Discovery of fluorescent properties in mitoxantrone hydrochloride injection for tracing: application for sentinel lymph node biopsy in breast cancer.

Introduction
Various tracers have been employed for sentinel lymph node (SLN) biopsy (SLNB), such as radionuclide, visible dye, and indocyanine green. International guidelines currently recommend the combined use of radioisotope tracing and visible dye as the standard method to reduce the false-negative rate (FNR) in SLN detection [1]. However, tracer use varies regionally due to constraints such as legislative barriers, safety concerns, and resource requirements. In China, radionuclide tracers are not routinely approved for SLNB; instead, methylene blue (MB) and indocyanine green (ICG) are widely adopted for their practicality, rapid imaging, low cost, and safety. Although MB provides good visibility and is cost-effective, its lymphatic targeting remains limited [2, 3]. Subcutaneous injection of MB has been shown to tend to diffuse on the body surface [4, 5], which can affect the aesthetic outcomes of patients undergoing breast-conserving surgery. Additionally, MB may stain secondary lymph nodes (LNs) and cause allergic reactions. ICG offers rapid imaging time (5–15 min) and detects more LNs than other methods, but its high sensitivity can lead to over-removal of nodes to reduce the FNR [6–8], and thus potentially increase complications like lymphedema. Recent advances include receptor-targeted fluorescent agents [9] and hybrid magnetic-fluorescent tracers for multimodal detection [10]. Hence, it is of great value to develop novel tracers to improve the precision and safety of SLNB.
Mitoxantrone hydrochloride injection for tracing (MHI) is a novel domestically developed tracer recently approved by the National Medical Products Administration (NMPA) for regional lymphatic drainage tracing in thyroid surgery and SLN tracing in breast cancer [11]. Preclinical pharmacology studies have revealed that MHI demonstrates a high affinity for the lymphatic system following subcutaneous or subserous administration [12]. Regarding the underlying mechanism [13], mitoxantrone forms a uniform acidic solution that, upon interstitial injection, transitions into ~ 100 nm nanocrystals under local pH conditions. These nanocrystals selectively enter lymphatic capillaries (via 120–500 nm gaps or cytophagocytosis) but not blood capillaries (with 30–50 nm endothelial spaces). Transported via lymphatic flow, they accumulate in lymph nodes, providing durable visual staining for effective tracing.
A phase I clinical trial has assessed the safety and tolerability of MHI for lymphatic tracing during SLNB in breast cancer patients [14], with no dose-limiting toxicities observed even at a dose of 2.0 mL (5 mg/mL). Preliminary studies indicate that MHI enables highly successful SLN tracing via effective visual staining, supporting its use as a safe and reliable alternative technique [15, 16]. Notably, mitoxantrone, a component of MHI, has been known to exhibit intrinsic fluorescence properties as a chemotherapeutic agent since 1988 [17]. However, whether MHI can generate detectable fluorescence signals for imaging of lymphatic drainage and be detected on the body surface remains unexplored.
This study reported for the first time that MHI enabled clear visualization of lymphatic drainage under a near-infrared fluorescence imaging system, allowing for dynamic and real-time monitoring. We also determined the optimal MHI concentration and injection timing for breast cancer SLNB, refining the procedural protocol. As a dual-functional agent offering both visible and fluorescent staining, MHI may simplify the procedure, improve surgical accuracy, and enhance clinical outcomes.

Methods

Patient enrollment
A total of 103 female patients with clinically lymph node (LN)-negative breast cancer who underwent SLNB followed by completion ALND at Qilu Hospital of Shandong University from January 2023 to February 2024 were included in this study. The inclusion criteria were as follows: (1) aged 18–70 years; (2) clinically negative axilla based on ultrasound and physical examination; (3) no definitive surgical contraindications based on routine preoperative assessments; and (4) no significant abnormalities in hematological, hepatic, or renal functions. The exclusion criteria were as follows: (1) previous breast or axillary surgery; (2) administration of neoadjuvant chemotherapy; (3) diagnosis of inflammatory breast cancer; (4) pregnancy or breastfeeding; (5) severe allergy to ICG, MB, mitoxantrone, or anthracycline antibiotics; and (6) mental or cognitive impairments.
The surgical approaches for the breast and axilla were determined through a collaborative decision-making process involving the patients, their families, and the healthcare providers, with written informed consent obtained from all patients before surgery. All the enrolled patients opted for mastectomy. For those undergoing axillary lymph node dissection (ALND), intraoperative lymphatic imaging was initially performed using various tracers. The blue-stained or fluorescent SLNs identified along the lymphatic drainage were defined as true SLNs (true-SLNs), labeled, and sent separately for routine paraffin pathological examination. ALND was then routinely performed, and the FNR was calculated by dividing the number of false negative SLNs by the number of all patients who truly have axillary metastasis (False Negative + True Positive). This study was approved by the Ethical Committee of Qilu Hospital of Shandong University.

