Background
Despite low incidence of lymph node (LN) metastases in renal cancer (RCC), dissemination into lymph nodes portends extremely poor prognosis [
1,
2]. In low risk and high risk cancers, the detection rate with a routine pathohistological staining is only 4% and 10%, respectively and lymph node dissection (LND) in low risk cancers has no proven survival benefit [
3,
4]. Furthermore, lymphatic drainage in renal tumors is unpredictable and drainage outside local retroperitoneal lymph nodes may be one of the reasons for the lack of a survival advantage with conventional LND [
5‐
7]. Sentinel lymph node biopsy (SLNB) typically consists of a preoperative lymphoscintigraphy to indicate the anatomical location of the first tumor-draining lymph nodes, i.e., the sentinel node(s), combined with a γ-probe guided surgical biopsy of these nodes. SLNB may improve staging of RCC especially in patients who have aberrant drainage within or outside retroperitoneal lumbar lymph node basins and might have a role in future translational research of tumor immunology and biology of early metastasis. It has been suggested in prostate cancer that extended pelvic lymph node dissection combined with SLNB increases the yield of nodal lymph node metastases, especially in high risk disease [
8]. Due to proven efficacy in staging procedures in many cancer types and less associated morbidity of SLNB compared to LND, SLNB is routinely used in melanoma, breast, head-neck, vulvar, cervical, and penile cancer and actively studied in many other tumor types [
9,
10]. Feasibility of SLNB procedure in renal tumors has been studied and confirmed earlier by our group and others [
11,
12]. However, both groups had high rate of non-visualization of sentinel lymph node (SN) on scintigraphy and single-photon emission computed tomography with computed tomography (SPECT/CT). Also, the procedure is currently lacking standardization. These factors may hamper widespread applicability and adoption to further study its use in clinical practice. Therefore, the aim of our study was to determine incidence and predictive factors of non-visualization of SN on scintigraphy and SPECT/CT in renal tumors, evaluate detection rate at surgery, and propose a standardized protocol for future studies.
Results
Of the 78 patients, 5 patients were excluded because of ineligibility. Therefore, 73 patients were available for final analysis (Fig.
1). The characteristics of the patients and the procedure are shown in Table
2. The majority of patients (68.5%) had low or intermediate risk clear cell RCC. Median time between the injection and early and late imaging with visualization and non-visualization was 33 min (IQR 19–51 min), 28 min (IQR 15–65 min) and 149 min (IQR 175–188), 182 min (IQR 139–194) respectively. Median tracer dose was 209 mBq (IQR 187–222) and 212 (IQR 196–218) with visualization and non-visualization, respectively. Visualization of SNs on imaging was 63% [95% CI 50–73%]. All SNs visualized on planar images were also visualized on SPECT/CT. A total of 80 SNs in 46 patients were visualized, and the mean number of SNs on imaging was 1.1 (IQR 0–2).
In six patients, no SN could be harvested. In three patients, there was no activity detected with a γ-probe in vivo nor ex vivo and a selective dissection of lymph nodes was performed in the area of the SNs as visualized on SPECT/CT. In three patients, radioactive SNs were detected with a γ-probe but these were not harvested due to severe obesity and risky dissection or non-accessibility through the exposure. On the contrary, in five patients with non-visualization on pre-operative imaging, radioactive SNs were detected with a γ-probe the next day at surgery. Therefore, SNs were detected and harvested in 45 patients (61% [95% CI 49–72%]). The mean number of harvested SNs was 2.12 (IQR 0–3) and mean number of non-SNs was 3.01 (IQR 0–4). A total of 155 SNs were removed. Two patients had occult SN metastases (one was with visualized SN and the other had non-visualization).
