Highlights of the 36th EANM Annual Congress 2023, from hometown Vienna, Austria: “A SYMPHONY OF SCIENCE”

Introduction

Between the 9th and 13th of September 2023, the international nuclear medicine community gathered in Vienna, hometown to the headquarters of the European Association of Nuclear Medicine (EANM), for its 36th Annual Congress. EANM encompasses 40 European National Societies et al. with 20 non-European affiliated societies. The 36th Annual Congress, chaired by Professor Valentina Garibotto, brought together a total of 7659 healthcare professionals, researchers, leading experts, policymakers and industry representatives. Out of 127 countries from around the world, 7047 participants gathered on-site and 612 online. Twenty countries were represented by over 100 participants, ranging to 20 countries sending their sole best delegate. China, with 202 participants, was the largest non-European country represented.

A total of 144 live on-site oral sessions and 15 e-Poster sessions were supplemented with 85 uploaded e-Posters on the Virtual Platform. In addition to scientific talks, participants could join innovative sessions specifically devoted to exchange and discussion, similar to the 2022 EANM conference [1]. The exhibition hall, covering 4195 m², hosted 151 companies and associations, organising an additional 12 satellite symposia. All activities were captured on a daily basis at EANM Live online TV, presented by Roy Sheppard. The EANM’23 Virtual Platform, which could be reached up to 3 months after the Congress, received a total of 589,164 views.

Out of the 2438 abstracts, 1991 were selected for presentation at the congress. The abstracts were selected based on their scientific merit by many reviewers. The scientific contributions were then divided over oral presentations and e-Posters. The content of the abstracts clearly showed that nuclear medicine diagnostics, therapy, and theranostics continue to increase; new tools and tracers are rapidly being established; and research and innovation consistently provide novel tools for validation and testing within the setting of clinical care. Notably, around 300 prospective studies (including a large number of clinical trials) were presented. Out of 356 top-rated abstracts, the authors of this review selected 26 to be highlighted in the opening ceremony. This review provides a summary of these scientific abstracts, describing the latest advancements, innovations, and breakthroughs in nuclear medicine as presented at the Highlights session of the EANM’23 Congress. The Highlights lecture ceremony was musically accompanied by the Vienna-based string quartet KK-STRINGS, who offered a musical journey through classical and modern music in the spirit of the Symphony of Science (presenting pieces from Antonio Vivaldi: Spring from the Four Seasons, 1720; Wolfgang A. Mozart: Divertimento in D, 1772; Johann Strauss II: Voices of Spring Waltz, 1882; and Stevie Wonder: You are the Sunshine of my Life, 1972).

Oncology highlights—the great classics

For oncology, the selected abstracts represented various topics focussing on different diagnoses and therapies and ranging from purely diagnostic to theranostic applications employing both radionuclide and “classic” targeted therapies. Despite the huge body of evidence available for the use of [¹⁸F]FDG in lymphomas and lung cancer, these topics still represent a field of active investigation. A Danish nationwide cohort study on the appropriate use of [¹⁸F]FDG positron emission tomography (PET) in non-small cell lung cancer (NSCLC) patients after curative treatment was presented by Guldbrandsen et al. They analysed data from 13,746 NSCLC patients diagnosed between 2007 and 2020 in the Danish Lung Cancer Registry who underwent curative-intent treatment. Multivariable regression was used to analyse trends in the use of [¹⁸F]FDG PET imaging over time and to identify factors associated with the use of [¹⁸F]FDG PET imaging during the first 2 years of surveillance after curative therapy. The analyses demonstrated a substantial increase in the use of [¹⁸F]FDG PET, mainly associated with advanced stage and treatment with definitive chemoradiation or stereotactic body radiation therapy (SBRT). On the one hand, these findings further underline that [¹⁸F]FDG PET is increasingly important in managing these patients, on the other hand, further research will be needed to determine how this increase in the use of [¹⁸F]FDG PET may have impacted patient outcomes and healthcare system efficiency. Voltin et al. addressed the issue of the divergent outcomes of approved products in real-world patients. They analysed a set of 88 patients, from four German university hospitals and one Italian centre, who underwent [¹⁸F]FDG PET before CAR T-cell therapy with tisagenlecleucel (n = 62) or axicabtagene ciloleucel (n = 26). Extra-nodal involvement emerged as the most relevant of conventional tumour and patient characteristics with metabolic tumour volume (MTV) further improving outcome prediction. Notably, while the most suitable MTV cut-off among patients receiving tisagenlecleucel was 11 mL, a markedly higher threshold of 259 mL showed optimal predictive performance in those undergoing axicabtagene ciloleucel treatment. The study has practical repercussions as metabolic tumour burden might also be used to define the extent of bridging therapy needed individually for each construct. In the accepted abstracts, the performance of new tracers has also been evaluated in clinical scenarios in which [¹⁸F]FDG PET is more frequently used. A correlation between immunohistochemistry and [⁶⁸ Ga]Ga-Trivehexin PET-CT imaging of αvβ6-integrin in patients of head and neck squamous cell carcinoma and pancreatic adenocarcinoma was evaluated by Da et al. from the Fortis Memorial Research Institute in Gurugram (India). In this pioneer first in human study, they demonstrated a good tracer uptake in 24/29 patients, and αvβ6-integrins expression was higher along the tumour margins than in the centre, which was consistent with the pattern of [⁶⁸ Ga]Ga-Trivehexin uptake.

