Ann Thorac Surg 2007;84:995-999
© 2007 The Society of Thoracic Surgeons
New Technology
Virtual Mediastinoscopy for Safer and More Accurate Mediastinal Exploration
Hiroyuki Shiono, MDa,b,*,
Meinoshin Okumura, MDa,
Noriyoshi Sawabata, MDb,
Tomoki Utsumi, MDb,
Masayoshi Inoue, MDb,
Masato Minami, MDb,
Noriyuki Tomiyama, MDc,
Hikaru Matsuda, MDd,
Yoshiki Sawa, MDa,b
a Medical Center for Translational Research, Osaka University Hospital, Osaka, Japan
b Division of Thoracic and Cardiovascular Surgery, Department of Surgery, Osaka University Graduate School of Medicine, Osaka, Japan
c Department of Radiology, Osaka University Graduate School of Medicine, Osaka, Japan
d Hyogo College of Medicine, Hyogo, Japan
Accepted for publication March 19, 2007.
* Address correspondence to Dr Shiono, Medical Center for Translational Research, Osaka University Hospital, 2-15 Yamada-oka, Suita, Osaka, 565-0871, Japan (Email: hishiro{at}surg1.med.osaka-u.ac.jp).
 |
Abstract
|
|---|
Purpose: Virtual endoscopy can theoretically produce images of hollow organs from computed tomographic scanning by discriminating walls with the air space. We produced virtual images of the mediastinum (ie, virtual mediastinoscopy based on positron emission tomography and computed tomography scanning data to visualize lymph nodes and great vessels similar to cervical mediastinoscopy).
Description: Virtual images from 5 patients with positive mediastinal positron emission tomography findings were produced using computer software designed for virtual endoscopy. Visualization of lymph nodes and vessels was done based on positron emission tomography-computed tomography and enhanced computed tomographic scanning data, respectively.
Evaluation: Virtual mediastinoscopy clearly showed three-dimensional relationships between active nodes and surrounding structures. Great vessels, such as the innominate artery and azygos vein, which require assessment during a mediastinoscopy, were visualized in virtual movies. Further, perspective views in the craniocaudal direction based on surgeon-orientation, simulated actual views were obtained during cervical video mediastinoscopy.
Conclusions: Virtual mediastinoscopy provided realistic images of the mediastinal anatomy, and has the potential to make cervical mediastinoscopy and other mediastinal explorations safer, as well as more accurate.
|
Introduction
|
|---|
During mediastinoscopy procedures, surgeons often experience difficulty discriminating the lymph nodes from surrounding tissues. In addition, great vessels, including the azygos vein and pulmonary arteries, must be accurately recognized to avoid injury and serious complications [1].
|
Technology
|
|---|
Virtual endoscopy is a new evolving technology based on computer processing of three-dimensional (3-D) image data provided by computed tomographic (CT) scans. The 3-D endoscopic images produced allow for detailed examinations of the structures [2, 3], as well as enabling precise preoperative planning and serving as a navigational tool [4].
We conducted the present pilot study in an attempt to determine the technical feasibility of virtual mediastinoscopy based on images obtained by integrated positron emission tomography (PET) and computed tomography, as well as multidetector-row CT images, which were used to reveal the locations of the lymph nodes and great vessels during a cervical mediastinoscopy.
|
Technique
|
|---|
Five patients scheduled for a clinically indicated standard cervical mediastinoscopy as a histologic examination at Osaka University Hospital (based on positive CT or PET-CT findings of examinations of the mediastinal lymph nodes) were enrolled after obtaining written consent. The PET-CT examinations were performed using a scanner (Discovery ST16-Elite [GE Medical Systems, Tokyo, Japan]) at MI Clinic. The CT images with a slice thickness of 1.25 mm were reconstructed at 1.00-mm intervals and fused with PET image data. Virtual images of the mediastinum were reconstructed based on these PET-CT data using computer software (Volume Viewer Plus, Voxtool 6.2.0n [GE Medical Systems, Tokyo, Japan]). A variable threshold value in the range of –500 Hounsfield Units was used for reconstruction of the virtual bronchial wall. The cone angle of the virtual endoscope was adjusted within a range of 50°. "Fly-through sequence" images were constructed from the sternal notch to the carina along the anterior surface of the trachea, and were recorded as movie files. Virtual images of the branches of the aortic arch, azygos vein, and pulmonary arteries were also produced with the same software based on CT scans with contrast injection using methods routinely performed for lung cancer staging.
