Introduction

Gastric cancer (GC) ranks fourth in tumor incidence and second in cancer death, and it occurs most commonly in East Asia areas [1]. In China, over 70% of GC patients are diagnosed at advanced stages, and this limits their effective treatments [2]. What’s worse, the resistance of gastric cancer toward chemotherapy renders this disease with no effective therapeutic options. The overall survival (OS) rate for 5 years is less than 20% [3,4,5].

Human epidermal growth factor receptor 2 (HER2) is overexpressed in 7–34% of GC patients [6]. Its inhibitor, trastuzumab, has been demonstrated to increase OS for HER2-positive advanced gastric cancer (AGC) and gastroesophageal junction cancer (GEC) patients [6, 7]. On October 2010, FDA approved the combination therapy of trastuzumab and chemotherapy as the first-line treatment of HER2-positive AGC and GEC. Essentially, accurate determination of HER2 status in vivo is essential in effective treatment. The most frequently used analytical methods of HER2 levels are immunohistochemistry (IHC) and fluorescent in situ hybridization (FISH). However, pathological samples taken from patients may not represent the large bulky mass of lesions, and the absence of tumor may result in false negative diagnosis. In addition, tumor heterogeneity may exist between primary tumors and metastases, among different metastases, and even in primary tumor alone, and within the same lesion before and after treatment [6, 8,9,10,11,12]. It is difficult to use IHC/FISH to accurately reflect systemic lesions, to monitor HER2 levels, and to measure heterogeneity of tumors [13]. To overcome these problems, a systemic in vivo techniques, such as positron emission tomography (PET) or single-photon emission computed tomography (SPECT) imaging, are required to monitor HER2 expression in patients noninvasively.

Many SPECT/PET imaging agents have been developed to determine HER2 expression and localize the primary tumors and/or metastases. In 2006, Perik et al. reported that 111In-DTPA-trastuzumab could predict the cardiotoxicity and therapeutic efficacy of trastuzumab [14]. In 2009, Dijkers et al. reported that 89Zr-trastuzumab PET imaging agent exhibited excellent tumor uptake and visualization of HER2-positive metastases, and its detection sensitivity was equivalent to that of 111In-trastuzumab; 89Zr-trastuzumab gave better images than 111In-trastuzumab because of high spatial resolution and sensitivity of PET equipment [15, 16]. Many PET radionuclide-labeled trastuzumab agents have been suggested to be effective for HER2-positive tumor detection, such as those labeled with 124I and 64Cu [17,18,19,20,21,22].

Our laboratory has a long-standing interest in radio-immunotherapy (RIT) to treat solid tumors using antigen-selective tumor-targeting radiolabeled antibodies and the corresponding PET imaging agents. We have successfully developed 64Cu-NOTA-trastuzumab as PET/CT imaging agent in clinical imaging of patients with gastric cancer [20]. It exhibited comparable detection sensitivity as that of 18F-FDG, even in liver metastases. The long physical decay half-life and low liver background of 124I-labeled agents were demonstrated to be useful to observe small lesions of tumor. Here, we describe our synthesis and evaluation of 124I-trastuzumab in patient-derived xenograft (PDX) mouse models of GC. In addition, it was successfully applied in PET/CT imaging of GC patients in a clinical trial with Ethics Approval License: 2018KT48. In the clinical studies, we evaluated the feasibility and plausible conditions for PET imaging of 124I-trastuzumab to detect HER2-positive lesions in HER2-positive GC patients and to differentiate quantitatively between HER2-positive and HER2-negative GC lesions.

Materials and methods

Preparation of 124I-trastuzumab

Trastuzumab was labeled with 124I by the N-bromosuccinimide (NBS) method. The details of the radiolabeling and quality control conditions are described in the supplemental materials.

Micro-PET imaging of 124I-trastuzumab in PDX mouse models

All animal experiments were performed in accordance with the guidelines of the Animal Care and Use Committee of Peking University (Ethics Approval License: 2015KT08). The details to establish the mouse models can be found in the supplemental materials. Micro-PET images were acquired at 2, 24, 48, 72, and 96 h post-injection on a SuperArgus PET scanner (Sedecal, Spain) with 14.8 MBq 124I-trastuzumab or 124I-hIgG1 in HER2-positive and HER2-negative PDX mouse models [23, 24]. HER2 levels in PDX xenograft tumors were confirmed by IHC stain: HER2-positive models were HER2 3 + , HER2-negative models were HER2 1 + .

