Application of the long axial field-of-view PET/CT with low-dose [¹⁸F]FDG in melanoma

Positron emission tomography (PET) integrated with computed tomography (CT) is a whole-body imaging technique combining functional and morphological information in one procedure. PET/CT, mainly with the radiotracer 2-deoxy-2-[fluorine-18] fluoro-D-glucose ([¹⁸F]FDG), the workhorse of molecular imaging, is nowadays a standard imaging method applied in the clinical routine of oncology for several indications, including diagnosis, staging, response evaluation, targeting of radiation therapy, disease control and recurrence detection [1, 2]. Particularly, in malignant melanoma, [¹⁸F]FDG PET/CT is considered the modality of choice when there is clinical suspicion of locoregional or systemic metastases at initial staging, disease recurrence, and for treatment response assessment [3‐7].

In most types of tumors, the imaging field of view (FOV) for PET/CT studies is from the skull base to upper thigh (torso imaging), which covers most of the relevant portions of the body in many oncological diseases [1]. However, due to the propensity of malignant melanoma to metastasize anywhere in the body, true whole-body (total body) imaging, i.e., from the top of the head through the feet, is recommended as the standard of care for the assessment of the malignancy [8]. This approach, however, leads to a non-negligible extension of the scanning time with a subsequent decrease in throughput, which is important for busy nuclear medicine departments.

Significant developments have occurred in the field of molecular imaging since the original description of the prototype, fused PET/CT scanner by Beyer et al. [9]. Most recently, long axial field of view (LAFOV) PET/CT systems, such as the Biograph Vision Quadra (Siemens Healthineers) and the uEXPLORER (United Imaging), have become available, rendering digital total-body PET/CT as the current state of the art, enabling, apart from the achievement of larger anatomical coverage, an increase in system sensitivity [10‐12]. Indeed, preliminary results from the clinical application of these systems show that they lead to improved image quality and lesion quantification resulting from higher sensitivity, alternatively allowing for a significant reduction in acquisition time or low-dose examination protocols [13‐15].

In the present study, we investigated melanoma patients with the new LAFOV Biograph Vision Quadra PET/CT after application of low-dose [¹⁸F]FDG. Our aim was to determine an appropriate acquisition time range for low-dose oncological imaging by analyzing different emulated scan times and the quality of the resultant PET images.

In standard clinical practice, the achievement of high quality PET images that can overcome the inherent poor sensitivity of current whole-body PET scans requires either an increase in the radiotracer injection dose or a prolongation of the acquisition time, both methods with radiation exposure- and logistical consequences [15, 21]. In this context, the recent advent and clinical introduction of ultra-extended FOV PET/CT scanners represents an evolution in molecular imaging with potentially major clinical and financial implications, while enabling, at the same time, a reduction of the associated radiation burden [22]. Importantly, it has already been shown that low-dose or even ultra-low-dose radiotracer activities achieve a comparable image quality to conventional PET, and meet clinical requirements in different tumors [13‐15, 23, 24]. Based on this experience, we, herein, aimed to determine an appropriate acquisition time range for low-dose oncological PET/CT imaging with the new LAFOV Biograph Vision Quadra PET/CT (Siemens Healthcare) after application of 2.0 MBq/kg, and not 3.0 MBq/kg as per EANM guidelines [1], [¹⁸F]FDG activity in a cohort of melanoma patients.

The major findings of our analysis are the following: firstly, the reduction of PET acquisition times up to 5 min in low-dose [¹⁸F]FDG LAFOV PET/CT imaging is feasible in melanoma and can be safely performed in terms of visual analysis, applied both in a binary fashion (normal-pathological) and based on calculation of lesion detection rate. In particular, the shortening of PET acquisition times to 6 min was associated with absolutely no potential clinical consequences in the context of patient staging or restaging in the present cohort for any of the studied patients. Further, a decrease of both SNR and TBR is observed by reducing the PET acquisition time, although the change in absolute tumor SUV_mean values is small and clinically rather non-relevant.

The Biograph Vision Quadra PET/CT system essentially is comprised of an axial concatenation of the equivalent of four Biograph Vision PET 600 scanners, providing an axial FOV of 106 cm, equipped with silicon-based photomultiplier (SiPM)-based photon detectors characterised by superior timing resolution. These characteristics enable improved TOF estimation and efficient photon detection, a high spatial resolution, and an increased sensitivity by a factor of five when compared to the preceding Biograph Vision 600 system [25‐28]. The results of the visual analysis of our study show that this increased sensitivity of the new scanner, may allow for a reasonable decrease in acquisition time by 5 min (i.e., from 10- to 5-min), even when applying low dose examination protocols. Further, absolutely no differences were observed between the reference of 10-min PET protocol and the 8-min and 6-min acquisitions regarding classification of scans as normal or pathological, while the differences in lesion detectability rate were minimal and, importantly, without any potential clinical consequence for any patient of the cohort. This means that all patients would have been correctly characterized in terms of staging and restaging of the disease even when applying a 6-min, low-dose PET/CT protocol. Notably, with regard to the proportion of pathological scans, a decrease of the acquisition times even up to four minutes did not lead to any significant change, as reflected by the respective 95%-CI for the differences between various acquisition prortocols.

