| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Radiation Oncology (Dr. Piet), Academic Medical Center, Amsterdam, the Netherlands; and the Departments of Radiation Oncology (Drs. Lagerwaard, van Sörnsen de Koste, Slotman, and Senan) and Pulmonary Diseases (Dr. Kunst), VU University Medical Center, Amsterdam, the Netherlands.
Correspondence to: Frank J. Lagerwaard, MD, PhD, Department of Radiation Oncology, VU University Medical Center, De Boelelaan 1117, 1007 MB Amsterdam, the Netherlands; e-mail: fj.lagerwaard{at}vumc.nl
Abstract
Background: The diagnostic yields with transbronchial needle aspiration (TBNA) for mediastinal nodes are highly variable. Nodal positions, as assessed on a breath-hold conventional CT scan, do not account for nodal motion. We studied nodal motion on four-dimensional (4D) CT scans.
Methods: A total of 47 mediastinal nodes were identified on 4D CT scans performed for radiotherapy planning in 25 patients with lung cancer. Nodes were mainly located at stations 4R, 4L, 7, and 2R, and each identified node was contoured in all 10 phases of the 4D CT scan. Nodal motion was correlated with changes in carina position.
Results: The mean (± SD) nodal diameter was 10.2 ± 4.0 mm; and the mean nodal volume was 1.8 ± 2.3 mL. Movement was maximal in the craniocaudal axis (mean length, 4.7 ± 2.3 mm), and the corresponding mean mediolateral and ventrodorsal movements were 2.8 ± 1.9 mm and 2.4 ± 1.8 mm, respectively. The mean three-dimensional displacement of the nodal center was 6.2 ± 2.9 mm, and it exceeded 10 mm in five nodes. The nodal mass was constantly present in only 25 ± 14% of the region encompassing all nodal positions. The mean variation in craniocaudal distance between all nodes and the carina position during respiration was 5.3 ± 2.1 mm (range, 2.2 to 10.5 mm).
Conclusions: Both nodal motion and the varying distance between the carina and nodal position may explain the lower diagnostic yields for TBNA procedures performed without real-time guidance.
Key Words: lung cancer • lymph node motion • staging • transbronchial needle aspiration
The accurate staging of mediastinal lymph node disease is important in the diagnosis and treatment of lung and esophageal cancer. The standard imaging technique for staging non-small cell lung cancer (NSCLC) is a CT scan but false-positive and false-negative rates are unacceptably high.123 Fluorodeoxyglucose-positron emission tomography (PET) scanning has a higher sensitivity of 84% and a specificity of 89%; the sensitivity improves to 93%, and the specificity to 95% using combined CT-PET scanning.456 As false-positive rates range from 10 to 15%, it is strongly recommended that positive PET scan findings should be confirmed by cytopathology.678
Mediastinoscopy remains the "gold standard" for mediastinal staging, but access to all nodal stations is not possible; it has a complication rate of approximately 2.5%.9 The growing experience with minimally invasive techniques such as biopsy during endoscopic ultrasound fine-needle aspiration and transbronchial needle aspiration (TBNA) without or with endobronchial ultrasound (EBUS) guidance indicates that such approaches can spare patients more hazardous surgical staging investigations. Adding staging with endoscopic ultrasound fine-needle aspiration to mediastinoscopy identifies more patients with either mediastinal tumor invasion (T4) or lymph node metastases (N2/N3) compared with staging by mediastinoscopy alone.10
However, the diagnostic yield of TBNA can be highly variable, with a metaanalysis11 reporting the pooled sensitivity and specificity to be only 39% (95% confidence interval, 17 to 61) and 99% (95% confidence interval, 96 to 100), respectively. Nodal motion may explain the lower yields when TBNA is performed without real-time ultrasound imaging. Nodal location is often estimated from the distance between the carina and a visible lymph node, as measured on a diagnostic CT scan. This approach can be inaccurate if breathing-induced nodal motion is substantial. Mediastinal nodal motion is of interest in radiotherapy planning as conventional elective radiation fields are replaced by involved-field radiotherapy for lung tumors12 and in patients with Hodgkin disease.13 However, the limited data on nodal motion are from studies1415 that used repeat spiral CT scans or fluoroscopy of calcified nodes. Respiration-correlated CT scanning or four-dimensional (4D) CT scanning has been in clinical use in radiotherapy treatment planning since 2003 for both thoracic tumors1617 and healthy organs.1819 With a 4D CT scan, spatial and temporal information on organ motion and shape can be acquired by recording respiratory signals during cine-scan acquisition.20 We retrospectively evaluated 4D CT studies that had been performed for radiotherapy planning in order to assess mediastinal nodal motion during quiet respiration.
Methods and Materials
4D CT scans of the thorax have been routinely performed for both stereotactic radiotherapy and respiration-gated irradiation of lung cancer at the VU University Medical Center since 2003.1617 Consecutive 4D CT scans were reviewed, and 25 patients were identified in whom visible mediastinal lymph nodes could be clearly demarcated from vascular structures and other lymph nodes. A total of 47 distinct nodes were identified in 14 patients with stage I NSCLC who underwent stereotactic radiotherapy after a negative staging with an fluorodeoxyglucose-PET scan, and in 11 patients who were treated with respiration-gated radiotherapy for either stage III NSCLC or limited disease small cell lung cancer.
