The history of nuclear medicine (and molecular imaging)

Nuclear medicine first became recognised as a potential medical speciality in the early 1950s of the last century when Seidlin reported on the success of radioactive iodine (I-131) in treating a patient with advanced thyroid cancer [1, 2]. The use of the rectilinear scanner contributed to the widespread clinical use of diagnostic nuclear medicine and resulted in planar, somewhat spotty images which constituted of plotted dots [3]. Importantly, this methodology already acted on the tracer principle, i.e. the use of a radioactive compound (radiopharmaceutical) to visualise in vivo processes and to reveal physiological and pathophysiological changes. In those days, nuclear medicine often served as an alternative approach or an add-on tool to the radiological morphological imaging techniques [4]. Demonstrating, for example, a blood–brain barrier disruption using brain diethylenetriaminepentaacetic acid (DTPA) scans as a sign for a possible brain tumour was certainly — at the time — less invasive when compared to the radiological pneumoventriculography, which used air injected in the ventricular system as an indirect contrast for adjacent brain tumours. 99mTc colloid liver scans, for example, added to the detection of hepatic defects in case of primary liver tumours or liver metastases [5]. Due to the clinical breakthroughs over the following decades in radiology, in computed tomography (CT) and in magnetic resonance imaging (MRI) with increasing resolution and accuracy, the clinical dissemination of these imaging techniques was rapid. In addition, this development resulted in a reduction of the nuclear medicine scans of the aforementioned type. In parallel, nuclear medicine itself as an imaging field saw the development of single-photon emission computed tomography (SPECT) [6] and positron emission tomography (PET) [7, 8] imaging modalities, benefitting from the ever-increasing computation power and continuous improvements in hardware detection systems, such as new crystals and digital detectors [6, 9, 10]. The clinical deployment of SPECT was boosted by the availability and development of 99mTc generators and pharmaceutical kits [11], allowing more efficient and cost-effective in-house institutional radiopharmaceutical labelling. Similarly, clinical PET imaging disseminated rapidly in the hospital environment with the breakthrough and growing availability of fluorodeoxyglucose (18F-FDG) [12] for oncological indications. Due to fast and accurate computational capacity, nuclear medicine was able to make quantum leaps, made possible by, for example, iterative image reconstruction rather than filtered back projection, combined with corrections for time of flight, depth of interaction and movement [7, 13]. As a next step, the nuclear clinical imaging systems converged with radiological imaging into hybrid systems, as SPECT-CT, PET-CT or PET-MRI, adding the advantages of radiological systems to the nuclear imaging techniques, with the latest addition being total body PET-CT [14]. Today, developments are continuously pushing the existing frontiers, including, for example, artificial intelligence, nanotechnology and secured giant cloud-based data servers for global shared data warehouses combining imaging data with clinical, histopathological and patient outcome datasets.

The rise of molecular imaging

Molecular imaging, developed by ‘the father of molecular imaging’, the late Dr. Sanjiv Gambhir, is defined as ‘the visualization, characterization, and measurement of biological processes at the molecular and cellular levels in humans and other living systems’ [15]. Molecular imaging typically includes two- or three-dimension imaging as well as quantification over time [15] and differs from other more traditional imaging methods in that tracers are used to image molecular targets and pathways occurring within an area of interest, in contrast to imaging that differentiates qualities, such as density (CT), reflectance (ultrasound) or water content/properties (MRI). The pioneering role of PET in the field of molecular imaging relied heavily on its inherent strengths, which include a larger range of radiopharmaceuticals, moreover, introducing radionuclides of biological atoms, higher sensitivity and resolution of the detection system and more accurate quantification as compared to SPECT imaging. The early clinical applications of PET emerged in the 1980s, especially in the field of neurology, followed by cardiology and later on in oncology in the 1990s [16, 17]. In neurology, PET was initially used to characterise cerebral perfusion and assess brain tumours non-invasively [18,19,20]. In cardiology, the development of [O-15]-H2O, a radioactive variation of regular water, led to the clinical and currently gold standard application of quantitative evaluation of myocardial blood flow using PET [21]. Additionally, PET became more widely accepted as a clinical application due to its role in oncology for staging and treatment monitoring [22].

While the nineteenth century in medicine was dominated by the focus on infectious diseases, in high-income countries in the last century, healthcare focused especially on oncology as detailed, for instance, in the Pulitzer-winning book The Emperor of All Maladies: A Biography of Cancer written by S. Mukherjee [23]. The discoveries in tumour biology and detailed insights into oncological cellular pathways, as illustrated in the seminal Hallmark of Cancer article by Hanahan and Weinberg in 2011 [24], shifted the focus towards precision medicine and the need for molecular imaging, providing an additional boost for the domain of nuclear medicine.