Oncology—new development for diagnostics

Several abstracts covered novel diagnostic applications of PET imaging in breast cancer patients. [¹⁸F]FDG PET can be limited for detection of small metastases, lymph node metastases, and in patients with particular tumour types or receptor expression, for example, in patients with lobular or with low-grade oestrogen receptor-positive breast cancer. Other PET tracers than [¹⁸F]FDG have been proposed to overcome these limitations and to broaden the applicability of diagnostic PET imaging in patients with different clinical presentations of breast cancer.

For example, Knip et al. analysed patients with stage II/III or locoregional recurrent oestrogen receptor-positive breast cancer in a prospective multicentre clinical trial to compare the influence of histological subtype and tumour grade on [¹⁸F]FDG and oestrogen receptor-targeting [¹⁸F]FES PET. A total of 41 patients with 43 tumours were included, and suspect lesions were verified by pathology or conventional imaging. The authors showed that [¹⁸F]FES-PET is more accurate in lobular and grade-1 ductal tumours, whereas for grade 2 and grade 3 ductal tumours, [¹⁸F]FDG-PET and [¹⁸F]FES-PET performed equally well. The results indicate that both histological subtype and tumour grade can influence the selection of the optimal PET tracer and that a selection based on tumour biology may be necessary to increase the clinical value of PET imaging in heterogeneous diseases, such as breast cancer.

A different approach is the usage of pan-cancer tracers like radiolabelled fibroblast activation protein inhibitors (FAPIs) which may be more specific than [¹⁸F]FDG and can be applied in different tumour biologies. Guo et al. investigated the diagnostic accuracy of [⁶⁸ Ga]Ga-FAPI-46 PET-CT in patients with primary and metastatic breast cancer in comparison with breast ultrasound and [¹⁸F]FDG PET-CT. They included 30 patients without further subclassification for hormone receptor status who underwent PET-CT for diagnosis of suspected breast lesions at initial staging or for detection of recurrence. SUV_max, tumour-to-background ratio, lesion detection rate, and TNM classification were evaluated. [⁶⁸ Ga]Ga-FAPI PET-CT showed higher tracer uptake and tumour-to-background ratios than [¹⁸F]FDG PET-CT in different anatomical categories. Moreover, it detected more lesions than [¹⁸F]FDG PET-CT and breast ultrasound. This can lead to upstaging in TNM classification and potentially lead to a change in management in a relevant fraction of patients.