Factors associated with non-visualization
Table
1 shows factors for visualization, univariable and multivariable analysis. In univariable and multivariable analysis, an increased risk of non-visualization was associated with patient age, showing a trend toward older patients having higher risk of non-visualization (
p = 0.008 and 0.002, respectively). Whereas including age categories with below and above median age of 59 years did not show statistical significance in univariable analysis (
p = 0.11). Other factor associated with non-visualization in univariable analysis was kidney/liver activity on scintigraphy at late planar imaging; however, this factor did not reach statistical significance in multivariable model.
Discussion
SLNB in general has a role in multiple tumors in studying lymphatic drainage as well as having importance in staging with an advantage of being a more sensitive and less aggressive procedure than LND [
9,
14‐
17]. SLNB has proven efficacy in breast, head-neck, vulvar, cervical and penile cancer, and melanoma and is actively studied in many other tumor types [
9,
18,
19]. The first feasibility and outcome studies of SLNB in renal tumors have been published showing that the procedure is safe; however, further studies in a larger cohort for clinical or research utility in high risk cancer may be limited due to a significant rate of non-visualization. Two groups that studied feasibility of renal SN with lymphoscintigraphy and SPECT/CT reported non-visualization rates varying between 25 and 73% [
11,
12]. Although inclusion criteria, timing of imaging, imaging, and mapping techniques were comparable, distinct methods for radiotracer injection were used and also radiotracer dose, imaged area, and modalities for SN detection varied (Additional file
1: Table S1). A few factors of Sherif et al.’s group that could theoretically have lowered their visualization rate were low volume of injected radiotracer and imaging only the abdominal area, whereas our group included thoracic lymphoscintigraphy, where 20% of SNs were found. Although, Sherif et al. also used blue dye intraoperatively, they detected only 12.5% of SN with this technique, which is in line with earlier reported significantly lower sensitivity of blue dye (50–70%) compared to radiotracer [
12,
20]. Today, our study has the largest renal tumor SLNB cohort with 73 patients presenting 37% of non-visualization on scintigraphy and SPECT/CT imaging with intratumoral injection. In the study of Sherif et al. [
12] consisting of 13 patients, the radioisotope was injected peritumorally. Non-visualization on scintigraphy and SPECT/CT was reported in 8 out of 11 (73%) of cases. They detected most of the SNs using a γ-probe, whereas in our study most of the SNs were visualized by scintigraphy combined with SPECT/CT, and the minority was detected with γ-probe only (
n = 5 (18%)). Compared to SLNB in other primary tumor sites, which have scintigraphy and SPECT/CT visualization rates of 82–100%, SLN visualization in renal tumors is lower [
9,
21].
It has been suggested in other tumor types that non-visualization of lymphatic drainage can be related to patient age, BMI, size of the tumor, high nodal tumor load, neoadjuvant therapy, occlusion of lymphatics, radioisotope type, dose and volume, and lymphovascular invasion [
9,
14‐
16,
22]. Including all the possible factors (Table
2), we acknowledged some specific aspects that might be associated with non-visualization regarding kidney physiology. Due to high vascularity of the kidney and vascular tumors, inappropriate radiotracer injection into a highly vascularized area of the renal tumor may cause a wash-out phenomenon which might be related to non-visualization. However, our results were not conclusive about the association between non-visualization and wash-out phenomena in multivariable analysis.
The only aspect which demonstrated association with non-visualization in our group was age with a tendency toward having more non-visualization with older age. Age has been associated with false negative rate in melanoma and breast cancer patients and it has been speculated that it may be caused by degeneration of the lymphatic system resulting in decreased rate of lymph flow with increasing age or secondary to variable or sluggish lymphatics in older patients [
14,
23‐
25]. In fact, change in lymphatic function of SNs in older patients and lymphatic aging have been demonstrated earlier [
14,
25]. This has been shown by decline in radiotracer transit with older age in melanoma patients with intraoperative lower counts of radioactivity measured with a γ-probe. In addition, there has also been association between age and lower number of metastatic SNs which further suggests an alteration of lymphatic function with aging [
14]. However, due to lacking records of radioactivity count measurements and low number of SN metastatic cases, further analysis to substantiate an association in our cohort is unavailable.
