Supplementary Figure S1B) experienced a median (IQR) of 1.0 (1.0–1.0) (range 1–3) progression events after a median (IQR) of 10.1 (6.9–13.4) months (time from initial diagnosis to first progression event). At the time of inclusion in the study or death, 63.4% (128/202) of the patients had the same NSCLC stage as their initial diagnosis, NM-NSCLC was downstaged and upstaged for six (3.0%) and three (1.5%) patients, respectively, while 32.2% (65/202) were upstaged to Stage IV NSCLC [72.3% (47/65) with Stage III at initial diagnosis; Figure 1(a). Distant metastasis (among the 65 metastatic patients) was first documented a median (IQR) of 10.8 (7.8–16.7) months after initial NSCLC diagnosis.

The length of retrospective observation period in the NM-NSCLC subpopulations was numerically longer among patients with Stage I/II NSCLC at initial diagnosis than Stage III NSCLC patients (hereafter referred to as Stage I/II NSCLC and Stage III NSCLC patients) and among patients whose primary tumor had been resected than unresected patients (hereafter referred to as resected and unresected patients; Figure 1(b)). This trend is also reflected in the much higher proportion of deceased patients among Stage III NSCLC patients (45.8%) and unresected patients (49.4%) compared with Stage I/II (13.1%) and resected (20.2%) patients, respectively (Supplementary Figure S1A). In addition, rate of progression was nearly double in the Stage III NSCLC subpopulation compared with the Stage I/II NSCLC subpopulation (any event: 49.2% versus 25.0%; distant event: 39.8% versus 21.4%; Supplementary Figure S1B).

Resected patients were followed up for a median of 16.9 months after surgery during which time nearly a third of resected patients (30.3%; 36/119) experienced a postoperative progression event, a median of 10.1 (8.0–14.6) months after surgery. The postoperative progression event was distant in 77.8% (28/36) and locoregional in 22.2% (8/36) of the cases.

3.2. Patient and disease characteristics

Disease staging was based on the TNM 8th Edition criteria of the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC) in 94.0% (188/200) of evaluable patients, while 7th Edition criteria were used for the remaining 6.0%. Patient and disease characteristics at initial diagnosis, overall and by NSCLC stage at initial diagnosis and primary resection status are summarized in Tables 1 and 2 respectively. Overall, differences in patient and disease characteristics between resected and unresected subpopulations were relatively consistent with those between Stage I/II and Stage III NSCLC subpopulations, as expected given that unresected patients mostly comprised of Stage III NSCLC patients (94.0%; 78/83) (Tables 1 and 2, Figure 1(a)).

Table 1.

Patient characteristics at initial diagnosis, in the overall NM-NSCLC population and per disease stage and resection status.

	Overall (N = 202)	Stage I/II (Ν = 84)	Stage III (Ν = 118)	Resected (Ν = 119)	Unresected (Ν = 83)
Sociodemographic characteristics
Male, % (n/N)	73.3 (148/202)	72.6 (61/84)	73.7 (87/118)	73.1 (87/119)	73.5 (61/83)
Urban area, % (n/N)	70.6 (142/201)	71.4 (60/84)	70.1 (82/117)	71.4 (85/119)	69.5 (57/82)
Age in years, median (IQR)	68.8 (62.1–74.0)	70.4 (59.9–74.1)	67.5 (62.6–73.5)	67.1 (60.1–73.6)	70.0 (63.3–74.2)
Aged <70 years, % (n/N)	54.0 (109/202)	48.8 (41/84)	57.6 (68/118)	57.1 (68/119)	49.4 (41/83)
BMI in kg/m2, median (IQR)	26.1 (23.2–29.3)	26.6 (24.5–29.4)	25.9 (22.4–29.3)	26.2 (24.3–29.4)	25.9 (22.1–28.4)
Ever smokers, % (n/N)	92.4 (183/198)	90.1 (73/81)	94.0 (110/117)	90.5 (105/116)	95.1 (78/82)
Current smokers, % (n/N)	42.4 (84/198)	40.7 (33/81)	43.6 (51/117)	42.2 (49/116)	42.7 (35/82)
Smoking history of ever-smokers in pack-years, median (IQR)	50.0 (37.5–69.0)	50.0 (30.0–76.5)	50.0 (37.5–67.5)	50.0 (33.8–80.0)	54.0 (40.0–62.5)
Medical history and comorbidities
≥1 clinically significant medical conditiona,b or surgical procedure, % (n/N)	67.3 (136/202)	65.5 (55/84)	68.6 (81/118)	64.7 (77/119)	71.1 (59/83)
≥1 ongoing medical condition/comorbidity, % (n/N)	60.4 (122/202)	61.9 (52/84)	59.3 (70/118)	58.0 (69/119	63.9 (53/83)
≥2 ongoing medical conditions/comorbidities, % (n/N)	35.6 (72/202	29.8 (25/84)	39.8 (47/118)	32.8 (39/119)	39.8 (33/83)
Clinically significant medical conditionb or surgical procedure reported in ≥ 20% of the patientsc, % (n/N)
Hypertension	32.7 (66/202)	32.1 (27/84)	33.1 (39/118)	35.3 (42/119)	28.9 (24/83)
Diabetes mellitus	19.3 (39/202)	20.2 (17/84)	18.6 (22/118)	17.6 (21/119)	21.7 (18/83)
COPD	16.8 (34/202)	9.5 (8/84)	22.0 (26/118)	12.6 (15/119)	22.9 (19/83)
History of other primary malignancy, % (n/N)	5.4 (11/202)	4.8 (4/84)	5.9 (7/118)	5.9 (7/119)	4.8 (4/83)
Disease-related symptoms
Presence of any NM-NSCLC-related symptoms, % (n/N)	69.5 (132/190)	57.5 (46/80)	78.2 (86/110)	66.1 (74/112)	74.4 (58/78)
Reported in ≥ 20% of the patients with symptomsc, % (n/N)
Cough	46.2 (61/132)	41.3 (19/46)	48.8 (42/86);	43.2 (32/74)	50.0 (29/58);
Bloody sputum	25.0 (33/132)	17.4 (8/46)	29.1 (25/86)	24.3 (18/74)	25.9 (15/58)
Fatigue	18.9 (25/132)	34.8 (16/46)	10.5 (9/86)	25.7 (19/74)	10.3 (6/58)
ECOG PS, % (n/N)
0	61.4 (116/189)	69.5 (57/82)	55.1 (59/107)	69.3 (79/114)	49.3 (37/75)
1	34.4 (65/189)	29.3 (24/82)	38.3 (41/107)	29.8 (34/114)	41.3 (31/75)
≥2	4.2 (8/189)	1.2 (1/82)	6.5 (7/107)	0.9 (1/114)	9.3 (7/75)
Cardiorespiratory fitness, % (n/N)
Poor	8.7 (15/172)	6.5 (5/77)	10.5 (10/95)	5.0 (5/101)	14.1 (10/71)
Moderate	62.8 (108/172)	54.5 (42/77)	69.5 (66/95)	58.4 (59/101)	69.0 (49/71)

