HER2-targeted dual radiotracer approach with clinical potential for noninvasive imaging of trastuzumab-resistance caused by epitope masking

Rationale: The decreased HER2-accessibility by epitope masking is a primary trastuzumab-resistance mechanism. In this study, we developed a HER2-targeted dual radiotracer approach to predict the HER2-trastuzumab engagement noninvasively. Methods: Two novel HER2-specific VHHs, MIRC208 and MIRC213, were acquired by immunizing alpaca with human HER2 protein, and were site-specifically labeled with 99mTc. Biodistribution and SPECT/CT imaging studies were performed in mice bearing HER2-positive and HER2-negative tumors. The HER2 binding sites of 99mTc-MIRC208 and 99mTc-MIRC213 were investigated by cell binding and SPECT/CT imaging studies. We evaluated the therapeutic predictive ability of our dual-radiotracer imaging approach for trastuzumab treatment in mice bearing MUC4-positive tumors (trastuzumab-resistant JIMT-1 and 87MUC4) and MUC4-negative tumors (trastuzumab-sensitive 7HER2 and NCI-N87). The preliminary clinical studies of 99mTc-MIRC208 were performed in two patients with HER2-positive breast tumors. Results: 99mTc-MIRC208 and 99mTc-MIRC213 clearly visualized HER2-positive tumors, but not HER2-negative tumors. 99mTc-MIRC208 competes with trastuzumab for HER2-binding while 99mTc-MIRC213 recognizes HER2 on an epitope that is not masked by MUC4. The SPECT/CT studies with 99mTc-MIRC208 and 99mTc-MIRC213 clearly showed that the MUC4-negative and trastuzumab-sensitive 7HER2 and NCI-N87 tumors had very similar tumor uptake with the SUV208/SUV213 (2 h) ratios of 1.11 ± 0.17 in 7HER2 and 1.25 ± 0.22 in NCI-N87. However, the MUC4-positive JIMT-1 tumors showed the decreased SUV208/SUV213 (2 h) ratio (0.63 ± 0.07), which correlated well with the low response rate to trastuzumab therapy. The SUV208/SUV213 (2 h) ratio was reduced to 0.72 ± 0.02 in MUC4-expressing NCI-N87 cells, and resulting in the decreased trastuzumab sensitivity, further supporting the correlation between the SUV208/SUV213 (2 h) ratio and trastuzumab-sensitivity. The primary and metastatic HER2-positive lesions of patients were clearly visualized by 99mTc-MIRC208 SPECT at 2 h post injection. Conclusion: Overall, we demonstrated that the dual radiotracer imaging strategy is a valid noninvasive approach for the cancer patient selection before trastuzumab therapy. 99mTc-MIRC213 SPECT is utilized to quantify the tumor HER2 expression and screen HER2-positive cancer patients, while 99mTc-MIRC208 SPECT is used to determine the HER2-accessibility of trastuzumab. The SUV208/SUV213 (2 h) ratio is an important biomarker to determine the responsiveness of trastuzumab therapy.


Introduction
Human epidermal growth factor receptor 2 (HER2) is an important therapeutic target for many types of cancers [1,2]. Targeted therapy with trastuzumab (Traz), a humanized IgG monoclonal antibody against the HER2 extracellular domain (ECD), has become a mainstay for patients with HER2-positive breast cancer (BC) and gastric cancer (GC) [3,4]. However, only a small fraction of cancer patients is responsive to the Traz therapy due to the inherent and/or acquired resistance mechanisms [1,2,[4][5][6]. One of the resistance mechanisms is the decreased binding of Traz to HER2 caused by epitope masking [1,[7][8][9][10][11]. It was reported that the membrane-associated mucin 4 (MUC4) can dimerize with HER2 and sterically impede the Traz-binding of its epitope on HER2 receptors, leading to significant resistance to Traz therapy [8,9,12]. In a recent study, it was found that MUC4 overexpression in tumors is strongly correlated with shorter disease-free survival (DFS) in the HER2-positive BC patients receiving adjuvant Traz treatment [7]. Therefore, determination of the HER2-accessibility is essential for improvement of the response rate prior to Traz treatment [2].