Other accepted determinants that have shown association with non-visualization in different tumor types are tumor location in the organ and method of injection (intratumorally, peritumorally) [
26‐
28]. Comparison of intra-and peritumoral hybrid isotope injection in prostate cancer was studied recently, showing more non-visualization with intratumoral injection than peritumoral; however, more pN positive patients were detected with intratumoral method [
26]. Detection of SNs has also been better with periareolar compared to intratumoral radiotracer injection in breast cancer [
29]. However, considering the kidney’s high vascularity, these tumor-bearing organs are not comparable and in fact injecting radiotracer into both the kidney and tumor could cause substantial radiotracer wash-out. Aspects of injection, e.g., injection site, number of depots, etc., were not associated with non-visualization in our study; however, it has been described that most of the lymphatics in kidney cancers with sinus vein invasion locate peritumorally and less are seen within the tumors [
30]. This would suggest that Sherif et al. with peritumoral injection [
12] should have had higher rate of visualization; however, they detected most of the SNs by γ-probe only and non-visualization rate with SPECT/CT was even higher than in our series.
Another important consideration regarding radiotracer injection is tumor location in the kidney (anterior, posterior, upper, intermedial, lower pole). Even though our results did not demonstrate difference between the tumor location and visualization, ultrasound-guided intratumoral injection can be technically challenging especially for anterior location. In addition, dynamics of the kidney caused by breathing can challenge the precision of the injection.
Non-visualization could also be related to lymphatic drainage directly into the thoracic duct without interfering retroperitoneal lymph nodes, whereas posterior lumbar lymphatics are more inclined to this route [
5,
13,
30‐
32]. Thoracic duct might be also the path for a lympho-venous connection and subsequent hematogenic metastases mainly into lungs which is one of the most common metastatic sites in renal cancer and may occur without concurrent retroperitoneal lymph node metastatic involvement. An indication for the alternative drainage from the posterior lumbar lymphatics directly into the thoracic duct would be an association of non-visualization with posterior radiotracer depots, multiple injections, or larger tumors, because in all these occasions probability of radiotracer drainage into the posterior lumbar lymphatics are theoretically higher [
13]. However, due to methodology used, we were not able to confirm this in our study. We were also unable to confirm the route on imaging in the cases with non-visualization because scintigraphy and SPECT/CT imaging probably are not sensitive enough to detect minimal amounts of radioactivity in the thoracic duct.
In view of the experience from the studies performed by the two different groups [
11,
12], it appears that the majority of SNs are visualized 2–4 h after tracer injection and in case of non-visualization, delayed imaging can be considered. Imaging should consist of planar scintigraphy of the trunk including the thoracic cavity because 20% of sentinel lymph nodes are visualized in this area [
5], combined with SPECT/CT of the area of interest. Intraoperative SN detection rates with γ-probe and -camera were comparable to preoperative lymphoscintigraphy and SPECT/CT, indicating that timing between imaging and operation in our study was optimal.
99mTc nanocolloid was used as a radiotracer in both studies; however, whether its combination with ICG could improve intraoperative detection rate with an impact to overall detection rate and also reduce the time spent on SN harvesting remains to be studied. Concerning the dosage, according to depot activity measurements, 225 MBq
99mTc appears to be adequate. Injection site peri- or intratumorally is debatable; however, we believe that injection into periphery of the tumor avoiding necrotic areas could result in better distribution of the tracer because high vascularization of the kidneys may cause a wash-out phenomenon. Nevertheless, when high-risk tumors are studied, additional peritumoral injection can be considered due to a more peri- than intratumoral lymphatic distribution [
30].
Our study has a number of limitations. We enrolled mostly low risk patients, who have low incidence of lymph node metastatic disease. In addition, the study was not primarily designed to detect non-visualization, thus it has the limitations inherent to any retrospective data analysis. Besides, this study is not able to explain a predominant cause for non-visualization. Finally, the number of patients is low which may have an effect on the overall statistical power in our study.