Physica Medica xxx (2014) 1e5
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Technical notes
Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients R. Matheoud a, R. Giorgione b, S. Valzano a, G. Sacchetti c, E. Colombo b, M. Brambilla a, * a
Medical Physics Department, University Hospital e Novara, Novara, Italy Dermatology Department, University Hospital e Novara, Novara, Italy c Nuclear Medicine Department, University Hospital e Novara, Novara, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 July 2013 Received in revised form 13 March 2014 Accepted 14 March 2014 Available online xxx
The aim of this study was to determine the suspension level for the sensitivity of an intraoperative scintillation gamma probe in the detection of the sentinel lymph node (SLN) in melanoma patients. Thirty-eight consecutive patients with melanoma were enrolled in the study during a 12-month period and underwent lymphatic scintigraphy after the peritumoral intradermal administration of about 14 MBq of 99mTc-nanocolloids. The SLNs were successfully removed during the surgical intervention about 4 h later. To identify and localize the SLN, a scintillation NaI(Tl) collimated probe was used. Predictably, the probe sensitivity decreased as the photopeak energy window was progressively narrowed, from 6.9 0.7 counts per second (cps)/kBq (designated as the ‘optimum,’ or ‘OPT,’ sensitivity) to 2.5 0.3 cps/kBq (LOW sensitivity) and to 1.4 0.2 cps/kBq (VLOW sensitivity). Maximum lymph node count rates (cps) were determined for the foregoing energy windows prior to skin incision (PREOPT, PRELOW, PREVLOW, respectively) and in vivo after incision (INVOPT, INVLOW, INVVLOW). Forty-three SLNs were removed with a mean source-to-detector distance of 46 24 mm (min 12 mm, max 92 mm). Four SLNs could not have been detected using PRELOW. This figure would have decreased to 34, with nine undetectable lymph nodes, with PREVLOW. One SLN could not have been identified using INVLOW and four could not have be identified using INVVLOW. In the clinical scenario of SLN detection in melanoma patients, a system sensitivity of 2.5 cps/kBq represents a suspension level, that is, a level under which the equipment must be suspended from clinical use and the poor performance must be investigated. Ó 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Intraoperative gamma probe Sentinel lymph node Melanoma Quality control
Introduction Intraoperative probes are now an important, well-established technology in the management of cancer, since they enable surgeons to localise small tumors or lymph nodes to be removed in a surgical procedure. The most established type of intraoperative probe is the non-imaging gamma probe and the most common application is the detection of the sentinel lymph node (SLN) [1]. The technique of lymphatic mapping and SLN biopsy for melanoma has emerged during the last two decades as a minimally invasive approach to evaluate regional lymph node basins in patients with intermediate- and high-risk primary cutaneous
* Corresponding author. Tel.: þ39 321 3733369; fax: þ39 321 3733327. E-mail address: [email protected] (M. Brambilla).
melanoma and has changed the approach to the clinically negative lymph node basin in melanoma [2]. SLN biopsy distinguishes patients without nodal metastases, who can avoid nodal basin dissection with its associated risk of lymphedema, and those with metastatic involvement who might benefit from additional therapy. This procedure represents a significant advantage as a minimally invasive procedure, considering that only an average of 20% of melanoma patients with Breslow thickness between 1.5 and 4 mm harbour metastasis in their SLN and are therefore candidates for elective lymph node dissection procedures [3]. Scintillation gamma probes may be quantitatively characterized by several performance parameters, such as energy resolution and spatial resolution [4,5]. Among them, one of the most relevant parameters is probe sensitivity (or efficiency), that is, the detected count rate per unit activity.
http://dx.doi.org/10.1016/j.ejmp.2014.03.006 1120-1797/Ó 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
2
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
The National Electrical Manufacturers’ Association [6] and a consortium of the Italian Association of Medical Physics, the Italian Association of Nuclear Medicine, the Italian Study Group for Radioguided Surgery and Immunoscintigraphy, and the National Task Group for Breast Cancer (Italy) [7] have promulgated detailed protocols for measurement of sensitivity of intraoperative probes. Unfortunately, these standards do not establish minimum performance levels or minimum acceptance criteria for quality control tests. In addition, such criteria cannot be found in the recently published Radiation Protection Report No. 162 of the European Commission [8], which has established a set of 347 criteria for acceptability of about 30 medical radiological instruments. The most common criterion is a “suspension level” for a measurement of a performance parameter. Equipment failing to meet the suspension-level criterion should either be designated for specified limited applications or suspended from clinical use completely pending appropriately adjustment or repair. For the most common application, that of the detection of the SLN, a minimum requirement of a 5 cps/kBq sensitivity for a gamma probe has been recommended [9]. This value was derived from assumptions regarding the typical injected activity, the time between injection and SLN excision and the radionuclide uptake of the SLN typically encountered in breast cancer surgery but was not based on experimental measurements in a clinical scenario. The purpose of the present study was to determine the suspension levels for the sensitivity of an intraoperative scintillation gamma probe in the detection of the sentinel lymph node in melanoma patients, based on experimental measurements performed both in vivo and in vitro. Material and methods Patient population The study ran from March 2012 to March 2013 and involved, prospectively, 38 patients (16 women, 22 men; aged 34e78 y; mean age, 55 y) with melanoma. The inclusion criteria for performing lymphatic mapping were a Breslow thickness either greater than 1 mm ordwith some risk factors (ulceration, high mitotic rate, Clark level IV or V)dless than 1 mm. The primary lesion was on the lower