A standard cervical mediastinoscopy was performed under general anesthesia in the usual manner with a mediastinoscope (Wolf, Knittlingen, Germany) linked to a video camera. Navigation was done in real time, with the virtual images moved forward or backward on the display by dragging with a mouse as the mediastinoscope moved. These images were then compared with the video mediastinoscopic findings.
The Ethics Committee of Osaka University Hospital gave approval for this study involving 5 patients who gave informed consent and waived the need to obtain patient consent for publication of the results.
|
Clinical Experience
|
|---|
Virtual mediastinoscopic images were successfully reconstructed before performing a cervical mediastinoscopy in all 5 patients without any complications (Table 1). Multiple enlarged and radioactive mediastinal lymph nodes were found in CT and PET-CT scans of patient 1. The craniocaudal fly-through sequence images used for virtual mediastinoscopy clearly visualized the radioactive lymph nodes in that case. Further, the 3-D relationships of the targets and surrounding structures were clearly shown (Figs 1A,
2A,
3A). Easy access to all target nodes through the scope was accomplished with the help of the virtual images, and incisional biopsies were sequentially and accurately performed (Figs 1B, 2B, 3B). All of the sampled nodes were determined to be metastatic adenocarcinoma.
View larger version (67K):
[in this window]
[in a new window]
|
Fig 1. (A) A portion of a fly-through movie produced for virtual mediastinoscopy obtained from positron emission tomography-computed tomographic images of patient 1. (B) Corresponding view during cervical video mediastinoscopy in the same patient. The most cranial node (station 1), which should first be assessed during a cervical mediastinoscopy, was shown to be the shortest in the virtual mediastinoscopy images, based on the orientation of the surgeon. The nodes appearing next (2R and 2L) are shown in the same view as in (A). (LUL = left upper lobe; RUL = right upper lobe; TR = trachea.)
|
|
View larger version (69K):
[in this window]
[in a new window]
|
Fig 2. Virtual mediastinoscopy images showing three-dimensional relationships between the two paratracheal nodes (2L and 2R) and surrounding structures, including the trachea. (A) Simulation of the view of the surgeon during video mediastinoscopy. (B) A suction tube can be seen at station 2R. (TR = trachea; 6 = para-aortic node.)
|
|
View larger version (92K):
[in this window]
[in a new window]
|
Fig 3. (A) Virtual image and (B) corresponding operative view at the mid-tracheal level. The virtual view clearly shows lymph node 4R beyond 2R. (LMB = left main bronchus; TR = trachea; 6 = para-aortic node.)
|
|
The CT scans of patient 2 demonstrated a pretracheal lymph node with a short-axis diameter of 1.5 cm that was located slightly cranial, whereas the left tracheobronchial node had a diameter of 1 cm. The radioactivity of these lymph nodes in the PET scan images was slightly higher than the background. Virtual mediastinoscopy clearly visualized both lymph nodes and the surrounding structures (Fig 4A). Accurate biopsies of the lymph nodes were then performed through an endoscope and revealed no malignancy.
View larger version (73K):
[in this window]
[in a new window]
|
Fig 4. (A) Paratracheal node 2R in virtual mediastinoscopy image and (B) operative view during cervical mediastinoscopy in patient 2. The nodes, which were radioactive, but proven not to be malignant, were soft and buried in fat tissue (B). (TR = trachea.)
|
|
A right tracheobronchial lymph node was found to be enlarged and abnormally positive in PET scan images of patient 3. The node was visualized using virtual mediastinoscopy from the PET-CT data diagonally at the anterior of the lower trachea (Fig 5C) and slightly on the cranial side of the azygos vein (Fig 5B). Other virtual images produced at the same time from enhanced CT images clearly showed the innominate artery, which must be assessed early during the dissection procedure (Fig 5A) as well as the azygos vein and pulmonary artery in the area of dissection near the carina (Fig 5B). The 3-D information obtained from the virtual images of both lymph nodes and great vessels was helpful during the operation (Fig 5D). A histologic evaluation after the cervical mediastinoscopy revealed metastatic adenocarcinoma.
