This study assessed the performance and safety of DSNB in patients with penile cancer and cN0 inguinal stage per groin. Primary endpoints were sensitivity of the procedure, disease-free survival (NPV) and DSNB-related adverse events. Secondary scan-related endpoints were visualization of a SN (detection rate), number of SNs and the number of counts at the most active SN.

The detection rate in this study was 91% and 96% per procedure and groin, respectively, and is therefore generally in line with available literature. Wever et al. reported a detection rate of 92% in 644 patients and Sedigh et al. reported a detection rate of 80% in 35 patients [9, 12]. Valdés Olmos et al. reported a detection rate of 97% in their patient cohort; however, unilateral visualization of the SN was considered a successful procedure, which was not the case in this study [13]. In patients with non-visualization, repetition of the DSNB procedure could be an option, as was done in two patients in this study. In these patients, a SN could be identified on the repeat DSNB. Likewise, Sahdev and colleagues identified 20 out of 166 patients with unilateral non-visualization of the SN. Of these 20 patients, 7 underwent repeat DSNB at a later date, with 6 having successful visualization of the SN [14].

In this study, the number of SNs on SPECT/CT was significantly higher than on static planar imaging. SPECT/CT provides 3-dimensional data facilitating the evaluation of the groin and the underlying SNs.

Moreover, due to the lower image contrast, SNs with limited tracer uptake above the superimposed background cannot be visualized by static planar imaging and only by SPECT/CT. This study in one of the largest patient cohorts with penile cancer showed that SPECT/CT outperforms static planar imaging in detecting SNs per groin. This is in analogy to the study of Saad et al., in which the addition of SPECT/CT proved to be significant in 76% of their patient cohort. SPECT/CT provided superior images with increased nodal yield, more precise localization, a clearer distribution and drainage pattern, and a significant reduction in false extranodal hot spots observed on conventional static planar imaging [15]. A previous study by Jimenez-Hefferman et al. also evaluated this comparison in a large group of 1508 patients, yet only seven patients of this cohort had penile carcinoma. In their study, SPECT/CT also had an influence on the drainage territory of the malignancy; however, this was not the case in this study, since the penis always drains to the inguinal nodes [16]. In 1 groin, a SN was suspected on the static planar image without confirmation on the SPECT/CT (non-visualization). In this patient, a modified ILND was performed which showed 12 negative LNs and the patient remained disease-free during follow-up. Possibly, the focus of increased tracer uptake on the planar static image could represent stasis of tracer in a lymphatic vessel.

In turn, the number of resected lymph nodes was significantly higher than the number of SNs on SPECT/CT. This discrepancy is likely a result of the surgical procedure, which uses besides a gamma probe, also blue dye and palpation of the groin to find SNs.

When analyzing the counts of the SN with the highest uptake, a significantly higher value was obtained for SPECT/CT compared to static planar imaging. Since both imaging modalities are acquired immediately one after the other, it seems unlikely that more tracer would have entered the SNs by the time of the SPECT/CT as explanation of this phenomenon. However, static planar images are not corrected for body attenuation and SPECT/CT has a higher signal-to-noise ratio. Furthermore, no geometric mean could be calculated to correct for this phenomenon on the static planar images given that only anterior acquisitions were available. More data are needed regarding the correlation of SN characteristics on SPECT/CT and their associated pathologically proven positivity. Currently, no data are available about the counts of SNs on SPECT/CT and their associated tumor load.

In this manuscript, the dynamic images were not included in the semi-quantitative analysis of the SN. These dynamic images are mainly important to objectify the visualization of the SN. Additional static planar and SPECT/CT imaging are subsequently performed to localize the SNs. The duration of the dynamic acquisition represents the time to visualization after injection of the radioactive tracer. In this study, a SN could be seen after an average of 22 min. An alternative procedure workflow could be to inject the patient at the nuclear medicine department and acquire a static planar image 60 min after injection. If SNs are visualized, additional SPECT/CT can be performed, if not, additional injection and late imaging is required.

