**3. Discussion**

18F-FDG PET/CT is regarded as a reliable outcome predictor and an elective imaging technique for treatment response evaluation of MM due to its ability in differentiating active from inactive sites of the disease [30]. Three independent, easily attainable with routine PET/CT parameters have been recognized to adversely affect both PFS and overall survival (OS). In particular, the presence of more than three focal 18F-FDG-avid lesions, a SUVmax > 4.2 of the lesions and the presence of extramedullary disease (EMD) are associated with an adverse outcome [12,31]. Moreover, the complete remission of the 18F-FDG-avid lesions after therapy has been shown to confer superior PFS and OS; contrarily, the persistence of pathologic findings on PET/CT after treatment is associated with a worse prognosis [12–14,31,32].

Thus far, the vast majority of PET/CT studies in MM were restricted either to descriptive analyses, mainly through the identification of focal, hypermetabolic lesions, and/or semi-quantitative analyses of parameters derived from static imaging of focal lesions. Little light, however, has been shed on the interpretation and prognostic value of the diffuse bone marrow involvement—irrespective of the presence of focal lesions—on PET/CT. Moreover, the quantitative aspect of PET, which is feasible only after performance of dynamic scanning, has only been scarcely utilized, due to the routine application of conventional, static protocols.

In the present case, we monitored by means of dynamic and static PET/CT a patient with an initially asymptomatic SMM, who demonstrated a transition to symptomatic myeloma, and was subsequently successfully treated. The patient did not show any typical signs of myeloma involvement on PET/CT, i.e., focal, hypermetabolic lesions, at any phase during the course of the disease. However, at transition from SMM to symptomatic MM, a diffusely increased 18F-FDG uptake in the bone marrow was observed; this was accompanied by a marked increase of both the semi-quantitative (SUV values) and the quantitative, pharmacokinetic parameters, derived from bone marrow of the iliac crest. Importantly, after the successful therapeutic intervention, the diffuse uptake in the bone marrow as well as the semi-quantitative and quantitative parameters showed a pronounced remission. This response was also confirmed by the long-term, clinical follow-up of the patient.

Our findings suggest, firstly, that in untreated MM, a diffuse 18F-FDG uptake in the bone marrow may reflect an actual bone marrow infiltration by plasma cells. Particularly in patients suffering from SMM, the appearance of a diffuse hypermetabolic bone marrow pattern—regardless of the concurrent emergence of focal lesions—during the course of the entity may reflect transition to symptomatic disease and should, therefore, lead to further investigation. We are, indeed, aware of the several causes leading to a false-positive, diffuse, homogeneous, bone marrow 18F-FDG uptake on PET/CT, such as severe anemia, previous administration of granulocyte colony-stimulating factor (G-CSF), chemotherapy or erythropoietin [33]. However, in the present case, all potential causes of a false-positive bone marrow 18F-FDG uptake could be excluded from the patient's history.

Secondly, SUV values not only from myeloma lesions—as consistently highlighted by previous studies—but also from random bone marrow samples, may be used to monitor disease transition and response to treatment. This could be particularly helpful in the follow-up of myeloma patients negative for 18F-FDG-avid focal lesions. The metabolic state of the bone marrow as evaluated by SUV calculations and/or in comparison to reference organs has recently been put into focus of MM research, rendering promising results as a potentially prognostic factor [22,34]. The present findings are in support of this direction.

Finally, the information acquired after the application of full dynamic PET/CT was in line with the respective qualitative (visual) and semi-quantitative (SUV) findings during the different phases of the disease. The dynamic 18F-FDG PET/CT protocol offers the unique advantages to investigate the tracer accumulation over time through generation of the respective time activity curves (TACs) as well as to extract pharmacokinetic indices that reflect dedicated parameters of the tracer's metabolism, such as perfusion, transport or phosphorylation. This quantitative aspect is a major advantage of PET/CT, which is neglected when using conventional, static, whole-body protocols (usually 60 min post-injection) and descriptive analysis as the only diagnostic tool. In the present case, quantitative, dynamic PET/CT showed that the transition of asymptomatic SMM to symptomatic disease was accompanied by a marked increase of 18F-FDG accumulation in the bone marrow over time, compared to baseline PET/CT. Moreover, several quantitative 18F-FDG parameters, including the regional blood volume (VB), the tracer influx rate (Ki), the carrier-mediated transport of the tracer from plasma to bone marrow (K1), the phosphorylation rate of 18F-FDG in the bone marrow (k3) as well as the degree of tracer heterogeneity—reflected by the parameter fractal dimension (FD)—showed a distinct increase. Contrarily, a pronounced decrease of the respective TAC and pharmacokinetic parameters were observed after the successful therapeutic intervention. These results are in line with previous findings of our group regarding the potential role of dynamic PET in MM prognosis and treatment response evaluation [22,35–37]. By complementing the information offered by conventional imaging with the multiparametric, pharmacokinetic data extracted by dynamic PET/CT, the diagnostic certainty of the reading physician could be enhanced, particularly in patients with ambiguous findings. Moreover, our understanding of the pathophysiology of the disease and its response to treatment can be improved.

Although these findings could sugges<sup>t</sup> the wider usage of dynamic PET/CT in MM, more data, preferably derived from large prospective studies, are warranted to prove the potential benefit of the modality. Moreover, we note some practical considerations related to the possible implementation of dynamic PET/CT in clinical routine: firstly, it is more timeconsuming than conventional PET/CT, since it requires in most cases a 60-min acquisition, followed by the conventional, static, whole-body PET/CT acquisition. This may lead to patient discomfort as well as to logistical issues in a nuclear medicine department. Furthermore, data interpretation is challenging, based on sophisticated software tools and application of the—rather complex—compartment modeling and fractal analysis. According to the previous, for the time being, qualitative and semi-quantitative analysis will remain the main evaluation tools of PET/CT in MM. However, dynamic PET/CT could be applied in selected cases, for example in the context of clinical trials in MM, adding significant quantitative information and reducing inter-observer variability. Moreover, the recent advent of new PET/CT scanners, which allow dynamic studies over several bed positions by using a continuous bed movement, will facilitate the use of dynamic PET protocols and reduce the whole acquisition time, making dynamic PET/CT an attractive and cost-effective approach in oncological imaging [38].