However, breast cancer was not the only tumour type which was investigated by FAP-targeted imaging. Chen et al. compared the diagnostic performances of [⁶⁸ Ga]Ga-FAPI PET-CT and [¹⁸F]FDG PET-CT in diagnosing lymphomas also aiming to characterise the influences of FAP and glycolytic markers on tracer uptake by involved lesions. A total of 186 participants underwent dual-tracer imaging, and [¹⁸F]FDG was characterised by a higher staging accuracy (98.4%) than [⁶⁸ Ga]Ga-FAPI (86.0%). Moreover, 52 FAPI + /FDG − lesions and 2939 FAPI − /FDG + lesions were observed. GLUT1 and hexokinase-2 were overexpressed in both lymphoma cells and the tumour microenvironment, while FAP was only expressed in stromal cells. Similarly, Watanabe et al. carried out a head-to-head comparison of [⁶⁸ Ga]Ga-FAPI PET-CT, [¹⁸F]FDG PET-CT, and contrast-enhanced CT in 241 patients with various solid tumours. Scans were reviewed by one blinded central reader. Clinical use and accuracy indicators were calculated by means of histopathological validation from 629 lesions. Per-patient, per-region, and per-lesion detection rates were significantly higher in [⁶⁸ Ga]Ga-FAPI PET-CT compared to [¹⁸F]FDG PET-CT for total, primary, abdominopelvic, liver, and other metastases (p < 0.05). Sub-analysis also showed that [⁶⁸ Ga]Ga-FAPI PET-CT outperforms [¹⁸F]FDG PET-CT in terms of accuracy indicators. [⁶⁸ Ga]Ga-FAPI PET-CT was also the subject of the study from Lanzafame et al. who focused their attention on various bone and soft-tissue type sarcomas. They analysed the scans of 200 patients and showed higher maximum SUV values particularly for soft-tissue type sarcomas and myxoid liposarcomas using [⁶⁸ Ga]Ga-FAPI compared to [¹⁸F]FDG (Fig. 1). Overall [⁶⁸ Ga]Ga-FAPI PET-CT showed higher clinical performance scores compared to [¹⁸F]FDG PET-CT. Correlation between [⁶⁸ Ga]Ga-FAPI uptake and histopathologic FAP expression showed a moderate positive trend. With a SUV_max > 10 cut-off for all regions, FAP-directed radioligand therapy eligibility was found for about half of the patients.

Fig. 1

In patients with sarcoma, [⁶⁸ Ga]Ga-FAPI PET has a higher tumour uptake and detection rate and equal accuracy when compared with [¹⁸F]FDG PET. FAP expression can be also used to candidate patients for radioligand therapy (courtesy of Helena Lanzafame et al.). A modified version of the figure appears in the final publication resulting from the presented study [2]

Besides FAPI PET, PSMA-targeted PET imaging and radioligand-based theranostics in patients with prostate cancer are and have been a major topic in all major nuclear medicine conferences over the recent years and already deserve to be regarded as a great classic. Certainly, EANM’23 should not be missed in this trend, and exciting and groundbreaking results on PSMA theranostics were presented. Hofman et al. showed the 54-month follow-up of the proPSMA trial, in which the authors found baseline [⁶⁸ Ga]Ga-PSMA-11 PET-CT to be prognostic for treatment failure in men with intermediate-to-high risk prostate cancer. A total of 251 patients (ISUP grade groups 3–5, PSA ≥ 20 ng/ml or clinical stage ≥ T3, M0 disease and a median follow-up of 41.0 months) were included. Patients were treated with curative-intent surgery or radiotherapy with or without androgen deprivation, and treatment failure was defined by biochemical failure, initiation of salvage therapy, or development of distant metastatic disease. In the analysis, regional nodal status assessed by [⁶⁸ Ga]Ga-PSMA-11 PET-CT was prognostic for medium-term oncologic outcomes (for freedom from treatment failure: HR 2.1, p = 0.01), outperforming conventional imaging in which nodal status was not prognostic.

These data are another example of the clinical high value of PSMA-targeted PET-imaging in different stages of prostate cancer. However, the symphony of theranostic PET tracers for this disease does not end with PSMA. A different opportunity, for example, for patients without PSMA expression, is the use of gastrin-releasing peptide receptor (GRPR)-targeting molecular imaging. Duan et al. presented the final results from a prospective phase II/III imaging trial in which they investigated PET-MRI using the GRPR-antagonist [⁶⁸ Ga]Ga-RM2 in patients with biochemical recurrence of prostate cancer [3]. One hundred patients with rising PSA levels after prostatectomy or radiotherapy and non-contributory conventional imaging (negative CT, MRI, and/or bone scan) were included and the diagnostic performance of [⁶⁸ Ga]Ga-RM2 PET-MRI was compared to that of MRI alone. They found a better diagnostic performance of [⁶⁸ Ga]Ga-RM2 PET-MRI with significantly higher detection rates, sensitivity, and accuracy, indicating the highly promising characteristics of GRPR-targeted PET-MRI. Figure 2 shows detection rates of [⁶⁸ Ga]Ga-RM2 PET/MRI in comparison to MRI alone for different serum PSA level groups and image examples of a patient with a subcentimeter, non-specific lymph node, which was ruled negative on MRI, but positive on [⁶⁸ Ga]Ga-RM2 PET-MRI and subsequently revealed to be a lymph node metastasis.