FDG-PET emerged as a game changer in expanding the scope of PET imaging, especially in clinical oncology, and its appearance was heralded in a landmark article by the late Henry Wagner in 1991: ‘Clinical PET, its time has come’ [25]. Presently, a broad spectrum of PET tracers are being developed to visualise molecular biomarkers and used in clinical or research settings in many clinical domains, resulting in novel avenues, for example, immuno-PET (e.g. 89Zr-immuno-PET) to measure target engagement of therapeutic antibodies and tailoring of therapy with monoclonal antibodies [26] or amyloid PET (e.g. 11C-Pittsburgh Compound-B PET) to detect amyloid depositions in the brain [27]. Other mostly radiological molecular imaging tools, such as MR spectroscopy, hyperpolarized MRI or MR chemical exchange saturation transfer (CEST) imaging, may be considered smaller players in terms of the scope of future applications and still not close to becoming clinical routine.

Finally, iodine radionuclides and related diagnostics and therapy demonstrated from the early start of nuclear medicine the potential of radionuclide therapies and theranostics (i.e. diagnostic combined with therapeutic capacity) [28]. Theranostics, nowadays, is a rapidly growing field in clinical nuclear medicine using, for example, alpha emitters such as 213Bi-DOTATOC receptor-targeted alpha-radionuclide therapy for neuroendocrine tumours [29] and prostate-specific membrane antigen (PSMA)-directed theranostics in prostate cancer [30]. All the assets of nuclear medicine and the era of targeted (molecular) medicine proved to be a match made in heaven and contributed significantly to the rapid worldwide dissemination of nuclear medicine, initially in many high-income countries but presently on a global scale.

In 2002, the European Journal of Nuclear Medicine (EJNM) changed its branding into the European Journal of Nuclear Medicine and Molecular Imaging (EJNMMI), claiming the representation and embodiment of the unique selling point proposition of molecular imaging as ‘molecular imaging has its roots in nuclear medicine and in many ways is a direct extension of our existing discipline’ [31]. In 2012, the Society of Nuclear Medicine recognised the growing diversity in the field by adding ‘molecular imaging’ to its title [32]. At preclinical conferences, especially such as the European Society for Molecular Imaging (ESMI) and similar societies, this embodiment started to happen on a smaller scale and continued within the EANM as reflected by the initiation of a committee on Translational Molecular Imaging and Therapy in 2007.

Nuclear medicine and molecular imaging in Groningen as an illustration of history

The aforementioned overview of a global history of molecular imaging was also reflected in the local developments in Groningen, the Netherlands. Groningen had pioneered PET since the early 1970s. In 2005, a new department was constituted in Groningen, called ‘the Department of Nuclear Medicine and Molecular Imaging’, arguably one of the first nuclear medicine departments to adopt the molecular imaging label. This new department emerged from the fusion of the previous clinical department of conventional nuclear medicine and the PET research facility. Given the need for collaboration within a single group, which united the strengths of one speciality within a single centre while performing the four classical university hospital tasks of patient care, research, teaching and training, a new name was needed. This new name symbolised a shared future while giving credit to feelings, achievements and assets of the previous entities. ‘’Nuclear medicine’’ needed to stay in the title, as it reflected the name used for the speciality and was internationally recognised by authorities, patients and referring physicians. Hence, the PET entity was represented by adding ‘’molecular imaging’’, reflecting the continuous effort at the molecular forefront of its radiochemistry to translate growing insights in molecular pathways at the benefit of in vivo imaging. This was especially demonstrated by the immuno-PET work driven by medical oncology, in collaboration with pharmacists and nuclear medicine collaborators, presenting, for example, results from a first-in-human study assessing the feasibility of imaging with 89Zr-labeled atezolizumab (anti-PD-L1) to predict response to PD-L1 blockade in several tumour types [33]. After the configuration of a joint medical imaging centre, consisting of the Department of Nuclear Medicine and Molecular Imaging and the Department of Radiology, the ‘TracerLab’ in Groningen was founded as a synergistic initiative between these departments, together with the Stratingh Institute for Chemistry from the Faculty of Sciences and Engineering. The TracerLab aims at developing and applying innovative chemistry for molecular imaging in general, including PET, MRI and, last but not least, optical imaging. Multidisciplinary collaboration further allows to translate the imaging tracers in the chain from bench to bedside.