Single-domain variable region (VHH) is the smallest antibody fragment derived from camelid heavy chain-only antibodies. The VHH possesses many unique features (e.g. good antigen affinity, fast renal clearance and high tissue penetration), which make it better suited than the full-length antibody as the targeting moiety for radiotracer development [28,29]. In this study, we generated two HER2-specific radiotracers 99m Tc-MIRC213 and 99m Tc-MIRC208. 99m Tc-MIRC213 was a Traz-noncompetitor useful for selection of the HER2-positive patients and quantify the tumor HER2 expression. 99m Tc-MIRC208 is a Traz-competitor useful for the determination of HER2-accessibility of Traz because its epitope on HER2 is masked by MUC4. The combination of 99m Tc-MIRC213 and 99m Tc-MIRC208 could be a powerful noninvasive tool to select responders to Traz therapy before treatment initiation.

Generation of novel HER2-specific 99m Tc radiotracers
A library containing 42 individual VHHs were generated from an alpaca immunized with human HER2 ECD protein ( Figure 1A). These VHHs could be divided into 10 groups based on the full-length sequence homology. We randomly selected one VHH from each group for production and purification in an E. coli expression system ( Figure 1B). The purity of 10 selected VHHs were verified by SDS-PAGE ( Figure  1C). The surface plasmon resonance (SPR) experiment was performed to evaluate the binding affinity of 10 selected VHHs against the recombined human HER2 ECD protein. Traz and a VHH against other receptor were chosen as the positive and negative controls, respectively, to validate the method ( Figure 1D and Figure S1). We found that all 10 VHHs showed excellent binding affinity for HER2. MIRC244 had a relatively low affinity for HER2. MIRC213 has the best HER2-binding affinity with Kd value at the 10 -9 M level. Other VHHs bind to HER2 protein with the HER2-affinity in 10 -8 M range. Since VHHs with different sequences of complementarity determining region 3 (CDR3) are likely to recognize different target epitopes, these VHHs were further divided into 3 classes on the basis of the alignment of the CDR3 sequence. MIRC208, MIRC213 and MIRC220 were selected as the lead compounds for subsequent studies ( Figure 1C-D).
MIRC208, MIRC213 and MIRC220 were genetically engineered to contain a C-terminus Sortase A recognition tag, LPETG. The G4EC oligopeptide was conjugated to the C-terminus of the modified VHHs using Sortase A-mediated transacylation to generate the expected VHH-ECs bioconjugates for 99m Tc-radiolabeling ( Figure 2A). Successful bioconjugation was verified by LC-MS analysis ( Figure S2). The purity of VHH-ECs was confirmed by SDS-PAGE ( Figure 2B). The 99m Tc-labeling was achieved by reacting VHH-EC bioconjugate with the 99m Tc-GH (GH = glucoheptate) intermediate at 37 °C for 0.5 h to yield the corresponding 99m Tc radiotracers 99m Tc-MIRC208, 99m Tc-MIRC213 and 99m Tc-MIRC220 (Figure 2A). Their radiochemical purity prior to purification was 89.23 ± 2.58%, 90.12 ± 1.22% and 89.92 ± 2.91%, respectively ( Figure S3A). All three radiotracers were stable for >6 h in saline at 25 °C and retained their integrity in mice blood at 6 h post injection ( Figure S3B). The in vitro binding assays were performed to test their HER2 affinity ( Figure 2C). The results clearly showed that 99m Tc-MIRC208, 99m Tc-MIRC213 and 99m Tc-MIRC220 are all able to bind specifically to HER2 ECD and the HER2-expressing 7HER2 BC cells rather than the other receptors in the HER family, murine HER2 ECD and HER2-low MCF-7 cells ( Figure 2C-D and Figure  S4).