Fig. 2

Top: The detection rates of [⁶⁸ Ga]Ga-RM2 PET-MRI were greater than MRI across all PSA ranges; the figure was reprinted with permission from the final publication resulting from the presented study [3]. Bottom: Image example of a 65-year-old man with biochemically recurrent prostate cancer. PSA at the time of PET-MRI was 7.9 ng/mL. ⁶⁸ Ga-RM2 PET-MRI (axial PET (A); axial fused PET-MRI (B); maximum intensity projection [MIP] (D), respectively) shows a positive lymph node interaortocaval (red arrows). MRI (axial MRI (C)) shows a subcentimeter, non-specific lymph node, which was ruled negative. Subsequent lymph dissection revealed lymph node metastasis compatible with prostate cancer (courtesy of Duan et al.) [2, 3]

Oncology—new development for therapy and theranostics

Other abstracts presented in the EANM’23 Highlight lecture were focused on tumour diseases that have, in recent years, received less attention in nuclear medicine theranostics. In some of the abstracts, molecular imaging using novel tracers was applied for the stratification of other targeted therapies than radionuclide therapies.

One of those abstracts was the one by Thiec et al., in which the authors describe the application of [⁶⁴Cu]Cu-ATSM PET-CT as a response predictor to neoadjuvant therapy in patients with locally advanced rectal cancer. Preliminary results from 24 patients with T3 or T4 rectal adenocarcinoma with or without regional lymph node involvement enrolled in a multicenter pilot study were presented. All patients underwent baseline [⁶⁴Cu]Cu-ATSM and [¹⁸F]FDG PET-CT, and most patients underwent additional preoperative [¹⁸F]FDG PET-CT. SUV_max, SUV_mean, tumour-to-muscle-ratio, hypoxic tumour volume (HTV), and hypoxic burden (HB: HTV x SUV_mean) were derived from [⁶⁴Cu]Cu-ATSM and ΔMTV and ΔTLG from [¹⁸F]FDG PET-CT images in rectal tumour VOIs. The authors found that a larger HTV correlated with a poorer metabolic response (ΔTLG). These findings are a first step that holds promise towards future clinical applications of this novel theranostic PET tracer.

Even the great classic of [¹⁸F]FDG PET was not neglected in the EANM’23 highlights as a potential predictor of therapy response. Seban et al. used pre-therapeutic [¹⁸F]FDG PET to predict complete response to neoadjuvant chemotherapy with or without addition of the anti-PD1 immune checkpoint inhibitor pembrolizumab in patients with early triple-negative breast cancer. They found that the basic [¹⁸F]FDG PET markers total MTV and tumour SUV_max were strong predictors of complete response after neoadjuvant chemotherapy regardless of the addition of immunotherapy. Moreover, thyroid uptake in [¹⁸F]FDG PET (indicated by thyroid SUV_max) was a potential predictor of immune checkpoint inhibitor-induced hypothyroidism after neoadjuvant immunotherapy.

Other studies presented at the EANM’23 highlights combined [¹⁸F]FDG PET with novel tracers to derive clinically valuable parameters. Chen et al. used dual-tracer PET ([¹¹C]C-acetate and [¹⁸F]FDG PET) to derive molecular imaging biomarkers in patients with advanced hepatocellular cancer. They described results from a prospective trial including 75 patients who were randomised into two groups treated with lenvatinib or nivolumab and underwent dual-tracer follow-up PET-CT. The authors found, that within the lenvatinib group, patients with [¹¹C]C-acetate-avid tumours had a significantly longer progression-free survival (8.0 versus 4.0 months), whereas within the nivolumab group, this was the case for patients with [¹⁸F]FDG-avid tumours (9.0 versus 3.0 months). Therefore, the authors proposed dual-tracer PET as a metabolic differentiator for a molecular imaging-guided therapy regime to treat [¹¹C]C-acetate-avid advanced-stage hepatocellular cancer with lenvatinib and [¹⁸F]FDG-avid tumours with nivolumab. This study is another example of defining parameters from molecular imaging to select patients for stratified non-radionuclide therapies (i.e., variation on the theranostic theme).