extremity in 12 patients, on the upper extremity in 3 and on the trunk in 23. The mean Breslow thickness was 2.2 mm (range, 0.5e9.1 mm). The general exclusion criteria were wide local excision of a previous tumor and patients who did not provide consent to participate. Twenty-four patients had a superficial spreading melanoma, 11 nodular melanoma, 1 acrolentiginous melanoma, 1 Spitz Nevus and in 1 patient the lesion could not be classified. In 34 patients, the Breslow thickness was less than 4.0 mm. Ulceration was present in 13 lesions. The study was approved by the Institutional Ethics Committee and informed consent was obtained from all the patients. Lymphoscintigraphy Injection procedure. The same day of surgery, all patients had lymphoscintigraphy to locate the SLN. A mean administered activity of 13.9 2.5 MBq of 99mTechnetium-nanocolloid (GE Healthcare) in a volume ranging from 0.2 to 0.5 ml was injected adjacent to the biopsy wound or primary lesion when still present. The tracer was injected intradermally at 2e5 points, depending on the length of the wound. The tracer was always injected about 0.5e 1.0 cm away from the scar or the tumor margin. The needle was inserted as tangentially as possible to the skin surface. The injection
site was covered with a sticking plaster to prevent leakage of tracer through the punctured skin. Imaging. Images were obtained using a dual-head gamma camera with low-energy, high-resolution collimators. Dynamic acquisition of 60 frames of 20 s in a matrix of 128 128 was started immediately to visualize the progress of lymphatic flow. If necessary, acquisition was continued for 45 min to adequately identify the SLN. Subsequently, anterior and lateral static images with an acquisition time of 300 s in a 256 256 matrix were obtained. In addition to the standard anterior and lateral images, oblique views of 30e45 were obtained whenever the site of injection obscured the SLN. Static images were repeated at 2 h and, when nodes were not clearly depicted, at 4 h after injection. SLNs (first-echelon nodes) were defined as receiving drainage directly from the injection site. Non-SLNs (second-echelon nodes) receive drainage from the SLNs, but were not evaluated in the present study. Scintigraphy was used to determine the draining lymph node basin, to distinguish SLNs from non-SLNs, to determine the number of SLNs, and to mark their location on the skin. Sensitivity measurements Sensitivity of the collimated probe was measured following the protocol promulgated by a consortium of Italian Association of Nuclear Medicine, the Italian Study Group for Radioguided Surgery, and the National Task Group for Breast Cancer [7]. For assessment of sensitivity performance of the probe in different operation settings, a fixed source of a 57Co pen (w3.7 MBq on the reference date; w37 kBq on the date of measurement) was used. The source was used to simulate the level of activity uptake in the target (i.e., about 0.2% of initial injected activity of 18.5 MBq). A high-activity source was not used because the dead time of the system might have caused considerable underestimation of the measured sensitivity. The source was positioned in air at the centre of and in contact with the probe’s sensitive area and counts were acquired. For each probe, thirty 10-s measurements were taken. Measurements were repeated for different modes of operation (i.e., by varying the width of the energy window). Using the standard energy window for 99mTc, that is, 99e 348 keV, the probe sensitivity was 6.9 þ 0.7 cps/kBq (OPT). By narrowing the energy window to 56e227 keV and to 112e199 keV the sensitivities obtained were 2.5 þ 0.3 cps/kBq (LOW) and 1.4 þ 0.2 cps/kBq (VLOW), respectively. The sensitivities were measured according to the protocol in Ref. [7], that is, by using a 57 Co point source of 0.04 MBq placed in close contact to the collimated probe. Surgery Surgery was performed on the same day after scintigraphy (mean time between injection and SLN excision of 3.7 0.75 h). Before incision, the accuracy of the skin mark was verified with a gamma detection probe (MR100-LVR15, Nucleomed) and the count rate (cps) (PRE) was recorded. The count rate (cps) was also measured at another site within the same lymph node basin to establish a background value. The site for the incision was chosen on the basis of the probe measurements and scintigraphy. A small incision of a several centimetres was made. The gamma probe was then used to measure the count rate (cps) in the SLN (INV). The count rate of the SLN was also measured after excision (ex vivo) (EXV). For an ex vivo background, the instrument table of the scrub nurse was used. Finally, the wound was scanned with the probe for SLNs that might have been overlooked, due to limited resolution of lymphoscintigraphy. Non-SLNs that were readily accessible were removed at the discretion of the operating surgeon. The
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
measurement of PRE, INV, and EXV count rates for the lymph node and for the background were repeated in each patient with the three levels of probe sensitivity. Minimum detection count rate If significant fluctuations in the count rate measurement are only due to counting statistics, which follow a Gaussian distribution, then the minimum detection activity can be estimated. According to the Currie’s equation [10], the mean net count rate (ND) corresponding to the minimum detectable activity that achieves false-negative and false-positive rates of no greater than 5% can be calculated as follows [11]:
ND ¼ 4:653
qffiffiffiffiffiffiffiffiffiffi NBkg þ 2:706
(1)
3
zeolites as radioactive sources of different shapes, dimensions and activity concentrations for lesion simulation in Nuclear Medicine imaging has been previously described [12,13]. The average density of the chabazite was previously reported to be 1.73 g cm3. They had irregular shapes and their volumes were in the range of 0.2e1.5 mm3 (Fig. 1A). Sources were allowed to decay after preparation in order to provide a range of count rates similar to those typically found in SLNs. Sources were placed in a cylindrical phantom and immersed in water at depths ranging from 0 to 10 cm from the surface where the collimated probe was suspended (Fig. 1B). In order to avoid release of activity, the zeolites were wrapped with heat-sealed plastic film before being inserted in the water-filled phantom [12].