In our study, DSNB had a sensitivity of 79% and three false negative groins were reported. An explanation for the false negative results could be that the SN was not detected due to malignant invasion, possibly due to tumor blockage limiting tracer uptake [17]. Besides hematoxylin and eosin staining, immunohistochemistry should be performed in the evaluation of the SN, since it can detect micro-metastatic disease smaller than 2 mm [11]. Using DSNB, Valdés Olmos et al. reported a sensitivity of 89% for detection of inguinal lymph node metastases in 74 patients with penile cancer [13]. Hadway et al. showed a sensitivity of 95% and FNR of 5%; however, this was in a follow-up period of 11 months [18]. A recent systematic review and meta-analysis published in 2022 showed a pooled sensitivity of 87% (82–91) in 2893 patients (28 studies) [11]. In another systematic review and meta-analysis by Zou et al., a pooled sensitivity and NPV of 88% (95% CI 84–90%) and 99% (98–99%), respectively, were found [19]. In a large prospective study in 264 patients (500 groins), a sensitivity of 92% and 95% was found for the DSNB alone and with ultrasound, respectively [20]. The sensitivity in our study is lower in comparison with these data, which may be attributed to the fact that 12 groins were excluded as a result of positive US and pre-DSNB (core needle) biopsy, leading to a low number of true positives compared with earlier studies. Besides, two true positives of the DSNB were excluded from the analysis as a result of local recurrence. Last, this series includes data from the initiation of the use of DSNB in penile cancer and a learning curve effect may contribute to low sensitivity compared to high volume centers with established extensive experience.

In our study, the FNR was 21%, which is higher than reported in a previously published study by Dimopoulos et al. (5.1%) in 151 patients, a study in 264 patients by Lam et al. (5%) and a study by Leijte et al. (7%) in 323 patients [20,21,22]. One of the landmark studies on this topic published in 2005 by Kroon et al. showed a FNR of 16% [23].

The complication rate of surgery in our study was low (1% of patients), which is much lower than the complication rate in the study by Dimopoulos et al. (complication rate 21%) and more in line with the results of Leitje et al. (complication rate 4.7%) [21, 22]. However, the study by Dimopoulos recorded their complications prospectively and in their study, the majority (10.7%) of complications were Grade I, due to groin swelling requiring no intervention [21]. Our retrospective study is conducted within a concentrated care model and likely underestimates the total number and grade of complications as patients may have presented elsewhere with complications not captured in our records.

There are some limitations to this study. This study was retrospective and future prospective studies could provide more insights on the performance of DSNB in penile cancer. In this study, a period of 18 months was used as a reliable interval for a SN. However, a recent systematic review showed that a 12-month period also shows reliable ground truth [11]. If this is considered for this study, it would not affect the false negative results, but it would affect the true negatives given that four local recurrences could then be excluded, achieving eight supplementary true negative groins and also a higher NPV. Moreover, the sensitivity and FNR were calculated taking into account the information of only the DSNB procedure and not of the entire diagnostic work up (clinical examination + US + core needle biopsy + DNSB). As a result, there is an underestimation of the sensitivity and overestimation of the FNR since 12 groins were already initially scheduled for an ILND and thus 12 additional true positives were not taken into account. Secondly, a very widely geographical patient group were referred to our center due to its expertise in the topic and some interesting information was therefore not available in the medical report: (1) the type of surgery on the primary tumor, possibly explaining the number of local recurrences; (2) follow-up of complications, possibly resulting in an underestimation of complication rate. Counts of the SN reported by the gamma probe per-operatively were not included in this study, since a longer period of time elapsed between SPECT/CT acquisition and resection, so comparison of the amount of counts would be irrelevant since more tracer could have entered the SNs. Also, exact comparison of individual SNs between SPECT/CT and resection is not feasible.