Patient enrollment

This study was performed under a single-center prospective imaging protocol and was approved by the Medical Ethics Committee of Peking University Cancer Hospital (Ethics Approval License: 2018KT48). The patients included in this study had pathologically confirmed HER2-positive (IHC 3 + or IHC 2 +/FISH +) or HER2-negative (IHC 2 +/FISH-, IHC 1 +) advanced GC/GEC and were aged between 20 and 75 years. Patients’ specimens for pathological examination came from gastroscopy or surgical procedures, mainly at primary lesions. According to the size and invasion range of the primary tumor, we took 1 ~ 4 specimens in tumor bodies, 2 ~ 4 specimens in tumor-free gastric walls at a distance from the tumor, and 1 ~ 2 upper and 1 ~ 2 lower broken ends, respectively. The main exclusion criteria were that the patients were without congestive heart failure, severe liver or kidney dysfunction, pregnant or lactating, had no known hypersensitivity to trastuzumab or could not lie in the scan bed for more than 1 h. Two HER2-positive (IHC 3 +) patients and four HER2-negative (IHC 2 +/FISH-) patients from June 2018 to December 2018 were included. All patients provided written informed consent before participating in the study.

124I-trastuzumab and 18F-FDG PET/CT imaging in GC/GEC patients

Patients underwent 18F-FDG PET/CT 3 days before 124I-trastuzumab PET/CT imaging studies. For 124I-trastuzumab PET imaging, a patient was initiated a thyroid iodine uptake blockade by oral administration of Lugol’s potassium iodide (ten drops each time, 3 times a day) 3 days before and continued for 7 days after the administration of the radiotracer. To reduce the nonspecific uptake in the liver, patient 1 and patient 6 were co-injected with 5 mg “cold” trastuzumab and 124I-trastuzumab, and the other four patients were co-injected with 10 mg “cold” trastuzumab and 124I-trastuzumab. For all patients, 74 ± 17 MBq 124I-trastuzumab was injected intravenously. PET imaging was performed at 1, 24, 48, 72, and 96 h post-injection for patient 2 and at 1, 24, and 48 h post-injection for the other 5 patients.

Image analysis

Tumors uptake analyses were restricted to lesions of ≥ 1 cm diameter on a PET image of 18F-FDG PET/CT relative to the adjacent tissues. After the imaging features of the lesions were analyzed, the regions of interest (ROI) were drawn to measure the standard uptake value (SUV). Based on 18F-FDG imaging results, a total of 29 lesions were identified and were analyzed correspondingly by PET imaging of 124I-trastuzumab. The analysis showed that the SUVmax of 18 lesions from HER2-positive patients (patient 1 and patient 3) was higher than 11 lesions from HER2-negative patients (patient 2, 4, 5, 6). So, we observed lesions with 18F-FDG high uptake and 124I-trastuzumab high uptake in HER2-positive lesions, and lesions with 18F-FDG high uptake and 124I-trastuzumab low uptake in HER2-negative lesions. The mean SUV of solid organs (liver, spleen, and kidney) was calculated by the layered boundary delineation method. The mean SUV of solid organs, such as liver, spleen, and kidney, was calculated using the layered boundary delineation method. The mean SUV of blood was obtained from the PET images of the cardiac ventricles, and the heart wall was visualized as a region of relatively low uptake adjacent to the ventricles. The internal radiation dose was calculated on the basis of radioactivity data from the blood, urinary excretions, and normal tissues of the heart, liver, spleen, kidneys, and other parts of the body at each time point using the OLINDA/EXM software.

Radiation dosimetry

The organ radioactivities and their data analysis are described in the supplemental materials. The absorbed radiation doses in each individual organ were calculated using the OLINDA/EXM software [16, 25, 26].

