A Portable γ-Camera for Intraoperative Detection of Sentinel Nodes in the Head and Neck Region

MATERIALS AND METHODS

Surgical Procedure

Fifteen of the patients underwent surgery 18–24 h after injection (the next morning), and the other 10 patients underwent surgery the same afternoon (<6 h after injection). Radioguided sentinel lymphadenectomy was assisted by a hand-held γ-probe (Neoprobe; Johnson & Johnson Medical) and a portable γ-camera (Sentinella, S102; Oncovision). This camera is equipped with a pinhole collimator and a CsI(Na) continuous scintillating crystal. The field of view is 4 × 4 cm when the distance to the imaging plane is 3 cm, and 20 × 20 cm when the collimator is placed at a distance of 15 cm. The intrinsic spatial resolution of the camera is 1.8 mm, and the extrinsic spatial resolution is 7 and 21 mm for distances of 3 and 15 cm, respectively. Detection sensitivity depends on the distance from the detector to the imaging plane, being 19,135 and 1,111 cpm/MBq for distances of 3 and 15 cm, respectively. More technical details can be found in the report of Sanchez et al. (6).

Before incision, the sterile-wrapped detector was placed 5–15 cm above the marked sentinel node levels to provide an overview of all the nodes in relation to the injection area (Fig. 1). At least 1 overview image (acquisition time, 1 min) was made. Concurrently, the location of each sentinel node was determined by placing the laser pointer exactly above the sentinel node signal. Acquisition continued until certainty was reached about the location of the node, and the time to visualization was recorded. After excision of the nodes, a new image was acquired and compared on-screen with the image before excision. The acquisition time was at least 1 min, and the camera was placed at the same position and distance as for the image acquired before excision. If hot spots remained near the injection area or near previously marked sentinel node levels, these nodes were regarded as additional sentinel nodes and removed. Nodes that were preoperatively defined as second-echelon nodes were left in situ.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Surgeon positions sterile-wrapped head of camera (A) above surgical field using laser pointer (B). On screen of γ-camera, sentinel node is visualized just cranial to red cross of laser pointer, and second-echelon node is visualized caudal to cross of laser pointer (C). Head of portable γ-camera can be placed above sentinel nodes to exactly localize each node (C). When head of camera is placed farther from patient, entire surgical field can be visualized, providing overview of injection area and all radioactive nodes.

RESULTS

A median of 3 sentinel nodes per patient was visualized (mean, 2.8; range, 1–7) on preoperative images, and all were visualized on the screen of the portable γ-camera intraoperatively. Superficial nodes with strong tracer uptake were depicted within 10 s. Deeply located and radioactively weak nodes took 30–50 s to visualize. All sentinel nodes were visualized within 1 min.

Eight of the 70 sentinel nodes were in difficult locations: near the injection area, next to another sentinel node, or deep within the parotid gland. These nodes could be localized more efficiently with the addition of the portable γ-camera.

In a separate case, a preoperatively visualized node that could not be found with the probe could be excised after localization with the portable γ-camera.

With imaging after excision, 9 additional radioactive nodes were identified in 6 patients. These were regarded as possible sentinel nodes and removed. These 9 nodes had not been visualized preoperatively. One additional sentinel node was discovered behind the injection area in neck level II, after removal of the primary lesion. Additional nodes were removed because substantial radioactivity remained after excision of a sentinel node in the other 5 patients (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Postexcision imaging in 2 patients. In patient with melanoma of left cheek, portable γ-camera images show injection area and 2 sentinel nodes before incision (A). Both sentinel nodes could be removed with help of portable γ-camera and γ-probe. Postexcision imaging showed remaining radioactivity at site of excision of most caudal sentinel node (B). This hot spot had not been separately visualized on preoperative images. Because substantial remaining radioactivity was visualized near injection area, this additional hot spot was considered to be sentinel node instead of second-echelon node. Because its location allowed for easy removal, it was excised. All sentinel nodes appeared to be tumor-negative. In another patient with temporal melanoma, cluster of 2 sentinel nodes was visualized caudal to injection area (C). After excision of 2 radioactive nodes, only background radioactivity and radioactivity at site of injection remained (D), confirming adequate sentinel node excision. No nodal metastases were found on pathologic examination.

One (parotid) sentinel node was clearly visualized with the portable γ-camera but was not excised because of the risk of damaging the facial nerve.

Six patients had 1 tumor-positive sentinel node on pathologic examination. One of these positive nodes was an additional node that had not been separately identified on preoperative images but was identified with postexcision imaging.

