With the emergence of immune checkpoint inhibitor therapies, which are based on the engagement of the host immune system to battle against cancer, the identification of primary as well as secondary lymphoid organs via non-invasive in vivo imaging has gained increasing importance in recent years [45]. In a translational study, Schwenck et al. were able to identify responders to immune checkpoint inhibitor-based cancer immunotherapy in tumor-bearing experimental mice and metastatic melanoma patients based on responder-specific metabolic changes in glucose metabolism within the bone marrow and spleen non-invasively in vivo determined by whole-body 18F-FDG-PET/CT imaging [45]. This essential role of the secondary lymphatic organs, such as the TDLNs, is further supported by a study from Fransen et al., highlighting that immune cell homing and immune cell activation within the TDLN upon anti-PD-L1 mAb therapy is pivotal for an efficient treatment response [11]. Another recent publication by Wu et al. indicated that activated T cells, which are characterized by high glucose metabolism and lactic acid production, create acidic niches within LNs and thereby suppress T-cell effector functions associated with reduced IFN-γ or IL-2 cytokine release. T cell functioning is thus subject to immune regulation by acidification of paracortical zones within the LNs [46]. These studies accentuate the special importance of secondary lymphatic organs, such as TDLNs, and consequently the need to identify TDLNs by non-invasive in vivo imaging in a preclinical as well as clinical setting. In recent years, several preclinical and clinical studies have focused on labeling lymphatic vessels and thus TDLNs by injection of Isosulfan Blue, Patent Blue V, Evans Blue, technetium-99 m (99mTc) or 18F-FDG in close proximity to the respective tumor [47,48,49]. The respective TDLNs were subsequently identified macroscopically by the eye or by optical imaging (Isosulfan Blue, Patent Blue V, Evans Blue), SPECT (99mTc) or PET/CT as well as PET/MRI (18F-FDG) [47,48,49,50].

Thorek et al. successfully identified lymphatic vessels and LNs by 18F-FDG PET/Cerenkov imaging and demonstrated the potential of PET/Cerenkov-guided LN resection [28]. Cerenkov-guided TDLN identification would be a useful, promising, and innovative approach. Later, Lockau et al. successfully identified even LN metastases using in vivo 18F-FDG PET lymphography in B16-F10 melanoma-bearing mice [49]. Our study focused on the validation of the suitability and the comparison of the two fluorescent dyes, Patent Blue V and IRDye® 800CW PEG for OI as well as of 18F-FDG for combined PET/MRI to identify the TDLNs in experimental mice with a solid subcutaneous MC38 adenocarcinoma tumor located on the right flank.

The OI dyes Patent Blue V and IRDye® 800CW PEG were not qualified for sensitive and reliable in vivo identification of TDLNs, given that we were only able to visualize TDLNs in one out of five Patent Blue V-injected and in only one out of five IRDye® 800CW PEG-injected experimental mice.

Patent Blue V was suitable for reliable and sensitive in situ and ex vivo identification of the TDLNs, namely, the accessory and axillary LNs (Fig. 2). In sharp contrast, IRDye® 800CW PEG was not qualified for reliable and sensitive in situ and ex vivo identification of the TDLNs. One reason for the lack of IRDye® 800CW PEG accumulation in TDLNs might be the higher molecular weight compared to Patent Blue V (25,000–60,000 g mol−1 vs. 582.7 g mol−1 [44, 51]) but also other molecular characteristics such as hydrophobia or molecular charge can have an impact. Gretz et al. showed that molecules with a lower molecular weight enter the LN cortex and disperse through the reticular network due to improved tissue penetration, whereas molecules with a high molecular weight are unable to access the cortex [52, 53]. Contrary to our results, data from Krishnan et al. revealed that IRDye® 800CW PEG is qualified for the in vivo identification of lymphatic vessels and draining LNs as well as blood vessels when injected intravenously into patients with oral squamous cell carcinoma [54].