The rise of optical molecular imaging

Driven by the need for clinical answers in the new era of precision medicine, the development of molecular imaging tools beyond PET and SPECT is rapidly expanding in terms of both targets and accessibility. The first intraoperative fluorescence optical imaging procedure was carried out in 1948 when intravenous fluorescein was used to visualise intracranial neoplasms during neurosurgery [34]. Efforts to increase specificity led to the first-in-human study assessing the feasibility of intra-operative tumour-targeted fluorescence molecular imaging real-time during surgery in 2011, where folate conjugated to fluorescein isothiocyanate was used as a targeted optical imaging agent to detect ovarian cancer and possible peritoneal metastases [35]. As put forward in this editorial and related review, optical imaging is no longer restricted to preclinical studies and is already translated into routine clinical care by both surgery- and endoscopy-focused groups in phase I–III clinical trials. Illustrating the rapid rise of optical imaging research until now, the review in this same issue of the journal had to use a metanarrative approach including artificial intelligence to filter more than 6000 presently available studies, to finally address 58 publications on clinical studies relevant within the scope of the review.

Similar to the interest of medical oncologists in PET-CT and the development and validation of novel radiopharmaceuticals, surgical oncologists and gastroenterologists, amongst other clinicians such as ophthalmologists, are particularly keen on optical imaging as an additional method to address their specific challenges in diagnosis before and during surgical or endoscopic treatment. In Groningen, for example, the strength of optical molecular imaging was translated into preclinical and clinical work by surgeons and gastroenterologists, in close collaboration with disciplines like pathology and hospital pharmacy, amongst others [35,36,37,38]. Due to the rapid deployment of their research and clinical translation into routine daily clinical care, these are now in need of a partner to embrace the clinical support, development of optical tracers, standardisation, validation and thus methodological development, also towards regulatory agencies like the FDA and EMA. Consequently, in Groningen, the aforementioned separate activities and experts in the field of molecular imaging are now functionally merging for more sustainability, innovative synergism and economies of scale towards this ambition.

While developing this innovative yet challenging imaging field, optical-imaging-adopting clinicians need partners who can support them with a proven platform and longstanding experience on how to adopt a sound and reliable methodological development in clinical adaptation [39]. Recent articles in Lancet Oncology demonstrate the socioeconomic need for a concerted effort of imaging and oncologists on a global scale [40, 41]. Optical imaging may benefit from classical imaging stakeholders to represent and propagate its methodology towards regulators and reimbursement policies. As an experienced multidisciplinary speciality within the same field of molecular imaging, nuclear medicine is well suited to help address these needs, having gone through the same ordeal. The beating heart of nuclear medicine in terms of ongoing innovation is radiochemistry, in essence providing biologically active molecules labelled with a radionuclide, while in optical imaging, the label is a fluorophore, creating the common ground for merging specialities. Radiochemists, trained to develop labelling methods for complex bioactive molecules, are instrumental in creating molecular avenues for optical tracer developments. Radiopharmacists have the expertise to take this role as imaging modality-spanning pharmacists to implement in-house production of optical tracers. Medical physics, as the backbone of nuclear medicine, has a longstanding experience and focus on quantification and standardisation, which still presents a major challenge in the upcoming field of optical imaging. Finally, nuclear medicine physicians are used to thinking and discussing with referring physicians in terms of molecular pathway. Along the same lines, biologists in our preclinical facilities may help accommodate the translation to their colleagues in surgery, gastroenterology and other fields to come. As a result, the above provides the instant glue between optical and nuclear medicine, a bond to be cherished.

Similarities, differences and complementarity between optical imaging and nuclear medicine and molecular imaging