Dosimetry and radiobiology were the focus of several other abstracts. Tronchin et al. modelled the effect of daughter migration on dosimetry estimates for ²²⁵Ac in targeted alpha therapy. This is very valuable and timely information given the concern that the decay energy of ²²⁵Ac can break the bond to the targeting vehicle, thereby releasing free daughter radionuclides in the body. The aim of the researchers here was to develop a multilevel compartment model that includes individual biokinetics for the daughter isotopes. They showed that, when compared to the situation if the daughter isotopes decayed at the target site, [²²⁵Ac]-DOTATATE is internalised in tumour cells, due to migration, kidneys receive a 12% increase in absorbed dose, while tumour dose is decreased by a factor 5. Regarding radiobiology, Tamborino et al. provided scientific proof of the relative biological effectiveness of peptide receptor radionuclide therapy (PRRT) compared to external beam radiotherapy (EBRT) in neuroendocrine tumours (NET). The researchers developed a model that provides realistic cellular dosimetry calculations from a cluster-forming scenario. Results show that cellular localization of the receptor, cluster size, and uptake heterogeneity (caused by cluster formation) are the main sources of uncertainty for absorbed dose calculations. Taking this into account, the researchers show that PRRT requires an absorbed dose three times higher than EBRT to achieve an equivalent effect. Theranostics was obviously one of the leading actors of EANM’23 both in educational and scientific sessions. Handula et al. discussed the need to improve the bioavailability of the radioligand to achieve higher tumour uptake (and consequently enhance therapeutic efficacy) in patients candidates for radionuclide therapy of NETs. They reported the development and evaluation of SSTR2-antagonists bearing palmitic fatty acid to improve the pharmacokinetic properties of DOTA-JR11. They demonstrated that the new DOTA-JR11 analogues were successfully synthesised and radiolabeled with ¹⁷⁷Lu exhibiting good binding affinity to SSTR2 and showing acceptable cell uptake. Next Generation theranostics based on the Tetrazine ligation was the target of the abstract by Battisti et al. They presented a new radiolabeling technology, which provides just a single isomeric product from the tetrazine ligation within 20 min, without the need for any purification. They demonstrated that the tetrazine ligation can be used to label ¹⁸F/²¹¹At to any targeting vector using this platform, producing ready-to-inject radiopharmaceuticals within minutes without the need for purification. Accordingly, this methodology holds the promise of labelling any targeting vector with ¹⁸F/²¹¹At.

Last but not least, Hofman et al. presented the results of Lutectomy, a phase 1/2 study evaluating dosimetry, safety, and potential benefit of pre-surgery [¹⁷⁷Lu]Lu-PSMA-617 therapy in patients with high-risk localised prostate cancer. The primary objective of the study was to measure tumour radiation absorbed dose. Patients with hormone-refractory prostate cancer (HRCaP) (PSA > 20 ng/mL, ISUP grade group (GG) 3–5, ≥ cT2c or N1), high tumour uptake (SUV_max ≥ 20) on [⁶⁸ Ga]Ga-PSMA-11 PET-CT, and scheduled for radical prostatectomy (RP) were eligible and were divided into two cohorts. Cohort A (n = 10) received 1 cycle of 5 GBq [¹⁷⁷Lu]Lu-PSMA, and Cohort B (n = 10) received 2 cycles of 5 GBq [¹⁷⁷Lu]Lu-PSMA followed by RP scheduled 6 weeks after treatment completion. Radiation dose was estimated using post-therapy SPECT/CT at 4, 24, and 96 h using a voxel-based 3-phase exponential clearance model with partial-volume correction. The study suggests that [¹⁷⁷Lu]Lu-PSMA before RP in men with HRCaP delivered high radiation doses with few toxicities and without compromising surgical safety. Significant PSA reductions were achieved and authors concluded that [¹⁷⁷Lu]Lu-PSMA in this setting warrants further evaluation.