Results
where NBkg is the background count rate.
In vivo
In vitro measurements
Tables 1 and 2 present the characteristics of the patient population and of the primary tumors and their lymphatic drainage, respectively. The gamma probe system under evaluation in this study showed good precision with a coefficient variation of less than 1%. The mean (standard deviation) background count rates were nearly zero (0.2 0.1 cps), when no radioactive source was around, while increasing to 13.1 3.7 cps when measured at another site within the same lymph node basin of the SLN (Table 2). According to Eq. (1), an SLN was considered detectable, with statistical reliability and within a clinically realistic count time per site of 30 s, when the net count rate over the SLN exceeded the level of 20 cps. Lymphoscintigraphy detected at least one radioactive SLN in 38 of the 38 clinically involved basins. Lymphatic drainage to a second nodal basin outside the clinically involved basin was found in five cases. Overall, 43 SLNs were removed with a mean source-todetector distance of 46 24 mm (min 12, max 92 mm), as measured during the lymphoscintigraphy. The volume of the excised SLNs averaged 0.48 0.41 cm3. No statistically significant correlations were found between SLN probe count rates and SLN depth or volumes.
SLNs were simulated using Bowie chabazites (Fig. 1), a kind of natural zeolites which are hydrated crystalline aluminosilicates with absorptive properties. The methodology followed to use
Table 1 Characteristics of the patients.
Figure 1. Samples of Bowie chabazite zeolites used for SLNs simulation (A). Experimental set-up. The black arrow indicates the zeolite suspended by the white support at different depths in the phantom, the white arrow indicates the probe placed orthogonally to the phantom surface to measure the count rate of the zeolite (B).
Variable
SLN biopsy N ¼ 43(%)
Age (years)
56 12
Sex Male Female
22 (58%) 16 (42%)
Tumor location Arm Trunk Leg
3 23 12
Tumor histology Acral lentiginous Nevus associated Nodular Superficial spreading Breslow thickness (mm)
1 1 11 24 2.2 1.8
Clinically involved basin Axilla Groin Leg
27 9 2
Second nodal basin Axilla Groin
4 1
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
4
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
Table 2 Lymphatic drainage distribution depending on localization of primary lesion of cutaneous melanoma. Localization of primary lesion
Inguinal drainage
Upper extremities Lower extremities Trunk Total
Axillary drainage
Other drainage
3 9 1
28
2 (leg)
10
31
2
As expected, mean PREOPT were significantly higher than PRE(232 265 versus 103 115 cps; p < 0.001). In turn, PRELOW were significantly higher than PREVLOW (67 71 cps; p < 0.001). Mean INVOPT were significantly higher than INVLOW (641 897 versus 332 562 cps; p < 0.001). Again, INVLOW were significantly higher than INVVLOW (158 154 cps; p < 0.001) (Table 3). Four SLN (9.3%) could not have been detected using PRELOW, since the count rate was below 20 cps. This figure would have been increased to nine undetectable SLN (21%) using PREVLOW (Fig. 2A). One SLN could not have been identified using INVLOW, since the count rate was below 20 cps. This figure would have been increased to four unidentifiable SLN using INVVLOW (Fig. 2B). LOW
In vitro The count rate of the simulated SLNs with varying activities and depth are reported in Table 4. With an activity level of 1 MBq in the target, the measured count rates were always above the trigger level of 20 cps, irrespective of the probe sensitivity and of the distance from the source to the probe. On the other hand, with an activity level of 0.1 MBq in the target, the count rate fell below 20 cps at a depth of 2e4 cm when the sensitivity of the probe was 2.5 cps/kBq, indicating a suboptimal performance. Discussion Clinically acceptably performance parameters for gamma probe systems have been suggested and published by various researchers. However, detailed information on the acceptability of the equipment remains limited. Specification of minimum acceptabe performance parameters in general and sensitivity in particular provide practical guidance for the clinical utilization of such devices. As stated by the European Commission [8], “ Failure to meet a suspension level will establish
Figure 2. Frequency distribution of measured count rate (cps) in melanoma patients before surgical excision (A) and intraoperatively (B) for the different levels of probe sensitivity. Vertical line represents the count rate level of 20 cps above which an SLN was considered detectable, with statistical reliability and within a clinically realistic count time per site of 30 s.