Pathological protocol for SLN evaluation
All SLNs were evaluated according to our institutional standard protocol. Each node was serially sectioned along its long axis at 2-mm intervals. The cut surfaces were examined macroscopically for visible metastases, and the number, size, and appearance of each node were recorded. For intraoperative assessment, all the blocks were subjected to frozen sectioning, followed by hematoxylin and eosin (H&E) staining and pathological evaluation. In cases where the lymph node was too small or morphological features were inconclusive, immunohistochemistry (IHC) for cytokeratin (CK) was performed to aid diagnosis. For final postoperative confirmation, all SLNs were formalin-fixed, paraffin-embedded, and re-examined with H&E staining.

Exploration of the optimal application protocol for MHI
Using the “MB” mode (excitation: 670 nm; emission: 690 nm) of a near-infrared fluorescence imaging system (Jinan Micro Intelligent Technology Co., Ltd.), a spectrum similar to that of MB, we first demonstrated that intracutaneously injected MHI produced clearly visible fluorescent images of lymphatic vessels (Fig. S1-2).
To determine the optimal MHI concentration for fluorescence tracing, 32 breast cancer patients (Cohort I) were randomly allocated into four subgroups receiving different MHI concentrations: undiluted (5 mg/mL), 1:2 dilution, 1:5 dilution, and 1:10 dilution. After anesthesia, 0.2 mL of MHI at various concentrations was intracutaneously injected at five locations: at the 3, 6, 9, and 12 o’clock positions around the areola and directly on the tumor surface. The fluorescent lymphatic vessels were observed in real-time on the monitor screen using a near-infrared fluorescence real-time imager. Subsequently, the lymphatic vessels stained with fluorescence and/or visible dye were dissected. The LNs directly indicated by these lymphatic vessels, stained with visible dye and/or fluorescence, were designated as the true-SLNs and then sent separately for routine pathological examination (Fig. S3). The imaging rate and time, detection rate, and FNR were recorded and analyzed. Subsequently, the optimal concentration determined from this initial cohort was applied to another 31 breast cancer patients (Cohort II) to further validate its tracer effect.

Comparison of lymphatic tracing efficacy among MHI, MB, and ICG
To compare the lymphatic imaging effects of MHI with MB or ICG, 40 patients (Cohort III) were allocated into two groups. In short, in the MB group, 0.2 mL of MB was injected intracutaneously at five locations: at the 3, 6, 9, and 12 o’clock positions around the areola and on the tumor surface. The lymphatic drainage pathways were mapped using the “MB” mode of the near-infrared fluorescence real-time imager, and the imaging time was duly recorded. In the MHI and ICG group, MHI was first injected intracutaneously at five locations, followed by visualizing and mapping the lymphatic drainage pathways using the “MB” mode. Subsequently, 0.2 mL of ICG was intracutaneously injected at the same sites, and lymphatic imaging was performed using the “ICG” mode. After that, the lymphatic vessels were dissected, and the fluorescent or blue-stained LNs were identified as true-SLNs. The imaging time, detection rate, and postoperative pathology of axillary LNs were recorded. The consistency of lymphatic vessels and SLNs identified across different imaging modes was compared. The FNR was calculated only in patients who underwent ALND and had successfully detected SLNs.

Results

Exploration of the optimal MHI concentration for SLNB
The baseline characteristics of breast cancer patients included in this study are summarized in Table 1. In Cohort I, MHI injection visualized fluorescent lymphatic vessels in 28 patients (87.5% imaging rate) and identified blue-stained and fluorescent SLNs in 30 patients (93.75% detection rate), with an overall FNR of 11.11%. The median imaging time was 2 min across all groups (Table 2). Representative images for the four concentration groups are shown in Fig. 1. Both the undiluted MHI (5 mg/mL) and the 1:2 dilution groups achieved 100% SLN detection and lymphatic vessel imaging rates, with comparable imaging times, indicating stable tracing capability. However, one false-negative case occurred in the 1:2 dilution group. Surgeons also noted that blue-staining visibility was superior in the undiluted group compared to the 1:2 dilution.

In Cohort II, fluorescent lymphatic vessels were observed in all patients (100% imaging rate). SLNs were detected in 29 patients (93.5% detection rate), including 15 positive SLNs and one false-negative case, resulting in an FNR of 6.25%.

Of the 39 patients receiving undiluted MHI, all exhibited surface-visible fluorescent lymphatic vessels, with an SLN detection rate of 94.9% and an FNR of 5.26%. This indicates that undiluted MHI provides stable fluorescent and blue-stain dual-modal tracing.