Clinical

While PET imaging allows for a standardised, non-invasive and highly sensitive full-body scan, optical imaging provides real-time, wide-field imaging with high spatial resolution (Fig. 1). Although the sensitivity and specificity of the available nuclear imaging modalities as SPECT-CT and PET-CT are superior to those of any other current clinical imaging modality, optical imaging has the potential to visualise biomarkers at the cellular and molecular level in a real-life environment without the need for ionising radiation and its inherent risks. It thereby facilitates the improvement of early detection and staging of disease in a multitude of imaging sessions. While PET is primarily used preoperatively for, for example, diagnostic or therapy response purposes, optical imaging is mainly used as a supportive imaging tool before and during interventional procedures. In optical imaging, the imaging data flows back and forth in real-time, i.e. it is experienced by the operator while performing a procedure or within minutes of imaging, for example, a specimen on the back-table. In contrast, in PET imaging, the information is obtained during the scanning procedure and not used in a real-time setting in the operating theatre or during an endoscopic procedure. Optical imaging has already shown to increase surgical efficacy in head and neck cancer [42], breast cancer [43] and colorectal cancer [44] and to improve (early) disease detection by endoscopy in, for example, oesophageal cancer [38] by using fluorescent labels linked to compounds that target disease-specific biomarkers. Furthermore, it has been applied into the field of infectious disease [45,46,47] and cardiovascular medicine to detect bacteria in implants and vulnerable atherosclerotic plaques [48,49,50]. Optical imaging has also been employed for decades to obtain detailed images from within the eye to assess retinal perfusion with fluorescein or indocyanine green (ICG) by using a dedicated split lamp and more recently to assess the retinal layers using optical coherence tomography (OCT), a cross-sectional tomographic imaging technique that measures light back-scattered or back-reflected within the tissue [51]. Finally, there is a crucial role for ‘closed-field’ ex vivo imaging during optical guided procedures, where the surgical specimen is imaged on the back-table in the operating room, or in the pathology examination room, within a black-box, thereby allowing for pathological assessment of the excised tissue immediately after resection, with greater control over optical parameters such as illumination intensity, field homogeneity and geometry, camera distance and the lack of interference of ambient background light [52]. Closed-field imaging will enable more accurate and standardised operating procedures for the visualization of excised tissue and thereby provides a more reliable and accurate comparison of imaging data obtained by the observers, individual centres and study results, which is of need for regulatory approval. These preliminary clinical results show the potential of optical imaging to guide interventions and improve (early) disease detection in real-time. Furthermore, in vivo fluorescence imaging can be used to predict therapeutic outcomes, as shown for cancer and Crohn’s disease [53, 55, 56].

Fig. 1
figure 1

The advantages (green) and drawbacks (red) of human PET imaging (left) and optical imaging (right)

Full size image

Optical imaging is also well-suited for theranostic purposes since it can facilitate the detection of disease, determine the target location and dosimetry and monitor therapeutic effects in situ [53]. Exemplary is a self-reporting and self-adapting theranostic nanoparticle based on poly(lactic-co-glycolic acid) designed to encapsulate camptothecin as an anticancer drug and a caspase-3 activatable fluorescent peptide [54]. The fluorescence is turned off due to the fluorescence resonance energy transfer (FRET) effect and can be turned on upon reaction with caspase-3, an important protease in apoptosis, to achieve visualised therapeutic reporting for precise tumour treatment [54]. Of special interest is also the potential of a theranostic combination of optical imaging and photopharmacology, an emerging field in which light is used to activate drugs selectively and reversibly in real-time [57].

Chemistry

A fluorescent tracer’s design must consider its pharmacological properties, toxicity profile, chemical/metabolic stability and in vivo imaging performance. In general, the chemical challenges in nuclear and optical imaging are similar; while introducing the imaging label, maintaining the functionality and favourable pharmacokinetics of the targeting moiety is vital [58]. In the earlier evolutionary stages of nuclear molecular imaging, radiochemists have developed valuable expertise in this respect. However, it must be noted that there is a key structural difference between the labels used in nuclear and optical imaging: while radionuclides are often typically single atoms (11C, 18F) already present in many chemical structures, and their introduction to the targeting molecule can therefore be often achieved in a near-seamless way, the fluorophores used in optical imaging are typically large, planar, lipophilic molecules, whose incorporation into the targeting molecule has the potential to change its properties significantly. A limitation of PET imaging is that it relies solely on the accumulation of tracers, and the signal in this modality is continuously detected and cannot be activated, e.g. through the catalytic activation of an imaging probe by a specific enzyme resulting in the dequenching of the fluorescence signal. Conversely, in optical imaging, this can be achieved when a quenched fluorophore conjugated on a cleavable backbone encounters a molecular target, such as a proteolytic enzyme with a specific cleavage site activity [59, 60] for the backbone, so the quenched fluorophore becomes activated ones the enzyme is presented at the site of interest, e.g. proteases such as matrix metalloproteases [61]. In a similar off/on strategy, fluorescent tracers can be designed to be quenched (non-activatable) in the off-state and activated, for example, by an acidic microenvironment using an ultra-pH sensitive amphiphilic polymer [36].