Table 4 Count rate characteristics of simulated SLNs using the molecular sieves. Table 3 Count rate characteristics of the SLN. Variable Source-to-detector distance (mm)
SLN
Activity (MBq)
Depth in water (mm)
OPT count rate (cps)
0.1
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
134 45 27 14 8 6 563 195 87 48 29 20 1816 692 398 176 87 55
46 24
Count rate before incision (cps) PREOPT PRELOW PREVLOW Background
232 103 67 13.1
Count rate intraoperative (cps) INVOPT INVLOW INVVLOW
641 897 332 562 158 154
Count rate ex vivo after excision (cps) EXVOPT EXVLOW EXVVLOW Background
766 372 158 0.2
265 115 71 3.7
1065 649 161 0.1
0.5
1
99 26 13 8 4 3 250 98 40 23 10 7 685 188 171 66 21 16
LOW count rate (cps) 100 35 21 13 7 5 199 92 44 29 20 13 1109 429 254 136 68 44
98 21 13 8 4 3 69 12 11 5 6 5 304 123 68 40 13 7
VLOW count rate (cps) 73 25 17 10 6 4 126 65 32 20 14 8 667 292 188 98 50 32
82 18 12 5 4 3 45 22 12 3 4 1 333 57 39 7 7 8
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
that the operation of the equipment involved is sufficiently poor to raise an alarm indicating action is required.” Tiourina et al. [4] evaluated four commercially available surgical gamma probes. They reported a wide range of absolute sensitivities in air measured at 1 cm from the calibrated source both for noncollimated (from 0.83 to 26.4 cps/kBq) and for collimated probes (from 5.0 to 21.8 cps/kBq). Wengenmair et al. [9] tested more then 50 probe/collimator combinations with 13 measurement systems from 8 manufacturers. Their results indicate marked differences in performance. In particular, for collimated probes, sensitivity ranged from 2 to 23 cps/kBq. They recommended a minimum requirement of a 5 cps/ kBq sensitivity for a gamma probe. This suspension level was derived from assumptions for breast cancer surgery regarding the typical injected activity (80 MBq), the time between injection and SLN excision (1 day) and the radionuclide uptake of the SLN (0.05e0.005%). Yu et al. [11] performed an acceptability study on the Navigator system, which had two cadmiumezincetelluride probes of 11-mm and 14-mm diameters. Both probes had internal collimators, and an external collimator was also provided for the 14-mm probe. The sensitivity of the 11-mm probe was 4.8 cps/kBq and the sensitivity of the 14-mm probe was 6.0 cps/kBq. When the external collimator was placed onto the 14-mm probe, its sensitivity decreased by 35% to 3.9 cps/kBq. According to the results of the present study, intraoperative gamma probes used in the search of SLNs in melanoma patients and having an absolute sensitivity equal to or lower than 2.5 cps/ kBq should be suspended from medical use. The results of the present study must be interpreted in the light of some limitations. First, we do not have data available in the range between probe sensitivities of 2.5 cps/kBq, which represents a suspension level, and 6.9 cps/kBq, which represents optimal probe performance. Accordingly, the suspension level might be actually higher than the one determined in the present study. Second, we only evaluated the clinical scenario of detection of SLN in melanoma patients. We cannot guarantee that the same conclusions apply in other clinical scenarios, such as the detection of breast SLNs [14]. Conclusions The evaluation of the gamma probe systems is important and should be performed at acceptance and during quality control
5
programs. The definition of an appropriate suspension level for intraoperative gamma probe sensitivity of 2.5 cps/kBq has been proposed in the present work, based on experimental measurements performed simulating the clinical scenario of SLN detection in melanoma patients.
Conflict of interest None to declare.
References [1] Heller S, Zanzonico P. Nuclear probes and intraoperative gamma cameras. Semin Nucl Med 2011;41:166e81. [2] Boland GM, Gershenwald JE. Sentinel lymph node biopsy in melanoma. Cancer J 2012;18:185e91. [3] Mariani G, Erba P, Manca G, Villa G, Gipponi M, Boni G, et al. Radioguided sentinel lymph node biopsy in patients with malignant cutaneous melanoma: the nuclear medicine contribution. J Surg Oncol 2004;85:141e51. [4] Tiourina T, Arends B, Huysmans D, Rutten H, Lemaire B, Muller S. Evaluation of surgical gamma probes for radioguided sentinel node localisation. Eur J Nucl Med 1998;25:1224e31. [5] Zanzonico P, Heller S. The intraoperative gamma probe: basic principles and choices available. Semin Nucl Med 2000;30:33e48. [6] National Electrical Manufacturer’s Association. NEMA standards publication NU-3-2004: performance measurements and quality controls guidelines for non-imaging intraoperative gamma probes. Rosslyn, VA: National Electrical Manufacturer’s Association; 2004. [7] Chiesa C, Toscano F, Mariani M, Bombardieri E. Physical performance parameters of intraoperative probes. In: Approach G, et al., editors. Radioguided surgery: a comprehensive team. New York: Springer; 2008. pp. 12e22. [8] EC. European commission (2012) criteria for acceptability of medical radiological equipment used in diagnostic radiology, nuclear medicine and radiotherapy. Luxembourg: Office for Official Publications of the European Communities; 2012. [9] Wengenmair H, Kopp J. Gamma probes for sentinel lymph node localization: quality criteria, minimal requirements and quality of commercially available systems. Available from: http://www.klinikum-augsburg.de/index.php/ fuseaction/download/lrn_file/gammaprobes.pdf [accessed 28.05.2013]. [10] Currie LA. Limits for qualitative detection and quantitative determination, application to radiochemistry. Anal Chem 1968;40:586e93. [11 ] Yu SK, Ma KM, Wong KN, Leung J, Leung LC. Intraoperative gamma probe for Sentinel node localisation: evaluation study. J HK Coll Radiol 2005;8:40e8. [12] Bailey DL, Snowdon G, Cooper RG, Roach PJ. The use of molecular sieves to produce point sources of radioactivity. Phys Med Biol 2004;49:N21e9. [13] Matheoud R, Secco C, Ridone S, Inglese E, Brambilla M. The use of molecular sieves to simulate hot lesions in 18F-fluorodeoxyglucose- positron emission tomography imaging. Phys Med Biol 2008;53:N137e48. [14] Trifirò G, Lavinia Travaini L, De Cicco C, Paganelli G. Sentinel node detection and radioguided occult lesion localization in breast cancer. Phys Med 2006;21(Suppl. 1):20e3.
