European guidelines for the surgical management of pure ground-glass opacities and part-solid nodules: Task Force of the European Association of Cardio-Thoracic Surgery and the European Society of Thoracic Surgeons

TABLE OF CONTENTS

• ABBREVIATIONS AND ACRONYMS  2

• ABSTRACT  3

• PREAMBLE  3

• 1 INTRODUCTION  4

•  1.1 Definition of GGO and nodule classification  4

•  1.2 Characteristics of subsolid nodules  4

• 2 METHODOLOGY  4

• 3 BACKGROUND  5

•  3.1 Definition of a ground-glass opacity  6

•  3.2 Assessment of the malignant risk of a subsolid nodule  8

• 4 MANAGEMENT OF SINGLE PURE GROUND-GLASS OPACITIES (NON-SOLID NODULES)  12

•  4.1 The role of a non-surgical biopsy  12

•  4.2 Surgery in pure GGO  12

•  4.3 Lymph node strategy  12

• 5 MANAGEMENT OF THE PART-SOLID NODULE  13

•  5.1 The role of a non-surgical biopsy  13

•  5.2 Mutational status  14

•  5.3 Lymph node strategy  14

• 6 MANAGEMENT OF MULTIPLE GROUND-GLASS NODULES  15

• 7 HOW TO IDENTIFY A NON-PALPABLE GROUND-GLASS OPACITY  15

•  7.1 Preoperative techniques for localization and marking  15

•  7.2 Injectable agents  16

•  7.3 Intraoperative techniques for localization and marking  17

• 8 MINIMALLY INVASIVE SURGERY APPROACH  17

• 9 SUBLOBAR VERSUS LOBAR RESECTION  17

•  9.1 Pulmonary function  18

•  9.2 Oncological outcomes  18

• 10 FUTURE DIRECTIONS AND GAPS IN KNOWLEDGE  19

• 11 KEY MESSAGES  19

• SUPPLEMENTARY MATERIAL  19

• CONFLICTS OF INTEREST  19

• ACKNOWLEDGEMENTS  20

• FUNDING  20

• AUTHOR CONTRIBUTIONS  20

• Reviewer information  20

• REFERENCES  20

ABSTRACT

Pulmonary nodules are often encountered by respiratory physicians, radiologists and thoracic surgeons alike. The European Association of Cardio-Thoracic Surgery and the European Society of Thoracic Surgeons established a multidisciplinary collaboration of clinicians (task force) with expertise in managing pulmonary nodules with the aim of producing the first joint comprehensive review of the scientific literature, with a specific focus on the management of pure ground-glass opacities and part-solid nodules.

The scope of the document has been defined by the European Association of Cardio-Thoracic Surgery and European Society of Thoracic Surgeons governing bodies and focuses on 6 main areas of interest agreed upon by the task force. These include managing solitary and multiple pure ground-glass nodules and solitary part-solid nodules and identifying the non-palpable lesion, the role of minimally invasive surgery and the decision making behind sublobar versus lobar resection.

The literature revealed that, with the increasing use of incidental computed tomography scans and computed tomography lung cancer screening programmes, the detection of early-stage lung cancer is going to increase, with a higher number of potential cancers presenting on the ground-glass and part-solid nodule spectrum. Given that the gold standard for improved survival is surgical resection, there is an urgent need for comprehensive characterization of these nodules and for guidelines that are directed towards their surgical management. It is recommended that standard decision-making tools be used to determine the risk of malignancy and thus guide referral for surgical management and that decisions for surgical resection be made in a multidisciplinary setting with equitable consideration given to radiologic characteristics, the evolution of the lesion, the presence of a solid component, patient fitness and patient co-morbidities. Given the recent surge in robust level I data comparing sublobar and lobar resection with the release of Japanese Clinical Oncology Group 0802 and Cancer and Leukemia Group B 140503 data, a global overview of an individual case must be adopted into clinical practice. This set of recommendations is based on the available literature; however, close collaboration in the design and conduct of randomized controlled trials is still of utmost importance to answer further questions in this rapidly evolving field.

PREAMBLE

Clinical practice guidelines summarize and assess all relevant evidence on a specific topic at the time of their creation, with the goal of assisting physicians in selecting the best management strategies for individual patients with a given condition. These guidelines take into consideration the impact on patient outcomes as well as the risk–benefit ratio of different diagnostic or therapeutic methods. Although these guidelines do not replace textbooks, they complement them and cover topics pertinent to contemporary clinical practice. They serve as a vital tool to aid physicians in making decisions in their daily practice. However, in essence, although these recommendations serve as a valuable resource to guide clinical practice, their application should always be tailored to the needs of the individual patient. Each patient's case is unique, presenting its own set of variables and circumstances. The guidelines are a tool designed to support, but not supersede, the decision-making processes of physicians, based on their knowledge, expertise and understanding of their patients' individual situations. Furthermore, these guidelines should not be interpreted as legally binding documents. The legal responsibilities of healthcare professionals remain firmly grounded in applicable laws and regulations, and the guidelines do not alter these obligations.

The European Association for Cardio-Thoracic Surgery (EACTS) and the European Society of Thoracic Surgeons (ESTS) selected a task force composed of professionals working in the field of this particular pathological condition. In an effort to maintain transparency and uphold integrity, all experts involved in the development and review of these guidelines provided declarations of interest, detailing any possible conflicts. These declarations have been collected into a single file available on the EACTS website (https://www.eacts.org/resources/clinical-guidelines). Any changes to these declarations during the writing process must be immediately reported to the EACTS and ESTS. The EACTS and ESTS provided all financial support for this task force, with no involvement from the healthcare industry.

Building upon this collaborative work, the clinical practice guidelines committees of the EACTS and ESTS oversaw the creation, refinement and approval of these new guidelines. A comprehensive review of the draft was carried out by an external panel of experts in the field. Their feedback informed the necessary revisions. After this thorough review and updating process, the final document received approval from all the experts on the task force and the governing bodies of the EACTS and ESTS. This approval made it possible for the guidelines to be published in the European Journal of Cardio-Thoracic Surgery.

These guidelines, endorsed by both the EACTS and the ESTS, represent the official viewpoint on this topic. They show a commitment to ongoing improvement, as regular updates will be made to keep the guidelines relevant and useful in the constantly evolving field of clinical practice.

1 INTRODUCTION

1.1 Definition of GGO and nodule classification

Ground-glass opacity (GGO) is a descriptive term referring to an area of hazy increased attenuation in the lung on computed tomography (CT) scans with preservation of bronchial and vascular structures. The term ground glass derives from the industrial technique in glassmaking whereby the surface of the normal glass is roughened by grinding it [1].

Pulmonary nodules have been classified as solid nodules (with no GGO component) and subsolid nodules (SSNs) (with a GGO component). Subsolid nodules are staged pure GGOs (non-solid nodules) and part-solid nodules [2].

According to the Fleischner Society Recommendations 2017 and the American College of Radiology Lung Imaging Reporting and Data System (Lung-RADS) 2022, all CT scans of the thorax in adults should be reconstructed and archived with contiguous thin sections (< 1.5 mm, typically 1.0 mm) to enable accurate characterization and measurement of small pulmonary nodules using lung window images. To calculate the nodule mean diameter [3], one should measure both the long and short axes to 1 decimal point in millimetres and report the mean nodule diameter to 1 decimal point. The measurements of the long and short axes may be in any plane to reflect the true size of the nodule. Volumes, if obtained, should be reported to the nearest whole number in mm³. The mean diameter is the mean of the longest diameter of the nodule and its perpendicular diameter.

Using the consolidation's maximum diameter divided by the nodule's maximum diameter, the consolidation-to-tumour ratio (CTR) has been obtained. This parameter has proven to be essential in the characterization of SSNs [4].

Using CTR, lung nodules can be classified as follows (Fig. 1):

• Pure ground-glass opacity (non-solid nodules): CTR = 0;

• Part-solid nodule: 0 < CTR < 1;

• Solid nodule: CTR = 1.

1.2 Characteristics of subsolid nodules

Previous articles have stated that lung adenocarcinomas with a CTR ≤ 0.5 could be non-invasive [5–7]. Even though SSNs are more likely to be malignant compared to incidentally detected solid nodules, subsolid lesions have a better prognosis because they may correspond to preinvasive and invasive lesions with a slow progression. Pure GGOs (especially those < 5 mm) are generally non-invasive lesions and are histologically represented by atypical adenomatous hyperplasia or adenocarcinoma in situ (AIS). Part-solid lesions are minimally invasive lesions corresponding to minimally invasive adenocarcinoma (MIA) or invasive lesions (lepidic-predominant or non-lepidic-predominant adenocarcinoma) (Table 1, Supplementary Table 1) [8, 9].

Given the recent trial data [6, 10] on outcomes following surgical resection of subsolid nodules less than or equal to 2 cm, now is an opportune time to formally address how this heterogeneous set of lesions should best be managed.

2 METHODOLOGY

To provide clinical practice guidelines for healthcare professionals involved in the surgical management of GGO, a task force of internationally recognized experts in the fields of thoracic surgery, clinical epidemiology and biostatistics was selected by the governing bodies of the EACTS and the ESTS following the processes detailed in the EACTS methodology manual for clinical practice documents [11].