In the experimental setup of our study, the accumulation in the TDLNs might be significantly slower due to the high molecular weight of IRDye® 800CW PEG in combination with the s.c. administration route. As stated above, this finding is related to the observation that molecules with a high molecular weight distribute within the draining LN less strongly than molecules with a low molecular weight [52]. The clinically approved NIR dye indocyanine green with a molecular weight similar to that of Patent Blue V (775 g mol−1) [55] has been extensively and successfully applied for the in vivo identification of TDLNs in several animal species as well as in humans [56,57,58,59]. In contrast to IRDye® 800CW, which is conjugated to PEG, indocyanine green is an unconjugated fluorophore resulting in a lower molecular weight, which might lead to a fast and, therefore, more pronounced uptake of the contrast agent in the TDLN. Similar to other NIR dyes, indocyanine green is characterized by reduced background autofluorescence and increased tissue penetration [58], representing an advantage over Patent Blue V based on the emission of light in the visible light spectrum. Nevertheless, the OI of Patent Blue V was applicable to identify the TDLNs in situ and ex vivo in a fast and cost-effective manner. Taken together, Patent Blue V, IRDye® 800CW PEG, and indocyanine green accumulate via the lymphatic vessels within the TDLN and therefore do not provide functional information on the immune cell activation state or glucose metabolism.

In addition to the two evaluated OI dyes, we evaluated whether s.c. injection of 18F-FDG near the MC38 adenocarcinoma together with PET/MRI is a qualified and sensitive tool to identify the TDLNs. In contrast to Patent Blue V and IRDye® 800CW PEG, which indirectly label the lymph stream and thus the draining LN, the radiotracer 18F-FDG is taken up by cells with enhanced glucose demand and glucose metabolism, such as tumor cells, resident cells, and immune cells. LNs such as TDLNs are mainly composed of immune cells (T cells, B cells, etc.). 18F-FDG in the TDLNs (without LN metastasis) is taken up almost exclusively by the mentioned immune cells. In this regard, it is important to consider that activated T cells within the TDLN directed against tumor-associated antigens exhibit enhanced glucose metabolism and thus take up more 18F-FDG than LNs without activated T cells [60]. In our study, non-invasive in vivo PET/MRI 30 min after s.c. injection of 18F-FDG near the tumor at the right flank exhibited enhanced 18F-FDG uptake within the right axilla of the MC38 tumor-bearing mice. In contrast to Patent Blue V, where the accessory axillary and the proper axillary LN were identified as TDLNs by ex vivo OI, ex vivo 18F-FDG biodistribution and autoradiography analysis exclusively identified the accessory axillary LN as the main TDLN (Fig. 4C–E).

The difference between in situ/ex vivo Patent Blue V OI biodistribution analysis and ex vivo 18F-FDG biodistribution and autoradiography analysis might be due to the different characteristics of the two applied imaging agents, as 18F-FDG is taken up predominantly by immune cells with an increased metabolic glucose demand, suggesting that immune cell homing and activation rather occurs in the accessory axillary LN [61, 62]. In contrast to non-invasive in vivo OI, combined non-invasive in vivo 18F-FDG-PET/MRI allows the detection of functional TDLNs with high glucose metabolism and furthermore provides anatomical information on the exact location of the LN at high spatial resolution [47]. Compared to 18F-FDG-PET/MRI, non-invasive in vivo Patent Blue V OI was not applicable to differentiate between accessory axillary and proper axillary LNs and thus required ex vivo quantification. As the most common clinically applied PET tracer, 18F-FDG has also been applied by Singh et al. to identify sentinel LNs in patients with metastatic malignant melanoma. In this context, Singh et al. reported that non-invasive i.v. (anterior cubital vein) preoperative 18F-FDG-PET/CT imaging cannot serve as a substitute for lymphoscintigraphy with 99mTc-nanocolloid (sensitivity: 100%) due to the low sensitivity of 18F-FDG-PET (sensitivity: 14.3%; 95% CI, 2.5 to 44%) [63], which might be associated with the low diameter of the metastatic nodules, as indicated by Crippa et al. In this patient cohort, only 23% of LN metastases with a diameter less than 5 mm were identified by 18F-FDG-PET. In contrast, 83% of LN metastases with a diameter of 6–10 mm and 100% of LN metastases with a diameter greater than 10 mm were identified by 18F-FDG-PET [64].