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Technical notes
Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients R. Matheoud a, R. Giorgione b, S. Valzano a, G. Sacchetti c, E. Colombo b, M. Brambilla a, * a
Medical Physics Department, University Hospital e Novara, Novara, Italy Dermatology Department, University Hospital e Novara, Novara, Italy c Nuclear Medicine Department, University Hospital e Novara, Novara, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 July 2013 Received in revised form 13 March 2014 Accepted 14 March 2014 Available online xxx
The aim of this study was to determine the suspension level for the sensitivity of an intraoperative scintillation gamma probe in the detection of the sentinel lymph node (SLN) in melanoma patients. Thirty-eight consecutive patients with melanoma were enrolled in the study during a 12-month period and underwent lymphatic scintigraphy after the peritumoral intradermal administration of about 14 MBq of 99mTc-nanocolloids. The SLNs were successfully removed during the surgical intervention about 4 h later. To identify and localize the SLN, a scintillation NaI(Tl) collimated probe was used. Predictably, the probe sensitivity decreased as the photopeak energy window was progressively narrowed, from 6.9 0.7 counts per second (cps)/kBq (designated as the ‘optimum,’ or ‘OPT,’ sensitivity) to 2.5 0.3 cps/kBq (LOW sensitivity) and to 1.4 0.2 cps/kBq (VLOW sensitivity). Maximum lymph node count rates (cps) were determined for the foregoing energy windows prior to skin incision (PREOPT, PRELOW, PREVLOW, respectively) and in vivo after incision (INVOPT, INVLOW, INVVLOW). Forty-three SLNs were removed with a mean source-to-detector distance of 46 24 mm (min 12 mm, max 92 mm). Four SLNs could not have been detected using PRELOW. This figure would have decreased to 34, with nine undetectable lymph nodes, with PREVLOW. One SLN could not have been identified using INVLOW and four could not have be identified using INVVLOW. In the clinical scenario of SLN detection in melanoma patients, a system sensitivity of 2.5 cps/kBq represents a suspension level, that is, a level under which the equipment must be suspended from clinical use and the poor performance must be investigated. Ó 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Intraoperative gamma probe Sentinel lymph node Melanoma Quality control
Introduction Intraoperative probes are now an important, well-established technology in the management of cancer, since they enable surgeons to localise small tumors or lymph nodes to be removed in a surgical procedure. The most established type of intraoperative probe is the non-imaging gamma probe and the most common application is the detection of the sentinel lymph node (SLN) [1]. The technique of lymphatic mapping and SLN biopsy for melanoma has emerged during the last two decades as a minimally invasive approach to evaluate regional lymph node basins in patients with intermediate- and high-risk primary cutaneous
* Corresponding author. Tel.: þ39 321 3733369; fax: þ39 321 3733327. E-mail address: [email protected] (M. Brambilla).
melanoma and has changed the approach to the clinically negative lymph node basin in melanoma [2]. SLN biopsy distinguishes patients without nodal metastases, who can avoid nodal basin dissection with its associated risk of lymphedema, and those with metastatic involvement who might benefit from additional therapy. This procedure represents a significant advantage as a minimally invasive procedure, considering that only an average of 20% of melanoma patients with Breslow thickness between 1.5 and 4 mm harbour metastasis in their SLN and are therefore candidates for elective lymph node dissection procedures [3]. Scintillation gamma probes may be quantitatively characterized by several performance parameters, such as energy resolution and spatial resolution [4,5]. Among them, one of the most relevant parameters is probe sensitivity (or efficiency), that is, the detected count rate per unit activity.
http://dx.doi.org/10.1016/j.ejmp.2014.03.006 1120-1797/Ó 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
2
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
The National Electrical Manufacturers’ Association [6] and a consortium of the Italian Association of Medical Physics, the Italian Association of Nuclear Medicine, the Italian Study Group for Radioguided Surgery and Immunoscintigraphy, and the National Task Group for Breast Cancer (Italy) [7] have promulgated detailed protocols for measurement of sensitivity of intraoperative probes. Unfortunately, these standards do not establish minimum performance levels or minimum acceptance criteria for quality control tests. In addition, such criteria cannot be found in the recently published Radiation Protection Report No. 162 of the European Commission [8], which has established a set of 347 criteria for acceptability of about 30 medical radiological instruments. The most common criterion is a “suspension level” for a measurement of a performance parameter. Equipment failing to meet the suspension-level criterion should either be designated for specified limited applications or suspended from clinical use completely pending appropriately adjustment or repair. For the most common application, that of the detection of the SLN, a minimum requirement of a 5 cps/kBq sensitivity for a gamma probe has been recommended [9]. This value was derived from assumptions regarding the typical injected activity, the time between injection and SLN excision and the radionuclide uptake of the SLN typically encountered in breast cancer surgery but was not based on experimental measurements in a clinical scenario. The purpose of the present study was to determine the suspension levels for the sensitivity of an intraoperative scintillation gamma probe in the detection of the sentinel lymph node in melanoma patients, based on experimental measurements performed both in vivo and in vitro. Material and methods Patient population The study ran from March 2012 to March 2013 and involved, prospectively, 38 patients (16 women, 22 men; aged 34e78 y; mean age, 55 y) with melanoma. The inclusion criteria for performing lymphatic mapping were a Breslow thickness either greater than 1 mm ordwith some risk factors (ulceration, high mitotic rate, Clark level IV or V)dless than 1 mm. The primary lesion was on the lower extremity in 