The EACTS/ESTS strived to ensure diversity in forming the writing group and adequate transparency in disclosing any relationships with industry and other entities. The chairperson was entirely free of relevant conflicts of interest (COI) from 1 year before the task force was formed until the publication of the document. Disclosure of any COI was required from the other task force members and the peer reviewers if any change occurred, following the COI policies of both societies [11]. After the EACTS had reached a consensus on the project's scope, the task force committee developed the final table of contents, which received full support. A rapid systematic literature review was then conducted in collaboration with a biomedical information specialist (see Acknowledgements) to ensure that the sections and recommendations accurately reflected the most recent data available. A set of specifically developed search terms (Supplementary Material) was used. The approach was based on the standardized Population, Intervention, Comparison, Outcome and Time (PICOT) questions. To conduct a rigorous and comprehensive review, we collaborated with a team of experts, including an informatics medical specialist, a clinical epidemiologist (M.M.) and chapter leaders, to perform a systematic literature review. These experts actively participated in identifying and analysing the relevant studies, and their input was crucial in ensuring the review's thoroughness and accuracy. Additionally, all authors contributed their expertise and insights to the review process by actively discussing applicable studies. Crucial evidence was summarized in the newly developed EACTS/ESTS evidence tables. To maintain the relevance of the clinical practice guidelines to modern clinical practice, we focused on more recent data for our evidence synthesis while also including essential publications, regardless of their publication age. The present document did not include studies in languages other than English. After methodological quality was assessed with attention to the study type and quality, prioritizing randomized controlled trials (RCTs), meta-analyses of RCTs if available over observational studies and case series, treatment recommendations and explicative text were written following the process defined in the EACTS Methodology Manual for clinical guidelines [11]. The medical evidence was critically appraised for quality by writing task force members with the assistance of a clinical epidemiologist if needed.

The task force members collaborated closely on writing all the chapters. According to the EACTS/ESTS policies for dealing with COI, each task force member was asked to report any change in COI immediately before meeting and voting and was allowed to vote on expert statements only in the absence of a COI for the particular topic. Although the voting margin was 75%, a total of 90% of statements were voted on by 100% of the task force members, whereas the remaining 10% of statements received 92% positive votes from task force members. Due to the high variability in economic parameters and the lack of data on cost-effectiveness, cost analyses were not considered or delivered. The drafted document underwent rigorous external validation by anonymous reviewers selected by the editor-in-chief of the European Journal of Cardio-Thoracic Surgery and 3 rounds of revision by the task force members before obtaining the approval for publication by the editor-in-chief of the European Journal of Cardio-Thoracic Surgery and the governing bodies of EACTS and ESTS.

The level of evidence (LoE) and the class of recommendation (COR) were weighted and graded according to predefined scales, as outlined in Tables 2 and 3. The COR denotes the strength of the recommendation by weighing health risks and benefits related to the particular intervention. The LoE reflects the quality of the evidence that supports the specific recommendation based on the quantity and consistency of data from clinical trials and other research studies. For each recommendation, the COR and LoE were critically appraised to account for the unique features of patients with GGOs to resolve the disparity between published evidence and its applicability to the general patient population.

3 BACKGROUND

The management of lung cancer is undergoing rapid evolution as multiple therapy modalities continue to evolve. Nevertheless, the gold standard for optimal survival for patients with this disease remains surgical resection at an early stage. To this end, there has been a transformation in our ability to perform parenchymal-sparing operations safely. There has been a shift in the diagnostic pathway for this condition, which does not routinely generate symptoms until much later in the disease process. The use of CT scanning for minimal symptoms and now of lung cancer screening for asymptomatic patients at risk of lung cancer is transforming the percentage of patients with lung cancer presenting at the earliest stage. This percentage has moved from a quarter of all patients to three-quarters of all patients presenting at an early stage, leading for the first time to significant improvements in the survival of patients with this condition, which is the leading cause of cancer deaths in Western society [10].

However, with this transformation in the number of CT scans being performed to look for evidence of lung cancer, we are faced with new diagnostic challenges as we identify large numbers of small abnormalities that may be at risk of developing into malignant processes.

In particular, SSNs, which include pure GGOs and PSNs, are found in approximately 17% of patients undergoing CT lung cancer screening [12]. It is important to note that both these categories of nodules are difficult to biopsy when small and thus most often remain a radiologic diagnosis while under surveillance. The range of histologic possibilities for these nodules includes benign inflammatory change, which will disappear on subsequent imaging. Atypical adenomatous hyperplasia, AIS and MIA are considered premalignant histologic diagnoses, whereas lepidic predominant adenocarcinoma and invasive adenocarcinoma are malignant conditions with the potential to metastasize.

Thus, clinicians faced with this massive number of SSNs require a framework to determine each nodule's risk and progression to malignancy so that they may present their patients with a range of options, including continued observation, needle biopsy or excisional biopsy.

In addition, whereas pure GGOs (non-solid nodules) have a low risk of malignancy, PSNs have a higher risk of malignancy than their solid counterparts. However, when malignant, they tend to exhibit more indolent behaviour, with slower growth and lower metastatic potential [13].

3.1 Definition of a ground-glass opacity

Before we embark on decisions concerning the management of these SSNs, it is essential to define what constitutes an SSN. The International Association of Staging in Lung Cancer has addressed the definition of a GGN in the 8th edition of the guidelines, which is a specific publication dedicated to SSNs and their classification [14]. They state, “In non-mucinous lung adenocarcinomas, ground glass versus solid opacities CT findings tend to correspond to lepidic versus invasive patterns seen pathologically. However, this correlation is not absolute; so, when CT features suggest non-mucinous AIS, MIA and lepidic predominant adenocarcinoma, the suspected diagnosis and clinical staging should be regarded as a preliminary assessment and are subject to revision after pathological evaluation of resected specimens.”

Part-solid adenocarcinomas present with a solid component and a GGO on CT scans. At the pathological examination, the solid component usually corresponds to the invasive part, and the GGO, to the lepidic part. To define the tumour (T) category by tumour size, only the size of the solid component on CT or the invasive component at the pathological examination is considered because the size of the solid/invasive component determines prognosis. However, documentation of both the size of the solid component/invasive part and the whole tumour, including the ground-glass and lepidic components in radiologic and pathologic reports, is recommended.

Because the literature on non-mucinous AIS deals with tumours smaller than 2 or 3 cm, there is insufficient evidence to support the occurrence of 100% disease-free survival in patients with tumours larger than 3.0 cm. Therefore, these tumours should be classified as lepidic predominant adenocarcinomas, and histologic sampling is essential for these tumours. Reports in the literature also state that the invasive size of SSNs is a better predictor of survival than total tumour size. Thus, for nodules less than 3 cm in total tumour size, the solid component alone is used to determine the T descriptor. See Figure 2 for a complete description of the staging of SSNs in the 8^th edition of the International Association for the Study of Lung Cancer (IASLC) guidelines [14].

In terms of how an SSN is measured, the IASLC notes that the Fleischner Society and the American College of Radiology Guidelines have recommended that the overall size of the ground-glass and/or solid components be based on the bidimensional average of the long and short dimensions to estimate risk. Still, for the Classification of Malignant Tumours (Tumor, Node, Metastasis) staging purposes, the single largest dimension measured on CT sections using thin sections (≤ 1.5 mm, typically 1.0 mm) and multiplanar reconstructions is used [15]. Following recommendations from the Flesichner Society and from the American College of Radiology, nodules [3], including the solid portion of PSNs, should be measured on lung windows using a high-spatial frequency (sharp) filter, although soft tissue windows can help when evaluating changes in nodule density over time [16]. Regardless, some authors believe it is best to measure the solid component on the mediastinal windows or on both the lung and mediastinal windows of the CT scan rather than on the lung windows [17, 18]. Controversies exist with regard to the optimal window setting. Certain studies have used electronic calipers to measure the diameters of all the SSNs on a lung window, whereas other studies reported on the use of both lung and mediastinal window settings to measure the SSN and calculate the tumour-disappearance ratio.

Increasing CTR is correlated with poorer survival, which again supports the policy of using the solid component of the nodule to assess the T stage [17]. Although pure GGOs must have a CTR ratio of zero, this ratio can be used to describe differences in the solid component of the PSNs and has been used extensively in the literature, including recently in the Japanese Clinical Oncology Group (JCOG)0802 trial, where a CTR ratio of more than 0.5 was an inclusion criterion for entry into this trial due to the prognostic impact of this criterion [6].

Regarding the equivalence of nodule size using either 2-dimensional measurements or 3-dimensional volumetric analysis, the United Kingdom Lung Screening trial defined 4 categories. Category 1 represents a benign nodule with a diameter of less than 3 mm, equivalent to a volume of less than 15 mm³. Category 2 corresponds to a diameter of 3–4.9 mm, which is equivalent to a volume of 15–49 mm³. Category 3 involves a diameter of 5–9.9 mm, which is equivalent to a volume of 50–500 mm³. Lastly, category 4 represents a diameter greater than 10 mm, which is equivalent to a volume exceeding 500 mm³ [19]. Volumetric analysis is considered superior to 2-dimensional analysis but requires software that may not be available to all institutions when making decisions on nodules.

