Exploration of tumor size measurement methods in preoperative breast cancer assessment using whole-body silicon photomultiplier PET: feasibility and first results

• DOI: 10.1007/s11604-024-01533-3
• Publication Date: 2024-02-12T07:02:02Z
• Authors: Hiroyuki Kuroda; Takeshi Yoshizako; Nobuhiro Yada; Tomomi Kamimura; Nobuko Yamamoto; Mitsunari Maruyama; Rika Yoshida; Mizuki Fukuda; Yuko Kataoka; Masayuki Itakura; Yasushi Kaji
• Keywords: Adult; Aged, 80 and over; Silicon; Breast Neoplasms; Middle Aged; Tumor Burden; 03 medical and health sciences; 0302 clinical medicine; Fluorodeoxyglucose F18; Positron Emission Tomography Computed Tomography; Positron-Emission Tomography; Preoperative Care; Humans; Feasibility Studies; Original Article; Female; Whole Body Imaging; Radiopharmaceuticals; Retrospective Studies; Aged

Abstract Purpose Whole-body silicon photomultiplier positron emission tomography (WB SiPM PET) could be used to diagnose breast cancer spread before lumpectomy. We aimed to investigate the method of measuring the tumor size by WB SiPM PET as a basis for diagnosing breast cancer spread in the breast.

Materials and methods We retrospectively reviewed 35 breast cancer lesions in 32 patients who underwent WB SiPM PET/CT in the prone position as preoperative breast cancer examinations from September 2020 to March 2022. In all cases, a 20-mm spherical VOI was placed in the normal mammary gland to measure the mean standardised uptake value (SUVmean) and the standard deviation (SD) of 18F-fluorodeoxyglucose (FDG) uptake. We prepared four types of candidates (SUVmean + 2 SD, SUVmean + 3 SD, 1.5 SUVmean + 2 SD, 1.5 SUVmean + 3 SD) for thresholds for delineating tumor contours on PET images. On the semiautomatic viewer soft, the maximum tumor sizes were measured at each of the four thresholds and compared with the pathological tumor sizes, including the extensive intraductal component (EIC).

Results The lesion detection sensitivity was 97% for WB SiPM PET. PET detected 34 lesions, excluding 4-mm ductal carcinomas in situ (DCIS). PET measurements at the '1.5 SUVmean + 2 SD' threshold demonstrated values closest to the pathological tumor sizes, including EIC. Moreover, '1.5 SUVmean + 2 SD' had the highest concordance (63%).

Conclusions The study demonstrated that among various PET thresholds, the '1.5 SUVmean + 2 SD' threshold exhibited the best performance. However, even with this threshold, the concordance rate was limited to only 63%.