^aPast or ongoing.

^bOther than NSCLC.

^cIn any of the examined populations.

BMI: Body Mass Index; ECOG: Eastern Cooperative Oncology Group; IQR: Interquartile Range; n: number of patients with variable; N: total number of patients with available data; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; PS: Performance Status.

Table 2.

Disease characteristics at initial diagnosis, in the overall NM-NSCLC population and per disease stage and resection status.

	Overall (N = 202)	Stage I/II (Ν = 84)	Stage III (Ν = 118)	Resected (Ν = 119)	Unresected (Ν = 83)
Disease Characteristics
Method of initial diagnosis, % (n/N)
Histology	88.6 (179/202)	98.8 (83/84)	81.4 (96/118)	90.8 (108/119)	85.5 (71/83)
Cytology	8.4 (17/202)	1.2 (1/84)	13.6 (16/118)	5.9 (7/119)	12.0 (10/83)
Histology & cytology	3.0 (6/202)	-	5.1 (6/118)	3.4 (4/119)	2.4 (2/83)
Primary tumor localization reported in ≥ 20% of the patientsa, % (n/N)
Right upper lobe	45.0 (91/202)	42.9 (36/84)	46.6 (55/118)	40.3 (48/119);	51.8 (43/83)
Left upper lobe	22.8 (46/202)	25.0 (21/84)	21.2 (25/118)	21.8 (26/119)	24.1 (20/83)
Right lower lobe	18.3 (37/202)	15.5 (13/84)	20.3 (24/118)	16.8 (20/119)	20.5 (17/83)
Primary tumor histology, % (n/N)b
Adenocarcinoma	57.9 (117/202)	66.7 (56/84)	51.7 (61/118)	65.5 (78/119)	47.0 (39/83)
SCC	39.1 (79/202)	29.8 (25/84)	45.8 (54/118)	31.1 (37/119)	50.6 (42/83)
Adenosquamous cell carcinoma	2.5 (5/202)	2.4 (2/84)	2.5 (3/118)	2.5 (3/119)	2.4 (2/83)
Mucus adenocarcinoma	0.5 (1/202)	1.2 (1/84)	-	0.8 (1/119)	-
Tumour differential grade, % (n/N)
1/2	52.4 (75/143)	52.2 (36/69)	52.7 (39/74)	55.2 (53/96)	46.8 (22/47)
≥3	47.6 (68/143)	47.8 (33/69)	47.3 (35/74)	44.8 (43/96)	53.2 (25/47)
Angiolymphatic invasion, % (n/N)	8.6 (13/152)	6.7 (5/75)	10.4 (8/77)	9.7 (10/103)	6.1 (3/49)
Pleural invasion, % (n/N)	16.6 (26/157)	15.8 (12/76)	17.3 (14/81)	18.7 (20/107)	12.0 (6/50)
Number of lymph nodes dissected, median (IQR)	2.0 (0.0–8.0)	6.0 (2.0–12.0)	0.0 (0.0–4.0)	6.0 (2.0–13.0)	0.0 (0.0–0.0)

^aIn any of the examined populations.

^bNon-squamous cell carcinoma was adenocarcinoma in all cases.

IQR: Interquartile Range; n: number of patients with variable; N: total number of patients with available data; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; SCC: Squamous Cell Carcinoma.

The proportion of patients with available data whose disease was detected incidentally for purposes other than lung cancer screening was 18.2% (35/192). All incidentally detected lung cancer cases were detected through imaging studies, the majority of which involved detection of a “malignant solitary pulmonary nodule” (Figure 2). Diagnostic and staging procedures are summarized in Supplementary Table S1.

Figure 2.

(a) Rate of incidentally detected NSCLC and (B) Incidental findings in the overall NM-NSCLC population and in the subpopulations by disease stage at initial diagnosis.

Excluding 10 patients (three with Stage I/II and seven with Stage III NSCLC) for whom method of detection of lung cancer was unknown.

n: number of patients with variable; N: total number of patients with available data; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer.

Among the overall NM-NSCLC study population, 51.0% (103/202) of patients underwent any NSCLC-related molecular testing in the context of initial diagnosis, and in particular a median of 3.0 days before initial NSCLC diagnosis (IQR: −9.0 to 1.0 days), whereas an additional 5.4% underwent such testing ≥3 months after initial diagnosis and before first disease progression. In the context of initial diagnosis, the NSCLC-relevant biomarkers (druggable targets) of EGFR mutation and PD-L1 expression were tested in 23.8% (48/202) and 43.6% (88/202; Figure 3). Most frequent methods used for EGFR testing were next-generation sequencing (NGS; in 43.8%) and polymerase chain reaction (PCR; in 39.6%) (Supplementary Table S2). PD-L1 expression was assessed by immunohistochemistry in 93.2% of the tested patients, NGS in 5.7% and PCR in the remaining one patient (1.1%), as recorded by the investigators. Of the patients overall, 17.8% (36/202) had both EGFR mutation and PD-L1 expression tested, none of whom were positive for both markers. EGFR mutation was observed in 6.3% (3/48) of the population (Figure 3(a)). PD-L1 expression was positive (i.e., TC ≥ 1%) in 57.1% (8/14) and 56.7% (34/60) of the tested Stage I/II and Stage III NSCLC patients, respectively (excluding indeterminate results; Figure 3(b)). The frequency, findings and timing of molecular testing, as well as PD-L1 expression cutoffs, in the overall NM-NSCLC population and the subpopulations by disease stage at initial diagnosis are provided in Figure 3 and Supplementary Table 3.

