Diagnostic performance of dedicated breast positron emission tomography in cases of histology-unknown abnormal uptake, and parameters associated with breast cancer

Objectives: Ring-shaped dedicated breast positron emission tomography (dbPET) is one of the available high-resolution molecular breast imaging modalities, and it can detect small breast cancers. It is known to exhibit higher resolution and detectability of small lesions than whole-body positron emission tomography computed tomography (PET/CT) [1, 2]. In addition, the features of F18-fluorodeoxy glucose (FDG) uptake corresponding to known malignancies have been analyzed [1, 3]. The aim of the current study was to evaluate the diagnostic performance of dbPET in cases of histology-unknown abnormal uptake and define parameters associated with malignancy.

Methods: A total of 627 women underwent FDG PET/CT and dbPET at our hospital between April 2015 and September 2017 for pre-therapeutic (n = 46) or post-therapeutic (n = 121) evaluation of breast cancer, further examination of suspicious lesions detected via another modality but not diagnosed as malignant (n = 21), and cancer screening (n = 439). Women with histories of other malignancies (n = 32) and metastasis (n = 15) were excluded. We analyzed 40 cases of abnormal uptake on dbPET in 37 breasts of 35 women, not including lesions in cases of known breast cancers. Two nuclear medicine physicians divided the morphological features of lesions determined on dbPET into one of three categories—focus (or foci), mass, or non-mass without any clinical information. Classification as “focus” indicated that lesions were a punctuate and too small to characterize (5 mm or less), and also included cases of multiple foci with a diffuse distribution. “Mass” indicated a focal uptake larger than focus (over 5 mm). “Non-mass” indicated uptake in an area that was not categorized as mass, and differed from normal surrounding glandular parenchyma and multiple foci with a linear or segmented distribution. In addition, quantitative values including the maximum standardized uptake value (SUV_max), SUV_peak, total lesion glycolysis (TLG), and metabolic tumor volume (MTV) of the lesions, and lesion-to-contralateral background ratio (L/B) were also obtained. The PET parameters and clinical factors associated with breast cancers and other lesions were compared using Pearson’s chi-square test or Fisher’s exact test, and the Mann-Whitney U test. Multivariate logistic regression analysis was also performed to define the factors associated with breast cancer.

Results: We explored 40 lesions in 37 breasts of 35women. Four women had a lesion in each of their two breasts, and a woman had two lesions in different segments of the same breast. Of these 40 lesions, 13 (32.5%) were determined to be breast cancers. Of the biopsies of these 13 breast cancers, 11 were ultrasonography (US)-guided, and 2 were stereotactic vacuum-assisted because they were not visible on US. The respective morphological features of the breast cancer group vs. the other group were 76.9% (10/13) vs. 40.7% (11/27) for mass, 15.4% (2/13) vs. 3.7% (1/27) for non-mass, and 7.7% (1/13) vs. 55.6% (15/27) for focus (p = 0.0122). SUV_peak (3.521 ± 2.387 vs. 2.001 ± 0.774, p = 0.0234), TLG (6.238 ± 15.319 vs. 0.851 ± 1.468, p = 0.0017), and MTV (0.977 ± 1.639 vs. 0.331 ± 0.559, p = 0.004), L/B (3.165 ± 2.916, p = 0.0432) also differed significantly between the two groups. Three quantitative parameters of dbPET (SUV_peak, TLG, and MTV) were highly correlated (correlation coefficient over 0.73), so they were assessed one by one in the multivariate analysis in order to avoid multicollinearity. Multivariate logistic regression analysis indicated that morphological categorization on dbPET was independently associated with malignancy. Conclusion: This study demonstrated that morphological features representing the size, number, and distribution of areas of abnormal uptake on dbPET were associated with breast cancers and may be useful for the diagnosis of histology-unknown lesions.

• Download figure
• Open in new tab
• Download powerpoint