Particularly in malignant melanoma, where true whole-body (total body) imaging and, therefore, two-bed position scanning with the Biograph Vision Quadra system is required, this shortening of PET acquisition times without a concomitant, clinically relevant decrease of diagnostic performance would be very practical for busy departments, since it would allow the execution of a greater number of exams and lead to an increase in patient throughput. At the same time, the shortened protocol would improve patient comfort and considerably prevent motion artifacts. Moreover, the reduction of the applied radiopharmaceutical dose represents an economic as well as a dosimetric advantage. These considerations take on even greater significance, when taking into account the potential need for serial PET/CT scanning of many of the patients with advanced melanoma who undergo immunotherapy in terms of treatment monitoring [29].

On the other side, the shortening of the PET protocol was associated with a decrease of both parameters used to objectively evaluate image quality, namely SNR and TBR. The observation that extending the acquisition times improves image quality is not new. In specific, the herein observed trend is in line with the results of previous studies also employing ultra-extended FOV PET/CT scanners, which similarly showed a decreasing SNR as PET acquisition times decreased in cohorts of different tumors studied with different tracers applied both at standardized and low doses [13‐15].

The parameter TBR also decreased progressively from PET-10 to PET-2. This finding seems to be inconsistent with previous studies on ultra-extended FOV PET/CT scanners, which showed an inverse relationship, i.e., higher TBR values in shorter acquisitions compared to standard acquisitions [13, 15]. It has to be noted, however, that in these studies, TBR was calculated as the ratio of SUV_max of tumor lesions divided by SUV_mean of the liver background, which was not the case in our analysis. In the present analysis TBR was defined as the ratio of SUV_mean of the melanoma lesions divided by SUV_mean of the liver. We prefered to calculate the mean tumor uptake since it incorporates information from multiple voxels, making it less susceptible to image noise compared to SUV_max [30]. This approach of SUV_mean-based calculation of TBR was also recently applied by Sari et al. in another study with the Biograph Vision Quadra system [31]. Indicatively, the application of the alternative formula (lesion SUV_max/liver SUV_mean) would have led to consistent results with the above-mentioned studies, namely a progressive increase of TBR as the acquisition times decrease (data not provided in the manuscript but available at request). Notably, in lesion-based analysis, the reduction of acquisition time to 6 min and 5 min led to no significant differences in TBR and tumor lesions’ SUV_mean, respectively.

Our study has some limitations. Firstly, this is a single-center retrospective analysis. Thus, a validation of the herein presented findings in larger patient cohorts ideally studied in the context of a multicenter, prospective trial would be required. Secondly, the use of histopathological findings as a reference to validate the detectability of tumor lesions is a more reliable method than referring to PET-10 images. However, in our study the vast majority of the PET-positive findings were not histopathologically confirmed, which is, obviously, not possible in the clinical setting and was, moreover, beyond the scope of this analysis. Nevertheless, since all included patients had a biopsy-confirmed melanoma, for which they received the respective treatment, and given the very high accuracy of [¹⁸F]FDG PET/CT in detection of the malignancy, most focal [¹⁸F]FDG-avid lesions, for which a benign aetiology could be safely excluded, were considered consistent with tumor involvement. Thirdly, the analysis was confined to the comparison of the 10-min PET scans, as recommended by the manufacturer, with five specific sub-divisions (2-min, 4-min, 5-min, 6-min, and 8-min) of the original acqusitions, and did not include even more acquisitions. However, we believe that the conclusions drawn with the present comparisons provide sufficient hints for clinical practice in the context of the purpose of this study. Finally, the study focused on the analysis of the scans covering the first bed position, i.e. from the head to the thigh. The PET scans performed in terms of the second bed position, covering the lower extremities, were not analyzed in the same fashion, since the original PET acquisitions lasted shorter (5 min). This will be however the topic of a future work of our group.

In an attempt to determine an appropriate acquisition time range for low-dose oncological imaging with the new LAFOV Biograph Vision Quadra PET/CT, we investigated a cohort of malignant melanoma patients after administration of body weight-adapted dose of 2.0 MBq/kg [¹⁸F]FDG. We could show that the reduction of PET acquisition times from head to upper thigh up to 5 min in low-dose LAFOV PET/CT imaging can be safely performed, leading to no significant changes in terms of visual analysis. Moreover, the shortening of the PET protocol from 10 to 6 min was associated with absolutely no potential clinical consequences for any of the enrolled patients. Finally, an increase in noise was observed by reducing the protocol from longer to shorter acquisition times, although in lesion-based analysis, the change in TBR and tumor SUV_mean values was non-significant up to 6-min and 5-min acquisitions, respectively. These findings may have significant implications for the diagnostic work-up of patients with melanoma, given the need for true whole-body imaging in this type of cancer.