Quantitative [89Zr]Zr-Trastuzumab PET and Diffusion- Weighted MRI for Characterization of Metastatic HER2-Positive Breast Cancer with PET/MRI

Abstract

Current methods for evaluating HER2 expression in breast cancer are invasive and fail to capture spatial and temporal heterogeneity between primary tumors and metastases. Nuclear imaging allows for whole-body, noninvasive assessment of human epidermal growth factor receptor 2 (HER2) and can distinguish between HER2-positive and HER2-negative lesions. MRI offers superior soft-tissue contrast and quantitative metrics, such as the apparent diffusion coefficient (ADC) from diffusion-weighted (DW) imaging, providing prognostic value. The goal of this study was to present quantitative imaging metrics from [⁸⁹Zr]Zr-trastuzumab PET and simultaneous DW-MRI to characterize HER2-positive metastatic breast cancer lesions. The secondary aim was to explore the utility of combining [⁸⁹Zr]Zr-trastuzumab PET with DW-MRI for intratumoral habitat mapping using multiparametric [⁸⁹Zr]Zr-trastuzumab PET/MRI to enhance characterization of lesions and assess the response to HER2-targeted therapy. Methods: Fifteen patients with confirmed HER2-positive breast cancer underwent simultaneous PET/MRI 5-7 d after receiving 77 ± 1.9 MBq of [⁸⁹Zr]Zr-trastuzumab. Whole-body ADC maps were generated from DW-MRI, and regions of interest in normal and malignant tissues were delineated. Imaging metrics included SUV for PET and ADC for DW-MRI. Threshold- and clustering-based methods were applied for intratumoral characterization through multiparametric mapping. Results: Malignant tissues exhibited significantly higher [⁸⁹Zr]Zr-trastuzumab uptake than did normal tissues. High uptake in the kidneys, liver, and blood pool complicated lesion identification near these tissues. ADC mapping improved lesion characterization in the brain, soft tissue, and bone. The diagnostic accuracy of [⁸⁹Zr]Zr-trastuzumab PET alone improved when combined with ADC mapping (area under the curve, 0.59 and 0.75, respectively). Multiparametric analysis revealed intratumoral heterogeneity, identifying distinct subregions with variable tracer uptake and diffusion characteristics. Conclusion: Combining [⁸⁹Zr]Zr-trastuzumab PET with DW-MRI offers a multiparametric imaging approach for characterizing HER2 expression and cellular density in HER2-positive metastatic breast cancer. Increased tracer uptake in malignant lesions and improved lesion characterization through ADC mapping highlight the potential of this combination for evaluating treatment response and tumor heterogeneity. Large-scale validation is needed to confirm these findings and support integrating [⁸⁹Zr]Zr-trastuzumab PET and DW-MRI into clinical management for better patient outcomes.

Keywords: DW-MRI; HER2 PET/MRI; HER2-positive metastatic breast cancer; [89Zr]Zr-trastuzumab; multiparametric.

Graphical abstract

FIGURE 1.

Physiologic uptake of [⁸⁹Zr]Zr-trastuzumab across various tissues, presented as SUV_mean ± SD (n = 12–15 per tissue). (A) Differential uptake of [⁸⁹Zr]Zr-trastuzumab in normal tissue was observed. (B) Representative whole-body [⁸⁹Zr]Zr-trastuzumab PET maximum-intensity projection images of 3 patients, 5–7 d after administration, display heterogenous accumulation of tracer in various organs.

FIGURE 2.

(A) Bar plots demonstrating overall increased [⁸⁹Zr]Zr-trastuzumab uptake observed in metastatic lesions compared with contralateral and normal organ uptake. (B) Representative T1-weighted MRI and fused [⁸⁹Zr]Zr-trastuzumab PET/MRI with primary or metastatic lesions outlined in white.

FIGURE 3.

Transverse and sagittal CT, T1-weighted anatomic and corresponding ADC maps of various breast cancer metastatic lesions, outlined in magenta. Decreased ADC signal observed within enhanced regions of brain lesions is indicative of cellularly dense tissue, surrounded by regions of increased ADC, characteristic of edema. MRI and ADC allowed for clear delineation of tumors in breast and soft-tissue metastases. ADC maps allowed for accurate placement of cardiac blood pool (BP) ROI and delineation of osseous lesions borders.

FIGURE 4.

Receiver operating characteristic (ROC) curves demonstrating predictive performance of [⁸⁹Zr]Zr-trastuzumab PET/MRI metrics for response to HER2-targeted therapy. ADC_mean showed highest individual predictive value (AUC, 0.74), whereas SUV_mean and SUV_max alone were not predictive. TBR_mean yielded modest improvement over raw SUV metrics, and its combination with ADC_mean produced slight increase in AUC (AUC, 0.75); however, overlapping 95% CIs suggest this was not statistically significant.

FIGURE 5.

(A) Representative longitudinal [¹⁸F]FDG PET maximum-intensity scans showed decreased tumor burden, followed by progression. (B) Representative [⁸⁹Zr]Zr-trastuzumab PET/MR images are shown. (C) Two multiparametric assessment techniques, with contours marking voxel classifications. Median-based thresholding (2-step method) demonstrates even split between regions based on metric intensity, with limited spatial coherence. In contrast, clustering (1-step method) identified distinct intratumoral clusters, reflected in voxel metric distribution.

FIGURE 6.

Quantitative multiparametric evaluations of breast-to-brain metastases using [⁸⁹Zr]Zr-trastuzumab PET/ADC compared with physiologic habitats: edema with restricted diffusion, tumor enhanced core, vasogenic edema, and residual. (A) Representative transverse [⁸⁹Zr]Zr-trastuzumab PET, anatomic T1-weighted, ADC, and T2/FLAIR sections of brain lesion are displayed, with overlay of multiparametric maps on using median intensity separation (2-step) and clustering (1-step) of ADC and SUV values (B). (C) Bar plots of SUV_mean and ADC demonstrated improved separation between 1-step, 2-step, and physiologic regions. Voxelwise ADC vs. SUV plots illustrate overlap of dual-intensity voxels.

FIGURE 7.

Qualitative analysis of large metastatic lesions revealed distinct patterns in ADC and SUV distributions. (A) SUV was more homogenous in responding tumors. (B and C) Nonresponding lesions showed homogenous ADC distributions skewed toward lower values, indicating increased cellular density, whereas responding lesions had more heterogeneous ADC distributions. HH = high PET/high ADC; HL = high PET/low ADC; LH = low PET/high ADC; LL = low PET/low ADC; RD = restricted diffusion.