high-grade neuroendocrine carcinoma. As most lung cancers are diagnosed at advanced stages, the differentiation of NSCLC and SCLC led to palliative management of lung cancer until the early 2000s. Thus, the diagnostic approach was focused on tissue acquisition by obtaining small samples to determine histopathological character of the tumor combined with imaging studies to perform a thorough tumor, node, metastasis (TNM) staging. From 1980s to early 2000s, the development in anti-cancer management (mainly the introduction of platinum-doublets as the main palliative modality for stage IV NSCLC) was driven by NSCLC histology not otherwise specifed (NOS) and advanced TNM staging [5].

With the introduction of new cytotoxic (pemetrexed) and biological agents (bevacizumab) having increasing ef cacy or toxicity depending on histology, respectively, signi cance of de ning NSCLC subtypes and insuf ciency of NSCLC NOS was recognized in the early 2000s [6, 7]. To this end, histochemical and immunohistochemical studies were used more widely to differentiate adenocarcinoma and squamous cell carcinoma in cytologic and small biopsy specimens more consistently. To de ne the minimally required immunohistochemical markers for differentiating squamous cell carcinoma and adenocarcinoma in previously NSCLC NOS-classi ed small samples, an international group of experts in medical oncology, pulmonology, pathology, and thoracic surgery developed the International Association for the Study of Lung Cancer/American Thoracic Society/ European Respiratory Society (IASLC/ATS/ERS) lung adenocarcinoma classi cation in 2011 [8].

This change in goals and requirements for the acquisition of tumor tissue continues to refect on lung cancer care and drug development (e.g. immune-checkpoint inhibitors of programmed death-ligand-1 for advanced squamous cell lung cancer) [9]. In parallel with the novel anti-cancer therapies targeting weaknesses in genomic basis of cancer, the importance of adequate tissue in diagnosing and treating NSCLC has increased considerably in the last 10 years.

Among the heterogeneous consequences of cancers (invasion and metastasis, induction angiogenesis, replicative immortality, resistance to cell

death, reprogramming of energy metabolism, evasion of immune surveillance, circumvention of growth suppressors, and sustained proliferative signaling), particularly sustained proliferative signaling is frequent in some NSCLC subgroups as it generally originates from genomic mutations in key oncogenes which encode for activated tyrosine kinases [2, 10]. Three major genomic events leading to the direct activation of tyrosine kinases in NSCLC are overexpression or ampli cation, mutation, and rearrangement with partner genes. The most prevalent oncogenes that are ampli ed, mutated, or rearranged in NSCLCs are KRAS, EGFR, ALK, ROS1, MET, ERBB2, BRAF in adenocarcinomas, and FGFR1, FGFR2/3/4, PI3KCA and DDR2 in squamous cell carcinomas [11–13]. The most prevalent and clinically applicable driver oncogenes in NSCLC care are EGFR mutations, and ALK mutations occurring through gene rearrangements [14–19].

The requirement for suf cient and high-­quality tissue sample for diagnosis, staging and treatment selection has increased in parallel with the growing minimally-invasive procedures for tissue acquisition. With the aim of standardizing the use of tissue for molecular diagnosis in lung cancer, molecular testing guidelines was published by the IASLC, Association for Molecular Pathology (AMP), and College of American Pathologists (CAP) in 2013 for selecting lung cancer patients to be treated with EGFR and ALK thyrosine kinase inhibitors [20]. Although rapid single gene assays are used currently as their use is prioritized in the guidelines, owing to the technological advances, comprehensive molecular pro ling by next generation sequencing (NGS) can be possible in routine clinical practice [21, 22].

Minimally Invasive Procedures

Mediastinoscopy

Mediastinoscopy is a surgical procedure performed under general anesthesia for exploring the superior mediastinum between the sternal notch and subcarinal area but can occasionally access the level of main bronchi [23].

As performed during EBUS-TBNA or EUS-­ FNA, the contralateral lymph nodes are exploredrst to rule out N3 disease and then exploration proceeds in a systematic way. Usually, the last to sample is the subcarinal lymph nodes as controlling bronchial artery and perinodal bleeding can be problematic. Cervical mediastinoscopy is conventionally considered to have a speci city and a positive predictive value of 100% as all lymph nodes are resected for histologic examination. However, the positive results are not corroborated by other methods. The median sensitivity and negative predictive value of conventional mediastinoscopy are reported as 78% and 91%, respectively [1–3] while video-mediastinoscopy has a median sensitivity and negative predictive value of 89% and 92%, respectively. Complications are rarely encountered (3%); serious bleeding requiring mediastinotomy occurs occasionally (0.4%) [24, 25], and mortality is low (0.5%) [26, 27].

