The most appropriate biopsy technique for PPLs can be a challenging clinical risk-benefit decision and factors such as tumour size, location, patient co-morbidities including emphysematous changes around the PPL, respiratory function, and the pre-test probability of malignancy must be taken into account. Broadly speaking, the three general management strategies are: (I) watchful waiting with serial imaging to detect interval change, including the use of PET scans; (II) minimally invasive diagnostic procedures (bronchoscopy and trans-thoracic needle aspiration (TTNA)); and (III) surgical excision for diagnosis and definitive management.

CT-guided TTNA is a common method of obtaining tissue and whilst its pooled sensitivity for malignancy of 90% is impressive (3,4) it also has a false negative rate of 20-30% (5). Furthermore it is complicated by minor pneumothorax in approximately 25% of cases and major pneumothorax requiring a chest tube in 5% of cases (6). Identified risk factors for pneumothorax include smaller lesion size, deeper location, proximity to fissures and the presence of emphysema (7). Concomitant emphysema increases the pre-test probability of malignancy (if related to smoking), reduces the patient's tolerability for pneumothorax, and increases the time that intercostal catheter is needed for pneumothorax resolution. Rates of hemorrhage with TTNA are substantially higher than for bronchoscopic biopsy (8).

Of those PPLs that undergo diagnostic surgical excision, almost 20% are subsequently found to be benign, unnecessarily exposing significant numbers of patients to operative risks and reduced residual lung function (9,10). Futhermore, surgical resection has significant costs to the health system.

The diagnostic yield from transbronchial lung biopsy (TBLBx) of PPLs has incrementally increased with the addition of radiological guidance. Sampling with standard bronchoscopy under fluoroscopic guidance is associated with an overall sensitivity for malignancy between 14-63% but this is highly dependent on lesion size (<2 cm=34%, >2 cm=63%), biopsy method (forceps biopsy=57%, brush=54%, wash=43%), and number of biopsies taken (4,11). CT guided transbronchial biopsies are associated with diagnostic yields of between 65-73% but concerns about radiation exposure and inefficient use of CT scan time [in one study, an average of 4.1 scans per patients were performed with a mean effective radiation dose of 0.55 mSv (12)] has prevented widespread adoption of this technique (13).

The advent of endobronchial ultrasound radial probe (EBUSRP), which provides a 360 degree ultrasound view of small airways and surrounding tissue, has significantly enhanced the bronchoscopic diagnostic yield for PPLs. The interface between the low density alveolated parenchyma and the solid tumour tissue is represented on ultrasound by a bright line, providing confirmation that the EBUS-RP is situated within the target lesion. With this technique Kurimoto et al. demonstrated diagnostic yields between 69-77%, irrespective of lesion size, and yield from lesions >3 cm was 92% (14). Paone et al. demonstrated the superiority of EBUS RP compared to conventional transbronchial biopsy (C-TBLBx) in a randomized study of 221 patients (97 EBUS TBLBx vs. 124 C-TBLBx). The sensitivity for diagnosis of malignant lesions was 78.7% in the EBUS TBLBx group compared to 55.4% in the C-TBLBx group (P=0.007); diagnostic yields were 69.2% and 78.4% for benign and malignant lesions in the EBUS TBLBx compared to 44.4% and 55.4% for benign and malignant lesions in the C-TBLBx group respectively. There was no difference in sensitivity between the two groups for lesions >3 cm, but sensitivity markedly decreased in the C-TBLBx group in lesions ≤2 cm (EBUS TBLBx=71% vs. C-TBLBx=23%, P<0.001) (15). When used with a guide sheath (GS), EBUS RP has the additional advantage of being able to tamponade the biopsy site, theoretically reducing the risk of significant hemorrhage. The diagnostic yield of EBUS is however limited by the lack of a real-time navigational system to guide the operator from the central airways to the peripheral target lesion.

The planning screen (Figure 5) consists of four panels, showing axial, coronal and sagittal CT views, and either a reconstructed three dimensional (3D) bronchial tree or virtual endobronchial view Functions such as zoom, pan, distance measurements, contrast, brightness and window level are accessible through a toolbar. The operator can navigate through the virtual endobronchial view as it would appear during bronchoscopy and the corresponding location of the virtual "tip of the bronchoscope" is tracked in all three CT views. Any point on any view can be selected and the corresponding location will be visible on the remaining views.