Statistical analysis

The data are presented as the mean ± SD. The micro-PET images were quantified using ROI-based quantification, and those data were analyzed using two-tailed, unpaired Student’s t-tests, and p values less than 0.05 as statistically significant. These statistical computations were performed using the Excel software program (Microsoft Corporation, Redmond, WA) or GraphPad software (GraphPad Prism 5).

Results

Synthesis, quality control, and micro-pet imaging of 124I-trastuzumab

The radiolabeling efficiency of 124I-trastuzumab was 95% ± 0.5% (n = 5), and the radiochemical purity after purification reached 99.2% ± 0.3% (n = 5). The specific activity was calculated to be 1.3 GBq/µmol. The results of quality control are shown in Table S1. Radio-TLC and HPLC analysis (Figure S1) of the radioligand revealed no aggregates, fragments, or radioactive impurities. The 124I-trastuzumab was produced as sterile and endotoxin-free injectable solution, which was stable both in saline and 5% human serum albumin (HSA) at 4 °C for up to 7 d (Figure S2).

The micro-PET imaging of 124I-trastuzumab versus 124I-hIgG1 in HER2-positive and HER2-negative PDX mouse models were performed at different time point after injection. Both 124I-trastuzumab and 124I-hIgG1 in HER2-positive PDX mouse models mainly resided in circulatory system within 2 h post-injection. Along with time, the tumor increased and could be visualized on the imaging of 124I-trastuzumab at 24 h with uptake at 11.91 ± 1.08 ID %/g and T/M ratio of 9.13 ± 2.52 (Fig. 1) and remained to be clearly visible up to 96 h post-injection. Low tumor uptake was observed during the entire development period using 124I-hIgG1, with uptake of 3.83 ± 0.09 ID %/g, n = 5, p = 0.0002 and T/M ratios of 3.49 ± 0.31 at 24 h (Fig. 1). In comparison, 124I-trastuzumab in HER2-negative PDX mouse models did not show clear tumor uptake, under the same conditions, with uptake of 5.05 ± 0.17 ID %/g n = 5, p = 0.0004 and T/M ratios of 2.73 ± 0.30 at 24 h (Fig. 1). Although the uptake of 124I-trastuzumab in HER2-negative PDX models was slightly higher than that of 124I-IgG1 in HER2-positive PDX models at 24 h post-injection, the T/M ratios of 124I-trastuzumab in the former were lower than those of 124I-IgG1 in the latter.

Fig. 1
figure 1

a Micro-PET images of 124I-trastuzumab in HER2-positive PDX models (124I-tras. (HER2 +)), 124I-hIgG1 in HER2-positive PDX models (124I-hIgG1. (HER2 +)) and 124I-trastuzumab in HER2-negative PDX models (124I-Tras. (HER2-)) at 2, 24, 48, 72, and 96 h post-injection; b Region of interest (ROI) analysis of the tumor uptake values in micro-PET images with 124I-Tras. (HER2 +), 124I-hIgG1. (HER2 +), and 124I-tras. (HER2-) models; c the tumor-to-muscle (T/M) ratios of 124I-trastuzumab in 124I-Tras. (HER2 +), 124I-hIgG1. (HER2 +), and 124I-Tras. (HER2-) models. Yellow arrows indicate the tumors. **, P < 0.05, ***, P < 0.001

Full size image

Patient characteristics

We had 6 patients in this study, with 2 patients HER2-positive based on IHC 3 +/FISH +  and 4 patients HER2-negative based on HER2 2 +/FISH- (Table 1). All 6 patients were stage IV cancer, and 5 patients had liver, bone, and lymph node metastases. Based on the selection criteria for lesions described above, we selected a total of 29 lesions (18 HER2-positive lesions and 11 HER2-negative lesions) for image analysis. No infusion-related reactions or adverse events were observed during the study.