Figure 3.

Frequency and findings of testing of (a) EGFR and (b) PD-L1 at initial diagnosis and throughout the observation period in the overall NM-NSCLC population and by disease stage at initial diagnosis.

^aEGFR results were based on the surgical specimen in only 12 of patients tested for EGFR at initial diagnosis (25.0%). ^bThe denominator includes positive, negative and indeterminate/unknown results, the latter being the case for 14 patients (two with Stage I/II and 12 with Stage III NSCLC). PD-L1 expression positivity was defined as TPS > 1%. EGFR: Epidermal Growth Factor Receptor; N: number of patients; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; PD-L1: Programmed death-Ligand 1; TPS: Tumour Proportion Score.

3.3. Initial treatment for NM-NSCLC, overall and by NSCLC stage at initial diagnosis

All study eligible NM-NSCLC patients (100.0%; 202/202) received treatment as part of initial management strategy. Time from initial NSCLC diagnosis to treatment initiation regardless of therapeutic modality ranged from 50 days before to 126 days (~4 months) after diagnosis with a median of 14.0 days, while surgery was performed earlier in the treatment pathway, followed by systemic therapy (Figure 4(a)). Radiotherapy (RT) was initiated later in the treatment pathway, specifically, a median (IQR) of 3.5 (2.3–5.2) months after initial diagnosis. This trend was evident across study populations (Figure 4(a)). Systemic therapy was the most frequent therapeutic modality among all NM-NSCLC patients (77.7%; 157/202) and among Stage III NSCLC patients (98.3%; 116/118; Figure 4(b)), which was administered as main therapeutic strategy (i.e., without surgery), either alone or in combination with RT and/or IO therapy, in 50.3% (79/157) and 65.5% (76/116) of the cases, respectively (Figure 4(c)). The second most frequent therapeutic modality among Stage III NSCLC patients was RT (59.3%; 70/118), administered as part of CRT in 81.4% (57/70) of these cases; cCRT in 48.6% (34/70) and sequential CRT (sCRT) in 34.3% (24/70). For Stage I/II NSCLC patients, surgery was the most frequent therapeutic modality (94.0%; 79/84) which was coupled with neoadjuvant and/or adjuvant systemic therapy in 48.1% (38/79) of the resected patients (Figure 4(a,b)). Only six (7.1%) Stage I/II NSCLC patients were treated with RT, three of whom as part of CRT (50.0%; Figure 4(b)).

Figure 4.

Initial management strategy in the overall NM-NSCLC population and by disease stage at initial diagnosis.

(a) Time from initial NSCLC diagnosis until the start of initial management strategy. (b) Frequencies of different therapeutic modalities. (c) Treatment setting in which systemic therapy was received among patients treated with such therapy and type of systemic therapy among unresected patients. (d) Type of systemic therapy received among patients treated with adjuvant therapy.

Box-plots depict median with interquartile range (Q1-Q3), including whiskers that extend from minimum to maximum values. ^aAn additional 13 patients (12 Stage III and one Stage I/II) had received ChT plus RT but were not recorded as “CRT” by the investigator. ^bExcluding 45 patients who had received one of the following treatment patterns: surgery only (41 Stage I/II patients; one Stage III patient), RT only (two Stage I/II patients) and surgery followed by RT (one Stage III patient). ^cAdjuvant and neo-adjuvant categories may or may not contain RT and/or IO. ^dFive patients (one Stage I/II and four Stage III) also received systemic therapy in the adjuvant setting. ^eOne patient received systemic therapy in the neoadjuvant setting but did not proceed to surgery. ^fOne patient received IO followed by RT, while the other patient received the following sequence: ChT ➔ RT (CRT) ➔ IO (consolidation) ➔ ChT (neoadjuvant) ➔ SUR. ChT:Chemotherapy; conc.: concurrent; CRT: Chemoradiotherapy; IO: Immune-Oncology (therapy); IQR: Interquartile Range; n: number of patients with variable; N: total number of patients with available data; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; Q1: 25% Percentile; Q3: 75% Percentile; RT: Radiotherapy; seq.: sequential; SUR: surgery; SYS: systemic therapy; TT: Targeted Therapy.

The combinations of systemic therapy types are presented in Figure S4C and S4D, for the overall NM-NSCLC population and in the subpopulations by disease stage at initial diagnosis, with “ChT only” being the most frequent systemic therapy pattern across settings. ChT was platinum-based in the majority of ChT cases (97.4%; 148/152). None of the Stage I/II NSCLC patients received IO therapy, while 22.4% (26/116) of the Stage III NSCLC patients were treated with IO therapy, administered in the consolidation or maintenance setting in 61.5% (16/26) of these patients, which followed CRT in most of these cases (81.3%; 13/16). Treatment with an EGFR TKI was reported in two patients overall, both EGFR-positive, while the other EGFR-positive patient was treated with surgery only.

3.4. Initial treatment for NM-NSCLC by primary tumor resection status

Median (IQR) time from initial NSCLC diagnosis until systemic therapy and RT among unresected patients treated with such therapy was 32.0 (19.0–45.0) and 104.0 (62.0–130.0) days, respectively. All but two of the unresected patients received systemic therapy (97.6%; 81/83), 67.9% (55/81) of whom were also treated with RT (Figure 4(c)). The combinations of systemic therapy types are shown in Figure 4(c). Of the unresected patients with Stage III NSCLC at initial NSCLC diagnosis (N = 78), 30.8% (24/78) were treated with systemic therapy only, whereas the remaining 69.2% (54/78) were treated with systemic therapy and RT. The most common reasons for not performing primary tumor surgical resection among patients with available data (N = 66), in descending frequency order were as follows: “stage III unresectable disease” in 56.1% (37/66), patient’s health and functional status in 30.3% (20/66), “patient’s refusal” in 10.6% (7/66), and other reasons for the remaining two patients (3%). Among reasons of health and functional status, the most common were “comorbidity” [in 12.1% (8/66) of the patients; with COPD being the most commonly reported comorbidity (4/8)], “poor cardiorespiratory fitness” [in 7.6% (5/66)], and “poor performance status” [in 4.5% (3/66)].