The German procedural instructions for the identification of TDLNs recommend the use of 99mTc-nanocolloid, which should be injected i.c. in close proximity to the melanoma, followed by non-invasive SPECT/CT and in situ lymphoscintigraphy during surgery. A meta-analysis of seventeen comparative 99mTc-nanocolloid-SPECT/CT and lymphoscintigraphy studies, including 1438 patients revealed a slightly higher sensitivity of SPECT/CT (98.28%) in comparison to lymphoscintigraphy (95.53%) [48]. Additionally, in most clinics melanoma patients will be i.c. injected near the tumor with Patent Blue V, as Patent Blue V is a highly recommended method for the identification of TDLNs with a high sensitivity (< 95%) and a low false-negative rate (5–10%) [65].

I.c. injection of 99mTc-nanocolloid and Patent Blue V within one patient with malignant melanoma can reveal contradictory results. It is possible that Patent Blue V strongly accumulates and macroscopically dyes the TDLN, whereas no or only a low 99mTc-nanocolloid concentration can be detected within the identical TDLN or vice versa. This observation is probably the consequence of the different molecular weights of both agents, as the molecular weight highly determines the grade of tissue distribution [53].

Based on German procedural instructions for nuclear medical sentinel LN diagnosis in patients suffering from melanoma, breast carcinoma, head, and neck cancer, or prostate carcinoma, the usage of 99mTc-nanocolloid represents the primary indication for TDLN identification [1,2,3,4].

In addition to immunotherapies, there is great interest in identifying TDLNs for sentinel LN resection to anticipate metastatic spread or detect LN metastases. In this regard, Bae et al. claim that 18F-FDG PET/CT is superior to CT/MRI or CT or MRI alone in terms of LN metastasis identification in patients with oral cavity squamous cell carcinoma [66]. In addition to these clinical studies, 18F-FDG is nonspecifically taken up by cells with elevated metabolic demand, including T cells and tumor metastases in the LN.

Identification of the TDLNs might be of special importance for upcoming novel therapeutic approaches with a focus on intranodal administration of immune checkpoint inhibitors or oncolytic viruses to improve the therapy outcome of cancer patients [31, 67]. Furthermore, immunotherapy would foster immune cell activation and, therefore, enlarge the TDLN and thus enhance the 18F-FDG uptake.

In summary, Patent Blue V-OI and 18F-FDG-PET/MRI identified the accessory axillary LN on the ipsilateral site of the tumor-bearing mice as the main TDLN.

For extensive continuative preclinical studies with a focus on the role of the TDLN during immune checkpoint inhibitor therapy, we recommend an initial cheap and easy investigation by injecting Patent Blue V s.c. at the respective areas near the tumor to enable in situ and ex vivo identification of the lymphatic drainage and, consequently, the TDLN. Secondly, we recommend functional three-dimensional quantitative high-resolution 18F-FDG PET/MRI investigations to reveal the drainage of 18F-FDG into the TDLN along with the associated glucose metabolism.

Combining Patent Blue V OI and 18F-FDG-PET/MRI would be very helpful in the preclinical setting. Thus, we recommend the first s.c. 18F-FDG-injections in proximity to the tumor followed by immediate (2 min.) PET/MRI over 30–60 min. Afterward, Patent Blue V should be s.c. injected into the experimental mice at the identical site of the 18F-FDG-injection. 5 min later, mice should be sacrificed and dissected and LNs of interest semi-quantitatively analyzed by OI and by γ-counting. These investigations could be linked to continuative flow cytometry analysis of the immune cell population of the TDLNs and NTDLNs.

Patent Blue V OI and visual identification of the TDLN and metabolic identification of the TDLN by 18F-FDG might provide two different yet essential pieces of information. Therefore, continuative studies are required to differentiate the role of the Patent Blue V accumulating TDLNs and the exclusively 18F-FDG accumulating TDLN (accessory axillary LN) from the role of the 18F-FDG non-accumulating but Patent Blue V accumulating TDLN (proper axillary LN) in regard of the efficacy of an immune checkpoint inhibitor-based cancer.

Our results demonstrate the feasibility of 18F-FDG-PET/MRI as a valid method for non-invasive in vivo and ex vivo identification of TDLNs. However, the additional use of Patent Blue V provides an additive value for macroscopic identification of the TDLN by the eye or by ex vivo optical imaging analysis. Both methods are valuable, easy to implement, and cost-effective. These findings might have broad preclinical and clinical implications for LN resections and the consequences for immune checkpoint inhibitor therapies, as reported by Fransen et al. [11].