Evaluation of 99m Tc radiotracers in tumor xenografts
99m Tc-MIRC208, 99m Tc-MIRC213 and 99m Tc-MIRC220 were evaluated for their capability to image HER2-positive tumors. Biodistribution studies were performed in 7HER2 tumor-bearing mice at multiple time points. It was found that all three 99m Tc radiotracers displayed a high tumor uptake ( Figure  3A). The maximum uptake of 99m Tc-MIRC208 and 99m Tc-MIRC220 in the 7HER2 tumors was observed at 1 h p.i. ( 99m Tc-MIRC208: 13.58 ± 1.73 %ID/g, 99m Tc-MIRC220 8.03 ± 0.95 %ID/g). 99m Tc-MIRC213 showed a slow binding rate for 7HER2 tumors and reached maximum tumor accumulation at 4 h p.i. (14.69 ± 1.31 %ID/g). The difference in tumor binding rates between three 99m Tc-VHHs may be caused by their different KD values. 99m Tc-MIRC208 and 99m Tc-MIRC220 bound to 7HER2 tumors faster because they showed greater KD values than 99m Tc-MIRC213. There was a low uptake in normal organs (except liver and kidneys) with rapid clearance for all three 99m Tc radiotracers. At 1 h and 2 h p.i., 99m Tc-MIRC208 had the kidney uptake of 352.68 ± 30.31 and 281.47 ± 31.80 %ID/g, respectively, which were significantly higher than that of 99m Tc-MIRC213 and 99m Tc-MIRC220 ( Figure 3B). 99m Tc-MIRC220 also showed lower tumor uptake and higher liver uptake at all the time points.  SPECT/CT studies were performed in the 7HER2 tumor-bearing model to illustrate the capability of 99m Tc-MIRC208, 99m Tc-MIRC213 and 99m Tc-MIRC220 as radiotracers to visualize HER2-positive tumors. It was found that all three 99m Tc radiotracers have significant uptake in the HER2-positive 7HER2 tumors ( Figure 3C) as early as 0.5 h after injection, which is in complete agreement with the results from biodistribution. 99m Tc-MIRC208 has more kidney uptake than other two 99m Tc radiotracers. High liver and low tumor uptake of 99m Tc-MIRC220 was also observed at all the time points. Therefore, 99m Tc-MIRC220 was eliminated from subsequent studies due to its unfavorable biodistribution.
SPECT/CT and biodistribution studies were also carried out in the HER2-negative MCF-7 tumor model to demonstrate the tumor specificity of 99m Tc-MIRC208 and 99m Tc-MIRC213 ( Figure 3D-E). Blocking studies were performed using excess VHH in 7HER2 tumor model. At 2 h after injection, there was no radioactivity could be detected in the HER2-negative MCF-7 tumors and a close to 90% reduction in tumor uptakes was observed upon the homologous VHH blockade ( Figure 3D-E). 99m Tc-MIRC208 and 99m Tc-MIRC213 bind to HER2 with different epitopes Human HER2 ECD consists of four subdomains I -IV. Traz recognizes the subdomain IV [3]. Traz blocking studies were performed to explore the epitopes recognized by 99m Tc-MIRC208 and 99m Tc-MIRC213. It was found that Traz significantly inhibited the binding of 99m Tc-MIRC208 to the HER2 ECD protein ( Figure 4A). In contrast, excess Traz had little effect on the uptake of 99m Tc-MIRC213. The VHH blocking assays were used as a positive control and revealed the HER2 specificity of 99m Tc radiotracers. Next, we conducted the same binding assays in 7HER2 cells ( Figure 4B). In agreement with the protein binding results, Traz significantly reduced the binding of 99m Tc-MIRC208 to 7HER2 cells, but not of 99m Tc-MIRC213. These results revealed the difference between 99m Tc-MIRC208 and 99m Tc-MIRC213 with respect to their HER2-binding epitopes. 99m Tc-MIRC208 may partially or fully bind to the subdomain IV of HER2 and compete with Traz while 99m Tc-MIRC213 binds to other subdomains ( Figure  4E).