Of the resected patients, 63.9% (76/119) were treated with systemic therapy (Figure 4(C)); of these, 84.2% (64/76) were treated with adjuvant, 9.2% (7/76) with neoadjuvant and 6.6% (5/76) with both neoadjuvant and adjuvant systemic therapy. Systemic therapy comprised mostly of ChT (97.4%; 74/76) (Figure 4(d)). The most common reasons for not prescribing adjuvant systemic therapy were “stage IA disease” (50.0%; 25/50), “stage IB disease with primary tumor ≤4 cm” (22.0%; 11/50), “patient preference” (14.0%; 7/50), “receipt of neoadjuvant therapy prior to surgery” (8.0%; 4/50), and other (6.0%; 3/50).

Systemic therapy was administered in combination with RT in 23.7% (18/76) of systemic therapy treated resected patients, as part of CRT in 77.8% (14/18) of these cases. The combinations of systemic therapy types are shown in Figure 4(d). Systemic therapy in the adjuvant setting comprised of ChT (± anti-VEGF or IO) in all but two patients who received an EGFR TKI only.

3.5. Patterns of initial management strategy

Patterns of therapeutic modality sequences as part of initial management strategy overall and in the subpopulations by disease stage and substage at initial diagnosis and primary tumor resection status are presented in Figure 5. The most frequent treatment sequences recorded among NM-NSCLC patients were: ChT in combination with RT in 26.7% (54/202) of the patients; surgery followed by ChT in 24.8% (50/202) of the patients [86.0% (43/50) of these were Stage II-IIIA]; surgery only in 20.8% [42/202; 85.7% (36/42) of these were Stage I]; and ChT alone in 11.9% [24/202; 70.8% (17/24) of these were Stage IIIB-C]. CRT was followed by IO in the consolidation or maintenance setting in 21.7% (13/60) of the CRT cases. A trimodality approach (surgery plus systemic therapy plus RT, in any order) was undertaken in 8.9% (18/202) of the patients overall, which comprised 19.0% (11/58) of the Stage IIIA NSCLC patients.

Figure 5.

Frequencies of therapeutic modality sequences as part of initial management strategy, in the overall NM-NSCLC population and by disease stage/substage at initial diagnosis and primary tumor resection status as initial management strategy.

Values inside bars indicate proportion of patients receiving the indicated therapeutic modality sequence (%); only values ≥ 5.0% are shown. The therapeutic modality sequences shown exclude information on consolidation/maintenance-IO; CRT was followed by consolidation/maintenance-IO in 23% of CRT-treated Stage III patients (13/57), two of whom had also undergone surgery.

^aThe patient with Stage III NSCLC but unknown substage belonged to the treatment group “CRT w/o surgery.” ^bOne patient received SYS in the neoadjuvant setting but did not proceed to surgery. ^cFor two patients (both Stage IIIA NSCLC), SYS also preceded SUR hence the setting of SYS was also classified as neoadjuvant. For four patients the following sequence was recorded SUR ➔ SYS (chemotherapy) ➔ RT, but was not specified as “CRT” by the investigators. Categories merged due to limited sample size. ^dIncluding three patients (two with Stage IIIA and one with Stage IIIB NSCLC) for whom SYS was also given after RT. Categories merged due to limited sample size. ^eIncluding three patients (one Stage IA and two Stage IIIA NSCLC) for whom SYS was also given after surgery. Categories merged due to limited sample size. ^fFor one patient (Stage IIIC NSCLC) the following sequence was recorded: SYS ➔ RT (CRT) ➔ SYS (consolidation) ➔ SYS (neoadjuvant) ➔ SUR; the setting of SYS for this patient was classified as “Other” hence not included in the calculations of neoadjuvant rates. ^g“RT only” for one Stage IB and one Stage IIB patient; “SUR ➔ RT” for one Stage IIIA patient. ^hSYS did not include ChT in a total of 5 cases (i.e., all SYS was ChT ±IO/TT). ChT:Chemotherapy; CRT: Chemoradiotherapy; IO: Immune-Oncology (therapy); n: number of patients with variable; N: total number of patients with available data; NM: Non-Metastatic; NSCLC: Non-Small Cell Lung Cancer; RES: resected; RT: Radiotherapy; seq.: sequential; SUR: surgery; SYS: systemic therapy; TT: Targeted Therapy; UNRES: unresected; w/o: without.

Differences in the patterns between Stage I/II and Stage III NSCLC closely resemble differences between resected and unresected patients, in line with the composition of the latter subpopulations in terms of NSCLC stage and substage. Specifically, 94.0% (78/83) of the unresected patients had Stage III at initial diagnοsis and the rest had Stage I/II, while 66.4% (79/119) of the resected patients had Stage I/II, 28.6% (34/119) had Stage IIIA NSCLC, and the remaining 5.0% (6/119) had Stage IIIB-C, with the Stage IIIA subpopulation showing the most variable treatment patterns across subpopulations by NSCLC substage.

3.6. Association of factors of interest with initial management strategy

The association of patient, disease and institution characteristics with initial management strategy of NM-NSCLC was examined with univariable and multivariable logistic regression analysis (Supplementary Figure S2 and S3). Multivariable regression analysis revealed that resected Stage I patients had 98% lower odds of receiving adjuvant therapy as part of initial management strategy than Stage II patients (Supplementary Figure S2B). Additionally, among Stage III patients, the factors significantly associated with receipt of CRT over ChT were sex, disease substage and institution type, by multivariable analysis (Supplementary Figure S3B).

3.7. Management upon locoregional progression

During the retrospective observation period, 10.4% (21/202) of patients experienced a locoregional and 32.3% (65/202) a distant progression event (Supplementary Figure S1B). Four of the patients with disease progression had been treated with CRT followed by IO in the consolidation or maintenance setting as part of the initial management strategy, with IO initiated more than 42 days after completion of CRT. Among CRT-treated unresected patients not receiving consolidation or maintenance IO, 15.6% of patients (5/32; excluding patients with ongoing CRT at study inclusion or death) experienced a progression event in ≤42 days after completion of CRT.

