Detection of HER2-Low Lesions Using HER2-Targeted PET Imaging in Patients with Metastatic Breast Cancer: A Paired HER2 PET and Tumor Biopsy Analysis

Abstract

Trastuzumab deruxtecan (T-DXd), a human epidermal growth factor receptor 2 (HER2)-targeted antibody-drug conjugate, demonstrated remarkable efficacy in previously treated patients with HER2-low metastatic breast cancer (mBC), marking a new therapeutic option for this patient population. Prior studies with HER2 PET using ⁸⁹Zr-radiolabeled antibodies were limited by high rates of imaging false positives for HER2-positive malignancy. In this retrospective study, we investigate whether these false positives (HER2-negative on pathology) could be explained by HER2-low lesions. Methods: A retrospective study was conducted of mBC patients who previously enrolled in 2 prospective HER2 PET imaging trials: NCT02286843 using ⁸⁹Zr-trastuzumab and ⁸⁹Zr-pertuzumab and NCT04692831 using ⁸⁹Zr-ss-pertuzumab. Patients were included if paired HER2 PET scan and biopsy were performed within a 2-mo period. Of 56 total patients, 23 patients met the inclusion criteria. Pathology results for biopsied lesions were collected, without repeat interpretation, and lesions were classified as HER2-positive, HER2-low, or HER2-0. SUV_max of biopsied lesions were compared between pathologic classifications to determine whether lesion uptake intensity could differentiate between HER2-positive and HER2-low lesions. Results: All prior false-positive lesions on HER2 PET scans from NCT02286843 were reclassified as HER2-low (instead of HER2-negative). In the ⁸⁹Zr-trastuzumab cohort, 3 lesions were HER2-positive (33%) and 6 were HER2-low (67%); in the ⁸⁹Zr-pertuzumab cohort, 2 were HER2-positive (29%) and 5 were HER2-low (71%). In the ⁸⁹Zr-ss-pertuzumab cohort (NCT04692831), 7 patients underwent recent biopsies of 8 total lesions demonstrating 1 HER2-positive (12%), 5 HER2-low (62%), and 2 HER2-0 lesions (25%). HER2 PET SUV_max of biopsied lesions were compared between HER2-positive and HER2-low lesions for the combination of all 3 radiotracer cohorts. HER2-low lesions had a significantly higher SUV_max (median, 12.7; interquartile range, 8.05) than did HER2-positive lesions (median, 6.4; interquartile range, 1.98; P = 0.01). Conclusion: HER2 PET imaging with ⁸⁹Zr-radiolabeled antibodies detects HER2-low lesions in addition to HER2-positive lesions in patients with mBC, suggesting its ability to visualize the entire spectrum of HER2 expression. All prior false positives on ⁸⁹Zr-trastuzumab and ⁸⁹Zr-pertuzumab PET scans were reclassified as HER2-low. Lesion SUV_max is not reliable in differentiating HER2-positive from HER2-low lesions; however, it may be useful in distinguishing lesions expressing HER2 from HER2-0 lesions.

Keywords: HER2; PET; breast cancer; pathology.

Graphical abstract

FIGURE 1.

Distribution of SUV_max according to HER2 status. Box plot follows Tukey convention.

FIGURE 2.

HER2-positive sternal lesion on HER2 PET imaging. 68-y-old woman with HER2-positive mBC with recent progression of disease on abemaciclib, trastuzumab, and fulvestrant. FDG PET and subsequent HER2 PET with ⁸⁹Zr-ss-pertuzumab sagittal maximum-intensity projection images (A and B, respectively) and axial fused PET/CT, CT, and PET images from FDG PET and HER2 PET demonstrated FDG-avid (C) and HER2 PET radiotracer-avid (D) left sternal lesion (blue arrows), with SUV_max of 34.4 on HER2 PET. This lesion underwent CT-guided biopsy (E) by interventional radiology. Representative hematoxylin and eosin (H&E) micrographs show poorly differentiated metastatic carcinoma (F) that was HER2 positive (IHC score, 3+) (strong complete membrane staining in >10% of tumor cells (G).

FIGURE 3.

HER2-0 and HER2-low bone lesions in 1 patient on HER2 PET. 47-y-old woman with HER2-positive cT2N1 left breast cancer diagnosed with de novo osseous metastatic disease on FDG PET. FDG PET and subsequent HER2 PET with ⁸⁹Zr-ss-pertuzumab maximum-intensity projection images (A and B, respectively) demonstrate multiple FDG-avid and HER2 PET bone lesions, left breast mass, and left axillary nodes. Axial fused PET/CT, CT, and PET images from FDG PET and HER2 PET demonstrated FDG-avid (C) and HER2 PET radiotracer-avid (D) left T12 lytic osseous lesion (blue arrows), with SUV_max of 9.4 on HER2 PET. This lesion underwent CT-guided biopsy (E) by neuroradiology. Representative hematoxylin and eosin (H&E) micrographs show moderately differentiated metastatic carcinoma (F) that was HER2-negative by IHC score of 0 (G). One month later, she underwent epidural tumor resection and decompression with posterior thoracic spine stabilization due to progression of T10 epidural disease and left T11–T12 facet fracture. Prior axial fused PET/CT, CT, and PET images from HER2 PET showed radiotracer-avid T10 lytic osseous lesion with epidural extension (H) (SUV_max, 11.8); pathology demonstrated HER IHC 1+ staining, consistent with HER2-low status (pathology slides not shown).

FIGURE 4.

HER2-0 liver lesion on HER2 PET. 63-y-old woman with HER2-negative right breast cancer after lumpectomy and adjuvant combination therapy with cyclophosphamide, methotrexate, and fluorouracil who subsequently developed HER2-positive right frontal lobe metastasis (IHC score, 1 to 2+; ISH, 2.0) and underwent craniotomy. While receiving paclitaxel, trastuzumab, and pertuzumab, she developed progression of disease with solitary FDG-avid segment 7 liver lesion, with few additional liver lesions seen on MRI. (A) FDG PET maximum-intensity projection image showed solitary FDG-avid segment 7 hepatic lesion (green arrow) and few FDG-avid bone lesions (red arrows). (B) Subsequent HER2 PET with ⁸⁹Zr-ss-pertuzumab maximum-intensity projection images demonstrated mild radiotracer avidity in liver and bone lesions. Axial fused PET/CT, CT, and PET images from FDG and HER2 PET demonstrated faint FDG avidity (C) and HER2 PET radiotracer avidity (D), slightly above or equal to liver background, in segment 3 hepatic metastasis (yellow arrows) which was biopsied, with SUV_max of 7.2 (liver background SUV_mean, 5.5). Additional liver lesions with slightly higher HER2 PET radiotracer avidity were not biopsied, including segment 5 lesion (E) (blue arrows), with SUV_max of 8.7. Segment 3 liver lesion underwent CT-guided biopsy (F) by interventional radiology. Representative hematoxylin and eosin micrographs showed moderately differentiated metastatic carcinoma (G) that was HER2-negative (IHC score, 0) (H).