SPECT/CT studies were performed in the 7HER2 xenografts model to further confirm the results from in vitro studies. Traz was injected 48 h prior to co-injection of 99m Tc radiotracer and a homologous VHH blocker ( Figure 4C-D). The SPECT/CT imaging studies were performed at 2 h p.i., because 99m Tc-MIRC208 and 99m Tc-MIRC213 showed similar uptakes in 7HER2 tumors at this time point ( Figure 3A-B, 99m Tc-MIRC208: 11.58 ± 1.39 v.s. 99m Tc-MIRC213: 10.56 ± 1.72, p = 0.73). The HER2 binding of 99m Tc-MIRC208 was significantly inhibited by excess Traz and there was ~90% reduction in the 7HER2 tumor uptake. In contrast, the 7HER2 tumor uptake of 99m Tc-MIRC213 was not affected by excess Traz. 99m Tc-MIRC208 and 99m Tc-MIRC213 did not compete with each other in HER2 binding.

Quantifying the tumor HER2 status and predicting resistance to Traz caused by MUC4 via a dual radiotracer approach
JIMT-1 is a well-known HER2-positive and MUC4-expressing BC cell line. The MUC4 receptor masks the epitope recognized by Traz [7-9, 30, 31]. Therefore, JIMT-1-bearing animal models were utilized to explore the impact of MUC4 on the HER2-binding of 99m Tc-MIRC208 and 99m Tc-MIRC213. The results revealed that MUC4 could not prevent 99m Tc-MIRC213 from binding to HER2, and 99m Tc-MIRC213 could detect the HER2 expression in multiple cancer cell lines, which was completely consistent with the HER2 protein analysis ( Figure  5A-B). The binding of 99m Tc-MIRC208 to JIMT-1 is much lower than that of 99m Tc-MIRC213, suggesting that MUC4 could mask the HER2-binding epitope of 99m Tc-MIRC208. Moreover, 99m Tc-MIRC208 and 99m Tc-MIRC213 accumulated similarly in the MUC4-negative cancer cells ( Figure 5A-C). The results from the cytotoxicity assay revealed that MUC4-negative 7HER2 and NCI-N87 cells were sensitive to Traz treatment, and JIMT-1 cells showed resistance to Traz, which is similar to HER2-negative MCF-7 did ( Figure 5D).
It has been reported that the increased tumor MUC4 expression could inhibit the HER2-binding of Traz and reduce the Traz-sensitivity [7,31,32]. Therefore, we transfected NCI-N87 cells to generate the MUC4-expressing GC cell line 87MUC4, and used it to establish animal models. Western blot studies demonstrated the increased MUC4 expression in 87MUC4 cells and tumor tissues ( Figure 6A and Figure S5). The overexpression of MUC4 had little impact on the HER2 expression. The elevated MUC4 significantly inhibited the binding of 99m Tc-MIRC208 to 87MUC4 cells, but not 99m Tc-MIRC213. The binding of 99m Tc-MIRC208 in 87MUC4 cells was less than that in NCI-N87 cells ( Figure 6B). The 99m Tc-MIRC208: 99m Tc-MIRC213 ratio in 87MUC4 was 0.66 ± 0.01, and was significantly less than that in NCI-N87 ( Figure  6C).