View larger version (125K):
[in this window]
[in a new window]
|
Fig 5. Virtual mediastinoscopy movie derived from enhanced computed tomographic scans of patient 3 showing the real time anatomy of the great vessels in the mediastinum in the orientation of the surgeon, including (A) the branches of the aortic arch, and (B) the azygos vein and pulmonary artery. (C) Radioactive node 4R was also visualized in virtual mediastinoscopy images obtained from positron emission tomography-computed tomographic scans and shown to be located on the cranial side of the azygos vein shown in B. (D) An incisional biopsy of 4R was performed safely and accurately. (AV = azygos vein; IA = innominate artery; LCC = left common carotid artery; LS = left subclavian artery; PA = pulmonary artery; RMB = right main bronchus; SVC = superior vena cava; TR = trachea.)
|
|
Bulky radioactive paratracheal nodes were found in patient 4, which may have been easily accessed even without the help of the virtual images. However, several biopsies were needed because of negative results from frozen specimens taken during the video mediastinoscopy procedure. Virtual mediastinoscopy provided helpful 3-D information regarding the relationship between the node and azygos vein, allowing avoidance of injury. Permanent samples tested after the operation revealed sarcoidosis.
In patient 5 the paratracheal lymph node was less than 1 cm in the findings of the CT scan; however it was revealed to be radioactive in PET-CT scans. The virtual images demonstrated the relationships between the node and right main bronchus (Fig 6A), as well as the azygos vein (Fig 6B). On the other hand, the target node on the cranial side of the vein was small and ambiguous in the fat tissue, and great effort was required to locate it. Bleeding from the fine vessels totaled 250 mL. The histologic diagnosis was a metastatic node of adenocarcinoma.
View larger version (84K):
[in this window]
[in a new window]
|
Fig 6. Virtual images showing a highly radioactive paratracheal node, (A) which was found located on the cranial side of the azygos vein when combined with (B) virtual images from enhanced computed tomographic scanning. (AV = azygos vein; LMB = left main bronchus; PA = pulmonary artery; RMB = right main bronchus; SVC = superior vena cava; TR = trachea.)
|
|
In all 5 patients the target lymph nodes were safely sampled and no major complications arose. Blood loss and operation time for the video mediastinoscopy procedures are shown in Table 1.
|
Comment
|
|---|
Virtual endoscopy procedures have mainly been used for examinations of cylindrical-shaped structures and cavities in the human body [2, 3]. In contrast, the mediastinum is not hollow, but rather it is filled with connective and fat tissues. The present is the first known report of virtual images of the anatomy of the mediastinum produced by distinguishing radioactive lymph nodes from other structures using integrated PET-CT scanning data. We found that virtual images derived from enhanced CT images were useful to easily discriminate important vessels, such as the branches of the aortic arch, superior vena cava, azygos vein, and pulmonary artery, allowing them to be protected from injury, as well as other structures, when combined with virtual images of the target nodes.
The advantage of using virtual mediastinoscopy images during a mediastinoscopy is that they can provide not only 3-D structural information, but also anatomical information in real time (ie, four-dimensional information). The fly-through images can be freely manipulated on the display by dragging with a mouse and realistically showing what surgery personnel will encounter during the procedure. We believe that virtual mediastinoscopy would be much more helpful if the fly-through images that were obtained were able to automatically move back and forth as the mediastinoscope is introduced. A recently developed real-time navigation system is able to superimpose the tip of surgical instruments on video images using an optical 3-D position sensor [5, 6]. In addition, alterations in the anatomy by introduction of the mediastinoscope into the pretracheal space can be ignored, because the scope primarily moves along the long axis of the trachea. Therefore synchronization is rather simple and should be one of the future applications of virtual mediastinoscopy. Particularly in cases with multiple lymph nodes, such as the present patient 1, synchronous virtual mediastinoscopy would improve accuracy in navigation and might be able to reduce blood loss and operation time in patients with a very small target, such as the present patient 5.
A potential limitation of our procedure is that virtual images of lymph nodes and vessels are obtained separately from PET-CT and enhanced CT scans, respectively. An integrated PET-CT procedure can be performed with an injection of vascular contrast material for the purpose of producing virtual images. However, because the patients were referred after having undergone conventional contrast-enhanced CT for staging, we could not ethically justify the use of contrast material in the present cases and combined PET scan data with unenhanced CT scan data. Future software development may allow PET images to be fused with enhanced CT images obtained separately.