All but one of the patients experiencing a locoregional progression event overall (95.2%; 20/21) received any pharmacologic and/or non-pharmacologic treatment in the post-progression setting (Supplementary Figure S4A). Treatment involved “systemic therapy only” in 65.0% (13/20) of these patients, “both systemic therapy and RT” in six patients (30.0%; 6/20) and “RT only” in one patient (5.0%; 1/20). The management strategies, including frequencies and patterns of systemic therapy types, upon locoregional progression in the overall NM-NSCLC population and the subpopulations by NSCLC stage at initial diagnosis are shown in Supplementary Figure S4B.

4. Discussion

Data from the EVIDENCE study provide insight into the real-world patient profile and initial management strategies in patients diagnosed with NM-NSCLC in Greece between January 2018 and December 2020, a period which precedes any major changes/updates in both European and local guidelines and in available treatment options for NM-NSCLC, except for the introduction of durvalumab as consolidation therapy for LA unresectable disease [6,7,16,17]. By obtaining data from 202 patients treated at 7 leading oncology and pulmonology clinics across diverse geographic locations and with a balanced representation of the public-academic and private-non-academic sectors, the study provides a generalizable dataset.

Considering that nearly 50% of the NSCLC patients present with Stage IV disease at initial diagnosis [23–27], and extrapolating the NM stage frequencies found in the EVIDENCE dataset in the overall NSCLC population, it may be estimated that 12%, 9% and 29% of NSCLC patients in Greece are diagnosed with Stage I, II, and III disease, respectively. The frequencies of Stage I and II diagnoses lie within the range of 12% to 22% for Stage I, and 6% to 10% for Stage II reported in other retrospective studies across different European countries and study periods, whereas that of Stage III is slightly higher compared with the reported 19% to 26% [23–26,28,29]; nonetheless, stage distribution seems to not have changed considerably over time, with respective rates of evaluable patients in Greece from the EPICLIN-Lung study estimated at 14%, 9%, and 35% in 2009 [5]. Additionally, our results are aligned with a 25% of NSCLC (comprised of adenocarcinoma and SCLC) having been initially diagnosed at Stage I/II during the period 2015 to 2017 based on the results of a hospital-based registry in Greece [30]. Furthermore, despite advances in screening methods, nearly a fifth (18%) of NM-NSCLC diagnoses in EVIDENCE were incidentally detected, a third of which (34%) were Stage III disease which is often not amenable to curative resection. Altogether these underscore the value of systematic lung cancer screening as part of the preventive healthcare strategy to increase early—rather than late-stage cancer diagnoses and reduce lung cancer mortality [31].

In EVIDENCE the initial management strategy varied across NM-NSCLC stages, as expected, and was largely aligned with contemporary at the time ESMO treatment guidelines (2017) [6] which preceded the era of (neo)adjuvant ChT-IO and TT [7,17]. Surgery was the predominant initial treatment for Stage I/II patients. Increasing use of systemic therapy was observed with advancing stage, including perioperative systemic treatment for Stage II patients and systemic treatment alone or in combination with RT for Stage III patients. Consistently, systemic therapy was mostly administered in the primary setting in the Stage III and unresected subpopulations and consisted mostly of PBC, whereas systemic therapy in the Stage I/II and resected subpopulations comprised mostly of adjuvant ChT. This is again in accordance with ESMO 2017 guidelines where PBC was recommended for resected stage II and III NSCLC with the setting of choice being the adjuvant. Slightly less than half of the resected patients (42%) did not receive adjuvant therapy, mainly due to Stage I disease. Multivariable regression analysis results further supported that Stage II NSCLC patients were more likely to receive adjuvant therapy compared with Stage I NSCLC patients. Furthermore, 9% of resected patients were treated with neoadjuvant therapy, all of whom had Stage III NSCLC at initial diagnosis, which is aligned with a review of real-world studies conducted in the US, Europe, Canada and Asia, suggesting neoadjuvant and adjuvant ChT use increases with increasing stage [32]. EVIDENCE also showed that, in the real world, there is still a large proportion (41%) of the NM-NSCLC population who are unable to undergo surgical resection, with most frequent reasons being Stage III unresectable disease (in 56%), patient’s health and functional status (in 30%), and patient’s preference (in 11%). Nearly 70% of the unresected Stage III NSCLC patients in EVIDENCE were treated with ChT plus RT, corroborating the recommendations for CRT in this patient population [6]. This is important especially considering that this population represented slightly more than a quarter of the overall NM-NSCLC population (27%) and in view of the anticipated expanded use of, the newly approved at the time, durvalumab in this setting [7].

Epidemiological studies reporting on the NM-NSCLC setting in Europe are scarce and among the available studies there is no overlap of data collection periods with the EVIDENCE study index period [24,25,28,32,33] apart from a study conducted in the Netherlands (2017–2019) [23] and one in Austria (2018–2021) [29]. Temporal variations in management strategies [24,25], as well as potential differences in treatment availability, definitions (such as CRT over ChT plus palliative RT) [25], local guidelines and reimbursement policies or reforms across countries, render any detailed discussion of similarities or differences with the previously published literature challenging.

Focusing on the aforementioned Dutch study of nation-wide registries and only on the 2018–2019 period, which overlaps with the EVIDENCE data, most frequent modalities among Stage 0-II NSCLC patients (Stage 0 patients representing only ~ 3%) were surgery or RT, in slightly less than 50% each, which differs compared with the vast majority (94%) of Stage I/II NSCLC patients in EVIDENCE who underwent surgery and the < 10% who were treated with RT [23]. Moreover, comparing the treatment modalities reported in EVIDENCE with earlier RWE European studies in the UK (index period: 2013–2017) [24], Denmark (2010–2016) [25] and Portugal (2015–2016) [28], and the 2009 EPICLIN-Lung study data (which, though old, is relevant due to inclusion of Greek data among other European countries) [5], a higher proportion of treated Stage I/II NSCLC patients had undergone surgery (94% versus a range of 37% to 68%) and a lower proportion had been treated with RT (7% versus > 23%). On the other hand, surgery and RT rates among Stage I/II in the more recent Austrian study were 83% and 16%, respectively [29], which are closer to the respective EVIDENCE rates. Of note, the rate of adjuvant systemic therapy among resected Stage I/II patients was almost double in the EVIDENCE cohort compared with the Austrian study (48% vs 25%) [29]. Overall, 12% of the EVIDENCE Stage I/II cohort did not receive standard therapies (five Stage I/II patients did not undergo surgery and five Stage II patients did not receive perioperative systemic therapy) [6].