There is clearly a difference in the prevalence of SSNs in Asia compared to Europe and North America. Yano et al. [20] in Japan documented the smoking status of 1400 of their patients over 30 years. They found that the incidence among never-smokers had risen from 13% to 30% of patients undergoing resection for non-small-cell lung cancer (NSCLC). In the JCOG0802 trial of more than 1000 patients who had primary lung cancer resections in Japan, 44% had no history of smoking [6]. The Flight Attendants’ Medical Research Institute pooled data from 13 cohort studies and 22 cancer registries to evaluate the incidence of lung cancer among never-smokers. Eight cohort studies were from North America or Europe, and 4 were from Asia. They found that Asians living in Korea and Japan (but not in the United States) had higher death rates from lung cancer than individuals of European descent and that lung cancer incidence rates were higher and more variable among women in East Asia than in other geographic areas with low levels of female smokers [21]. They also documented that lung cancer in non-smokers in North America is categorized as a ‘rare’ disease with fewer than 40,000 deaths per year or a cumulative risk of a non-smoker dying of lung cancer in the United States of 1.1% compared to 22% for a lifelong smoker.

Wang et al. compared lung cancer mortality in China to that in Australia [22]. Over 20 years, in Chinese males, the proportion of lung cancer deaths of all causes increased from 3.8% to 8.4%, whereas deaths of Australian males decreased from 7.5% to 6.8%. Smoking was the most critical risk factor for lung cancer among men in both countries in 1990 and 2019. Still, exposure to carcinogens in particulate matter was the second most important risk factor for lung cancer in Chinese men, but occupational exposure was most important in Australian men.

3.2 Assessment of the malignant risk of a subsolid nodule

The optimal management of PSNs has recently been investigated in several studies examining the implementation and management of lung-cancer-screening RCTs worldwide. These researchers performed low-dose CT scans on asymptomatic patients with various risk factors with long-term follow-up. They have amassed thousands of patients with a common prevalence of solid, part-solid and non-solid nodules. Thus, these reports best allow us to decide the optimal management of PSNs. Of note, there is no single optimal protocol: Every clinician and patient must weigh the risk that each nodule poses to the patient and the risk that an invasive test to obtain a histologic diagnosis of the nodule may pose. Both the Multicenter Italian Lung Detection (MILD) and the Dutch-Belgian lung cancer screening (NELSON) trials recently showed the characteristics of slow growth and indolent behaviour of SSNs. In the NELSON trial, the percentage of SSNs among 7135 volunteers was very low (3.3%), with a 63% disappearance rate at follow-up [23].

The MILD project is a lung cancer screening trial that prospectively randomized 4099 participants to a screening arm (n = 2376), with further randomization to annual (n = 1190) or biennial (n = 1186) low-dose computed tomography (LDCT) scans for a median period of 6 years or to a control arm (n = 1723) without intervention. The LDCT arm showed a 39% reduced risk of mortality at 10 years [hazard ratio (HR) 0.61; 95% confidence interval (CI) 0.39–0.95], compared with the control arm, and a 20% reduction of overall mortality (HR 0.80; 95% CI 0.62–1.03). The LDCT benefit improved beyond the 5th year of screening, with a 58% reduced risk of mortality (HR 0.42; 95% CI 0.22–0.79) and 32% reduction of overall mortality (HR 0.68; 95% CI 0.49–0.94) [31].

Silva et al. detected a total of 6541 nodules in 55.5% (1277/2303) of the screenees [11] and found SSNs in 16.9% (389/2303) of the screenees. During the 9.3 ± 1.2 years of follow-up, the HR of lung cancer diagnosis in subjects with SSNs was 6.77 (95% CI: 3.39–13.54), with 73% (22/30) of cancers not arising from SSNs (median time to diagnosis 52 months from identification of SSNs). Lung cancer-specific deaths in subjects with SSNs were significantly increased (HR = 3.80; 95% CI: 1.24–11.65) compared to subjects without lung nodules. Lung cancer arising from SSNs did not lead to death within the follow-up period [11].

These results showed the safety of active surveillance for conservative management of SSNs until signs of solid component growth appear and stressed the need for prolonged follow-up because of the high risk of other cancers [24].

A 2020 European Society of Radiology/European Respiratory Society statement paper on lung cancer screening suggests conservative management of SSNs [25].

Ricciardi et al. [26] recently published a review of the major European screening programmes, including the Italian lung study (ITALUNG [27]), the NELSON trial [28], the United Kingdom Lung Cancer Screening trial [29], the detection and screening of early lung cancer with novel imaging technology (DANTE) trial [30], the Danish lung cancer screening trial (DLCST) [31], the German lung cancer screening intervention trial (LUSI) [32], the MILD trial [33] and the CT screening for lung cancer study (COSMOS) [34]. The findings show that most SSNs retain an indolent course over many years: The decision-making process requires appropriate patient counselling and an understanding of the risk of malignancy, risk of treatment and expected benefits [26].

It should be noted, however, that a pure GGO represents a low-risk nodule, and a monitoring strategy may be optimal. For the 2022 update (Version 1), the National Comprehensive Cancer Network (NCCN) panel decided to continue using a cut-off for screen-detected nonsolid nodules of 20 mm and not to use the Lung-RADS 1.1 cut-off of 30 mm and recommended an earlier evaluation at 6 months for new non-solid nodules equal to or greater than 20 mm [2, 35].

In contrast, a PSN with both ground-glass change and a solid component often represents a higher risk nodule even than a solid nodule of comparable size. Still, it is interesting that these PSNs may have a more indolent nature than their solid counterparts, thus buying the clinician and patient the time to make a balanced decision as to the next steps in their pathway.

Several guidelines provide recommendations on managing pulmonary nodules based on the estimated risk of malignancy with several mathematical prediction models using both clinical and radiologic criteria [15, 36–38]. The Lung-RADS and NELSON-plus protocols appear to be the best prediction models [28].

McWilliams and co-workers published the first risk calculator—the Brock or the Pan Canadian Early Detection of Lung Cancer Study model—mathematically modelled to the outcome of screen-detected nodules in a large lung cancer screening trial [39]. In the Lung-RADS, published by the American College of Radiology, the Brock model was used to decide on the follow-up procedure in CT lung cancer screening. Lung-RADS is a classification proposed to aid with findings in low-dose CT screening examinations for lung cancer. It categorizes nodules into 4 risk categories, providing a percentage risk of malignancy and a proposed management plan for each category in order to standardize follow-up and management decisions. Similarly, the British Thoracic Society guidelines integrated the Brock model as a risk assessment tool for solitary non-calcified nodules [38].

The high performance of the Brock model in screening cohorts may not automatically be extrapolated to a clinical setting. Compared to routine clinical subjects, screening participants usually have more small nodules and are subject to specific inclusion and exclusion criteria (e.g. heavy smoking and age]. Nevertheless, a study by Chung et al. from the Netherlands recently showed the Brock model to have a very high negative predictive value capacity for discrimination of benign versus potentially malignant nodules in an unselected, heterogeneous clinical population [40].

The NELSON-plus protocol for LDCT scan-detected lung nodules stresses that mass doubling time could be a preferable measure of SSN progression. de Hoop and colleagues reinforced the indolent nature of invasive carcinomas derived from SSNs with a mass doubling time > 400 days [41]. Scholten et al., in their analysis of 7135 volunteers (median age at baseline of 58.0 years and median smoking history of 38.0 pack-years) with a median follow-up of 95 months, defined an increase of at least 30% of the mass of the nodule to be considered as growth: Using a cut-off of 30% growth in mass and/or volume of the solid component, no clinically relevant tumours were missed [22].

The British Thoracic Society and the American College of Chest Physicians (ACCP) guidelines recommend using the same diagnostic approach for incidentally found and screen-detected SSNs; Lung-RADS and NCCN guidelines refer specifically to screen-detected nodules; and Fleischner Society criteria are intended for incidentally found nodules only [42].

The Fleischner Society guidelines were first created in 2007; in 2013, the Society published guidance specific to SSNs [13]. They stated that the purpose of these recommendations was to reduce unnecessary follow-up examinations while providing the radiologist, clinician and patient greater discretion in making management decisions. Thus, they recommended a range of times rather than a specific interval for follow-up CT covering many scenarios (Figure 3). This change was made to recognize the multiple factors that determine the risk that cannot be easily incorporated into a summary table and the important role of patient preference for either more aggressive or more conservative management. Although they considered data from the National Lung Screening Trial, the NELSON, the International Early Lung Cancer Action Program, the Pan Canadian Early Detection of Lung Cancer Study model and the British Columbia Cancer Agency cancer screening trial, all of which support the use of less aggressive management of small nodules, they recognized that screening programmes have defined protocols to educate candidates about potential risks and the need for consistent monitoring, whereas incidentally identified nodules represent a separate population that requires a more varied approach to clinical management. They recommend no follow-up of any SSN 5 mm or less. They suggest that for nodules 6 mm or more, a CT scan should be performed to confirm persistence, and then interval scanning should be performed annually for up to 5 years if there is a solid component and every other year for pure GGOs. Multiple GGOs may be more likely to be benign, so a repeat scan at 3 to 6 months should be performed to confirm persistence; then management should be determined based on the most significant nodule.

Reproduced from MacMahon et al. [15] with permission from the Radiological Society of North America (RSNA).

The American College of Radiology has published Lung-RADS categorical criteria. It is one of the most extensive established categorization tools and the one in use in the United States [36]. It categorizes nodules into 4 risk categories, providing a percentage risk of malignancy and a proposed management plan for each category. The algorithm below is for a lung cancer screening programme (Figure 4).