A mediastinal biopsy, such as cervical mediastinoscopy, remains the gold standard for staging of patients with lung cancer, particularly in patients with positive PET scan findings [7], as false-positive PET scan results can not be ignored [8]. We have also reported the impact of residual mediastinal involvement after induction therapy [9], and recently adopted fiberscopic transbronchial needle aspiration, standard cervical mediastinoscopy, or exploratory thoracotomy procedures, or a combination thereof, as additional histologic examinations for patients with positive mediastinal PET results [10]. Virtual mediastinoscopy would be theoretically helpful during these invasive mediastinal exploration procedures, in addition to a standard cervical mediastinoscopy.
Although our series is small, and the results do not reveal whether virtual mediastinoscopy is actually helpful for objective criteria, such as blood loss and operation time, this real time-imaging technology has some additional applications. We believe that the concepts presented here have the potential to make all invasive procedures safer and their results more accurate.
|
Disclosures and Freedom of Investigation
|
|---|
This study was performed with the CT scanners and the computer software that were used for routine clinical examination in the hospitals. The authors performed a free and independent evaluation of this new technology, with no financial relationship with GE Medical Systems.
|
Disclaimer
|
|---|
The Society of Thoracic Surgeons, the Southern Thoracic Surgical Association, and The Annals of Thoracic Surgery neither endorse nor discourage use of the new technology described in this article.
|
Acknowledgments
|
|---|
We would like to express our thanks to Dr Seiki Hamada, Director, and Yasuo Iwamoto, Radiological Technologist, of Jinsenkai MI Clinic, as well as Masayuki Kudo of GE Medical Systems for their technical advice.
|
References
|
|---|
- Hammoud ZT, Anderson RC, Meyers BF, et al. The current role of mediastinoscopy in the evaluation of thoracic disease J Thorac Cardiovasc Surg 1999;118:894-899.[Abstract/Free Full Text]
- Liewald F, Lang G, Fleiter T, Sokiranski R, Halter G, Orend KH. Comparison of virtual and fiberoptic bronchoscopy Thorac Cardiovasc Surg 1998;46:361-364.[Medline]
- Ezzeddine D, Ezzeddine B, McKenzie R, Bhutani MS. Virtual gastroscopy: initial attempt in North American patients J Gastroenterol Hepatol 2006;21:219-221.[Medline]
- Asano F, Shindoh J, Shigemitsu K, et al. Ultrathin bronchoscopic barium marking with virtual bronchoscopic navigation for fluoroscopy-assisted thoracoscopic surgery Chest 2004;126:1687-1693.[Medline]
- Inoue T, Tamaki Y, Sato Y, et al. Three-dimensional ultrasound imaging of breast cancer by a real-time intraoperative navigation system Breast Cancer 2005;12:122-129.[Medline]
- Schramm A, Gellrich NC, Gutwald R, et al. Indications for computer-assisted treatment of cranio-maxillofacial tumors Comput Aided Surg 2000;5:343-352.[Medline]
- Toloza EM, Harpole L, McCrory DC. Noninvasive staging of non-small cell lung cancer Chest 2003;123:137S-146S.[Medline]
- Silvestri GA, Tanoue LT, Margolis ML, Barker J, Detterbeck F. The noninvasive staging of non-small cell lung cancer: the guidelines Chest 2003;123:147S-156S.[Medline]
- Sawabata N, Keller SM, Matsumura A, et al. The impact of residual multi-level N2 disease after induction therapy for non-small cell lung cancer Lung Cancer 2003;42:69-77.[Medline]
- Inoue M, Sawabata N, Takeda S, Ohta M, Ohno Y, Maeda H. Results of surgical intervention for p-stage IIIA (N2) non-small cell lung cancer J Thorac Cardiovasc Surg 2004;127:1100-1106.[Abstract/Free Full Text]
Related Article
-
Invited commentary
- Keith D. Mortman
Ann. Thorac. Surg. 2007 84: 1000.
[Extract]
[Full Text]
[PDF]
This article has been cited by other articles:
|
|
|
|
 
K. D. Mortman
Invited commentary
Ann. Thorac. Surg.,
September 1, 2007;
84(3):
1000 - 1000.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