Exploration of the optimal MHI injection time before SLNB
After injecting undiluted MHI, fluorescent lymphatic vessels were visualized within 5 min in 37 cases (94.9%), with specific timings as follows: 1 min in 15 cases (38.5%), 2 min in 13 cases (33.3%), 3 min in 5 cases (12.8%), 4 min in 2 cases (5.1%), and 5 min in 2 cases (5.1%). Additionally, in the remaining 2 cases (5.1%), fluorescence imaging of lymphatic vessels was achieved approximately 10 min after MHI injection (Table 3). These patients had undergone tumor biopsy before MHI injection, which may have disrupted lymphatic drainage. Next, the specimens dissected 15 min after injection were examined, and MIH was found in level I axillary LNs and even in level II or level III axillary LNs (Fig. 2).

Evaluation of lymphatic drainage patterns based on MHI fluorescence imaging
In Cohort I and Cohort II, 93 axillary-draining lymphatic vessels were mapped (Table 1). The number of fluorescent vessels varied by injection site (Fig. 3, Table S1), with a higher proportion arising from the tumor surface and from areas above or lateral to the areola. Analysis of tumor location further indicated that, irrespective of mass position, lymphatic vessels most frequently originated above and lateral to the areola (Table S2). Notably, the surfaces of masses located in the outer upper quadrant of the breast were the most prominent origin point. Consequently, based on preoperative fluorescence lymphatic mapping, the surgical incision should be placed as far as possible from these visualized vessels.

Comparison of the dyeing effect between MHI and conventional lymphatic tracers for SLNB
MHI was compared with commonly used tracers MB and ICG in Cohort III (Table S3). MB and MHI produced similar blue staining and exhibited comparable fluorescence wavelengths under real-time imaging. Due to these spectral and staining similarities, a self-comparison of lymphatic imaging between MB and MHI was not feasible. In contrast, the differing fluorescence wavelengths of ICG and MHI allowed effective evaluation using a self-controlled method.
The detection results for MB alone versus the combined use of MHI and ICG in Cohort III are summarized in Table 4. Among 19 patients with superficial lymphatic imaging under both “MB” and “ICG” modes (Fig. 4), lymphatic vessel pathways were consistent in 16 cases (84.2%). Additionally, MB injection showed significant dispersion, while MHI remained largely confined to the injection site with limited spread (Fig. 5a). Intraoperative dissection confirmed that MB diffused more widely within tissues. In contrast, MHI’s minimal dispersion maintained a clearer surgical field (Fig. 5b, Fig. S4). Additionally, MHI allows both visual and fluorescence detection, unlike ICG, which is visible only under fluorescence imaging. Therefore, MHI demonstrated SLN and lymphatic vessel imaging performance comparable to MB and ICG, along with visible blue staining, readily detectable fluorescence, and high lymphatic specificity, underscoring its potential as a suitable dual-modality tracer for SLNB.