Neurosciences—new development towards the disease modification era

The role of molecular imaging biomarkers to guide new disease-modifying treatments or new targets for the treatment of neurodegenerative disease was the focus of some of the best-rated abstracts for neurology. Peretti et al. investigated in Alzheimer’s disease (AD) patients the association between plasma glial fibrillary acidic protein with neurofibrillary tau tangles in the brain and cognitive decline independent of amyloid-B. A cohort of 123 subjects underwent amyloid- and tau-PET, blood collection, and mini-mental state examination (MMSE). Voxel-based linear regression was performed to assess the topographical correlation between tau SUVR and a plasma biomarker for neuroinflammation of the glial acidic fibrillary protein (GFAP). GFAP was associated with increased tau-PET uptake in the lateral temporal and inferior frontal lobes, independently of amyloid, gender, age, education, and MMSE scores (Fig. 3). Elevated plasma GFAP levels were also associated with accelerated cognitive decline, independently from tau and amyloid load. These results support neuroinflammation and astrogliosis as a relevant contributor to AD pathology, which can be monitored in blood, and suggest neuroinflammation as a potential target for future disease-modifying therapeutic trials targeting tau pathology. Neuroinflammation as a target for diagnostic or prognostic assessment and potentially for future therapeutic approaches was the topic of an increasing number of presentations during EANM’23. Zatcepin et al. explored the existence and accessibility of a microglia connectome in both mouse models and humans with AD and performed a translational assessment of microglial desynchronization as a potential AD biomarker. They depleted the microglia using CSF1R inhibition in wild-type mice and examined whether interregional correlation coefficients (ICC) of 18 kDa translocator protein (TSPO)-PET with [¹⁸F]GE-180 were affected by the lack of microglia. The link between dysfunctional microglia (TREM2-/-) and AD pathophysiology (β-amyloidosis, tauopathy) based on TSPO-PET was investigated in mouse brain, and then extended the methodology to analyse a human TSPO-PET dataset comprising of cognitively normal individuals and patients with prodromal and dementia stages of AD. Microglia-depleted mice demonstrated a dramatic reduction of TSPO-PET ICCs among all compartments of the brain. The analyses also revealed substantial alterations in TSPO-PET ICC in all the investigated AD models when compared to the age-matched controls. Patients with AD also showed a stage-dependent reduction of both the number of significant connections and microglia DI in six parietal and four temporal regions. An important novel tracer development potentially useful also in the field of neurodegenerative diseases was presented by Schou et al. who targeted imaging of PARP1 in the CNS. PARP1 is a key enzyme in DNA damage repair implicated in the pathophysiology of cancer and neurodegenerative disorders. These researchers show the results of the tracer development cascade. Subsequent PET studies in healthy volunteers demonstrated a crossing of the blood–brain barrier and high-affinity binding in the brain.

Fig. 3

Elevated glial acidic fibrillary protein (GFAP), a plasma biomarker for neuroinflammation, is associated with increased tau deposition in lateral temporal and frontal regions. Significant clusters are superimposed to an MRI template (courtesy from Peretti et al.) [2]

The use of new tracers outside the field of neurodegenerative diseases was also the focus of several scientific contributions. Xiao et al. aimed to set up and assess the value of quantitative PET (QPET) analysis to complement [¹⁸F]F-SynVesT-1 and [¹⁸F]F-FDG PET for the identification of epilettogenic zone in MRI-negative children with focal cortical dysplasia (FCD). [¹⁸F]F-SynVesT-1 is a radioligand for PET Imaging of Synaptic Vesicle Glycoprotein 2A. This study used a multivariate prediction model incorporating EEG, [¹⁸F]F-SynVesT-1, and [¹⁸F]F-FDG PET to enhance the precision of diagnostic tests designed to identify the location of EZ. QPET analysis confirmed a great potential for refining the precision of diagnostic tests utilised for EZ localization even when only [¹⁸F]F-FDG PET was used. Indeed, [¹⁸F]F-FDG PET confirmed to be a valuable tool in brain PET imaging in many presented studies also based on the large number of scans which can be made available to more sophisticated post-processing approaches. Ha et al. aimed to investigate changes in intrinsic brain networks during normal ageing. A total of 827 [¹⁸F]F-FDG PET-CT scans obtained from 357 healthy individuals were analysed. Independent component analysis (ICA) with a threshold of Z > 2 was used to generate masks of 21 intrinsic connectivity networks (ICN), and it was demonstrated that during normal ageing, intrinsic brain networks show a continuous decline in neural activity. The connectivity strength between functionally connected ICN-ICN was also gradually weakened. Notably, this ageing-related weakening of connectivity was demonstrated to accelerate during young adulthood, slow down in middle age, and accelerate again during old age.