SPECT studies were performed to calculate the SUV 208 /SUV 213 (2 h) ratio in the 87MUC4 tumor model. We found that the uptake of 99m Tc-MIRC208 in 87MUC4 tumors was significantly diminished, while there was little change in the uptake of 99m Tc-MIRC213 ( Figure 6D-E). The SUV 208 /SUV 213 (2 h) ratio (0.72 ± 0.02) in the 87MUC4 tumors was significantly lower than that in the NCI-N87 tumors ( Figure 6F). A Traz treatment study was performed to compare the therapeutic effect of Traz on tumor growth rates of 87MUC4 and NCI-N87 ( Figure 6G). The Traz therapy showed less inhibitory effect on the growth of 87MUC4 tumors than wild-type NCI-N87, indicating that increasing MUC4 expression could significantly decrease the Traz-sensitivity of NCI-N87. However, we found that Traz slightly inhibited the 87MUC4 tumor growth compared with saline group. This is maybe because the efficiency of overexpressing MUC4 receptor by lentivirus transfection is not 100%, resulting in a small number of wild and Traz-sensitive NCI-N87 cells in 87MUC4 tumors.   . (B) and (C) Both 99m Tc-VHHs showed similar accumulation in wild-type NCI-N87 cells, but 99m Tc-MIRC213 bound to 87MUC4 at significantly higher levels than 99m Tc-MIRC208 (n = 4). The error bars represent standard deviation. ***p < 0.001, student's paired t test. (D) SPECT/CT imaging of 99m Tc-MIRC213 and 99m Tc-MIRC208 in the same 87MUC4 tumor bearing mice. (E) Quantification of SPECT signals of 87MUC4 tumors from C (n = 3). The error bars represent standard deviation. ***p < 0.001, student's paired t test. (F) SUV208/SUV213 (2 h) in wild-type NCI-N87 and 87MUC4 tumors (n = 3). error bars represent SD. **p < 0.01, student's paired t test. (G) Tumor growth curves of 87MUC4 and wild-type NCI-N87 xenografts treated with saline or Traz. Tumor volumes are expressed as the mean ± SD (n = 7). **p < 0.01, ***p < 0.001, two-way analysis ANOVA followed by a Bonferroni post hoc test.

The preliminary clinical study on 99m Tc-MIRC208
Prior to the clinical study, we performed safety assessment studies in mice. Blood sample testing and H&E staining of major tissues from the mice receiving a single high dose of radioactivity indicated that both radiotracers had favorable toxicity profiles ( Figure  S9-S10). Therefore, we carried out a preliminary clinical study of 99m Tc-MIRC208 in the HER2-positive BC patients (ClinicalTrials.gov identifier, NCT045 91652). Two BC patients with HER2-positive tumors (HER2 IHC 3+) in the left breast have been enrolled up to date. In patient 01, 99m Tc-MIRC208 SPECT/CT clearly showed the HER2-positive primary tumor lesion and lymph node metastasis with a low background at 2 h p.i. (Figure 7A-B). The results of IHC staining showed HER2 is overexpressed in the primary tumor. In patient 02, the primary and large axillary lymph node metastatic lesions were clearly visualized by 99m Tc-MIRC208 SPECT/CT at 2 h p.i., but the pulmonary micro-metastatic lesions indicated by high 18 F-FDG uptake didn't show significant 99m Tc-MIRC208 accumulation. Both patients had no serious adverse events or obvious vital sign changes within a week after 99m Tc-MIRC208 administration.

