Youssef Chahid

Nuclear medicine examinations have become indispensable tools in modern medicine. These imaging techniques, such as positon emission tomography (PET) and single photon emission computed tomography (SPECT), provide invaluable insights into physiological processes within the human body, enabling early disease detection, staging, treatment planning, and treatment monitoring. However, the accuracy of nuclear medicine tools is not solely dependent on technology but is deeply influenced by patient-related characteristics, medication use, and analytical methods. A comprehensive understanding of these factors is essential for the accurate interpretation of nuclear medicine examinations.

This thesis is structured into three main parts, each addressing distinct aspects of factors that may affect nuclear medicine examinations, namely patient-related, medication-related, and analytical method-related factors.

In the first part of this thesis, the focus is on the exploration of which various patient-related factors can influence the outcomes of nuclear medicine examinations, with a specific focus on breast cancer patients. It is well known that accurate sentinel lymph node (SLN) staging is essential for both prognosis and treatment in patients with breast cancer [1]. However, preoperative lymphoscintigraphy may fail to visualize the SLN in some patients. Reported rates of SLN nonvisualization vary between 2% and 28% [2-6]. Different patient characteristics (body mass index (BMI), age) and tumour characteristics (size, location, palpability) have been found to be associated with SLN nonvisualization [2-6]. Although the injected dose of radiotracers and injection technique seem to be correlated with SLN non-visualization [2], information on the potential impact of the level of experience, for example, in radiotracer preparation or the level of experience of the administrator, is absent. In chapter 1, we describe a large retrospective study that we have performed which included 1886 SLN procedures to identify potential unknown independent factors associated with SLN nonvisualization on lymphoscintigraphy [7]. The SLN nonvisualization rate was 25.1% on lymphoscintigraphy at 4 hours post-injection. The SLN nonvisualization rate decreased to 9.4% after reinjection of the radiopharmaceutical. Multivariable analysis showed that age ≥ 70 years, BMI ≥ 30 kg/m2, and nonpalpable tumours were independent predictors of SLN nonvisualization. Tumour location, brand of radiopharmaceutical, injected dose and volume, experience of preparer, and administrator were not associated with SLN nonvisualization.

In conclusion, our results support the robustness of SLN detection on lymphoscintigraphy, independent of the experience of the radiotracer preparer or administrator. Our study demonstrated that reinjection of the radiotracer, in a peri- or intra-tumoural injection setting, is a viable option to enhance SLN visualization rates. Another potential strategy to reduce SLN nonvisualization rate to 2.0% may involve a multisite injection technique: i.e., simultaneous injection of peritumoral, subcutaneous, and subareolar regions [8]. Based on our results, it may be valuable to initiate future research studies examining the potential reduction of preoperative SLN nonvisualization in patients with non-palpable tumours, aged ≥ 70 years, or with a BMI ≥ 30 kg/m2 through the implementation of simultaneous intratumoral and periareolar injections. Additionally, it would be intriguing to explore whether these risk factors remain valid when SLN procedures are performed using hybrid tracers, such as indocyanine green (ICG) in combination with [99mTc]Tc-nanocolloid [9].

In chapter 2, we comment on the interesting findings of Quak et al. that age and BMI are significantly associated with breast density [10]. They showed that patients with a high-fat content in breast tissue were, based on univariate analysis, significantly more at risk of SLN nonvisualization compared to patients with less fatty breast densities [11]. Due to the lack of performing a multivariable analysis, it was not possible to properly assess the possible influence of breast density for SLN nonvisualization in relation to other risk factors such as natural ageing and BMI. For that reason, we conducted a large retrospective study, as described in chapter 3, to reproduce their claim that breast density is an independent predictor of SLN nonvisualization on lymphoscintigraphy [12]. However, after performing a multivariable analysis in 758 breast cancer patients, we concluded that only age ≥ 70 years, BMI ≥ 30 kg/m2, and nonpalpable tumours were independent predictors of SLN nonvisualization. Differences in tumour size, BI-RADS classification, or breast density were not significantly associated with SLN nonvisualization. Thus, parameters derived from mammography or breast MRI are not useful for predicting SLN nonvisualization on lymphoscintigraphy.

In conclusion, we have shown that interpretations based on univariate analysis can be misleading. Therefore, it is preferable to perform larger studies with multivariable analysis to adjust for confounders. Based on our findings, we conclude that parameters obtained from mammography or breast MRI do not enhance the previously described multivariable model, which incorporates age, BMI, and tumour palpability, for predicting SLN nonvisualization.

The second part of this thesis explores the influence of medications on nuclear medicine examinations. In chapter 4, we conducted a systematic review of the