Stage III NSCLC accounts for more than half of the non-metastatic and nearly a third of the overall NSCLC, with a balanced distribution between Stage IIIA and Stage IIIB-C patients, based on the results of the EVIDENCE dataset of Greek patients. Among LA NSCLC patients, the distribution of patient and disease characteristics is highly heterogeneous, definitions of resectability vary greatly across countries, guidelines and physician specialties and the range of treatment options is increasingly complex, hence no single optimal combination and sequence of modalities exists [34–37]. Findings of the EVIDENCE study showed that slightly less than two-thirds of Stage IIIA NSCLC patients underwent surgical resection (59%) and a trimodality approach was undertaken in a third of these patients (32%). The most variable patterns of treatment modalities in EVIDENCE were observed among Stage IIIA NSCLC patients, while the vast majority (90%) of Stage IIIB-C NSCLC patients did not undergo surgery as part of initial management strategy and mostly received systemic therapy either alone or in combination with RT. Resectability among Stage IIIA NSCLC patients was higher than the previously reported range of 15% to 37% in earlier European studies [24,25,28,33,38], whereas resectability for Stage IIIB was similar to previously reported rates of <11% [24,25,38]. With regard to the observed resectability rate among Stage IIIC NSCLC patients (25%), this cannot be appraised given the low number of patients in this subset (N = 12). Though the more recent Dutch and Austrian studies and the older EPICLIN-Lung study did not provide data per substage [5,23,29], it is noted that the proportion of CRT-treated Stage III NSCLC patients was similar between EVIDENCE and the Dutch cohort (i.e., approximately 50%) and higher than the Austrian cohort (29%), while 44% of the unresected Stage III EVIDENCE population did not receive CRT, which is the preferred option for unresectable stage IIIA/IIIB disease [6,7].

In terms of CRT sequencing, cCRT is the preferred approach based on evidence-based guidelines [6,7]. However, a third of the CRT-treated Stage III NSCLC patients in EVIDENCE was administered sCRT. Definitions of definitive versus palliative RT and timing of RT in relation to ChT lied with the investigator, which may in part underlie deviations of real-world treatment patterns from clinical practice guidelines. Institutional variations may also play a part, considering the fact that “non-academic institution” favored receipt of CRT among Stage III NSCLC patients based on the multivariable analysis performed. Notably, systemic treatment of Stage III NSCLC patients was initiated a median of 1.2 months after initial NSCLC diagnosis, while RT was initiated a median of 3.5 months, indicating a delay in initiation of RT. The delay in RT is unlikely to be driven by palliative or adjuvant RT among resected patients since, firstly, the delay was also evident among unresected patients (94% being Stage III), with a median of 1.1 months until systemic therapy and 3.4 months until RT; and secondly, nearly 80% of RT cases in EVIDENCE were defined as “CRT” by the investigator. These observations, in addition to the low rates of RT among Stage I/II patients compared with other European RWE studies described above, imply that access to certain treatment modalities may not be unified across the country and may impact treatment decision-making. Cross-country variations in definitions of cCRT also exist according to a 2020 survey of thoracic oncologists in different European countries (including Greece) with a focus on unresectable Stage III NSCLC [37]. In addition, “RT delay” (i.e., “RT is not immediately available”) comprised one of the reasons for administering induction ChT before CRT, while “access to therapy” was one of the factors that impacts decision to administer CRT [37]. The above observations altogether suggest that lack of standardized healthcare provision in NSCLC may in part underly heterogeneity in medical practices.

Systemic treatment other than ChT was observed in a minority of patients in EVIDENCE, as anticipated since the EVIDENCE study 3-year index period preceded the availability of TT and IO therapies in the NM-NSCLC setting apart from durvalumab in unresected Stage III NSCLC patients after cCRT. Although durvalumab was reimbursable in Greece during the study period [37], less than a quarter of CRT-treated Stage III NSCLC patients received IO in the consolidation/maintenance setting (23%; 13/57). The low proportion of patients receiving the newly approved at the time IO therapy [6,7], may be speculatively attributed to several factors. Past experience in the metastatic setting has shown that PD-L1 expression can be a predictive biomarker for NSCLC patients treated with such therapies [16,39,40], potentially influencing treatment decisions in the early days after approval in the NM setting. In addition, a lag in publications of real-world effectiveness and safety outcomes of IO use from the time of authorization may have led to an initial resistance to treatment adoption. Substantial RWE on this regimen is now available [41–43], with less restrictive eligibility criteria, also demonstrating that IO can be effective even when administered beyond 42 days after CRT completion [41,43,44]. Prior sCRT has also been observed in the real world [41,42]. In fact, eligibility criteria of the pivotal clinical trial have been barriers to implementing durvalumab in clinical practice according to the European survey of thoracic oncologists [37], among whom a Greek oncologist suggested that, at the time of the survey (2020), durvalumab implementation was restricted to PD-L1 positivity and cCRT [37]. Last but not least, potential delays and administrative challenges in the reimbursement application process through the national electronic pre-authorization medicines system may have served as barriers to adoption of IO therapy. International literature suggests that the uptake of IO in the NM-NSCLC setting has increased over the years [23,44,45], thus data from the present study will aid in assessing future local trends. As more IO therapies and combinations are approved, the evolving practices in cancer care should continue to be assessed in order to ensure the uptake of optimal treatments and uncover unmet treatment needs, which warrant further research.