Table 1 Patient characteristics
Full size table

The regional time–activity curves

The regional time–activity curves of 124I-trastuzumab in 5 patients were collected at 1, 24, and 48 h (Fig. 2). At 1 h, 124I-trastuzumab mainly resided in the circulatory system and well-perfused organs (including liver, spleen, and kidney). The SUVmax of 124I-trastuzumab in blood was 7.3–11.2, 3.6–11.9, and 2.2–8.2 at 1, 24, and 48 h, respectively, showing long-term high activity in vivo. Liver and kidney also had high radioactivity uptake due to the clearance of the radioligand, such as degradation into small molecular compounds in the liver, kidney, and other organs and excretion through the urine or feces. The SUVmax values were 4.6–7.4, 2.6–5.4, and 2.5–3.9 in the liver, 0.9–3.3, 0.5–4.5, and 0.4–3.6 in the kidneys, and 0.4–2.8, 3.3–14.7, and 3.9–19.5 in the urinary bladder, at 1 h, 24 h, and 48 h, respectively. The uptake in the kidney slightly increased at 1 h and decreased at 24 h and 48 h. The uptake in the urinary bladder was visualized at 1 h, increased rapidly at approximately 24 h, and decreased at 48 h. The spleen had high uptake values of 4.6–6.9, 2.2–6.3, and 0.8–4.6. The whole-body distribution of 124I-trastuzumab in patient 2 showed normal radiation dose compared with other analogous radioligands (Figure S5 and Figure S6). The uptake in the thyroid was a special concern for iodinated radioligand. 124I-trastuzumab showed low uptake at 1 h and 24 h but increased at 48, 72, and 120 h, with the SUVmax of 1.9–2.3, 4.1–8.4, and 4.8–11.0 at 1, 24 and 48 h post-injection, respectively. Within 24 h, the nonspecific uptake in the thyroid was pre-blocked, but the thyroid uptake increased after 48 h. Because the SUVmax values of patient 6 were 64.5 and 76.1 at 24 and 48 h, respectively, the regional time–activity curves of patient 6 were not shown.

Fig. 2
figure 2

Regional time–activity curves of 124I-trastuzumab in 5 patients. a The SUVs in the blood, thyroid, liver, spleen, kidneys, and bladder were calculated from ROI-based analyses of each tissue. The SUV in the blood was measured from a region of interest in the heart. In patient 6, the SUV in the thyroid was extremely high. b The SUVmax in the tumor and tumor-to-normal tissue ratio (TNR) in 5 patients

Full size image

Radiation dosimetry

Assessments of internal organ dosimetry were performed in 5 patients, with injected radioactivity at 74 ± 17 MBq. The highest absorbed dose was at the heart wall, followed by the liver, spleen, and kidneys (Table 2). The average effective dose was estimated to be 0.3011 ± 0.005 mSv/MBq.

Table 2 Organ-absorbed doses and effective dose estimated from whole-body 124I-Trastuzumab PET
Full size table

Time of lesion detection

The SUVmax values at left iliac metastasis were 5.1, 9.2, and 7.4, and those at right iliac metastasis were 6.5, 7.7, and 6.6 at 1 h, 24 h, and 48 h post-injection, respectively (Fig. 3a), maximized at 24 h. At 1 h, the radioligand mainly resided in the circulatory system and was not fully distributed. At 48 h, both radioactive decay and physiological metabolism of the radiotracer led to a low image quality because of insufficient radioactivity counting. The optimal time for tumor imaging was at 24 h.

Fig. 3
figure 3

a The 124I-trastuzumab images at 1 h, 24 h, and 48 h of patient 1 who had bilateral iliac metastases; b The 124I-trastuzumab PET image at 24 h and 18F-FDG PET image at 1 h of the HER2-positive lesions in patient 1 who had a right humeral metastasis, bilateral ilium metastases, and multiple spinal metastases; c The 18F-FDG PET image at 1 h and 124I-trastuzumab PET images at 1 h, 24 h, and 48 h of the HER2-negative lesions in patient 3 who had lymph node metastases. The red arrow indicates the tumors