Video-assisted mediastinoscopic lymphadenectomy (VAMLA) and transcervical extended mediastinal lymphadenectomy (TEMLA) –two technical variations of mediastinoscopy for systematic resection of mediastinal lymph nodesare not widely used although they have exceptional diagnostic performances. Both are performed through an incision similar to that used for mediastinoscopy to remove the lymph nodes systematically. In VAMLA, rst the subcarinal and right inferior paratracheal lymph nodes are removed en bloc, and then the left inferior paratracheal lymph nodes [28]. In TEMLA, mediastinal lymphadenectomy from the supraclavicular to the paraesophageal lymph nodes is performed using a sternal retractor to elevate the sternum. For removing the subaortic and para-­ aortic lymph nodes, a thoracoscope is used [29]. The sensitivity of VAMLA is about 100% while the sensitivity of TEMLA is higher than mediastinoscopy and EBUS-TBNA [30, 31]. However, VAMLA and TEMLA currently are not performed in the routine mediastinal staging of lung cancer owing to their invasiveness and high complication risks when compared with EBUS-­ TBNA and EUS-FNA that have comparable accuracy but less invasiveness [32]. These two

mediastinoscopic techniques are not included In the American College of Chest Physicians (ACCP) and European Society of Thoracic Surgeons (ESTS) Guidelines for staging lung cancer [1] but their use in clinical trials is encouraged [33].

Convex-Probe Endobronchial

Ultrasound-Guided Transbronchial

Needle Aspiration (EBUS-TBNA)

EBUS-TBNA -with lower morbidity and mortality, and higher cost-effectiveness than mediastinoscopyhave become the procedure of choice for diagnosis and staging of lung cancer [1, 34, 35]. Complications occur rarely; the rate of pneumothorax is 0.07–0.2% [36]. EBUS-TBNA is performed usually as an outpatient procedure by pulmonologists, interventional pulmonologists, or thoracic surgeons in a procedure or operating room, under moderate sedation or general anesthesia depending on the local practices, and by using a dedicated fexible bronchoscope with an ultrasound at the distal end. EBUS bronchoscope is advanced from the mouth, a laryngeal mask or an endotracheal tube to the distal trachea and then the US probe is apposed to the airway wall to reveal contiguous structures. The lymph node station is identi ed based on anatomic landmarks and a 21or 22-gauge needle is pushed through the airway wall into the target lesion under real-­ time visualization on US [2].

There is no consensus on the number of needle passes into each lesion but three passes and 15 excursions for each pass can usually obtain diagnostic material in more than 95% of cases [37]. The needle is withdrawn after each pass and some of the obtained sample can be placed and smeared on slides, and the rest of the sample can be placed in a preservative solution for cytologic analysis and cell block preparation, or the whole sample can be placed in the preservative solution.

2R and 2L (upper paratracheal), 4R and 4L (lower paratracheal), 7 (subcarinal), 10R and 10L (hilar), 11R and 11L (interlobar) and occasionally 12R and 12L (lobar) lymph node stations as

well as paratracheal and parabronchial masses next to the airway can be accessed by EBUS-­ TBNA [2]. Access to station 5 (subaortic) through a transpulmonary artery route [38] has been reported however it is not performed routinely and widely.

Endoscopic Ultrasound Guided Fine

Needle Aspiration (EUS-FNA)

EUS is used to guide transesophageal needle aspiration, another real-time ultrasound procedure. Posterior mediastinal sampling through the esophageal wall can be performed by EUS-­ FNA. The inferior pulmonary ligament (level 9), paraesophageal (level 8), subcarinal (level 7), and left paratracheal (level 4L) lymph nodes can be accessed and sampled by EUS-FNA whereas anterolateral paratracheal (levels 2R, 2L, and 4R) lymph nodes are dif cult to access and sample by EUS-FNA. As with EBUS-TBNA, EUS-FNA has a very low complication rate [39, 40]. EUS– FNA is unique in that it can access extramediastinal locations: the left lobe and a signi cant part of the right lobe of the liver as well as the left adrenal gland [2, 41].

With its above-mentioned strengths and weaknesses, EUS-FNA is complementary to EBUS-­ TBNA in diagnosing and staging lung cancer. The combination of EUS-FNA and EBUS-TBNA has a higher diagnostic performance (pooled sensitivity: 91%, pooled speci city: 100%) than each technique alone [1–3, 42].

Conventional Transbronchial Needle

Aspiration (TBNA)

The diagnosis and involvement of hilar and mediastinal lymph nodes in lung cancer can be determined cytologically or histologically by conventional transbronchial needle aspiration which is a safe and economical bronchoscopic method. However, as a “blind” method with no imaging guidance used during bronchoscopy, it can be adequate in staging if there are 1.5–2-cm or larger lymph nodes close to carina (paratra-

cheal and subcarinal) in the presence of a high pretest clinical probability of malignancy. Furthermore, as the performance of conventional TBNA in staging depends considerably on the prevalence of mediastinal involvement and operator skills, its accuracy varies widely. This method has a high false-negative rate and thus, cannot be considered as a de nitive mediastinal staging technique in routine practice [43, 44].

Radial-Probe EBUS-Guided

Procedures

There are two types of r-EBUS: miniaturized and ultraminiaturized versions for the assessment of central airway wall and contiguous structures [45], and for the detection of peripheral pulmonary nodules/masses [46], respectively.