12 patients, on the upper extremity in 3 and on the trunk in 23. The mean Breslow thickness was 2.2 mm (range, 0.5e9.1 mm). The general exclusion criteria were wide local excision of a previous tumor and patients who did not provide consent to participate. Twenty-four patients had a superficial spreading melanoma, 11 nodular melanoma, 1 acrolentiginous melanoma, 1 Spitz Nevus and in 1 patient the lesion could not be classified. In 34 patients, the Breslow thickness was less than 4.0 mm. Ulceration was present in 13 lesions. The study was approved by the Institutional Ethics Committee and informed consent was obtained from all the patients. Lymphoscintigraphy Injection procedure. The same day of surgery, all patients had lymphoscintigraphy to locate the SLN. A mean administered activity of 13.9 2.5 MBq of 99mTechnetium-nanocolloid (GE Healthcare) in a volume ranging from 0.2 to 0.5 ml was injected adjacent to the biopsy wound or primary lesion when still present. The tracer was injected intradermally at 2e5 points, depending on the length of the wound. The tracer was always injected about 0.5e 1.0 cm away from the scar or the tumor margin. The needle was inserted as tangentially as possible to the skin surface. The injection
site was covered with a sticking plaster to prevent leakage of tracer through the punctured skin. Imaging. Images were obtained using a dual-head gamma camera with low-energy, high-resolution collimators. Dynamic acquisition of 60 frames of 20 s in a matrix of 128 128 was started immediately to visualize the progress of lymphatic flow. If necessary, acquisition was continued for 45 min to adequately identify the SLN. Subsequently, anterior and lateral static images with an acquisition time of 300 s in a 256 256 matrix were obtained. In addition to the standard anterior and lateral images, oblique views of 30e45 were obtained whenever the site of injection obscured the SLN. Static images were repeated at 2 h and, when nodes were not clearly depicted, at 4 h after injection. SLNs (first-echelon nodes) were defined as receiving drainage directly from the injection site. Non-SLNs (second-echelon nodes) receive drainage from the SLNs, but were not evaluated in the present study. Scintigraphy was used to determine the draining lymph node basin, to distinguish SLNs from non-SLNs, to determine the number of SLNs, and to mark their location on the skin. Sensitivity measurements Sensitivity of the collimated probe was measured following the protocol promulgated by a consortium of Italian Association of Nuclear Medicine, the Italian Study Group for Radioguided Surgery, and the National Task Group for Breast Cancer [7]. For assessment of sensitivity performance of the probe in different operation settings, a fixed source of a 57Co pen (w3.7 MBq on the reference date; w37 kBq on the date of measurement) was used. The source was used to simulate the level of activity uptake in the target (i.e., about 0.2% of initial injected activity of 18.5 MBq). A high-activity source was not used because the dead time of the system might have caused considerable underestimation of the measured sensitivity. The source was positioned in air at the centre of and in contact with the probe’s sensitive area and counts were acquired. For each probe, thirty 10-s measurements were taken. Measurements were repeated for different modes of operation (i.e., by varying the width of the energy window). Using the standard energy window for 99mTc, that is, 99e 348 keV, the probe sensitivity was 6.9 þ 0.7 cps/kBq (OPT). By narrowing the energy window to 56e227 keV and to 112e199 keV the sensitivities obtained were 2.5 þ 0.3 cps/kBq (LOW) and 1.4 þ 0.2 cps/kBq (VLOW), respectively. The sensitivities were measured according to the protocol in Ref. [7], that is, by using a 57 Co point source of 0.04 MBq placed in close contact to the collimated probe. Surgery Surgery was performed on the same day after scintigraphy (mean time between injection and SLN excision of 3.7 0.75 h). Before incision, the accuracy of the skin mark was verified with a gamma detection probe (MR100-LVR15, Nucleomed) and the count rate (cps) (PRE) was recorded. The count rate (cps) was also measured at another site within the same lymph node basin to establish a background value. The site for the incision was chosen on the basis of the probe measurements and scintigraphy. A small incision of a several centimetres was made. The gamma probe was then used to measure the count rate (cps) in the SLN (INV). The count rate of the SLN was also measured after excision (ex vivo) (EXV). For an ex vivo background, the instrument table of the scrub nurse was used. Finally, the wound was scanned with the probe for SLNs that might have been overlooked, due to limited resolution of lymphoscintigraphy. Non-SLNs that were readily accessible were removed at the discretion of the operating surgeon. The
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
measurement of PRE, INV, and EXV count rates for the lymph node and for the background were repeated in each patient with the three levels of probe sensitivity. Minimum detection count rate If significant fluctuations in the count rate measurement are only due to counting statistics, which follow a Gaussian distribution, then the minimum detection activity can be estimated. According to the Currie’s equation [10], the mean net count rate (ND) corresponding to the minimum detectable activity that achieves false-negative and false-positive rates of no greater than 5% can be calculated as follows [11]:
ND ¼ 4:653
qffiffiffiffiffiffiffiffiffiffi NBkg þ 2:706
(1)
3
zeolites as radioactive sources of different shapes, dimensions and activity concentrations for lesion simulation in Nuclear Medicine imaging has been previously described [12,13]. The average density of the chabazite was previously reported to be 1.73 g cm3. They had irregular shapes and their volumes were in the range of 0.2e1.5 mm3 (Fig. 1A). Sources were allowed to decay after preparation in order to provide a range of count rates similar to those typically found in SLNs. Sources were placed in a cylindrical phantom and immersed in water at depths ranging from 0 to 10 cm from the surface where the collimated probe was suspended (Fig. 1B). In order to avoid release of activity, the zeolites were wrapped with heat-sealed plastic film before being inserted in the water-filled phantom [12].