Full size image

Comparison of direct lesion uptake between 124I-trastuzumab and 18F-FDG at 24 h

The correlation between the PET imaging of 124I-trastuzumab and 18F-FDG in HER2-positive lesions (Fig. 3b) was established in patient 1, who also underwent through GC resection to establish his IHC score of 3 + . 18F-FDG PET imaging was used to establish qualified tumor lesions, showing humeral metastasis, iliac metastases, and spinal metastases in this patient. In the 18 HER2-positive lesions established in the 18F-FDG PET image, the 124I-trastuzumab PET imaging showed higher uptake at 24 h than that at 48 h (SUVmax 7.83 ± 0.55 vs. 5.73 ± 0.47, p = 0.0065). The detection sensitivity of 124I-trastuzumab was 77.8% (14/18) at 1 h, 100% (18/18) at 24 h, but 88.9% (16/18) at 48 h. The correlation between the PET imaging of 124I-trastuzumab and 18F-FDG in HER2-negative lesions (Fig. 3c) was established in patient 3, who also underwent through GEC resection to establish his IHC score of 2 + , FISH-. 18F-FDG PET imaging was used to establish that the patient had lymph node metastasis, with SUVmax 6.7. The PET imaging of 124I-trastuzumab showed low uptake of SUVmax 1.9, 1.5, and 1.4 at 1, 24, and 48 h, respectively. In the 11 HER2-negative lesions established in the 18F-FDG PET image, the 124I-trastuzumab PET imaging showed lower uptake at 24 h than that with 18F-FDG (SUVmax 1.75 ± 0.29 vs. 6.46 ± 0.44, p < 0.0001).

PET imaging could help resolve ambiguity in CT images in identifying tumors. For one patient, we observed a 0.5 cm cortical defect in the left transverse process of the 5th cervical spine, which was not obvious on CT and was easily misdiagnosed (Fig. 4a). However, significant radioactivity uptake in both 18F-FDG PET imaging and 124I-trastuzumab imaging was observed at the same location, with SUVmax value of 5.31 and 7.09, respectively. Another lesion was observed in the third lumbar vertebra (L3), and the morphological changes of the lesions were also vague on the CT image. Both 18F-FDG and 124I-trastuzumab PET imaging showed significant radioactive accumulation with SUVmax value of 10.42 and 6.42, respectively (Fig. 4b).

Fig. 4
figure 4

a The 124I-trastuzumab PET image at 24 h and 18F-FDG PET image at 1 h of the HER2-positive lesions in patient 1 in the left transverse process of the 5th cervical spine; b The 124I-trastuzumab PET image at 24 h and 18F-FDG PET image at 1 h of the HER2-positive lesions in patient 1 in the third lumbar spine. The red arrows indicate the tumors

Full size image

Discussion

Radioligands based on 124I should be suitable to investigate lengthy biological processes since iodine-124 has relatively long physical half-life (t1/2 = 4.08 days), and it is a positron-emitting radionuclide with high-energy γ-emissions (0.6 MeV, 61%) and positron emissions (2.14 MeV, 24%). Direct radioiodination of surface-exposed tyrosine residues is effective and convenient way to synthesize these ligands, and several studies have shown the added iodine minimal impact on plasma clearance of the radiolabeled antibody.

Compared with the 89Zr- and 64Cu-labeled antibodies, 124I-trastuzumab achieved higher imaging contrast because of lower nonspecific uptake and better tumor-to-soft-tissue radioactivity ratio. However, after 124I-trastuzumab is degraded upon internalization, the released 124I-iodide may somehow diffuse freely from the tissues, resulting in the loss of specific signal in target tissues and in the increase in background signal over time. In contrast, when 64Cu- and 89Zr-labeled antibodies are processed, the radionuclides are trapped intracellularly in lysosomes and radioactivity remain in cell, leading to high uptake in normal clearance organs (liver and kidney) and high background uptake of other tissues that passively absorb antibodies [21, 22, 27]. When 124I cleaves with antibody upon degradation, it enters the blood circulation and gradually enters the thyroid and stomach organ, and it clears from kidney, instead of liver, leading to higher tumor to liver ratios, which are obviously beneficial in imaging resolution.