Discussion
Considering the crucial role that epitope masking plays in Traz-resistance, it is imperative to identify potential Traz-responders [7,8,30,33]. However, standard clinical procedures for HER2 examination fail to satisfy this clinical need [13][14][15][16]34]. In this study, we proposed a dual radiotracer approach using two HER2-targeted 99m Tc radiotracers against different epitopes. We demonstrated that this approach is a powerful tool for noninvasive imaging of the HER2-positive tumors, selection of appropriate patients for Traz treatment, and prediction of Traz-resistance caused by epitope masking. The 99m Tc-labeling method used in this study is site-specific, and has many advantages, such as the single product and minimal impact on target affinity, which are perfect for the design of radiotracers for clinical applications [35][36][37]. The final products could be established within only an hour with high radiopurity, HER2-binding affinity and solution stability (Figure 2 and Figure S3). 99m Tc-MIRC208 and 99m Tc-MIRC213 ( Figure 3) recognize two different epitopes of HER2 receptor, and provide different diagnostic information concerning the molecular characteristics of tumors ( Figure 4). 99m Tc-MIRC213 could be used to select HER2-positive patients and noninvasively quantify the tumor HER2 expression levels without the interference from MUC4. 99m Tc-MIRC208 was able to indirectly evaluate the HER2-accessibility of Traz. The SUV 208 /SUV 213 (2 h) ratio can be a biomarker for selection of potential responders before therapy initiation. The patients with " 99m Tc-MIRC213positive" tumors tend to show a higher SUV 208 /SUV 213 (2 h) ratio and would likely to benefit from Traz therapy. The patients with low SUV 208 /SUV 213 (2 h) might be non-responders, and should consider other HER2-targeted therapies without Traz. The radiotracers based on intact Traz, such as 64 Cu-Traz or 89 Zr-Traz, are the best molecular imaging tools for predicting HER2-accessibility of Traz [19,38,39]. However, these antibody-based radiotracers are not suitable for widespread clinical applications due to their poor pharmacokinetic characteristics and long-term lag between administration and after examination [24,25]. In this study, 99m Tc-MIRC208, a Traz-competitor, was used as a substitute for the radiolabeled Traz. High quality SPECT images could be acquired with 99m Tc-MIRC208 as early as 0.5 h post injection, which helps the rapid design of an optimal therapeutic strategy and increases patient compliance with the examination. However, 99m Tc-MIRC208 alone cannot predict the HER2-accessibility of Traz, as it is less hindered by MUC4 than trastuzumab. It has been reported that MUC4 can sterically prevent intact trastuzumab from binding to nearly 80% of HER2 receptors on the JIMT-1 cell membrane [9]. Because of the decreased molecular weight of VHH, only approximately 50% of HER2 is masked by MUC4 to 99m Tc-MIRC208 ( Figure 5C), which may lead to confusing results when the tumor expresses an extremely high HER2 level. Thus, if we want to use 99m Tc-MIRC208 to determine whether MUC4 masks the Traz epitope, we must know the HER2 expression level in tumor in advance. That is the exact reason why we utilize the SUV 208 /SUV 213 (2 h) ratios rather than the tumor uptake of 99m Tc-MIRC208 as the biomarker to predict the Traz-resistance before initiation of the Traz therapy.
Several other mechanisms also contribute to the Traz-resistance [2,40]. The Traz-mediated HER2 downregulation during treatment is an early prognostic indicator for therapeutic efficiency [41][42][43]. Previously, we found that 99m Tc-HYNIC-H10F, a peptide-based radiotracer against subdomain II of HER2, was useful to predict the therapeutic response by noninvasively monitoring the decline in tumor HER2 expression after Traz therapy [44]. 99m Tc-MIRC213 is a Traz non-competitor useful for noninvasive monitoring of tumor HER2 expression. Longitudinal monitoring of 7HER2-bearing mice undergoing Traz treatment revealed that the HER2 downregulation could also be detected noninvasively by 99m Tc-MIRC213 SPECT before any significant changes in the tumor size ( Figure S8). 99m Tc-MIRC213 has higher HER2 binding affinity and better tumor uptake than 99m Tc-HYNIC-H10F. Therefore, 99m Tc-MIRC213 has better sensitivity and accuracy in evaluation of the changes in HER2 expression. 99m Tc-MIRC213 could also be utilized to predict the treatment response of patients who have been defined as responders by our imaging approach and are being treated by Traz, which can further promote the design of personalized therapeutic approaches and improve patient management. 99m Tc-MIRC208 SPECT/CT studies were carried out in two HER2-positive breast cancer patients. At 2 h p.i., the HER2-positive primary tumor lesion and lymph node metastasis were clearly visualized with very low background. However, several micrometastatic lesions indicated by 18 F-FDG PET were not visualized by 99m Tc-MIRC208 SPECT, probably due to low HER2 expression in these lesions or poor sensitivity of SPECT as compared with PET. Since the pathological data of metastatic lesions are not available at the moment, this explanation remains speculative. It should be investigated in the future. High radioactivity accumulation was also observed in kidneys. No adverse events or abnormal vital signs were observed after injection, confirming that 99m Tc-MIRC208 is safe and well-tolerated in patients.