Results
where NBkg is the background count rate.
In vivo
In vitro measurements
Tables 1 and 2 present the characteristics of the patient population and of the primary tumors and their lymphatic drainage, respectively. The gamma probe system under evaluation in this study showed good precision with a coefficient variation of less than 1%. The mean (standard deviation) background count rates were nearly zero (0.2 0.1 cps), when no radioactive source was around, while increasing to 13.1 3.7 cps when measured at another site within the same lymph node basin of the SLN (Table 2). According to Eq. (1), an SLN was considered detectable, with statistical reliability and within a clinically realistic count time per site of 30 s, when the net count rate over the SLN exceeded the level of 20 cps. Lymphoscintigraphy detected at least one radioactive SLN in 38 of the 38 clinically involved basins. Lymphatic drainage to a second nodal basin outside the clinically involved basin was found in five cases. Overall, 43 SLNs were removed with a mean source-todetector distance of 46 24 mm (min 12, max 92 mm), as measured during the lymphoscintigraphy. The volume of the excised SLNs averaged 0.48 0.41 cm3. No statistically significant correlations were found between SLN probe count rates and SLN depth or volumes.
SLNs were simulated using Bowie chabazites (Fig. 1), a kind of natural zeolites which are hydrated crystalline aluminosilicates with absorptive properties. The methodology followed to use
Table 1 Characteristics of the patients.
Figure 1. Samples of Bowie chabazite zeolites used for SLNs simulation (A). Experimental set-up. The black arrow indicates the zeolite suspended by the white support at different depths in the phantom, the white arrow indicates the probe placed orthogonally to the phantom surface to measure the count rate of the zeolite (B).
Variable
SLN biopsy N ¼ 43(%)
Age (years)
56 12
Sex Male Female
22 (58%) 16 (42%)
Tumor location Arm Trunk Leg
3 23 12
Tumor histology Acral lentiginous Nevus associated Nodular Superficial spreading Breslow thickness (mm)
1 1 11 24 2.2 1.8
Clinically involved basin Axilla Groin Leg
27 9 2
Second nodal basin Axilla Groin
4 1
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
4
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
Table 2 Lymphatic drainage distribution depending on localization of primary lesion of cutaneous melanoma. Localization of primary lesion
Inguinal drainage
Upper extremities Lower extremities Trunk Total
Axillary drainage
Other drainage
3 9 1
28
2 (leg)
10
31
2
As expected, mean PREOPT were significantly higher than PRE(232 265 versus 103 115 cps; p < 0.001). In turn, PRELOW were significantly higher than PREVLOW (67 71 cps; p < 0.001). Mean INVOPT were significantly higher than INVLOW (641 897 versus 332 562 cps; p < 0.001). Again, INVLOW were significantly higher than INVVLOW (158 154 cps; p < 0.001) (Table 3). Four SLN (9.3%) could not have been detected using PRELOW, since the count rate was below 20 cps. This figure would have been increased to nine undetectable SLN (21%) using PREVLOW (Fig. 2A). One SLN could not have been identified using INVLOW, since the count rate was below 20 cps. This figure would have been increased to four unidentifiable SLN using INVVLOW (Fig. 2B). LOW
In vitro The count rate of the simulated SLNs with varying activities and depth are reported in Table 4. With an activity level of 1 MBq in the target, the measured count rates were always above the trigger level of 20 cps, irrespective of the probe sensitivity and of the distance from the source to the probe. On the other hand, with an activity level of 0.1 MBq in the target, the count rate fell below 20 cps at a depth of 2e4 cm when the sensitivity of the probe was 2.5 cps/kBq, indicating a suboptimal performance. Discussion Clinically acceptably performance parameters for gamma probe systems have been suggested and published by various researchers. However, detailed information on the acceptability of the equipment remains limited. Specification of minimum acceptabe performance parameters in general and sensitivity in particular provide practical guidance for the clinical utilization of such devices. As stated by the European Commission [8], “ Failure to meet a suspension level will establish
Figure 2. Frequency distribution of measured count rate (cps) in melanoma patients before surgical excision (A) and intraoperatively (B) for the different levels of probe sensitivity. Vertical line represents the count rate level of 20 cps above which an SLN was considered detectable, with statistical reliability and within a clinically realistic count time per site of 30 s.
