This study was approved by the Committee on Human Research at our institution. Informed consent requirement was waived. Records of 28 consecutive patients (15 men, 13 women; mean age 64 years) referred to our institution for percutaneous fiducial marker placement before initiation of cyber knife radiosurgery (Accuray, Sunnyvale, CA) between December 2004 and October 2009 were retrospectively reviewed. Indications for radiosurgery included unresectable non-small-cell lung carcinoma (n = 14 patients) and metastatic disease to the lungs from colorectal carcinoma (n = 5), soft tissue sarcoma (n = 5), melanoma (n = 2), renal-cell carcinoma (n = 1), and lymphoma (n = 1). One male patient with two nodules in the left lung underwent a computed tomographic (CT)-guided biopsy of one of the nodules 24 h before fiducial placement. This patient developed a pneumothorax at the time of biopsy, which was treated by insertion of a chest tube. The chest tube remained in place at the time of the fiducial placement. Data from this patient were therefore excluded from the analysis of pneumothorax-related risk factors, but were included in the analysis of fiducial migration.
At the request of our institution’s radiation oncologists, between one and three fiducials, each measuring 1.2 mm by 3 mm (Alpha-Omega Services, Bellflower, CA), were placed within or adjacent to each lesion. The number of fiducials used in each case was at the discretion of the interventional radiologist performing the procedure. In general, lesions greater than 3 cm in diameter were targeted with two or three markers, whereas one fiducial was used for nodules less than 3 cm in diameter. The fiducials did not have to be positioned within the lesion. Marker placement within 1 cm of the lesion was acceptable for targeting. If a fiducial migrated away from its intended location, additional markers were inserted until at least one marker remained within or adjacent to the target lesion.
A total of 59 fiducial markers were placed. In four patients, two lesions located in the same lung were targeted; one lesion was targeted in the remainder of the patients. Because marker migration occurred upon insertion in six patients, up to seven fiducials (range 1–7) were required to achieve durable lesion targeting with at least one marker. Between one and seven fiducials were used for each lesion (median 1), and between one and seven fiducials were used for each patient (median 2). Lung biopsies were not performed before fiducial placement.
In each insertion procedure, CT was used to guide a 15-cm, 18-gauge double-wall Chiba needle (18 patients; Cook, Bloomington, IL) or a 15-cm 19-gauge Temno guiding double-wall needle (10 patients; Cardinal Health, Dublin, OH) into the lung lesion (Fig. ). Choice of needle gauge was at the discretion of the physician operator. After the needle was placed in the lesion, the marker was delivered through the needle with a 0.035-inch straight guide wire (Long-Tip; Cook) or the needle stylet. After the procedure, patients were followed with chest radiographs at hourly intervals for 2 h. Chest tube insertion was performed in patients with a pneumothorax and one or more of the following clinical features: dyspnea, oxygen saturation of <92% on room air, pneumothorax exceeding 25% of hemithorax volume, or increasing pneumothorax on serial radiographs.
Patient records including CT scans of the thorax, chest radiographs, and clinic notes were retrospectively reviewed. Follow-up chest CT scans and/or chest radiographs were available for all patients. Follow-up period ranged from 1 to 46 months (mean 11.9 months). Preprocedure CT images were used to determine lesion volume and to look for radiographic evidence of emphysema (presence of bullae or parenchymal lucencies located in the typical centrilobular, panlobular, or paraseptal distribution pattern). Presence or absence of emphysema was also determined from chart reviews, which were focused on clinic notes, hospital discharge summaries, and on the results of pulmonary function tests. Patients were categorized as having emphysema if they carried this diagnosis in the medical record or if there was evidence of emphysema on CT. Postprocedural CT scans were also evaluated for the length of needle trajectory through lung parenchyma, fiducial marker position in relation to the lesion, and development of a pneumothorax. The number of needle adjustments within the lung parenchyma (defined as any needle manipulation after crossing the pleura and documented by the procedural CT) before deposition of each fiducial was also determined. Needle adjustments before crossing the pleura were not counted. Targeting of each lesion required a median of one adjustment (range 1–8). The location of the fiducial markers on follow-up CT scans and chest radiographs was compared to their location on the initial scans in order to determine whether fiducial migration had occurred.
Our study focused on the rates of pneumothorax and fiducial migration as well as on elucidation of potential risk factors that may lead to these complications. Patients with and without these complications were compared with regard to several factors including lesion volume, length of needle trajectory through the lung parenchyma, number of fiducials inserted in a given lesion, number of needle adjustments within the lung parenchyma required for fiducial positioning, lesion location (upper or lower lung zone), lesion location in relation to the pleural surface, preinsertion needle tip location with respect to the target lesion (inside or outside the target lesion), presence or absence of pneumothorax before fiducial insertion, presence or absence of emphysema (on the basis of CT and clinical records), violation of a fissure, and needle gauge. The associations of the above factors with fiducial placement-related complications including pneumothorax, chest tube placement, and fiducial migration were analyzed by multivariate generalized estimating equation models. This modeling approach appropriately accounted for the statistical dependence among multiple fiducials and multiple lesions in the same individual. All analyses were conducted by SAS software, version 9.2 (SAS Institute, Cary, NC). All results were reported as P values, with P = 0.05 serving as a threshold for statistical significance. Lesion volumes were calculated by multiplying the cross-sectional area of each section by the section thickness and adding the individual section volumes. Length of needle trajectory through the lung parenchyma was measured at the point of fiducial deposition as the distance between needle tip and pleural surface. Because 10 fiducials migrated immediately or soon after their insertion, this measurement was used to estimate the distance from the fiducial insertion site to the pleural surface.
Jeremy C. Durack, Jeremy Collins, Roy L. Gordon, Jeanne M. LaBerge, and Robert K. Kerlan, Jr.