In the NCCN guidelines version 1.2022, PSNs are divided according to a 6-mm cut-off. If the nodule is less than 6 mm, it is considered very low risk. Only an annual LDCT is recommended and is included in the Lung-RADS score 2. If the nodule is greater than 6 mm, with a solid component below 6 mm, the risk is Iow, and LDCT is recommended in 6 months (Lung-RADS score 3). If the solid component is between 6 and 8 mm, the risk becomes high (Lung-RADS score 4A); in the case of a solid component > 8 mm, the risk is very high (Lung-RADS score 4B) [3].

The NELSON trial used a study nodule management protocol as shown below to determine a management strategy for all nodules, including part-solid and non-solid nodules. Their original strategy is shown in Figure 5.

Many lung cancer screening trials have required a nodule management protocol. The studies and their management protocol are included in the literature review of the Swiss lung cancer screening programme (Table 3).

4 MANAGEMENT OF SINGLE PURE GROUND-GLASS OPACITIES (NON-SOLID NODULES)

The existing guidelines for the management of pure GGO focus heavily on radiologic assessment, particularly CT evaluation and surveillance. In light of the landscape of evolving techniques in both biopsy and surgical techniques, the need for further assessment with a non-surgical biopsy or surgical resection is under discussion.

4.1 The role of a non-surgical biopsy

The options available for the biopsy of pure GGO are increasing. Electromagnetic navigation bronchoscopy (ENB) is being increasingly practiced, and its performance has been assessed in several studies [43–45]. However, there is significant variation in the inclusion criteria, and they are certainly not specific to the assessment and biopsy of pure GGO. Generally, the yields for ENB of 65–84% are less than those for CT-guided percutaneous transthoracic biopsy (91%). The ACCP guidelines have favoured the role of the non-surgical biopsy.

In light of the recent improvements in surgical and anaesthesiological minimally invasive techniques able to offer complete removal of the lesion and reduce morbidity, the role of the preoperative biopsy may be limited to unfit patients or to patients with lesions located in anatomically difficult areas.

4.2 Surgery in pure GGO

There has been no clear and unified direction regarding when diagnostic surgical resection might be indicated for pure GGOs. There has also been further uncertainty about when the optimal timing of surgery might be when GGOs show growth. Generally, it has been performed when there is a high clinical suspicion of malignancy and/or an inconclusive biopsy result. Several case series have reported outcomes of patients undergoing surgical resection of solitary lesions [46–48]. The studies had different inclusion criteria with outcomes for a combined subsolid cohort (pure GGO). The surgical management also differed, including a lobectomy versus a sublobar resection (SLR), whereas others compared a segmentectomy against a wedge resection.

The findings among all case series generally demonstrated excellent long-term prognosis. Tsutani et al. reported a 3-year overall survival of over 98% for patients undergoing wedge resection or segmentectomy for GGO-dominant tumours [49]. Sublobar resection, once an operation for those who could not tolerate lobectomy, has recently returned to regular surgical practice. Indeed, the recent JCOG0802 trial has demonstrated that segmentectomies offer significantly better overall survival than lobectomies [50].

No study has guided us as to the optimal surgical approach for the resection of pure GGOs. However, with the excellent outcomes of limited resection, lobectomy would be best avoided to preserve lung tissue/volume. Generally, preserving pulmonary function should be considered an important priority in managing pure GGO. The form of resection might then be sensibly tailored based on the location of the GGO. A simple wedge resection for those immediately beneath the visceral pleura suffices, whereas more central lesions should be resected with segmentectomy. Minimally invasive surgery is known to be associated with improved early postoperative outcomes compared to thoracotomy. With respect to lung cancer, both robotic and thoracoscopic lung resections have been associated with lower morbidity and mortality rates. We would therefore recommend that the resection of a pure GGO be done via a minimally invasive surgery approach [51, 52].

Given the excellent general outcomes of patients with pure GGO, it is recommended that individual patient views be considered when considering resection. Shared decision making is an emerging concept by which the clinician actively shares scientific data with the patient. It has been defined by the National Institute of Health and Care Excellence as a process in which the care or treatment options are fully explored and appraised, along with the associated risks and benefits. The Institute of Medicine in the United States has also emphasized the importance of involving and supporting patient decision making [53].

There remains little evidence and understanding about measuring patient participation in the shared decision-making process with early-stage NSCLC. However, for example, results from the NELSON screening trial showed that subjects who could not make an informed decision to participate in the screening trial did not experience worse quality of life (QoL) than those who did. However, because the prognosis of pure GGO is generally very good with or without surgical resection, informing the patient about the options available would be recommended, given the risks associated with surgery.

When surgical intervention is planned for pure GGO, it is recommended that a thorough preoperative functional assessment be performed. It allows the surgeon to discuss the risks of resecting pure GGO and optimizing patients at higher risk. In contrast to lung cancer, the time to surgery can be extended given the excellent outcomes. This time allows for smoking cessation and the implementation of a pulmonary rehabilitation programme. A thorough assessment helps identify the apparently well patient with an underlying significant cardiorespiratory impairment who would present a high risk of major complications during the perioperative period and for whom it would be best not to proceed with resection of a pure GGO.

Although this is a European document, it is essential to note that many of the previously mentioned guidelines are from the West, with many recommendations based on evidence from Asia. Evidence shows that surgery for GGO may be offered more liberally in Asia. However, there is no evidence that surgery should be provided for pure GGO purely on a geographical or racial basis.

4.3 Lymph node strategy

Lung cancer screening will see a surge in the number of lung cancers presenting as pure GGOs. The presence of a GGO component has been associated with a favourable prognosis [54]. GGOs are a radiologic entity and can represent lesions on the adenocarcinoma spectrum; for those without a formal tissue diagnosis, the necessity to perform formal systematic lymph node sampling or dissection remains controversial. A prospective, multi-institutional study on image diagnosis was set to define early (noninvasive) adenocarcinomas of the lung (JCOG0201) among 545 patients with adenocarcinomas who underwent lobectomy and LND [4]. The JCOG0201 trial defined SSNs less than or equal to 3 cm in size with a CTR of 0.5 or less (i.e. part-solid lesions) as “radiological non-invasive”. In the JCOG0201 trial, the 5-year overall survival of radiologic noninvasive (121 patients, 22.2%) and invasive (424 patients, 77.8%) adenocarcinomas was 96.7% and 88.9%, respectively, and the difference was statistically significant (P < 0.001, log-rank test). The hypothesis that a consolidation/tumour ratio on thin-section computed tomography of 0.25 or less in cT1a (≤ 2.0 cm) could be used as a better radiologic criterion for a noninvasive pathology than a consolidation/tumour ratio of 0.50 or less in cT1a-b (≤3.0 cm) was not statistically significant (P = 0.259), with an overall survival of 97.1% and 92.4%, respectively [4].

In resected clinical stage IA adenocarcinoma presenting radiologically as pure GGO, numerous groups have shown the lymph node metastasis rate to be 0% on retrospective analysis (all pathological N0 disease) [55–61]. In resected GGO-predominant lesions (n = 129) on the adenocarcinoma spectrum, there was no difference in the 5-year recurrence-free survival between “no mediastinal lymph node evaluation” patients, “mediastinal lymph node sampling” patients and “mediastinal lymph node dissection” patients (100%, 92.9% and 93.8%, respectively, P = 0.89) [62]. There was no significant difference between the 3 groups regarding tumour location, mean size, positron emission tomography avidity, differentiation and pathological staging [62]. Using nomograms and multivariate modelling, pure GGO lesion status has been identified as a significant independent factor for lower risk of occult lymph node metastasis [58, 60, 63]. Given the low pathological propensity for lymph node metastasis in lesions presenting as pure GGO (CTR = 0), systematic lymph node dissection is not known to improve survival and requires further robust prospective data. However, systematic lymph node dissection is part of staging, and the current data are retrospective and small-scale.

In lesions without a preoperative diagnosis, frozen section analysis for confirmation of lung cancer can inform the need for systematic nodal dissection as part of routine intraoperative staging. In the absence of frozen section analysis, the risk of invasive cancer lesions should be assessed on a case-by-case basis and used to inform an intraoperative lymph node sampling strategy. Until we have large-scale prospective data, randomized or otherwise, we recommend adherence to the IASLC recommendations for intraoperative lymph node assessment for completeness of staging (3 mediastinal stations or nodes including subcarinal and 3 hilar nodes or stations) as a prerequisite for the assignment of the certainty of the stage for R0 resection.

5 MANAGEMENT OF THE PART-SOLID NODULE

The critical studies in this area are those performed by the JCOG [5, 6, 64, 65] trial. In the JCOG0804 trial, which was a cohort study of 333 patients in Japan who had a nodule with a CTR ratio of under 0.25 and a total size under 2 cm in the peripheral third of the lung, 341 underwent a wedge resection and 56, a segmentectomy; all patients were followed for 5 years. Half of the nodules were pure GGOs. None of these patients underwent positron emission tomography, endobronchial ultrasonography or mediastinoscopy. In 305 patients, the final diagnosis was adenocarcinoma, indicating that many of these nodules were invasive cancers. Still, a 99.7% 5-year survival showed excellent prognosis with the management plan of wedge resection with a clear margin [64].

In the JCOG0201 trial, out of 811 patients who were enrolled from 31 Japanese institutions with a tumour of less than 3 cm in the peripheral half of the lung, 545 were evaluated. The tumour was resected via a lobectomy, and the patients were followed for a median of 7 years. The researchers found that patients with a CTR of less than 0.5 had a 5-year survival of 96.7%, and patients with a CTR of less than 0.25 had a 5-year survival of 97.1%, compared to 88.9% for a CTR above 0.5 [5].