Cardiovascular—new clinical development

In the last decades, myocardial perfusion scintigraphy has developed into a well-established technique, and the field has moved in recent years towards the need for absolute quantification of myocardial perfusion with either CZT SPECT or PET-CT. This has led to the interest in potential new PET myocardial perfusion tracers. [¹⁸F]F-flurpiridaz (flurpiridaz) is such a novel PET myocardial perfusion imaging tracer and has been the subject of clinical trials over the last decade. Knuuti et al. studied the diagnostic efficacy and safety of flurpiridaz PET for the detection and evaluation of coronary artery disease (CAD) defined as > 50% stenosis by quantitative invasive coronary angiography (ICA) in a prospective international multicenter clinical trial. A total of 578 patients (age, 63.7 + 9.5 years) with suspected CAD from 48 clinical sites in the USA, Canada, and Europe were evaluable for the study. Patients underwent flurpiridaz PET and rest-stress 99mTc-labelled SPECT before ICA. Flurpiridaz showed significantly higher sensitivity and specificity compared to SPECT in the overall population (respectively 80.3% and 63.8% for PET vs 63.8% and 61.7% for SPECT (p < 0.0004), as well as in women, in patients with BMI ≥ 30 kg/m², and in diabetics. ROC-area under the curves for PET was significantly higher than SPECT in the overall population (0.80 vs 0.68, p < 0.0001), in women, and in patients with BMI ≥ 30 kg/m². Flurpiridaz was also superior to SPECT for the assessment of defect extent/severity (p < 0.0001). Image quality (% excellent or good) was significantly better than SPECT for rest images, pharmacological stress images, and treadmill exercise images. Radiation exposure was less for flurpiridaz than SPECT. The authors therefore concluded that flurpiridaz has promise as a new tracer for detection and evaluation of CAD not only in the general population but particularly also in women and patients with BMI ≥ 30 kg/m². Not only perfusion imaging has gained new insights but especially assessment of cardiovascular inflammation in patients with arteriosclerosis, both in the stable and acute phase of the disease, has stimulated the field. These data have shown that the residual cardiovascular risk after initial treatment is associated with inflammatory activity. In this light, the results of the Phase II LOMI Trial are of interest. Kosmala et al. assessed the cardiac lymphatic system as a marker of inflammatory response in patients after acute myocardial infarction by using C-X-C motif chemokine receptor 4 (CXCR4)-directed PET-CT using [⁶⁸ Ga]Ga-PentixaFor. The preliminary results of this study showed that, in 20 patients after acute myocardial infarction, CXCR4-directed imaging provided information on the extent of local and systemic inflammation, but more importantly, the immune response also identified subjects at risk for cardiac functional decline. Although preliminary, these data may allow for earlier and potentially more appropriate identification of patients in whom inflammatory-driven on-set of ischemic-based heart failure may be prevented.

Arteriosclerosis is primarily an inflammation-driven disease that evolves from initial inflammation to calcified plaques. Like inflammation, arterial calcification burden is a strong predictor of cardiovascular disease events, but there are several factors that may contribute to the calcification. This implies that knowledge of potential influencing factors may give rise to new treatment strategies. Bellinge et al. looked at the potential contribution of hypoxia within the development of arterial calcification using [¹⁸F]F-Fluoromisonidazole ([¹⁸F]F-FMISO) PET-CT in 41 patients (mean age was 64 ± 9 and 85% were males). The authors established that thoracic aortic [¹⁸F]F-FMISO wall activity is associated with arterial calcification burden. However, the relationship was dependent on the method of quantification. This means that further studies are required to better understand [¹⁸F]F-FMISO as a measure of hypoxia of the arterial wall and its role as a potential new treatment target.

The knowledge and thereby the awareness of transthyretin cardiac amyloidosis (TTR-CA) as assessed with basic nuclear medicine techniques has increased in the last decade. These data do not only illustrate the incidence of disease and potential targets for treatment but are also focussed on disease outcomes. Prognostic markers for outcome are important as they may help to identify patients at the highest risk for events and thereby may even guide treatment strategies. Zhao et al. studied the prognostic value of right ventricular (RV) [^99mTc]Tc-DPD tracer uptake in patients with TTR-CA by using quantitative SPECT-CT. In 53 patients with scintigraphy-confirmed TTR-CA based on Perugini’s uptake (grades 2–3) myocardial, SUVmax was normalised to bone activity (nSUVmax) or blood pool (wSUVmax), respectively. The primary outcome was the occurrence of major adverse cardiac events (MACEs). RV uptake was an independent predictor of MACEs in a univariable Cox analysis (HR:1.12, 95% CI, 1.02–1.22; p < 0.05). Therefore, RV uptake appears to represent a more advanced stage of disease burden in TTR-CA and could serve as a new prognostic marker in TTR-CA patients with worse outcomes. Interestingly these data are in line with RV [¹⁸F]F-FDG uptake in patients with cardiac sarcoid.