124I-trastuzumab gave higher imaging contrast than 64Cu-NOTA-trastuzumab because of lower nonspecific uptake and better tumor-to-soft-tissue ratios. Compared with our previously reported 64Cu-NOTA-trastuzumab, we observed that both 124I- and 64Cu-labeled trastuzumab gave specific signals in PDX mouse tumors and high-contrast images [20]. However, they differ in tumor and background uptake, resulting in different levels of imaging contrast in the PDX mouse model. For example, 124I-trastuzumab gave lower liver uptake at 3.62 ± 1.18 percent injected dose per gram [%ID/g] than the residualizing 64Cu-NOTA-trastuzumab (7.87 ± 0.52%ID/g, p < 0.0001 for each at 24 h, and in consequence, the tumor to liver ratios reached 1.93 ± 0.017 for 124I-trastuzumab at 24 h and 0.95 ± 0.029 for 64Cu-NOTA-trastuzumab at 14 h (p < 0.0001, n = 3), and 2.26 ± 0.069 for 124I-trastuzumab at 72 h and 1.14 ± 0.094 for 64Cu-NOTA-trastuzumab at 36 h (p < 0.0001, n = 3)). In addition, the tumor-to-muscle ratios reached 9.87 ± 0.095 for 124I-trastuzumab at 24 h and 10.59 ± 0.025 for 64Cu-NOTA-trastuzumab at 14 h (p = 0.003, n = 3), and 17.27 ± 0.76 for 124I-trastuzumab at 72 h and 7.13 ± 0.49 for 64Cu-NOTA-trastuzumab at 36 h (p < 0.0001, n = 3). In the Xenograft mouse model, we confirmed that the HER2 levels in tumors of HER2-positive models were HER2 3 + and those in HER2-negative models were HER2 1 + [20], with details in the supplemental materials.

We had successfully applied 124I-trastuzumab in human clinical PET imaging. In order to observe the complete pharmacokinetics of 124I-trastuzumab in vivo, we did PET imaging at 1, 24, 48, 72, and 96 h post-injection for patient 2. After 124I-trastuzumab degrades upon internalization, the released 124I-iodide may diffuse freely within the tissues, resulting in the loss of signal in target tissues and in the increase in background signal over time. Better contrast images of tumor were obtained at 24 and 48 h than those at 72 h and 96 h, confirming the degradation hypothesis described above. The effective dose of 124I-trastuzumab was estimated to be 0.3011 ± 0.005 mSv/MBq. For a normal injection dose of 74 ± 17 MBq, the internal radiation dose of 124I-trastuzumab absorbed by patient was estimated to be 17.16–27.4 mSv per PET scan, not significantly different from the estimated dose of 7.0 mSv for injected dose of 370 MBq per 18F-FDG PET scan.

18F-FDG PET imaging could be used to detect lesions based on the changes in glucose metabolism in tumor cells, while 124I-trastuzumab PET imaging could help to monitor the HER2 expression levels in lesions, resulting in accurate diagnosis and early treatment. We measured the tumor SUVmax values in both HER2-positive and HER2-positive patients using both 124I-trastuzumab and 18F-FDG PET imaging (Fig. 5). Stronger PET signals were observed using 124I-trastuzumab in HER2-positive lesions than those in HER2-negative lesions, with SUVmax of 7.83 ± 0.55 versus 1.75 ± 0.29, p < 0.0001 at 24 h. If we use a SUVmax cutoff value of 4.0 between HER2-positive and HER2-negative, the diagnosis sensitivity was 94.4%, and the specificity was 100%. More samples are needed to verify an exact threshold. With 18F-FDG, all 29 lesions in both HER2-positive and HER2-negative patients had SUVmax values greater than 4, and the range of SUVmax values in HER2-negative lesions overlaps with that of SUVmax values of the HER2-positive lesions (SUVmax 4.4–8.9, 4.7–14.1). In HER2-positive lesions, the SUVmax values of 124I-trastuzumab were 3.0–10.5, and those of 18F-FDG were 4.7–14.1. In HER2-negative lesions, the average of SUVmax values of 124I-trastuzumab was 1.75 ± 0.29, and that of 18F-FDG was 6.46 ± 0.44 (p < 0.0001). Thus, 18F-FDG PET imaging cannot be used to differentiate HER2-positive tumor tissues even with scores of 3 + from the test of the patients. The CT images only cannot detect the small and less morphological changes in lesions of patients, while the combination of PET imaging using 18F-FDG and 124I-trastuzumab could be used to identify abnormal metabolic signals in tissues with less morphological changes.