In the JCOG1211 trial, a total of 390 patients were recruited for a prospective cohort study of patients with a subsolid lesion with a CTR of less than 0.5. A segmentectomy was performed to investigate the safety and efficacy of this approach. The results of this trial are awaited.

After 4 years of recruitment into the landmark RCT of segmentectomy versus lobectomy, JCOG0802, due to the excellent prognosis of patients in the JCOG0201 trial, the JCOG investigators elected to change their entry criteria for the trial from a CTR of 0.25 to a CTR of 0.5 [5, 6]. Also, the lesions had to be in the outer one-third of the lung and less than 2 cm in size. Therefore, 62 patients with a CTR below 0.5 in this RTC and 40% of all lesions in the study were SSNs. Ninety percent of the lesions were adenocarcinomas; 6% of the lesions had either N1 or N2 positive nodes. The study found a 3.5% better lung function in the segmentectomy group, a 5.1% higher recurrence rate in the segmentectomy group and a 3% higher 5-year overall survival. This study demonstrated the safety of segmentectomy in this particular cohort of patients, half of whom had SSNs, and indicated the importance of lymphadenectomy. Equivocal rates of postoperative complications were seen between the 2 groups (27% vs 26%, respectively), indicating the segmentectomy approach to be a safe alternative to lobar resection.

5.1 The role of a non-surgical biopsy

The ability of the CT-guided biopsy to assess the invasive nature of a PSN has been investigated, as well as the discrepancy between biopsy pathology and surgical pathology. A PSN has been reported as a significant predictor for a false-negative biopsy [adjusted odds ratio (adjOR) = 3.95, 95% CI 1.21–12.85] [66].

A retrospective study of 318 patients with peripheral PSNs who underwent a CT-guided biopsy followed by surgical resection showed that the overall concordance rate between biopsy and surgical pathology was 64%. Better concordance was found with small tumours (≤ 2 cm) in predicting either the predominant histologic diagnosis (χ² = 7.091, P = 0.008) or high-grade adenocarcinoma, micropapillary and/or solid subtype (χ² = 22.301, P < 0.001) [67].

A prospective study of 9,239 patients demonstrated that the overall accuracy, sensitivity, specificity and positive and negative predictive values were 91.1% (95% CI, 90.6–91.7%), 92.5% (95% CI, 91.9–93.1%), 86.5% (95% CI, 85.0–87.9%), 99.2% (95% CI, 99.0–99.4%) and 84.3% (95% CI, 82.7–85.8%). Moreover, the authors reported a diagnostic failure rate of 8.9% (95% CI, 8.3–9.4%), with the presence of subsolid lesions as one important independent risk factor for failures (adjOR = 1.81, 95% CI 1.32–2.49), similarly as lesion size 1.1–2 cm (adjOR = 1.75, 95% CI 1.45–2.11) [68].

In a study involving 354 patients undergoing CT-guided core biopsies for lung nodules (64 PSNs), the sensitivity in the case of PSN was reported as 89.3%, with a specificity of 100%. The biopsies of part-solid lesions required more time (P = 0.021). No significant differences were found in the diagnostic yields in terms of sensitivity, specificity, accuracy, positive predictive value and negative predictive value for solid and part-solid lesions. Although the incidence of post-procedural haemorrhage was significantly higher for part-solid lesions (P = 0.016), the occurrence of symptomatic major haemorrhage (P = 0.86) and pneumothorax (P = 0.11) was not significantly different between solid and part-solid lesions [69].

The ACCP guidelines have favoured the role of the non-surgical biopsy [70]. Anyway, as previously stated, the role of a preoperative biopsy may be limited to unfit patients or to patients with lesions located in anatomically difficult areas.

In light of the very low morbidity related to the MIS approach with the widespread use of segmentectomy, the role of the non-surgical biopsy is confined to non-fit patients and to atypical locations.

5.2 Mutational status

The NSCLC presenting as SSNs exhibited a lower tumour mutation burden than solid nodules, with a less active immune environment in GGO components and immune pathways, decreased expression of immune activation markers and less infiltration of most immune-cell subsets [71].

A study of 864 [524 epidermal growth factor receptor-(EGFR) mutated] patients showed that the EGFR-mutated group had higher proportions of pure GGNs (4.1% vs 1.3%), GGO-predominant (23.7% vs 14.7%) and solid-predominant PSNs (37.2% vs 31.7%), whereas EGFR wild-type tumours presented predominantly as pure solid nodules (34.6% vs 52.2%, P < 0.0001). The EGFR-mutated group more frequently had a lepidic-predominant subtype than the wild-type EGFR group (20.2% and 11.9%; P < 0.0001) and presented a smaller tumour size and solid portion (P < 0.0001) with a higher GGO proportion (P < 0.0001) [72].

In adenocarcinomas presenting as PSNs, KRAS mutations were independently associated with worse relapse-free survival (RFS) (HR = 5.34, 95% CI 2.53–11.89; P = 0.001) and overall survival (OS) (HR = 2.63, 95% CI 1.03–6.76; P = 0.044) [73].

5.3 Lymph node strategy

Lymphadenectomy plays a pivotal role throughout the surgical treatment of NSCLC, representing a keystone in predicting long-term oncological outcomes and guiding postoperative strategies. To gain a correct staging of the nodal disease, the IASLC suggests dissecting at least 2R, 4R, 7, 10R and 11R stations for right-side lung cancer and 5, 6, 7, 10L and 11L stations for those on the left side [74]. In contrast, the NCCN recommends resecting at least 3 mediastinal LN stations to assess a correct NSCLC staging [75].

Several researchers have investigated the rate of nodal metastasis in PSNs. In lung cancer with a GGO component, unlike the solid counterpart, the role of lymphadenectomy is still controversial [76, 77].

In 2017, Flores et al. analysed the role of lymph node status in 203 screening-detected cases of NSCLC (151 patients with nodal dissection and 52 without) appearing as SSNs: The rate of positive nodes was 0.7% with a 10-year OS of 100%, regardless of LND [78].

Similar results were obtained by Ye et al.: When they compared 251 patients with invasive adenocarcinoma presenting as PSNs treated with surgical LND (sLND) versus 16 without sLND, no differences in long-term outcomes were found. Among 251 patients who underwent sLND, there were 2 cases of lymph node metastasis (1.2%) in patients with PSNs with a CTR < 0.5 and 5 cases (3.5%) in patients with PSNs with a 0.5 < CTR <1. No recurrences or deaths occurred among the 16 patients. The differences in the 5-year RFS and 5-year OS were statistically insignificant (P = 0.23 for 5-year RFS and P = 0.50 for 5-year OS) [55]. Likewise, Sakurai et al. showed that in 151 patients with lung cancer manifesting as PSN < 1 cm, no N1–N2 occurred [79].

Ye et al. analysed clinical stage T1aN0M0 lung adenocarcinomas appearing as GGO with minimal solid components (n = 118) and part-solid components (n = 13): 5 GGO predominant and 8 solid predominant, showing that all patients with PSN and minimal solid components were pathologically staged as N0. Only 2 (1.5%) of the patients with part-solid (solid parts > 5 mm, solid predominant) tumours had nodal involvement (1 patient had N1 and 1 had N2). GGO status had been revealed as being prognostic for nodal involvement at the multivariate analysis with an odds ratio of 44.2 and 95% CI between 7.9 and 247.4 [80].

Several large retrospective series recently clarified that the larger the solid component, the greater is the risk of lymph node involvement [81]. A study by Lee et al. of 593 part-solid adenocarcinomas showed that the solid-part size was prognostic for both univariate (HR = 1.10, 95% CI 1.03–1.18; P = 0.004) and multivariate analyses (HR = 1.06, 95% CI 1.01–1.12; P = 0.015]; the same was true for the number of LNs evaluated (HR = 1.05, 95% CI 1.02–1.09; P = 0.003] [82].

Investigating the CTR to differentiate predominantly GGO (CTR < 0.5) and predominantly solid (CTR > 0.5, but < 1) nodules, the authors of several large series agreed on the increasing risk of nodal metastases in the CTR > 0.5 group. Lin et al. demonstrated that predominant GGO lesions had a positive lymph node rate of 3.8% compared with that of the predominantly solid nodules, which had a lymph node rate of 6.9%. Recurrence rate [5 (2.4%) versus 21 (9.0%), P < 0.001] and disease-related death [3 (1.4%) versus 12 (5.2%), P = 0.013] were significantly lower in predominantly GGO compared to predominantly solid lesions [61].

Similar results were obtained in 372 patients with ground-glass-containing adenocarcinomas: more lymphovascular invasion and lymph node metastases were observed in the solid-predominant subgroup, which had significantly worse 5-year RFS [RFS = 89.2% (95% CI 84.1–94.3%) vs RFS = 98.0% (95% CI: 95.8–100%), P = 0.001) and 5-year OS [RFS = 96.7% (95% CI 93.8–99.6%) vs 100%, P = 0.023] than the GGO-predominant subgroup [83].

