3.A. Laboratory characterization of imaging properties
3.A.a. Energy resolution:
Averaged over seven trials the energy resolution was measured to be 21.5 ± 1.7 % (mean ± 95% confidence interval) at 140 keV.
3.A.b. Spatial resolution
2D: The phase-averaged FWHM of the width of the capillary image in each 2D projection image is plotted versus changing capillary-to-collimator separation in Figure 3. The average of seven trials is shown.
3D: Averaged over three scans, the FWHM of the image of the point source was 11.9 +/- 2.5 mm, 13.3 +/- 0.3 mm, and 14.2 +/- 1.8 mm in the coronal, sagittal, and axial planes, respectively. These three planes were defined in terms of a supine surgical patient.
3.A.c. Sensitivity
The 2D sensitivity of the SiPM gamma camera was found to be 171 ± 6.2 cps/MBq. 3D sensitivity is shown plotted versus source activity in Figure 4. The results show that the 3D sensitivity is uniform in this range of activity with a mean sensitivity of 203 ± 19.5 cps/MBq. The small difference between the 2D and 3D sensitivity is attributable to the use of a slightly more narrow energy window in 2D imaging compared to a wider energy window used for 3D scanning.
3.A.d. Node localization accuracy
The bar plot of Figure 5 compares the known separations between the camera and the eight simulated nodes to those reported by the imaging system. For the 16 trials (eight nodes and two post-scan camera positions each) the average error in the reported camera-to-node separation was 9.2 mm with a standard deviation of 2.7 mm. However, the separation reported by the system is actually that between the node and the input surface of the NaI(Tl) scintillator. That surface is separated from the camera’s outer surface by the thickness of the camera housing. Taking into account the combined 9 mm thickness of the collimator and camera housing, this corresponds to a true error of 0.2 ± 2.7 mm.
Averaged over all camera-to-phantom distances tested, the mean absolute error in the inter-node separation, computed from system-reported node locations, was 1.2 ± 0.34 mm (95% confidence interval).
3.B. Human Feasibility Study
A breakdown of the human study results in terms of the numbers of SLNs imaged and excised, along with the lymphatic basins in which they were located, is presented in Table 1 and Figure 8. Fifteen total nodes were excised from the five subjects. Of those, only one node (left groin, subject #5) was positive for malignancy.
Table 1. Number and location of SLNs imaged by lymphoscintigraphy, pre-operative probe fhSPECT, and pre-operative camera fhSPECT, and number and location of SLNs subsequently excised during surgery.
Case
|
Melanoma Site
|
Lympho-scintigraphy
|
Probe fhSPECT
|
Camera fhSPECT
|
Excised
|
3
|
Right lower abdomen
|
2 left groin
0 right groin
|
no scan
no scan
|
3 left groin
5 right groin
|
3 left groin
1 right groin
|
4
|
Midline epigastrium
|
1 left axilla
1 right axilla
|
1 left axilla
no scan
|
4 left axilla
3 right axilla
|
2 left axilla
1 right axilla
|
5
|
Lower back
|
4 left groin
5 right groin
|
1 left groin
1 right groin
|
2 left groin
4 right groin
|
2 left groin
2 right groin
|
6
|
Left heel
|
1 left knee
1 left groin
|
0 left knee
1 left groin
|
2 left knee
3 left groin
|
0 left knee
1 left groin
|
7
|
Right upper back
|
1 right axilla
|
no scan
|
5 right axilla
|
3 right axilla
|
Table 1 shows that for a number of basins (Case 3, left and right groins; Case 4, left axilla; and case 7, right axilla) lymphoscintigraphy identified fewer SLNs than were ultimately excised, suggesting false negatives via lymphoscintigraphy. However, for each of those basins the number of nodes identified by camera fhSPECT was equal to or greater than the number ultimately excised.
Probe fhSPECT scans were performed in Cases 4-6, but were precluded by insufficient pre-op time in Cases 3 and 7. For the same reason, the right axilla in Case 4 was not scanned. In the left axilla of Case 4, probe fhSPECT detected one fewer SLN than was excised. The same was true for both basins in Case 5. In Case 6 the number of SLNs detected by probe fhSECT matched the number excised in both basins.
For seven of the nine lymphatic basins imaged, the number of nodes identified by the investigational handheld camera fhSPECT system was either greater than or equal to the number visualized via lymphoscintigraphy. In Case 5, the number of nodes detected by lymphoscintigraphy exceeded that detected by camera fhSPECT in both the left and right axillary basins.
