To our knowledge, this is the first feasibility study evaluating the semi-quantitative parameters of prone SPECT/CT using [99mTc]Tc-sestamibi and comparing them with MBI-based semi-quantitative parameters in 18 patients with LABC. This study presents the first step towards a possible application of semi-quantitative parameters of prone SPECT/CT in LABC patients for prediction of response to NAC. Various semi-quantitative parameters were composed for early and delayed SPECT acquisitions (5 min p.i and 90 min p.i.) and MBI. No significant difference was observed between MBI and early SPECT semi-quantitative parameter FTV (p = 0.46). TBRmean and TBRmax were significantly higher for SPECT compared to MBI and showed greater variability between the measurements (p < 0.05).

Early SUV and TBRSPECT values were higher compared to late SUV and TBRSPECT, which probably is related to the clearance of [[99mTc]Tc-sestamibi via transmembrane transporter proteins (like P-gp and the multidrug resistance protein (MRP)). In this regard, [99mTc]Tc-sestamibi WOR is a promising predictive parameter for tumor non-responsiveness to NAC, as it reflects tumor multidrug resistance. Sciuto et al. reported high sensitivity and specificity for prediction of chemoresistance when applying a cut off WOR of 45% [19]. We were able to compose WOR derived from early and delayed SPECT/CT for potential future use in therapy response prediction.

The difference between FTVSPECT and FTVMBI might be explained by the FTVMBI calculations assuming a spherical tumor, while in clinical practice, tumors exhibit various shapes. FTVMBI data of four patients were excluded because the tumor was not completely within the field-of-view (located close to the chest wall) or the tumor showed diffuse growth, making realistic volume dilations unfeasible. These encountered limitations of MBI confirm the existence of challenges in achieving accurate tumor volume measurements when using MBI [12,13,14]. The increased variation between MBI and SPECT in tumors with higher average TBRmax and TBRmean values might be attributed to the absence of attenuation and scatter correction in MBI compared to SPECT. Photon counts within tumor's VOI are affected by surrounding tissue (axilla e.g.), tumor specifications, breast properties, and imaging settings [14]. Consequently, the same tumor may appear differently across different views or detectors, resulting in variations in VOI measurements, which might affect the TBR calculations. Moreover, the lower TBRMBI values are likely due to the higher septal penetration occurring by virtue of the 'near' contact imaging of the breast compared to the SPECT imaging. Therefore, the strengths of SPECT over MBI lie in its capacity for 3D imaging, especially for tumors located close to the chest wall and the clinical available SPECT attenuation and scatter correction, potentially composing more precise semi-quantitative parameters.

This study contains limitations. First, the limited number of subjects constitutes a major limitation and hence our results should be interpreted carefully. Although our study concerns only a small study population, we believe that our findings exhibit the complexity of assessing semi-quantitative SPECT/CT parameters and contribute to the knowledge of the application of [99mTc]Tc-sestamibi for response prediction in LABC. A dynamic study should be conducted to evaluate which model best suits [99mTc]Tc-sestamibi quantification and to relate the obtained pharmacokinetic measures to semi-quantitative measures obtained at different time intervals, hence examining their validity in this context. Second, the 42% threshold iso-contouring was utilized for the SPECT measurements based on a phantom study [16] since no protocols were available specifying the settings for quantitative SPECT with [99mTc]Tc-sestamibi for LABC. Visual evaluation was conducted for the contouring, with manual adjustments if necessary, making delineations observer-dependent and thus affecting their reproducibility. Although Collarino et al. showed in a phantom study that absolute SPECT/CT quantification of breast studies using [99mTc]Tc-sestamibi seems feasible (< 17% deviation) when 42% threshold iso-contouring is used for delineation of tumors (≥ 17 mm diameter) for various TBRmax (ranging from 9.6 to 3.3) [16], it is not clear how this 42% threshold iso-contouring would affect other semi-quantitative parameters in patients, such as FTV. Therefore, further investigation is necessary to determine which iso-contouring methods are most relevant and reproducible for clinically relevant semi-quantitative parameters in patients before applying quantitative SPECT for LABC in clinical settings. Third, outcome measures, such as pathologically confirmed therapy response, were not incorporated in this study. Before implementing response monitoring based on semi-quantitative SPECT with [99mTc]Tc-sestamibi in clinical practice, the clinical relevance of SPECT-derived semi-quantitative parameters needs to be assessed in a future large prospective clinical trial, including histopathological response to NAC as primary outcome measure and gold standard. For this, the practice SPECT quantification guidelines [20], which were not available during our data collection but which overall principles align with our study, could be considered. Furthermore, before classifying a change in a semi-quantitative parameter as a response, it is crucial to assess its test–retest variability. Additionally, it is worthwhile investigating for which tumor molecular subtypes these parameters are more consistent.