6 MANAGEMENT OF MULTIPLE GROUND-GLASS NODULES

Multiple pure pulmonary GGNs are defined as more than 2 GGNs (CTR = 0) found simultaneously in a single patient [84]. With the increasing use of CT screening for lung cancer, these findings are likely to become more frequent in our patient cohort. There is currently no standardized method of dealing with multiple GGNs. The main driving factor behind this is the lack of clinical data detailing the natural history, detection and treatment of a heterogeneous cohort. The Fleischner Society [15, 84] recommends using the dominant largest nodule to guide management, because it is the “prognosis-carrying” lesion. Herein lies a lack of robust level I and II evidence exploring the impact of early surgical resection of the dominant lesion compared to surveillance and follow-up only.

Retrospective data from a Korean group [85] evaluated the long-term outcomes in patients with multiple GGNs. Eighty-nine lesions were detected in 73 patients. Comparisons were drawn between those who underwent resection of all lesions versus those who underwent selective resection. No change in size or features of these lesions was noted at a median follow-up of 40.3 months, suggesting that surveillance may be an appropriate strategy. Indeed, some pure GGNs may satisfactorily resolve during the follow-up period, reflecting the spectrum of benign and malignant disease. The MILD trial data showed that the progression rate of GGNs to clinically significant disease was extremely low, again reinforcing an argument for surveillance in this setting [24].

Retrospective data from a Japanese series [86] has shown that, of 1,223 patients with resected NSCLCs, 67 patients (5.5%) had multiple part-solid lesions. These were stratified according to the presence of the ground-glass component into ground-glass predominant (CTR ≤ 0.5) [n = 24] and solid dominant (CTR > 0.5) [n = 43]. Five-year overall survival was significantly improved in the ground-glass predominant group compared to the solid dominant group (95.8% versus 68%; P = 0.009). Multivariable testing demonstrated that larger-sized nodules and the presence of a solid component conferred a poor prognosis, but the treatment of residual GGOs did not affect survival. Furthermore, work from the same group showed that patients with multiple GGOs and a solitary adenocarcinoma had a similar prognosis and that the outcomes of these patients were predicated on the biological behaviour of the dominant lesion [87].

There is currently no randomized or large-scale observational evidence assessing the direct impact of surgical resection of the dominant nodule versus surveillance only on overall survival for multiple GGOs.

7 HOW TO IDENTIFY A NON-PALPABLE GROUND-GLASS OPACITY

Minimal access approaches and the increased resection rates of pure GGOs have driven the need for localization methods. In this lesion subset, the success of finger palpation has been recorded to be as low as 24% [88]. This chapter highlights the various methods identified within the literature and provides recommendations for their uses. Although preoperative planning for anatomical sublobar resections is briefly discussed, the focus remains on identifying non-palpable GGOs.

7.1 Preoperative techniques for localization and marking

CT computed tomography lung segmentation for planning

Numerous platforms exist, both proprietary and open source, for the mapping of lung segments using the patient’s preoperative CT scan [89, 90]. This procedure aids in the identification of the lesion and its associated segment, along with anatomical variations that are often encountered. Depending on the software used, the relevant anatomy, such as the bronchial tree, can be displayed while hiding or reducing the opacity of the parenchyma or vasculature. These images are most often displayed and interacted with on 2-dimensional screens; however, 3-dimensional headsets are increasingly utilized. Despite being invaluable for complex anatomical sublobar resections, these planning techniques do not aid the intraoperative identification of intersegmental planes.

Electromagnetic navigation bronchoscopy lung mapping

This guidance tool utilizes the patient’s preoperative CT scan and calibrates the anatomy by analysing the disturbance caused by movements in the bronchoscope within an electromagnetic field. This technique has been utilized to obtain biopsies of peripheral nodules, which are otherwise difficult to access by percutaneous or traditional bronchoscopic approaches. More recently, this technique has been combined with fluoroscopy or cone-beam CT to increase the diagnostic yield. Furthermore, this technique is being utilized to place radio-opaque fiducials or dyes to aid in the intraoperative localization of non-palpable lesions. The ENB technique compared to the CT-guided technique demonstrated a reduced global time (143.4 vs 258.0 min, P = 0.002) [91]. Furthermore, ENB had a lower rate of pneumothoraxes in this propensity matched cohort. In addition, it can be performed by the surgeon in the same operating room just before the operation.

Marking devices

These devices are placed near the lesion of interest, allowing either direct visualization of the target area or visualization via imaging techniques.

Hookwire

Hookwire placement is the most common technique utilized for localizing pulmonary lesions [92]. It can be combined with dye marking [93]. This technique is suitable for subcentimetre lesions [94–96]. Furthermore, it has been utilized in lesions greater than 30 mm from the pleural surface [95–97]. Average success rates range from 93.6 to 100% [96, 98–100]. The mean time for the localization procedure ranges from 10 to 24 minutes. The greatest challenge with hookwire placement is the risk of a pneumothorax, pulmonary haemorrhage and patient discomfort. A pneumothorax is reported in 2.0 to 27.8% of cases. The presence of a pneumothorax increases the risk of wire dislodgement. However, the conversion rates to thoracotomy remain low at 0 – 4% [95, 101]. This situation is due to the visibility of the visceral puncture site [101]. This group demonstrated that a single puncture of the pleura is an independent protective factor for the development of a pneumothorax. Notably, conversions in these studies were not solely due to failure to localize but included the presence of adhesions and the management of complications. Pulmonary haemorrhages occur most commonly along the needle tract and, while frequent, rarely cause symptoms and do not prohibit the resection of lesions. Infrequent but catastrophic complications, such as air embolisms, have been reported [102].

Micro-coil

Micro-coils require a preoperative process similar to that of hookwire placement. Rather than a wire protruding from the skin, the coil is placed into the parenchyma adjacent to the lesion. Success rates are comparable at 91–100% [103–107]. Micro-coils take longer to place, given the increased complexity [108, 109] of the procedure. Hookwires and micro-coils provide similarly high success rates in the localization of lesions. Relative to the hookwire, micro-coil placement reduced complications of pneumothorax and haemorrhage [108–110].

Fiducials

Following navigation to the lesion, the placement of a radio-opaque fiducial marker can be utilized for intraoperative localization. The advantage of this technique is that it can be performed during the same anaesthetic event or in a delayed fashion for multiple nodules. The subgroup analysis of the NAVIGATE study demonstrated that 94.1% of markers remained in place for an average of 8.1 days postinsertion [111]. However, this procedure requires the presence of fluoroscopy or cone-beam CT intraoperatively to localize the lesion further.

7.2 Injectable agents

Dyes

Dyes, such as Patent Blue V, can be introduced around the lesion both via the percutaneous route and using ENB. Following placement of the dye, no additional intraoperative imaging techniques are required. A propensity matched study demonstrated a 98.7% success rate in localizing the nodule [112]. A retrospective comparison study comparing the hookwire and the use of methylene blue demonstrated similarly high success rates [113]. There was a trend towards reduced complications in the dye group; however, this trend failed to reach statistical significance. The critical caveat associated with this technique is rapid diffusion of the dye, preventing this technique from being used more than a few hours before the operation (recommended 3 h or less).

Indocyanine green

Indocyanine green (ICG) is used increasingly to identify intersegmental planes during sublobar resections. This technique has been used successfully to locate isolated lesions. Unlike dyes, this technique is not disadvantaged by the underlying pigmentation of the lung. A prospective study demonstrated high success rates via both percutaneous and endobronchial injection routes. The increasing depth of the lesion can reduce the success rate. Unlike intravascular injection of ICG, local injection near a lesion can remain on the lung surface for days, allowing for staged procedures [114]. In a retrospective study, the combined use of ICG and indigo carmine has been demonstrated to increase the detectability of lesions [115].

Lipiodol

Lipiodol injections with intraoperative fluoroscopy have similarly high success rates compared with hookwire insertion [110]. Although the lipiodol injection did take longer than the hookwire placement, the complication rates were significantly lower in the lipiodol group. Notably, the lipiodol group had a lower incidence of positive resection margins (2.38% vs 10.89%, P = 0.02).

Technetium-labelled human albumin macroaggregates

Computed tomography-guided injection of a suspension composed of 0.1 to 0.2 mL technetium-labelled human albumin macroaggregates and 0.2 to 0.3 mL of non-ionic contrast medium into or adjacent to the SPN is an effective method to localize nodules. During video-assisted thoracoscopic surgery (VATS)/robotic-assisted thoracoscopic surgery (RATS), the pulmonary area with the highest target/background ratio identified by an 11-mm diameter collimated thoracoscopic gamma probe is resected [116].

Medical adhesives

α-Cyanoacrylate is a fast-acting medical adhesive that polymerizes rapidly in contact with lung tissue. Unlike other injectable agents, it does not diffuse over time. The success rates for localization were 100% in the included studies [117–120]. This technique requires no further intraoperative imaging and aids the pathologist in localizing non-palpable lesions [117].

3-Dimensional printed navigational templates

The use of the 3-dimensional printed templates to aid in the preoperative placement of hookwires has significantly reduced the localization time and associated radiation dose [99]. This study highlighted the reduction in puncture attempts; however, this result did not translate into a statistically significant complication reduction.

7.3 Intraoperative techniques for localization and marking

Ultrasound

Ultrasound provides a radiation-free method of identifying lesions intraoperatively. The success rate for this procedure is lower than that for other techniques (86.9%) [121]. Limitations to this technique are realized with deeper lesions and with lesions with echo densities similar to those of the surrounding parenchyma.