In the central airways, miniaturized r-EBUS is used to guide conventional TBNA. However, it is not a real-time guidance as in EBUS-­ TBNA. Indications of r-EBUS in the central airways are early lung cancer diagnosis, lung cancer diagnosis and staging, evaluation of mediastinal invasion, diagnosis of a mediastinal or intrapulmonary lesion abutting airway wall, and guiding decisions regarding interventional bronchoscopic therapy, surgery or radiotherapy [45, 47–49]. The diagnostic yield of r-EBUS-TBNA for central lesions is 86% [47] while the accuracy of r-EBUS in showing airway involvement is 93% [48], and its utility in guiding interventional bronchoscopic therapy is 43% [49].

The ultraminiature r-EBUS is used to detect peripheral pulmonary nodules and masses or in ltrative lesions and to diagnose them by guiding procedures: TBB or cryo-TBB, TBNA, brushing and bronchoalveolar lavage. However, this is not real-time guidance. First, the guide sheath covered r-EBUS probe is advanced and the lesion is localized, then the probe is removed and biopsy procedures are performed through the guide sheath that is left in place and held in axed position [46, 50]. The diagnostic yield of r-EBUS-guided procedures in peripheral pulmonary lesions (70%; range: 46–92%) is higher than the yield of conventional bronchoscopic proce-

dures (40–62%) and comparable to those guided by navigational bronchoscopy techniques such as electromagnetic navigation bronchoscopy (ENB) (74%) or virtual bronchoscopy (67%). Using a multimodality approach can increase the diagnostic yield of r-EBUS: to 77% by combining r-EBUS, ultrathin bronchoscope and virtual bronchoscopy particularly in lesions located in the middle and outer 1/3 of the lung, up to 92% by combining guide sheath, fuoroscopy and virtual bronchoscopy, and from 67% to 88% by combining r-EBUS with ENB [46, 51].

Highest diagnostic yields are obtained by r-EBUS-guided TBNA (55–85%) or TBB (69– 72%) while r-EBUS-guided brushing or bronchoalveolar lavage has a yield of about 30–40%. Combining diagnostic procedures during r-EBUS, particularly TBNA and transbronchial biopsy increases the diagnostic yield [50–53]. A lesion size of greater than 2 cm, malignant nature, presence of bronchus sign, probe position within the lesion, and visibility by r-EBUS are the factors positively impacting the diagnostic yield of r-EBUS-guided bronchoscopic procedures [51]. However, using cryo-TBB instead of TBB does not increase the yield signi cantly (77% vs. 72%, respectively) [54].

Navigational Bronchoscopy-Guided

Procedures

As with the r-EBUS, navigational bronchoscopy can also be used to access pulmonary parenchymal lesions that are hard to reach. The term -ini- tially used synonymously with ENBcurrently involves peripheral bronchoscopy augmented by a computer-aided system using ENB-based guidance or non-ENB-based guidance. ENB depends upon the generation of an electromagnetic eld around the patient’s body, which allows a sensor to track within this eld. The volume within this electromagnetic eld is mapped to correspond with a three dimensional reconstruction of the patient’s anatomy using CT imaging. The sensor, then, can navigate through this volume while its position is shown on a virtual three dimensional map of the lung. Non-ENB-based platforms uti-

lize virtual bronchoscopy by using only chest CT to generate a rendering of the airway, fuoroscopy guidance for transparenchymal nodule access, or augmented fuoroscopic tomography and C-arm based tomography for navigating to and localizing nodules, respectively [55].

Navigational bronchoscopy has a signi cantly higher diagnostic yield (74%) than conventional bronchoscopy but comparable yield to r-EBUS as mentioned above. The lung cancer screening programs as well as increased use of cardiac and abdominal CT scans have increased the detection and thus, the importance of the management of pulmonary nodules, over 90% of which are found to be benign in nature after further assessment. Currently, various navigational bronchoscopy platforms are present to deal with the issue of managing these nodules. The main purpose of these systems is to assist with the acquisition of diagnostic tissue by a safe and minimally invasive method. However, the cost of navigational bronchoscopy is high, at least 5 times that of r-EBUS [46, 51, 55].

Besides the traditional navigational technologies mentioned above, there are several novel navigational technologies: fuoroscopic navigation, robotic-assisted bronchoscopy, and cone beam computed tomography.

In fuoroscopic navigation, digital tomosynthesis is incorporated to re-register the target during the procedure to lessen divergence between the location of the nodule on the pre-procedural CT and its actual location during the procedure.

Two currently available platforms of robotic bronchoscopy -the latest wave of navigational bronchoscopyenable the bronchoscopist to direct a bronchoscope via a controller apparatus that provides precision control [55].

Cone beam tomography provides reconstruction of the detailed and isotropic images of a speci ed anatomical area in the body. It uses a high-resolution two-dimensional detector for obtaining images. The C-arm in the system has to be rotated to obtain a three dimensional data set. The use of this technique in peripheral bronchoscopy is quite novel, the rst publication being in 2014. As it is rather an adjunctive imaging tool to assist in localizing lesions more precisely by