The EVIDENCE study also provides updated key demographic and clinical characteristics of the NM-NSCLC population in the real-world, which often differ from clinical trials. In view of the newly approved but also upcoming innovative treatments in NSCLC [17–22], data from the present study can aid in identifying the local and current patient population eligible for such treatments. The molecular profile of patients is increasingly important, including mutation and expression status of the druggable targets EGFR and PD-L1, respectively. A small proportion of NM-NSCLC patients were positive for EGFR in the tested EVIDENCE cohort (6%), which is lower than the range of 11% to 16% reported in earlier Greek studies [46–51]. Importantly, despite the fact that EGFR incidence has been reported to be similar across stages, big differences are documented in the current study between different stages. Nevertheless, the EVIDENCE EGFR-positivity rate cannot be considered robust, since results were based on a surgical specimen in only a quarter of the EGFR-tested patients. EGFR positivity may also be affected by other parameters, such as type of technology applied for the EGFR testing (single vs NGS testing), tissue quality & quantity (namely preanalytics & tissue stewardship) and poorly regulated diagnostic environment. The PD-L1 positivity rate in the EVIDENCE cohort was 48%, which is high compared to an earlier rate reported for Greek patients (36%) [52] but similar to a European cohort of Stage IIIB-IV (52%) [53] and lower than the recent Austrian real-world cohort of Stage I-III NSCLC (58%) [29]. Though disease stage distribution may differ across studies, this is not expected to affect prevalence estimates based on available literature [29,54,55] as well as our results showing a similar PD-L1 positivity rate between Stage I/II and Stage III NSCLC patients. Noteworthy, the PD-L1 testing rate among Stage III NSCLC patients was three times higher than the rate among Stage I/II patients, as expected given that only Stage III NSCLC patients had IO options available at the time. Testing rates were very low in certain cases, e.g., only a fifth of Stage I/II patients were tested for the EGFR or PD-L1 markers, rendering the estimation of patients who would benefit from targeted (neo)adjuvant treatment unreliable. Low EGFR testing rates may be partially due to later EMA approval of adjuvant EGFR TKI based on data from the ADAURA study, and the necessity of biomarker testing. Nonetheless, these estimates are potentially useful in understanding future trends in biomarker testing and in implementing corrective actions. The recent approval of genotype-matched therapeutics in the NM-NSCLC setting, the recent incorporation of the guideline-directed biomarker testing of PD-L1 and EGFR in the diagnosis and staging work-up, as well as the establishment of national programs for molecular testing [7], together may lead to an increase in the uptake of such testing in clinical practice which is necessary for the optimization of personalized treatment of NM-NSCLC.

Newer and improved treatments in the NM setting of NSCLC are much anticipated, given that median progression-free survival and time to recurrence of 1–2 years have been reported in the pre-IO and TT era [56,57]. In the present study, with median follow-up of 16 months in the NM setting, approximately 40% of the overall study population experienced a locoregional and/or distant progression event, with a third of the patients (32%) having advanced to Stage IV NSCLC at the end of the study observation period. These rates were almost double among Stage III over Stage I/II NSCLC patients, in line with the lower proportion of resected Stage III NSCLC patients compared with earlier stages (34% versus 94%). On the other hand, with a median follow-up of 17 months after surgery, nearly a third of resected patients experienced a progression event (30%), most of which were distant events (78%). This was in spite the fact that the majority of cases had undergone complete resection of the primary tumor, and the high ratio of earlier versus later initial stage of NSCLC among the resected patients in EVIDENCE (2:1). Treatments in the NM post-progression setting were also described as part of an exploratory endpoint of the EVIDENCE study, showing that almost half of treated patients (45%) received ChT-based treatment, a quarter were treated with CRT, another quarter with IO only, and, lastly, a minority (10%) with RT only. Albeit the small sample size and both recurrence and progression events being included in the present analysis, these observations are largely in line with ChT and CRT being the most commonly reported treatments upon locoregional recurrence in the real-world [57]. Considering that there is no clear guidance on therapies upon progression [6,10], and that re-recurrence rates range from 35 to 72% [57], the post-progression setting of NM-NSCLC represents an area of unmet need.

4.1. Limitations

As with any retrospective cohort study which is observational in nature, there are several limitations, including patient selection, confounding and information bias. As described in the study methods, certain measures have been taken, such as consecutive enrollment, to mitigate these limitations. In spite of relying on chart review of existing data, missing data did not exceed 20% for any of the variables with the exception of tumor differential grade (29.2%) and reasons for not performing surgery (20.5%), which should be taken into account in the interpretation of findings. The date of death was unknown for eight of the deceased patients, which however did not impact the estimation of the length of the retrospective period since imputed and as-observed analyses both resulted in 15.8 months (data not shown). The originally planned sample size was met, and available observations of interest were sufficient to address the primary outcome measures of initial NM-NSCLC management strategies in both the overall population and subpopulations by resection status and disease stage. On the other hand, certain NM-NSCLC substage subpopulations were represented by a limited number of patients (e.g., Stage IB: N = 18; Stage IIA: N = 13; Stage IIIC: N = 12), which could impact the precision of the estimates of the relevant outcome measures. In addition, in the absence of statistical significance tests, any between-subpopulation comparisons are purely descriptive. Furthermore, interpretation of the exploratory study outcomes of disease progression and molecular profile findings should take into consideration that as this was a real-world study no specific response assessment criteria or uniform laboratory testing, respectively, were enforced. In addition, given that only 21 patients experienced a locoregional progression event while on study (i.e., prior to distant metastasis, ICF signing or death), data on the post-progression management strategy in the context of NM-NSCLC setting are limited. Nevertheless, considering the adequate representativeness of the target population of NM-NSCLC patients in the country and that we captured data representing routine clinical care, not confined by strict and narrow clinical trial eligibility criteria, this study provides a robust dataset describing real-world patient, disease and treatment characteristics for NM-NSCLC.

5. Conclusion

This study of a cohort of 202 NM-NSCLC patients, comprising 41.6% Stage I/II and 58.4% Stage III patients showed that European guideline recommendations are largely taken up in routine care settings in Greece. Almost half of the patients underwent any NSCLC-related molecular testing at initial diagnosis, with overall positivity rates for EGFR mutation and PD-L1 expression documented at 6.3% and 47.7%, respectively. Surgery was the predominant initial management strategy for Stage I/II patients. Slightly less than two-thirds of Stage IIIA patients underwent surgery, while for Stage IIIB-C patients ChT alone or in combination with RT was the most frequent therapy. Data from the present study come to fill the gap of the currently limited and outdated RWE on local NM-NSCLC patient, disease and treatment characteristics. The findings of this study will serve as a baseline for future analyses investigating the changing treatment landscape as recently approved ICIs and late generation TKIs are incorporated into routine clinical practice in Greece.