Hybrid theatre procedures

Of the localization techniques, the most common issues are reduced success rate of localization as a result of displacement or diffusion of markers and the development of complications, such as pneumothorax. The increasing use of hybrid theatre suites has demonstrated an increased success rate due to reduced time between preoperative localization techniques and resection [94]. The use of the hybrid suite reduced the risk of hookwire displacement (0 vs 25%, P = 0.036) [122]. Furthermore, the pathway duration is quicker with greater patient comfort [123]. ENB used in the setting of a hybrid theatre also provides greater precision with immediate feedback regarding the accuracy of positioning [124].

8 MINIMALLY INVASIVE SURGERY APPROACH

Surgical resection of pure GGNs remains controversial and is divided into 2 (non-mutually exclusive) indications, which are diagnostic and therapeutic. Most groups reserve surgical intervention for high-risk subgroups (significant interval growth and/or development of a dominant solid component).

The issue is further complicated because it is not always possible to define a solitary pure GGN, and careful examination often reveals multifocal lung cancers. Given that there is no high-quality evidence on the indications of access concerning GGNs per se, we have assumed that the indirect evidence on outcomes with lung cancer, in general, would also hold true in this subset.

The VIOLET trial [51] is the largest RCT in thoracic surgery, assessing early postoperative outcomes between VATS and open surgery for lung cancer. The data confirmed earlier RCT evidence of lower pain and better health-related QoL [52] and showed that VATS as an access route for lobectomy for lung cancer was associated with better recovery and fewer in-hospital and long-term complications without any compromise to early oncological outcomes. Neither trial focused on the lesion of interest, namely the pure GGN. Furthermore, the study was predicated on suspected or pathological confirmed NSCLC and not on the radiologic diagnosis of presumed lung cancer, which is the primary criterion upon which the diagnosis of GGNs is based.

Historical randomized data comparing the 2 resection strategies did not demonstrate an oncological difference at long-term follow-up, nor did they allow for subgroup analysis for pure GGNs [125–127].

The question of whether minimally invasive surgery impacts long-term overall and disease-free survival in the setting of lung cancer remains to be formally answered, let alone the role of either technique in the setting of pure GGNs. However, the current randomized data indicate that minimally invasive surgery reduces the postoperative burden of morbidity for resected early-stage lung cancer and is associated with improved health-related QoL. This finding is certainly translatable to GGNs < 3 cm that undergo resection in select cases [128].

9 SUBLOBAR VERSUS LOBAR RESECTION

Lung cancer appearing as GGO-containing nodules on CT scans is more likely to represent a non-invasive histologic specimen, with a low risk of nodal metastasis and lymph vascular invasion [83, 129, 130].

Lobectomy has been considered the gold standard treatment for early-stage NSCLC, including GGO-containing lesions. The Lung Cancer Study Group trial published in 1995 reported significantly better survival after lobectomy than after SLR. The latter was thus reserved only for patients who could not tolerate a lobectomy [131]. The main concerns with this study are the inclusion of tumours larger than 2 cm, lack of discrimination between segmentectomy and wedge resection in the SLR arm, lack of differentiation among tumour histologic diagnoses and no analysis of the resection margins and intersegmental lymph nodes. With CT screening, increasingly more small-sized lesions, especially SSNs, have been detected. Thus, new evidence and guidelines for clinical practice are needed to improve the safety of surgery and decrease pulmonary function loss while retaining similar oncological outcomes.

Over the past 10 years, several studies have shown the non-inferiority of segmentectomy or even wedge resection over lobectomy in cT1a-b adenocarcinomas in terms of overall survival [132–138]. But most of these studies were biased due to their retrospective nature, lack of a proper control group and/or small case number. Still, the recent guidelines have shifted to recommend SLR as acceptable for GGO-dominant peripheral lesions ≤ 2 cm in size, which would correspond to AIS, MIA or T1a tumours according to the 8th edition of the IASLC staging manual.

Two large multicentre RCTs compared lobectomy and SLR for early-stage NSCLC, both designed to extend the indications of SLR to T1b tumours (Supplementary Table 2). Also, the JCOG0804 trial was a single-arm phase II study on wedge resection for small GGO-containing lesions [64]. The Cancer and Leukemia Group B (CALGB)140503 trial, a multicentre, noninferiority, phase 3 trial, has released its perioperative outcomes [139]. Out of a total of 697 patients, 340 underwent sublobar and 357 underwent lobar resections. Nearly 60% of the cases in the SLR arm had wedge resections. Although no significant difference in overall morbidity rates was noticed between the lobectomy (54%) and the SLR (51%) arms, the rate of grade 3 and above complications was much higher after segmentectomy (18.6%) than after wedge resection (11%) [10]. Similar results were found in the JCOG0802 trial, a randomized, multicentre, controlled, non-inferiority trial, in which 1106 patients were randomly assigned to receive a lobectomy (n = 554) or a segmentectomy (n = 552). This trial reported a longer operation time, more blood loss and a higher air leak rate in the segmentectomy arm compared to the lobectomy arm [138]. Considering the very low morbidity rate (5% grade 3 and above complications) in the JCOG0804 trial [64], it is reasonable to think that wedge resection is less traumatic and much safer than segmentectomy, thus more suitable for high-risk patients.

There is also a smaller randomized phase 3 multicentre trial (11 high-volume centres in Germany, Switzerland and Austria) that analysed a total of 107 patients with verified or suspected NSCLC up to 2 cm in diameter [140]. The subjects were randomly assigned to a segmentectomy (n = 53) or a lobectomy (n = 54). Five tumours (4.6%) had a GGO component: 3 in the segmentectomy arm (5.7%) and 2 in the lobectomy arm (3.8%). Also, this trial revealed comparable local control and not statistically different OS and DFS between a segmentectomy and a lobectomy.

9.1 Pulmonary function

All the trials mentioned above included good-risk patients who could tolerate a lobectomy. Although significantly more pulmonary function loss was noticed after segmentectomy than after lobectomy in the JCOG0802 trial, the absolute difference was merely 2–3%, which failed to reach the designated 10% end point [141]. CALGB140503 similarly showed an absolute difference between the lobectomy and segmentectomy arms of only 2% for both forced expiratory volume in 1 s and forced vital capacity, 6 months postoperatively (Supplementary Table 3) [10].

This result is probably due to the fact that postoperative pulmonary function loss is not in absolute proportion to the amount of lung volume resected and thus could not be accurately estimated by the traditional segment-counting method. This proposal has been proven by a recent prospective observational study showing that functional loss per lung unit resected was much higher after a segmentectomy than after a lobectomy. Although there might be some function-preserving benefit after a segmentectomy, this might not always be the case, especially when the number of segments removed was close to half of those in the corresponding lobe [142].

Given that pulmonary function loss was < 5% in the JCOG0804 trial [64], wedge resection is indeed a functional saving as an SLR. The presumption that segmentectomy helps preserve more function should be examined cautiously. Currently, no substantial evidence supports its use on a function-saving basis.

9.2 Oncological outcomes

The oncological outcomes of the JCOG0804 and 0802 trials have been published [10, 64, 141]. The former included GGO-dominant lesions ≤ 2 cm in size with a CTR ≤ 0.25, corresponding to cTis or cT1mi in clinical staging. The 99.7% 5-year RFS rate indicated that wedge resection would be enough for patients with such lesions [64]. The JCOG0802 trial compared the results of segmentectomy to lobectomy in patients with solid peripheral lesions or solid-dominant GGO-containing lesions (CTR > 0.5) sized ≤ 2 cm. For reasons unknown, the 5-year overall survival as the primary end point was significantly better (92-1-96-0) for segmentectomy (94.3%) than for lobectomy (91.1%). Overall, the RFS was similar between the 2 arms (segmentectomy 88.0% vs lobectomy 87.9%). This result would mean that a segmentectomy may now be preferable to a lobectomy for T1a-b tumours, especially those containing GGOs.

However, local relapse was significantly higher after segmentectomy (10.5%) than after lobectomy (5.4%). It is unclear whether LND was performed and whether the resection margins were adequate in the segmentectomy arm of the JCOG0804 trial. In a North American series of 715 SLRs for stage I NSCLC, 23% of the patients having a segmentectomy and 49% of the patients having an atypical resection (wedge) had not had a single node resected [143]. Moreover, a resection margin of 2 cm or equivalent to the lesion size is usually recommended for radical resection [144].

It is also worth noting that the inclusion criteria of the JCOG trials were based on CTR alone instead of on the tumour stage [64, 141]. The only criterion related to tumour size was a total lesion diameter of ≤ 2 cm. Both wedge and segment resections could thus be indicated in patients with cTis or T1mi lesions. Segmentectomy could either be indicated or contraindicated for T1b lesions sized around 2 cm, even though they are of similar clinical stages. In a retrospective study including 985 patients, all with GGO-containing lesions, RFS was significantly higher after lobectomy than after SLR when patients were matched by total lesion size. But RFS turned out to be similar between the 2 resection extent groups when patients were rematched by solid component size, regardless of the total lesion size [83]. It is thus a more acceptable practice to decide resection extent by tumour stage, i.e., solid component size, rather than by CTR alone.