DISCUSSION

Several portable and hand-held mini γ-cameras have been developed to provide intraoperative visualization of radiotracers (6,9,10). These cameras can give an overview of all radioactive hot spots within the surgical field. This overview can be compared with preoperative images to localize nodes and distinguish between first and higher echelons. In the example of Figure 3, the procedure could have been performed with the probe only, but the availability of both preoperative images and intraoperative images provided detailed information on the anatomic location of each sentinel node.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3. 

Comparison of preoperative images with intraoperative images. In patient with melanoma cranial to left ear, lateral planar lymphoscintigraphy after 2 h (A) showed 2 hot spots caudal to injection area and weak preauricular hot spot. All 3 hot spots had already been visualized on dynamic scans and on static planar images after 15 min. On 2-dimensional and 3-dimensional (B) SPECT/CT fusion, 2 sentinel nodes were localized just ventral to sternocleidomastoid muscle and 1 preauricular sentinel node was localized within upper part of parotid gland. In concordance with SPECT/CT, intraoperative images before incision showed injection area and 3 sentinel nodes in same relation to each other (C). After excision of first sentinel node, portable γ-camera showed remaining sentinel nodes (D), which were then localized and removed.

The overview before incision can also be compared with images obtained after the sentinel nodes (Fig. 2) and primary lesion have been excised, to detect additional sentinel nodes that had been overshadowed. In our series, this type of comparison led to excision of an additional sentinel node in 24% of the patients (including 1 positive node). In some cases, intense tracer uptake at the site of the sentinel node on preoperative images corresponds to a cluster of sentinel nodes. Imaging after excision of the first hot node can reveal the remaining sentinel node. Another reason for visualization of more sentinel nodes intraoperatively might be that more tracer has accumulated because of the longer time from injection to imaging. Although the portable γ-camera enabled detection of these nodes, they might also have been found by more extensive exploration with the γ-probe.

In our experience, weakening of the radioactive signal is limited, and intraoperative imaging takes less time if patients undergo surgery in a 1-d protocol. With sufficient measuring time, the portable γ-camera enables detection of weak hot spots.

The γ-camera can also be used to adjust skin marks intraoperatively, because the locations of nodes in a patient's definite surgical position may slightly differ from the preoperatively marked location.

Furthermore, if there is difficulty localizing a preoperatively found sentinel node, the γ-camera may be able to show the surgeon where the node is located. In our population, 11% of the nodes were localized more efficiently in this manner. In addition, a node that could not be found with the probe could also be excised. In the latter case, attenuation had probably prevented the probe from picking up the signal, whereas the node could be visualized and localized by continuous counting with the portable γ-camera.

Few authors have reported on their first experience with intraoperative localization of lymph nodes using these cameras (7,8,10–14). Motomura et al. showed that lymphoscintigraphy of sentinel nodes in breast cancer can reliably be reproduced in the operating room using a solid-state γ-camera (11). Kopelman et al. achieved exact intraoperative localization of sentinel nodes with a portable γ-camera in 9 pigs (7), whereas Mathelin et al. were able to localize sentinel nodes with a portable γ-camera in 11 breast cancer patients (12). Mathelin et al. could also successfully estimate the depth of the nodes preoperatively with this camera (12). Scopinaro et al. demonstrated that more nodes were removed when the hand-held γ-camera was used intraoperatively in addition to the γ-probe in breast cancer patients, and operation time was significantly shorter (13). Aarsvod et al., on the other hand, reported a longer operation time when an intraoperative γ-camera was used (14). We could not compare operation length, since the portable γ-camera was used in all patients. Acquiring intraoperative images takes only a few minutes, but if more nodes are found and excised because of intraoperative imaging, the procedure is likely to be prolonged. A recently published study from our institute showed the possibility of separate visualization of different radiotracers with the portable γ-camera, enabling the use of an iodine-source pointer as a guide for laparoscopic sentinel node seeking (8).

Because the γ-camera provides only a 2-dimensional view, the depth of nodes can be visualized only when the detector is turned and positioned at the side of the patient. This problem could be solved using a system with 2 separate detectors, although it would reduce the working space of the surgeon. Possibly in the future an integrated system that combines preoperative SPECT/CT images with intraoperative γ-camera guidance will enable intraoperative 3-dimensional navigation (15). Additionally, if reliable tumor-specific radiotracers were to be developed, a portable γ-camera could be able to guide excision of only the tumor-positive nodes (16).

CONCLUSION

Intraoperative visualization of sentinel nodes with a portable γ-camera is feasible and might improve intraoperative detection of sentinel nodes in the head and neck region. Sentinel nodes close to the injection area can be visualized. Determination of postexcision radioactivity might lead to detection of more sentinel nodes. Further research is warranted to assess the exact value of additional intraoperative imaging in these patients.