Three subjects (Cases 4, 5, and 6) were scanned with both probe fhSPECT and camera fhSPECT. For each nodal basin scanned in these three subjects, camera fhSPECT was able to visualize a greater number of nodes than probe fhSPECT. This difference could be attributable to the increased tomographic sampling available with the multi-detector element SiPM camera compared to the single detector element probe.
There were no nodal basins for which the number of SLNs detected via camera fhSPECT was fewer than the number ultimately excised. However, in seven basins the total number of hot spots imaged via pre-operative 3D scanning was greater than the number of nodes excised. Unfortunately, there was no way for nodes that were detected by 3D imaging, but not identified for excision conventionally, to be investigated during surgery or histologically, because the IRB-approved study protocol stipulated that node localization and excision were to be performed according to the current standard of care (guided only by lymphoscintigraphy and the non-imaging probe). Thus, the surgical team was blinded to the results of the 3D scans.
Nodes detected via imaging but not excised could fall into one of several categories. They could be nodes that were simply not detected by the non-imaging probe. Alternatively, they could be nodes that were detected by the probe but were not excised because their activity as measured in vivo with the probe was below the 10% threshold. It is possible for deeper lying nodes with in vivo activities below the 10% exclusion threshold to nevertheless have image-determined activities above the threshold because the attenuation correction algorithms built into the declipse®SPECT iterative reconstruction software can partially compensate for attenuation of node-emitted gamma rays by overlying tissue. In these cases, the activity determined from the 3D scan may be more accurate than that determined via the non-imaging probe. It is also possible for a node to be detected by the probe, and pass the 10% criterion, but be intentionally not excised because of surgical considerations such as a risk of complications associated with the required surgical path. It is also possible that the excess imaged nodes included higher echelon (non-sentinel) nodes.
In most institutions, radiotracer injection and lymphoscintigraphy are performed in the nuclear medicine clinic rather than the surgical pre-operative suite. Inter-site scheduling offsets and patient transfer times often dictate that injection/lymphoscintigraphy occur on the day prior to surgery, especially for patients scheduled for morning surgery. Studies comparing early (1-day) and delayed (2-day) SLNB protocols in breast cancer have demonstrated that the patterns of distribution of the sulfur colloid radiotracer in the lymph nodes and the results of SLNB are virtually identical for the early and delayed protocols [32,33]. However, the delayed protocol results in significant radioactive decay of the 99mTc (half-life = 6.0 hours) lymphatic tracer during the time interval between tracer injection and surgery, and thus very low (~ 2 MBq) SLN activity. Therefore, it was important to evaluate the ability of the investigational fhSPECT system to visualize SLNs with very low activity.
Table 2 shows, for each of the five cases reported here, the elapsed time between injection and lymphoscintigraphy (always 15 minutes post injection at our institution) and between injection and 3D scanning (probe fhSPECT and/or camera fhSPECT), along with the fractional reduction in the SLN activity. In three of the five cases, less than 15% of the SLN activity at the time of lymphoscintigraphy remained when the 3D scans were performed. In this study, for all but two basins (the left and right groins in case 5), the number of SLNs detected by camera fhSPECT was always equal to or greater than the number of SLNs detected by lymphoscintigraphy. The fact that this was true even when less than 15% of the activity remained during the 3D scan demonstrates the high sensitivity of camera fhSPECT.
Table 2. SLN radioactive decay for each of the five cases in this study. In the three cases shaded, less than 15% of the SLN activity present at lymphoscintigraphy remained at probe- or camera-fhSPECT imaging. The two 3D scans (camera fhSPECT and probe fhSPECT) were performed with no lapsed time between them, so their injection-to-scan time was identical). LS = lymphoscintigraphy; 3D scans = camera fhSPECT and (when performed) probe fhSPECT
Case #
|
Injected activity (MBq)
|
Injection-LS time (hrs)
|
Injection-3D scanning time (hrs)
|
Fraction of injected activity remaining at LS
|
Fraction of injected activity remaining at 3D scans
|
3
|
18.5
|
0.25
|
19.7
|
0.97
|
0.10
|
4
|
20.4
|
0.25
|
3.1
|
0.97
|
0.70
|
5
|
18.9
|
0.25
|
18.4
|
0.97
|
0.12
|
6
|
20.4
|
0.25
|
22.7
|
0.97
|
0.07
|
7
|
18.5
|
0.25
|
4.4
|
0.97
|
0.60
|