Table 4 Count rate characteristics of simulated SLNs using the molecular sieves. Table 3 Count rate characteristics of the SLN. Variable Source-to-detector distance (mm)
SLN
Activity (MBq)
Depth in water (mm)
OPT count rate (cps)
0.1
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
134 45 27 14 8 6 563 195 87 48 29 20 1816 692 398 176 87 55
46 24
Count rate before incision (cps) PREOPT PRELOW PREVLOW Background
232 103 67 13.1
Count rate intraoperative (cps) INVOPT INVLOW INVVLOW
641 897 332 562 158 154
Count rate ex vivo after excision (cps) EXVOPT EXVLOW EXVVLOW Background
766 372 158 0.2
265 115 71 3.7
1065 649 161 0.1
0.5
1
99 26 13 8 4 3 250 98 40 23 10 7 685 188 171 66 21 16
LOW count rate (cps) 100 35 21 13 7 5 199 92 44 29 20 13 1109 429 254 136 68 44
98 21 13 8 4 3 69 12 11 5 6 5 304 123 68 40 13 7
VLOW count rate (cps) 73 25 17 10 6 4 126 65 32 20 14 8 667 292 188 98 50 32
82 18 12 5 4 3 45 22 12 3 4 1 333 57 39 7 7 8
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
R. Matheoud et al. / Physica Medica xxx (2014) 1e5
that the operation of the equipment involved is sufficiently poor to raise an alarm indicating action is required.” Tiourina et al. [4] evaluated four commercially available surgical gamma probes. They reported a wide range of absolute sensitivities in air measured at 1 cm from the calibrated source both for noncollimated (from 0.83 to 26.4 cps/kBq) and for collimated probes (from 5.0 to 21.8 cps/kBq). Wengenmair et al. [9] tested more then 50 probe/collimator combinations with 13 measurement systems from 8 manufacturers. Their results indicate marked differences in performance. In particular, for collimated probes, sensitivity ranged from 2 to 23 cps/kBq. They recommended a minimum requirement of a 5 cps/ kBq sensitivity for a gamma probe. This suspension level was derived from assumptions for breast cancer surgery regarding the typical injected activity (80 MBq), the time between injection and SLN excision (1 day) and the radionuclide uptake of the SLN (0.05e0.005%). Yu et al. [11] performed an acceptability study on the Navigator system, which had two cadmiumezincetelluride probes of 11-mm and 14-mm diameters. Both probes had internal collimators, and an external collimator was also provided for the 14-mm probe. The sensitivity of the 11-mm probe was 4.8 cps/kBq and the sensitivity of the 14-mm probe was 6.0 cps/kBq. When the external collimator was placed onto the 14-mm probe, its sensitivity decreased by 35% to 3.9 cps/kBq. According to the results of the present study, intraoperative gamma probes used in the search of SLNs in melanoma patients and having an absolute sensitivity equal to or lower than 2.5 cps/ kBq should be suspended from medical use. The results of the present study must be interpreted in the light of some limitations. First, we do not have data available in the range between probe sensitivities of 2.5 cps/kBq, which represents a suspension level, and 6.9 cps/kBq, which represents optimal probe performance. Accordingly, the suspension level might be actually higher than the one determined in the present study. Second, we only evaluated the clinical scenario of detection of SLN in melanoma patients. We cannot guarantee that the same conclusions apply in other clinical scenarios, such as the detection of breast SLNs [14]. Conclusions The evaluation of the gamma probe systems is important and should be performed at acceptance and during quality control
5
programs. The definition of an appropriate suspension level for intraoperative gamma probe sensitivity of 2.5 cps/kBq has been proposed in the present work, based on experimental measurements performed simulating the clinical scenario of SLN detection in melanoma patients.
Conflict of interest None to declare.
References [1] Heller S, Zanzonico P. Nuclear probes and intraoperative gamma cameras. Semin Nucl Med 2011;41:166e81. [2] Boland GM, Gershenwald JE. Sentinel lymph node biopsy in melanoma. Cancer J 2012;18:185e91. [3] Mariani G, Erba P, Manca G, Villa G, Gipponi M, Boni G, et al. Radioguided sentinel lymph node biopsy in patients with malignant cutaneous melanoma: the nuclear medicine contribution. J Surg Oncol 2004;85:141e51. [4] Tiourina T, Arends B, Huysmans D, Rutten H, Lemaire B, Muller S. Evaluation of surgical gamma probes for radioguided sentinel node localisation. Eur J Nucl Med 1998;25:1224e31. [5] Zanzonico P, Heller S. The intraoperative gamma probe: basic principles and choices available. Semin Nucl Med 2000;30:33e48. [6] National Electrical Manufacturer’s Association. NEMA standards publication NU-3-2004: performance measurements and quality controls guidelines for non-imaging intraoperative gamma probes. Rosslyn, VA: National Electrical Manufacturer’s Association; 2004. [7] Chiesa C, Toscano F, Mariani M, Bombardieri E. Physical performance parameters of intraoperative probes. In: Approach G, et al., editors. Radioguided surgery: a comprehensive team. New York: Springer; 2008. pp. 12e22. [8] EC. European commission (2012) criteria for acceptability of medical radiological equipment used in diagnostic radiology, nuclear medicine and radiotherapy. Luxembourg: Office for Official Publications of the European Communities; 2012. [9] Wengenmair H, Kopp J. Gamma probes for sentinel lymph node localization: quality criteria, minimal requirements and quality of commercially available systems. Available from: http://www.klinikum-augsburg.de/index.php/ fuseaction/download/lrn_file/gammaprobes.pdf [accessed 28.05.2013]. [10] Currie LA. Limits for qualitative detection and quantitative determination, application to radiochemistry. Anal Chem 1968;40:586e93. [11 ] Yu SK, Ma KM, Wong KN, Leung J, Leung LC. Intraoperative gamma probe for Sentinel node localisation: evaluation study. J HK Coll Radiol 2005;8:40e8. [12] Bailey DL, Snowdon G, Cooper RG, Roach PJ. The use of molecular sieves to produce point sources of radioactivity. Phys Med Biol 2004;49:N21e9. [13] Matheoud R, Secco C, Ridone S, Inglese E, Brambilla M. The use of molecular sieves to simulate hot lesions in 18F-fluorodeoxyglucose- positron emission tomography imaging. Phys Med Biol 2008;53:N137e48. [14] Trifirò G, Lavinia Travaini L, De Cicco C, Paganelli G. Sentinel node detection and radioguided occult lesion localization in breast cancer. Phys Med 2006;21(Suppl. 1):20e3.
Please cite this article in press as: Matheoud R, et al., Minimum acceptable sensitivity of intraoperative gamma probes used for sentinel lymph node detection in melanoma patients, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.03.006