Development for new technological trends

Last but not least, the symphony of the latest scientific trends in nuclear medicine at the EANM’23 encompassed technological trends, which included both “classic” topics such as medical physics applications and novel radionuclides, as well as novel and “pop culture” images (post)processing using artificial intelligence. A contribution that was presented in the Highlight Lecture was the investigation by Brühlmann et al. into cyclotron-based production of Lanthanum-133 (¹³³La). ¹³³La is an attractive β⁺-emitter with similar chemical properties to ²²⁵Ac which may be used in theranostic applications. The authors proposed an optimization of the production method of ¹³³La which can yield high radionuclidic purity for theranostic purposes. The cyclotron-based production with solid targets may be scaled up to radioactivity amounts that are needed for clinical applications by increasing the irradiation time, thus, providing a prerequisite for future clinical applications. Other highlighted abstracts described new trends in radioembolization and radioguided surgery. Dietze et al. studied, in patients treated with hepatic radioembolization, whether the merging of the pre-treatment procedure and treatment into a single session would reduce the number of interventions from two separate interventions to one. A compact and mobile hybrid c-arm scanner able to perform simultaneous fluoroscopic x-ray, planar scintigraphic, and SPECT/CBCT imaging using a novel dual-layer detector was developed and tested. This allows for images to be acquired during the injection of [^99mTc]Tc-MAA to dynamically study particle distribution. This study showed the safety and feasibility of the procedure and similar dosimetric quality compared to the conventional clinical scanner outside the intervention theatre. Real-time x-ray/nuclear guidance and dosimetry during radioembolization interventions may in the future allow for single-session radioembolization treatments, important to the individual patient as well as for the pressure on our healthcare system. Translation of a drop-in beta-emission detection probe for robotic radioguided surgery (RGS) in several cancers is also a hot topic. Collamti et al. proposed an innovative approach to radioguided surgery, exploiting the intraoperative localization of beta-particle-emitting tracers (e.g. [^18F]FDG, [⁶⁸ Ga]Ga-PSMA, or [⁹⁰Y]Y-DOTATOC). They presented the design and clinical translation of a drop-in beta-emission detection probe for robotic guided surgery and they underscored the validity of this form of RGS, thereby providing a basis for the more widespread implementation of PET-tracers during radioguided robotic surgery.

From the extensive opus on the application of artificial intelligence, two technological approaches were presented during the EANM’23 Highlights ceremony. Xue et al. developed a wavelet-based deep-learning method to recover high-quality images from ultra-low-dose PET. For this purpose, they used total-body PET images of 550 patients that were acquired on total-body PET scanners from different manufacturers. Their novel approach was to perform denoising on the wavelet-decomposed high-frequency component of PET imaging, where the noise is mainly concentrated, in contrast to conventional deep-learning methods that are applied on the image domain and were used for comparison in the highlighted study. The novel technique outperformed the conventional deep-learning method and also the non-enhanced images. Fast computation and well-preserved image details may be the basis of future applications of deep-learning-based dose reductions in clinical PET imaging.

The second highlighted approach was by Spielvogel et al. from hometown Vienna who presented the development and multicenter validation of an artificial intelligence system to detect cardiac amyloidosis on scintigraphy images acquired with ^99mTc-labelled bone-seeking tracers in a standardised and accurate way. 19,201 scans from 16,059 patients originating from different centres using various scanners and tracers were included in the study. The artificial intelligence was developed using data from a single centre (Vienna) and validated on the remaining eight centres to detect the typical pattern of cardiac amyloidosis (i.e. a Perugini grade ≥ 2). Figure 4 provides an overview of the study design. The automatic approach allowed highly accurate and reliable prediction of cardiac amyloidosis with diagnostic performances comparable with nuclear medicine physicians. Moreover, the results were highly prognostic for survival and development of heart failure, thus providing a standardised and fast approach for reliably detecting cardiac amyloidosis which may in future be applied in clinical routine.

Fig. 4

Study design of an international multicenter trial to develop and validate an artificial intelligence system for detection of cardiac amyloidosis on scintigraphy images acquired with ^99mTc-labelled bone-seeking tracers (courtesy of Spielvogel et al.). The figure was created with BioRender.com and is used in the final publication resulting from the presented study [2, 4]