The LG can be aimed in different directions by turning the handle to one of eight preset positions, denoted by two arrows on the handle, then pulling on the neck of the catheter (Figure 3, 4). The degree of flex of the locatable guide is dependent on the amount of pressure applied to the catheter neck and importantly, the flex point is approximately 14 mm proximal to the LG tip.

In contrast, multivariate analysis from a study designed to evaluate predictors of success for EBUS RP guided TBLBx found that RP position relative to the lesion independently predicted yield (within lesion vs. adjacent to lesion vs. outside of lesion=83%, 61%, 4% respectively P<0.001) but underlying disease, lesion location, CT scan bronchus sign, operator or type of EBUS probe did not (26).

Only one study has investigated the effect of "bronchus sign" on ENB yield. 51 consecutive patients with pulmonary nodules from a single centre underwent ENB. Samples were obtained by alternating an aspirating needle with biopsy forceps and cytological samples immediately underwent rapid onsite cytopathology assessment (ROSE). Mean lesion size was 25 mm (15-35 mm), mean distance of 11 mm from pleura, 74% had bronchus sign and most lesions were situated in the right upper lobe. Of the 34 bronchoscopies which yielded a diagnosis (67%), 30 were associated with a bronchus sign (88%). Of the 17 non diagnostic bronchoscopies, only 8 had a bronchus sign (47%) (27). Multivariate analysis identified the presence of a bronchus sign as the only variable conditioning ENB diagnostic yield (P=0.005). This result may seem slightly surprising because ENB, unlike EBUS RP, demonstrates the relationship of the LG and EWC tip to the target lesions through real time computer generated images, akin to being able to look through the walls of the airways to the target lesion. By aligning the LG/EWC with the target lesion, one could theoretically deploy an aspirating needle through the EWC, directly into the lesion, irrespective of the presence of a leading bronchus. This is one of the theoretical advantages of ENB compared to EBUS RP.

Other factors conditioning ENB yield have been reported in several case series but none has been systematically studied. Whilst Gildea et al. and Wilson et al. found no difference in yield based on lobar location, Eberhardt's multimodality study found significantly lower yields from lower lobes in the ENB group when compared to EBUS RP and combined EBUS RP/ENB (29%, P=0.01), possibly due to exaggerated respiratory motion of lesions close to the diaphragm (21,24,28). Most studies show that diagnostic yield is independent of lesion size however only small numbers have been compared and multivariate analyses have not been performed. Several studies have incorporated fluoroscopy and ROSE but it's effect on diagnostic yield, procedure time, and number of samples taken remains unknown (19,24).

Although both conscious sedation and general anaesthesia (GA) have been used in published studies, the optimal method has not been tested. In two of Eberhardt's studies, there were no statistically significant differences in diagnostic yield according to anaesthetic technique [sedation vs. GA 67% vs. 76%, P=0.28, 64% vs. 70%, P=0.57 (21,29)]. Bertoletti found that ENB performed under an inhalational 50%/50% combination of nitrous oxide/oxygen mixture was efficacious and well tolerated (30). The optimal anaesthetic modality still needs to be identified but in the meantime, anaesthetic method will likely be determined by patient co-morbidities, lesion characteristics, operator experience and anaesthetist availability.

The benefits of using ENB to place fiducial markers for stereotactic radiotherapy is becoming increasingly recognised. Treatment for inoperable lung cancer using stereotactic radio surgery rather than external beam radiotherapy allows delivery of precisely targeted radiation to the tumour while minimising dosage to adjacent tissue, but fiducial markers first have to be accurately placed, either transthoracically, intravascularly, or bronchoscopically, to enable tracking of the tumour. Despite the latter being the safest method, bronchoscopically placed markers are essentially placed "blindly" because the tumour cannot be directly visualized. Using the navigational ability of ENB may improve placement accuracy. A case series of placement of fiducial markers using ENB assisted navigation showed accurate localization in 8 of 9 cases and a 90% retention rate one week later (35). In a separate study, 234 markers were successfully placed in 52 of 60 patients. At the time of Cyberknife planning, 215/217 (99%) of the coil-spring fiducials were still in place compared to only 8/17 (58%) of linear fiducial markers. Pneumothorax rate was 5.8% (36).