Funding Statement

This work was supported and funded by AstraZeneca. S.A. A.S., C.P., A.G., A.N., Z.P., and F.P. are employees of AstraZeneca.

Article highlights

• Real-world findings indicate that among patients diagnosed with NM-NSCLC between 1 January 2018 and 31 December 2020 in Greece, 23.3%, 18.3% and 58.4% were diagnosed with Stage I, II, and III disease, respectively.

• NSCLC was detected incidentally for purposes other than lung cancer screening in almost a fifth (18%) of the patients.

• Among the overall NM-NSCLC study population, EGFR mutation and PD-L1 expression were tested in 23.8% and 43.6%, respectively, with 6.3% and 47.7% overall positivity rate, respectively.

• There is still a large proportion (41%) of the NM-NSCLC population who are unable to undergo surgical resection, with the commonest reason being Stage III unresectable disease (in 56%). Resectability rates were 94%, 59% and 10% for Stage I/II, IIIA and IIIB-C NSCLC patients, respectively; 42% of the resected patients were not prescribed adjuvant therapy, mainly due to Stage I disease.

• Increasing use of systemic therapy was observed with advancing stage, including perioperative systemic treatment for Stage I/II patients (49%) and systemic treatment alone (20%) or in addition to RT (46%) for Stage III patients.

• The proportion of patients treated with RT was 7% and 59% among Stage I/II and Stage III patients. Nearly 70% of the unresected Stage III NSCLC patients in EVIDENCE were treated with ChT plus RT.

• The overall rate of progression was 39% (distant metastasis: 32%) and was nearly double in the Stage III NSCLC subpopulation compared with the Stage I/II NSCLC subpopulation (any event: 49% versus 25%; distant event: 40% versus 21%).

• Overall, the findings were generally consistent with European guidelines recommendations, and will be useful in understanding the uptake of new and innovative treatments in the future in this quickly evolving treatment landscape.

Author contributions

(I) Conception and design: A.S., Z.P.; (II) Administrative support: A.S., C.P., A.N., Z.P., F.P., A.K.; (III) Provision of study materials or patients: G.M., S.L., HL., V.G., D.M., S.A., A.C., D.S., M.L., EG.S., A.S., C.P., A.N., Z.P., F.P., A.K., K.N.S.; (IV) Collection and assembly of data: G.M., S.L., HL., V.G., D.M., S.A., A.C., D.S., M.L., EG.S., A.S., C.P., A.N., Z.P.,F.P., A.K., K.N.S.; (VI) Manuscript writing: G.M., H.L., V.G., S.A., A.S., C.P., A.N., F.P.; (VII) Final approval of manuscript: G.M., S.L., HL., V.G., D.M., S.A., A.C., D.S., M.L., EG.S., A.S., C.P., A.N., Z.P., F.P., A.G., K.N.S. All authors have read and agreed to the published version of the manuscript.

Disclosure statement

G.M. reports personal fees from AstraZeneca, during the conduct of the study. V.G. reports research grant from AstraZeneca, outside the submitted work. K.N.S. has received honoraria for participation in advisory boards and speakers’ bureau for AstraZeneca, Bristol Myers Squibb and Merck Sharp & Dohme. E.L. has received consulting fees from Roche, Novartis, Bristol Myers Squibb, Merck Sharp & Dohme, AstraZeneca, Takeda, GlaxoSmithKline, Merck, Amgen, Boehringer, Pfizer and Lilly and honoraria for lectures from Roche, Novartis, Bristol Myers Squibb, Merck Sharp & Dohme, AstraZeneca, Takeda, GlaxoSmithKline, AbbVie, Amgen, Boehringer Ingelheim, Pfizer and Lilly. A.C has received consulting fees and honoraria for lectures from AstraZeneca, Janssen, Bristol Myers Squibb and Merck Sharp & Dohme. D.M., S.A., D.S. and M.L. have received consulting fees and honoraria for lectures from AstraZeneca. E.S. has received honoraria for lectures from Roche, Novartis, Merck Sharp & Dohme, AstraZeneca, Amgen and Pfizer. A.S., C.P., A.G., A.N., Z.P., and F.P. belong to the staff of AstraZeneca. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Medical writing and editorial support were provided by Athena Georgilis of OPTIMAPHARM GREECE S.A. (CRO) and were funded by AstraZeneca.

Ethical conduct of research statement

The study was designed, conducted and reported in accordance with the ethical principles laid down in the Declaration of Helsinki, the Guidelines for Good Pharmacoepidemiology Practice (GPP) of the International Society for Pharmacoepidemiology, the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines where applicable, the EU General Data Protection Regulation (GDPR), and the local rules and regulations. All patients alive at the time of study conduct were required to provide a written informed consent to study participation. All participating sites were granted a waiver of consent for deceased patients by the Institutional Review Board of the study site.

In accordance with the current local regulatory requirements, the applicable study protocol and the patient’s Informed Consent Form (ICF), were reviewed and approved in writing by the competent site-specific Scientific Committees and/or Administrative Councils of the participating hospitals before the enrollment of any patient into the study and the performance of any study-related procedures. All Ethics Committees of the following participating hospitals approved the study protocol: Henry Dunant Hospital Center of Athens, General Hospital “G. Papanikolaou” of Thessaloniki, Metropolitan Hospital of Athens, Metropolitan General Hospital of Athens, University Hospital of Heraklion, University Hospital of Alexandroupolis, Sotiria Thoracic Diseases Hospital of Athens. Approval by a central Ethics Committee is not applicable for non-interventional studies in Greece.

Data sharing statement

The data presented in this study are contained within the article. The data are not publicly available due to restrictions that apply to the availability of the data (e.g., privacy or ethical). Datasets from this study may be available upon request from the corresponding author and provided upon approval from the sponsor and in accordance with data privacy and ethical provisions.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14796694.2024.2442295