Fig. 5
figure 5

a SUVmax values of the 124I-trastuzumab in HER2-positive (IHC 3 + or 2 + and FISH +) and HER2-negative (IHC 2 +/FISH-) patients; b SUVmax values of the 18F-FDG in HER2-positive and HER2-negative patients; c comparison of SUVmax values between 124I-trastuzumab and 18F-FDG images in HER2-negative patients; d comparison of SUVmax values between 124I-trastuzumab and 18F-FDG images in HER2-positive patients

Full size image

Although patient 1 received trastuzumab 4 days before the 124I-trastuzumab PET/CT scan, the lesions also showed clear uptake of 124I-trastuzumab. This shows that the radioligand could be useful to evaluate therapeutic responses, consistent with that of 89Zr-pertuzumab [27].

When iodinated radioligand is used, absorption into thyroid is always a concern. The SUVmax values of 124I-trastuzumab in the thyroid were 1.9–2.3, 4.1–8.4, and 4.8–11.0 at 1, 24, and 48 h post-injection, respectively. Within 24 h, the nonspecific uptake in the thyroid was well blocked, but after 48 h, the thyroid uptake was high. We attempted to block the thyroid uptake of 124I by orally administering Lugol’s potassium iodide, and the thyroid uptake was still observed; this finding indicates that more Lugol’s potassium iodide may be needed. We concluded that it might be difficult for patients to adhere to the scheduled dosage of Lugol’s potassium iodide and that the dosage of Lugol’s potassium iodide may not sufficient to block nonspecific thyroid uptake. When 124I is introduced into 124I-trastuzumab under oxidative conditions, it is generally randomly conjugated to tyrosine residues present in the antibody. Once 124I-trastuzumab is internalized and degraded, the released 124I-iodide or 124I-iodotyrosine may diffuse from the tissue, resulting in high uptake in normal thyroid tissues.

Liver uptake of 124I-trastuzumab may be modulated with therapeutic drugs. With a 10 mg cold trastuzumab preinjection, all patients had relatively low liver backgrounds at 24 h, and the SUVmax of the liver was 2.6, 2.9, 4.2, 4.4, and 9.9 in 5 trastuzumab-naive patients and 5.4 in one patient on trastuzumab therapy. Previous PET imaging studies on 89Zr-trastuzumab and 64Cu-DOTA-trastuzumab for HER2 patients demonstrated that 50 mg cold trastuzumab was needed to decrease the high liver uptake in trastuzumab-naive patients. Our results indicated that 10 mg cold trastuzumab preinjection was adequate for HER2 PET imaging in trastuzumab-naive patients, and 124I-trastuzumab may have the advantage of low liver uptake, consistent with the results obtained in the PDX mouse models.

The limitation of present human clinical PET study is the small number of patients with limited diversity in HER2 status. Among the 6 patients studied, only two types of patients were enrolled, HER2-positive (IHC 3 +) and HER2-negative (IHC 2 +/FISH-), which were measured using HE staining, IHC and FISH (Figure S7). Even for the limited number of patients, heterogeneity of HER2 status between primary and metastatic lesions was not established. In addition, the HER2 levels of all metastatic lesions through pathological examination were not attempted.

Conclusion

In this study, we have demonstrated that 124I-trastuzumab PET/CT imaging is safe without any adverse effect on patients and acceptable level of radiation dosage in whole body of human and can measure HER2 levels in HER2-positive patients with GC/GEC. The potential clinical applications of the imaging include assessing the HER2 status of lesions before therapy to guide selection of therapy, or monitoring the changes of HER2 levels during therapy to monitor the effect of therapy. In addition, HER2-targeted PET/CT imaging may also be used in prospective clinical trials of patients with GC/GEC to analyze the ability of 124I-trastuzumab to detect unsuspected HER2-positive metastatic disease arising from HER2-negative GC and to help predict the response of HER2-positive GC to HER2-targeted therapy. If combined with 18F-FDG imaging, 124I-trastuzumab imaging may also help the studies on the mechanism of tumor metastases and drug resistance.