Despite the outstanding results, there are several important limitations in this study. JIMT-1 is the only natural biological model we used to test the predictive validity of our dual radiotracer approach. More preclinical models that naturally co-express MUC4 and HER2 receptors should be enrolled in the future. Moreover, the Traz-resistance caused by other epitope-masking receptors, such as CD44/ hyaluronan, has yet to be investigated by our imaging approach. A large-scale clinical study of 99m Tc-MIRC208 SPECT and the clinical translation of 99m Tc-MIRC213 SPECT should be conducted in the future. The correlation between the SUV 208 /SUV 213 (2 h) ratio and the Traz-sensitivity in cancer patients remains to be explored in future clinical studies.

Conclusion
In summary, we present a proof-of-concept for the dual radiotracer approach to predict the Traz-resistance caused by epitope masking. We provide a powerful tool for assessment of the expression levels of both membrane and Traz-bound HER2. This tool could effectively stratify patients for Traz-therapy before treatment by rapidly screening HER2-positive patients and predicting their Traz-responsiveness, thus ultimately guiding personalized therapy and reducing toxicity and cost of unnecessary therapies that don't contribute to patient benefit.

Methods and Materials
All animal studies were performed according to the protocols approved by the Institutional Animal Care and Use Committee at Peking University. Detailed information on cell cultures, animal models, HER2-trageted VHHs generation, radiotracers preparation, SPECT imaging and tumor treatment is provided in Supplementary Information.

First-in human study with 99m Tc-MIRC208
The Institutional Review Board of Peking University Cancer Hospital & Institute approved this study (#2019KT114, NCT04591652), and all subjects signed a written informed consent. To data, two patients with clinical stage III invasive left breast ductal carcinoma were enrolled in this study. Patient 01 and patient 02 are women aged 55 and 65, respectively. The HER2 overexpression of primary tumors in both patients had been confirmed by IHC using the commercial test kit (IHC 3+). Both patients were enrolled without any prior treatment.
99m Tc-MIRC208 SPECT and 18 F-FDG PET/CT were acquired within 3 days for comparison. For 18 F-FDG PET/CT, patient fasted for at least 6 h before intravenous injection of 18 F-FDG at a dosage of 5.6 MBq/kg of body weight. 18 F-FDG PET/CT scan was performed 1 h post injection with a PET/CT scanner (Philips Medical Systems). 99m Tc-MIRC208 was freshly prepared in the morning of imaging day and the radiochemical purity was proved with ITLC method to be greater than 95% before use. 99m Tc-MIRC208 was given by intravenous injection and the injected activity was 14.2 MBq/kg. The total injection doses of patient 01 and patient 02 were 880.4 MBq and 823.6 MBq, respectively. A chest to abdomen SPECT/CT scans was performed at 2 h after 99m Tc-MIRC208 injection with a SPECT/CT scanner (Symbia T16; Siemens). A siemens workstation (MultiModality Workplace) was used for data processing. The images were evaluated and quantified by two experienced nuclear medicine physicians.

Statistical analysis
Detailed information on sample numbers and statistical tests used are described in the figure legends. Calculation was performed using GraphPad Prizm 8.0 software. The differences between two groups were tested with student's paired t test. Multiple comparison was done with two-way analysis of variance (ANOVA) followed by a Bonferroni post hoc test. P value of 0.05 or lower was considered statistically significant. n.s. indicates not significance.