One of the interesting findings from the JCOG0802 trial was that the local recurrence rate in the segmentectomy arm was almost double that in the lobectomy arm, although the OS and RFS remained similar. Thus, in SLRs, an adequate surgical margin is mandatory to decrease the risk of cancer relapse, especially local recurrence. El-Sherif et al. found a significantly higher risk of local recurrence when the surgical margin was less than 1 cm in patients having SLR. They thus recommended that an adequate surgical margin should be ≥ 1 cm [145]. Mohiuddin et al. found that, in 479 patients with early-stage NSCLC undergoing wedge resection, the risk of local recurrence could be significantly decreased by increasing the margin distance up to 15 mm [146]. In addition to margin distance, the ratio of margin distance versus tumour size was also proposed for the surgical margin decision. Sawabata et al. investigated the margin/tumour (M/T) ratio among patients with early-stage NSCLC. They found that a M/T ratio ≥ 1 was associated with negative margin cytological diagnoses, better recurrence-free survival and overall survival [147]. Moon et al. further proposed the margin distance/invasive component ratio in SLRs for GGO-containing adenocarcinomas by showing a significantly better prognosis when the margin distance/invasive component ratio was > 1 [148]. Therefore, a surgical margin ≥ 1 cm or an M/T ratio ≥ 1 should be secured in SLRs to ensure a clear resection margin.

10 FUTURE DIRECTIONS AND GAPS IN KNOWLEDGE

The major challenge in detecting pure GGO and PSNs as part of lung cancer screening programmes is in determining their malignant potential. The natural history of these lesions needs to be better understood to avoid overtreatment. This goal may be achieved by increasing the use of CT screening programmes worldwide. Pure GGO lesions have low malignant potential; however, tools are needed to aid prediction and guide resection. Molecular mutations can guide us in the future, but further studies must be conducted. The guidance for the surgical treatment of pure GGOs is based solely on results from cohort studies; no RCTs exist. The design and conduct of RCTs are needed to increase the evidence for guidance on resection, especially the need for lymph node assessment. Evidence for the treatment of PSNs is available from the results of the JCOG0802 trial [6] and the CALGB trial (NCT04944563) [139], which have provided considerable insight. Despite the survival superiority of segmentectomy, the incidence of local recurrence was higher in this group. The optimal M/T distance is unknown, and current recommendations are based solely on retrospective studies. Prospective studies should be conducted to explore the M/T distance with the lowest recurrence rate. The expected preservation of lung function with SLR was lower than expected in the JCOG0802 trial [6], and further studies on this issue are needed. The evidence for the superiority of VATS versus thoracotomy is provided by 2 high-quality RCTs. It may be assumed that the superiority of RATS is similar to that of VATS for a thoracotomy; however, level A evidence is needed. The different risks of malignancy observed between Asia and the West are not fully understood. Further studies on differences in molecular mutations and genetic dispositions should be conducted.

11 KEY MESSAGES

With the more frequent use of incidental CT scans and CT lung cancer screening programmes, the detection of early-stage lung cancer is increasing. Given that the gold standard for improved survival is surgical resection, there is a need for guidelines directing the treatment of early-stage lung cancer. This is the first European (EACTS/ESTS) guideline for surgical management of pure GGO (also called non-solid nodules) and PSNs. A task force of experts from EACTS and ESTS was gathered, and consensus on the recommendations was obtained. First, it is recommended that standard decision-making tools be used to determine the nodule risk for surgical management referral and that decisions for surgical resection be made by a multidisciplinary team, taking into account radiological characteristics, the evolution of the lesion, the presence of solid parts in a GGO and the patient’s risk factors. If surgery is deemed necessary, in the case of pure GGOs located in the peripheral third of the lung, wedge resection is recommended; for central pure GGOs, a segmentectomy is recommended. Lymph node sampling should be considered only in high-risk patients. For PSNs, a segmentectomy with LND is always recommended. Minimally invasive surgery (VATS or RATS) is preferred as a surgical approach for any resection. An adequate resection margin ≥ 1 cm or an M/T ratio ≥ 1 is recommended to ensure complete resection. For non-palpable SSNs, the use of localization techniques is recommended. In the case of multiple SSNs, the malignant risk of the dominant lesion should be considered, and a parenchyma-sparing strategy should be applied. Further resection should be based on accessibility. These recommendations are based on the available literature; however, a call for collaboration on the design and conduct of RCTs in the field is suggested.

European Association of Cardio-Thoracic Surgery/European Society of Thoracic Surgeons Committee Reviewers: Richard Milton^a, Mohsen Ibrahim^b, Pierre-Emmanuel Falcoz^c and Maninder Kalkat^d. The other reviewer wishes to remain anonymous.

^aDepartment of Thoracic Surgery, Leeds Teaching Hospitals Trust, Leeds LS9 7TF, United Kingdom ^bDepartment of Thoracic Surgery, Sant'Andrea Hospital, Sapienza University of Rome, Rome, Italy ^cDepartment of Thoracic Surgery, Nouvel Hôpital Civil, Hôpitaux Universitaires de Strasbourg, Cedex, France ^dDepartment of Thoracic Surgery, Heart of England NHS Foundation Trust, Bordesley Green East, Birmingham, United Kingdom

Disclaimer: A clinical guideline aims to apply to all patients with a specific condition. However, there will inevitably be situations where its recommendations are not suitable for a particular patient. While healthcare professionals and others are encouraged to consider these guidelines in their professional judgment, they do not override the responsibility of healthcare professionals to make decisions tailored to each patient's unique circumstances. Such decisions should be aligned with the latest official recommendations, guidelines from relevant public health authorities and applicable rules and regulations. It is important that these decisions are made in collaboration with, and agreed upon by, the patient and/or their guardian or carer.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

CONFLICTS OF INTEREST

René Horsleben Petersen received speaker fees from AstraZeneca, Medtronic, Medela and AMBU and serves as an advisory board member for AstraZeneca, Roche and MSD. Alessandro Brunelli serves as an advisory board member for AstraZeneca, Ethicon and MSD. Joel Dunning has received fees for proctorship from Intuitive and LivsMed and serves as a co-founder of the Cardiac Surgery Advanced Life Support (CALS) Center. Dominique Gossot has received speaker fees from Medtronic and Johnson & Johnson and consultancy fees from Delacroix-Chevalier. Peter B. Licht has received speaker fees from Medtronic, Ethicon and Johnson & Johnson. Eric Lim has received speaker fees from Medela; consultancy fees from Beigene, Roche and BMS; institutional research grants from AstraZeneca, Medela, Johnson & Johnson, Medtronic, Guardant Health, Lilly Oncology; company research grants from AstraZeneca, Boehringer Ingelheim, Medela, Johnson & Johnson, Medtronic and Takeda; holds patents P52435GB and P57988GB; and is the founder of My Cancer Companion, Healthcare Companion Ltd. The other authors have no disclosures.

ACKNOWLEDGEMENTS

We would like to extend our sincere gratitude to Christa Niehot, the project’s informatics medical specialist, who provided invaluable support in conducting the systematic literature review. Her expertise and guidance were instrumental in ensuring the accuracy and rigor of the review process.

FUNDING

This article was produced by and is the sole responsibility of the European Association for Cardio-Thoracic Surgery and the European Society of Thoracic Surgeons.

Author contributions

Milan Milojevic: Conceptualization; Methodology; Writing – review & editing. René Horsleben Petersen: Supervision; Validation; Writing – original draft, review & editing. Giuseppe Cardillo: Supervision; Validation; Writing – original draft, review & editing. Other authors: Writing – review & editing.

Reviewer information

Richard Milton, Mohsen Ibrahim, Pierre-Emmanuel Falcoz and Maninder Kalkat have indicated they would like to receive credit for their reviews.

REFERENCES

ABBREVIATIONS AND ACRONYMS

 
• ACCP

   American College of Chest Physicians

 
• adjOR

  adjusted odds ratio

 
• AIS

  adenocarcinoma in situ

 
• CALGB

  Cancer and Leukemia Group B

 
• CI

  confidence interval

 
• COI

  conflict of interest

 
• COR

  class of recommendation

 
• CT

  computed tomography

 
• CTR

  consolidation-to-tumour ratio

 
• EACTS

  European Association for Cardio-Thoracic Surgery

 
• EGFR

  epidermal growth factor receptor

 
• ENB

  electromagnetic navigation bronchoscopy

 
• ESTS

  European Society of Thoracic Surgeons

 
• GGN

  ground-glass nodule

 
• GGO

  ground-glass opacity

 
• HR

  hazard ratio

 
• IASLC

  International Association for the Study of Lung Cancer

 
• ICG

  indocyanine green

 
• JCOG

  Japanese Clinical Oncology Group

 
• LDCT

  low-dose computed tomography

 
• LND

  lymph node dissection

 
• LoE

  level of evidence

 
• Lung-RADS

  Lung Imaging Reporting and Data System

 
• MIA

  minimally invasive adenocarcinoma

 
• MILD

  Multicentre Italian Lung Detection

 
• M/T

  margin/tumour

 
• NCCN

  National Comprehensive Cancer Network

 
• NELSON

  Dutch-Belgian lung cancer screening trial

 
• NSCLC

  non-small-cell lung cancer

 
• OS

  overall survival

 
• QoL

  quality of life

 
• RATS

  robotic-assisted thoracoscopic surgery

 
• RCT

  randomized controlled trial

 
• RFS

  relapse-free survival

 
• sLND

  systematic lymph node dissection

 
• SLR

  sublobar resection

 
• SSN

  subsolid nodules

 
• T

  tumour

 
• VATS

  video-assisted thoracoscopic surgery

Author notes