The technique of 4D CT scanning and its use for evaluating nodal mobility has been described in detail previously.161920 Briefly, patients are scanned in the supine position during uncoached free breathing on a 16-slice CT scanner (Lightspeed scanner; GE Medical Systems; Waukesha, WI) with a slice thickness of 2.5 mm. No IV contrast agent was used as prior diagnostic CT scans using IV contrast material were available for all patients. Respiratory signals are recorded by an infrared camera that is mounted at the base of the CT scan couch using reflective markers that are placed on the upper abdominal wall of the patient. The respiratory signal file (Varian Medical Systems; Palo Alto, CA) and CT scan images are loaded into 4D software that sorts images into datasets from 10 specific respiratory phases (Advantage 4D; GE Medical Systems). Sorted CT scan images of the same respiratory phase are grouped together to obtain a series of three-dimensional (3D) scans, each representing a different phase of the respiratory cycle.
The phase-sorted datasets were imported into the stereotactic radiotherapy planning system (Brainscan, version 5.2; Brainlab AG; Heimstetten, Germany). Visible mediastinal nodes were manually contoured in all 10 respiratory phases by the same observer using standard window/level settings (width, 50 Hounsfield units; length, 500 Hounsfield units). Sagittal and coronal reconstructions were used to verify nodal contours, and a second observer checked all contours. Nodal volume was determined using standard software tools, and 3D coordinates of the center of mass of each node were obtained for each phase of respiration using the radiotherapy planning software (Brainscan; Brainlab AG). Nodal motion was derived from the displacement of the center of mass in the respiratory phases using these coordinates. An "intersecting volume," which indicates the region where the node was present during all phases of the respiratory cycle, was also manually contoured (Fig 1 ).
|
Results
A total of 47 mediastinal nodes in 25 patients were investigated. Mediastinal nodal locations21 are summarized in Table 1 , and the most frequent locations were 4R, 4L, 7, and 2R. The mean (± SD) short-axis diameter of all nodes was 10.2 ± 4.0 mm, and the mean volume was 1.8 ± 2.3 mL.
|
|
Mobility of Carina
The mean movement of the carina during quiet respiration averaged 6.5 ± 2.5 mm in the craniocaudal direction, and the 3D displacement vector of the carina in all 25 patients was 7.8 ± 2.3 mm.
Nodal Mobility Relative to the Carina
A substantial variation in the distance between the coordinates of the carina and mediastinal nodes was observed in the craniocaudal direction (mean movement, 5.3 ± 2.1 mm; movement range, 2.2 to 10.5 mm). The mean variation in 3D distance between the center of nodal mass and the carina was 4.7 ± 2.1 mm (Fig 3
), and even subcarinal nodes exhibited considerable movement relative to the carina.
|
The present study of 47 mediastinal nodes in 25 patients represents the largest analysis of mediastinal nodal motion to date. A previous study15 on the movement of 27 calcified intrathoracic nodes using fluoroscopy reported motion to be maximal in the craniocaudal direction, with a total displacement exceeding 10 mm in 22% of nodes. We have studied the mobility of lymph nodes in all available mediastinal stations on 4D CT scans, although TBNA cannot be performed for nodal levels 5, 6, and 8 nodal positions. In addition to respiration-induced changes in position, levels 5 and 6 nodal positions can be influenced by aortic movement, and level 8 nodes can be displaced by esophageal peristalsis.
The nodes analyzed in this study were small, with a mean short-axis diameter of 10.2 ± 4.0 mm. The relevance of smaller nodes was highlighted in a study correlating nodal size with the presence of metastases in 256 patients, where 44% of metastatic lymph nodes were found to be < 10 mm in diameter, and 12% of patients with nodal metastasis had no node exceeding 10 mm in size.22 The yield of TBNA without real-time ultrasound guidance was lower for nodes measuring < 20 mm, for which sensitivity was reported to be 63%, as opposed to 71% for nodes of 
20 mm.6 In contrast to the above data, a recent study by Herth et al23 using real-time EBUS-TBNA in 100 patients with mediastinal nodes of 
10 mm reported adequate punctures in all patients, and a sensitivity for detecting malignancy of 92%, underscoring the importance of real-time guidance.
Besides nodal motion, the 3D distance between the center of the nodal mass and the carina position varied by a mean of 4.7 ± 2.1 mm during quiet respiration. The possibility of even greater motion during a TBNA procedure cannot be excluded, and our findings serve to highlight the drawbacks of planning TBNA procedures solely on the basis of a conventional diagnostic CT scan. Newer integrated bronchoscopes with a linear array probe built into the bronchoscope allow EBUS-TBNA of mediastinal lymph nodes under direct ultrasound control. A recent prospective study comparing EBUS-TBNA with PET and CT scans found that EBUS-TBNA had an accuracy of 98%, and a sensitivity and specificity of 92% and 100%, respectively.24 It is notable that nodal size and location does not appear to influence the success of TBNA performed with real-time ultrasound control.2526 Although image-guided TBNA is only available in a limited number of medical centers worldwide, our findings support a greater role for this image-guided approach for the optimal preoperative staging of mediastinal nodes.
Footnotes
Abbreviations: EBUS = endobronchial ultrasound; 4D = four-dimensional; NSCLC = non-small cell lung cancer; PET = positron emission tomography; TBNA = transbronchial needle aspiration; 3D = three-dimensional
The authors have no conflicts of interest to disclose.
Received for publication December 9, 2006. Accepted for publication February 16, 2007.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