Discussion
SLNB has been widely accepted as the standard method for axillary LN staging [18], and the combined use of visible dye and radioactive tracers is recognized as the international gold standard for SLN detection [19]. However, the accessibility of radioactive tracers remains limited, with a proportion of only 60% in developed countries and no more than 5% in other regions [20]. Although MB is the most commonly used SLN tracer in China due to its good visibility, low cost, and ease of use, it has poor lymphatic targeting and a high dispersion potential [21]. Therefore, it’s urgently needed to develop more effective and reliable alternative techniques for SLN tracing and identification in breast cancer patients.
Mitoxantrone, an intravenous antineoplastic antibiotic, exhibits high lymphatic affinity and is primarily metabolized via the biliary system [22]. A previous study has revealed the safety and effective blue-staining capability of MHI for lymphatic tracing [15]. In this study, we demonstrated MHI exhibited a low dispersion potential, generally remaining localized at the injection site and thereby providing a clearer surgical field. We confirmed that MB and MHI produce visually comparable blue staining and share similar fluorescence wavelengths under real-time imaging, supporting MHI’s feasibility as a fluorescent tracer. In determining the optimal concentration, most patients achieved fluorescent lymphatic visualization within 3 min. One case with a 1:5 dilution showed delayed imaging (11 min), potentially due to higher BMI (27.4), impairing fluorescence penetration or prior biopsy in the upper outer quadrant, disrupting lymphatic drainage. While diluted MHI (1:2 or 1:5) achieved comparable fluorescence imaging rates, SLN detection rates, and FNR to the undiluted form, intraoperative and postoperative assessment confirmed that undiluted MHI provided superior blue-staining clarity. Therefore, undiluted MHI was identified as the optimal concentration. This was validated in an additional 31 patients, where undiluted MHI achieved a 100% lymphatic vessel imaging rate and a 93.5% SLN detection rate.
Based on our analysis, a 5 min staining interval after undiluted MHI injection was sufficient for SLN identification in most cases. Extending this time led to staining of additional nodes, potentially resulting in unnecessary excision. Therefore, a 5 min interval is recommended to optimize precision and minimize over dissection, though individual anatomical and metabolic variations should be considered. In prior work, we distinguished true-SLNs (first nodes receiving direct lymphatic drainage) from post-SLNs (downstream nodes). Our findings revealed that injection timing significantly affected the number of stained LNs [23, 24], underscoring the need for precise mapping to differentiate true-SLNs. Notably, in those cohorts, when true-SLNs were negative, no metastases were found in post-SLNs, indicating that metastasis rarely skips true-SLNs. MHI’s low dispersion and dual visible fluorescent visualization enable real time, high fidelity lymphatic mapping, making it an ideal tracer for true-SLN identification. Future studies will use MHI to further validate the true-SLN concept and assess whether true-SLN targeted biopsy can safely reduce the extent of dissection without compromising oncologic outcomes.
Analysis of the two cases with no identifiable SLNs revealed distinct lymphatic patterns: one showed drainage toward the internal mammary and subclavian regions, while the other, with an upper outer quadrant tumor, likely had lymphatic continuity disrupted by a prior biopsy incision. Among the 29 detected SLNs, one false-negative case yielded an FNR of 6.25%. Postoperative pathology in that case found only a single metastatic LN in level I, possibly due to lymphatic obstruction, as suggested by observed focal skin vascular proliferation. Mapping across all 63 cases showed that lymphatic vessels most frequently originated from areas above or lateral to the areola, regardless of tumor location. The tumor surface itself was a key origin point for upper-quadrant tumors. Therefore, preoperative MHI fluorescence imaging of lymphatic drainage should guide surgical incision placement relative to the tumor site.
In summary, data from all 39 cases injected with undiluted MHI showed that fluorescent lymphatic vessels were consistently visible on the body surface across all patients, with an SLN detection rate of 94.9% and an FNR of 5.26%. These results are consistent with a prior clinical trial analyzing data from 372 patients, which reports an SLN detection rate of 97.3% with undiluted MHI [15]. Li et al. have revealed that MB alone could achieve an SLN detection rate of 91% and an FNR of 13% [25]. Additionally, Guo et al. have reported an SLN detection rate of 97% using ICG alone, which further increases to 99.5% when ICG is combined with MB [26]. Hence, the SLN detection rate and FNR achieved with MHI alone are comparable to those achieved with MB or ICG alone. Moreover, the intraoperative use of both visible dye and fluorescence imaging after MHI injection offers a dual-confirmation method that enhances the accuracy of SLN identification. This approach minimizes the unnecessary removal of non-SLNs, thereby reducing the risk of postoperative lymphedema.
While MHI’s active component, mitoxantrone, is used systemically as chemotherapy, its tracer formulation differs entirely in dose, administration, and purpose, which is specifically designed and approved (NMPA) for lymphatic mapping. The injected dose is substantially lower than therapeutic levels, and prior clinical trials have confirmed its safety for SLNB, with no dose-limiting toxicities observed [14]. In this study, following clinical protocols, no local or systemic adverse events occurred. In practice, MHI requires only standard sterile handling; surgical specimens are routinely processed, and waste is disposed as non-hazardous. Although currently priced higher than ICG, MHI provides dual visible blue and fluorescent imaging in a single agent, which may improve procedural efficiency and overall cost effectiveness. Given its established safety profile and specific regulatory approval for lymphatic tracing, MHI presents a practical and efficient clinical alternative.
Several limitations should be acknowledged. First, the relatively small sample size may result in inadequate statistical analysis and prevent a precise determination of the optimal MHI concentration and imaging duration. Hence, a larger cohort is needed to validate these parameters. Second, although all enrolled patients underwent ALND, enabling FNR assessment, this may constitute overtreatment in those eligible for SLNB alone under current guidelines. Third, quantitative imaging systems were not used to measure signal intensity, which limits the objective assessment and comparative evaluation of tracer performance. Finally, patients receiving neoadjuvant chemotherapy or breast-conserving surgery were not included in this study, necessitating further validation of MHI application in these populations.

Conclusion
In this study, we demonstrated that MHI was a fluorescence tracer, which integrates the visual clarity of MB and the lymphatic specificity of ICG, making it potentially more suitable for SLNB. Furthermore, the optimal MHI concentration as a fluorescent tracer and the appropriate injection timing for precise SLNB were investigated. Most significantly, the high specificity of MHI, coupled with its dual capabilities in visible dye and fluorescence imaging, enables easier and more accurate identification of true-SLNs. These advantages may provide a substantial foundation for guiding precise treatment decisions for breast cancer patients.

Supplementary Information