Clearly, it remains challenging to find a single imaging method that can satisfy the increasing need for diagnostic information in all instances and clinical condition. Therefore, chemists can at times convert or even expand existing imaging agents to another modality as, for example, an antibody can be labelled with a radionuclide such as 89-zirconium (89Zr) or a near-infrared fluorescent dye such as IRDye-800CW. Exemplary for dual-modality imaging is the vascular endothelial growth factor (VEGF)-targeting antibody bevacizumab labelled with IRDye800CW or 89Zr-bevacizumab [62]. A hybrid tracer, for example, is [111In]In-DTPA-trastuzumab-IRDye800CW, a monoclonal antibody against the human epidermal growth factor 2 (HER2) receptor and dual-labelled trastuzumab with both a fluorophore (IRDye800CW) and a radioactive label (111In) used for multimodal imaging of HER2-positive breast cancer [63,64,65]. The design of such hybrid tracers can combine desirable features of two different imaging modalities and can potentially improve quantification and understanding of the underlying biological targets through the complementary information provided [66].

Physics

Optical imaging uses excitation light varying from the visible light (380–750 nm) (e.g. for fluorescein- or Cyanine 5.5-based tracers) to the near-infrared spectrum (750–1400 nm) (e.g. tracers employing ICG or IRDye-800CW as fluorophores) to excite fluorescent tracers from a ground state to an excited state, from which decay to a lower energy state occurs, through the emission of fluorescence at a lower energy, i.e. higher wavelength. On the other hand, PET measures two high-energy annihilation gamma photons produced after positron emission from a radionuclide-tagged tracer molecule. In optical imaging, the tissue penetration of the fluorescent signal is limited to the range of several centimetres at best, while in PET, there is a virtually unlimited penetration depth. However, spatial resolution of (pre)clinical PET is in the range of 1–4 mm [67], whereas optical imaging can reach few to hundreds of micrometres depending on the optical imaging modality used [53, 68]. Additionally, in contrast to PET, optical imaging uses non-ionising radiation and is easy to use and relatively inexpensive in terms of camera costs and tracer synthesis, although the cost-effectiveness still has to be confirmed by health economic studies (Fig. 1). While optical imaging outperforms PET in spatial resolution of subsurface disease, PET is superior to optical imaging in its non-invasive quantitative assessment of larger volumes. The latter is possible through such techniques as attenuation and scatter corrections, providing uniform images that allow for high reading confidence and accurate quantitative assessments. In optical imaging, such corrections resulting in accurate quantitative assessments are already implemented in preclinical small imaging optical fluorescence imaging systems but still need to be developed for clinical applications. As the fluorescence signal detected is influenced by optical tissue properties, such as absorption and scattering properties, wide-field fluorescence imaging alone does not necessarily reflect true tracer distribution. Several methods have been proposed to correct for the aforementioned optical properties, such as multi-diameter single-fibre reflection/single-ibre fluorescence (MDSFR/SFF) spectroscopy [69, 70]. However, such a quantification method has yet to be translated to wide-field fluorescence imaging.

Standardisation and implementation

In optical imaging, standardisation of imaging procedures is currently lacking, mainly due to a lack of uniform guidelines. The variety of camera systems, illumination sources, data processing methods and imaging procedures in use makes an adequate comparison between study results and observers challenging. The professional environment that was built for imaging techniques, such as PET, allowed the field of nuclear medicine to flourish in a clinical setting towards standard of care. To date, this is still lacking in the optical imaging field. In this respect, the optical molecular imaging community should benefit from the knowledge gained in clinical PET standardisation and quantification. If optical imaging were to be implemented into the standard of care, it needs an established uniform translational pathway. As nuclear imaging techniques, such as PET, had to tackle similar issues, we must join forces as a clinical molecular imaging community and share our lessons learned, irrespective of the applied imaging modality. The pathway developed to bring nuclear medicine and molecular imaging into the standard of care may provide the needed roadmap for advanced clinical translation of optical imaging and more recent innovations, such as clinical multispectral optoacoustic tomographic imaging. We, therefore, advocate more intensified multidisciplinary and international collaborations to facilitate the implementation of optical and optoacoustic imaging as the next step after X-ray CT, PET, SPECT, MRI and ultrasound imaging. Clinical questions need to be answered to improve patient care; the choice of the methodology is secondary, i.e. the methodology should be selected based on the clinical or research question.

Take-home message: cross-fertilization within nuclear medicine and molecular imaging

A unique opportunity exists to develop a stronger bond between the multidisciplinary nuclear medicine community and pioneering optical imaging clinicians, being major players in hospitals in need of accurate, reliable and reproducible imaging data similar for neurology, cardiology and medical oncology. The history of PET imaging development and the hybrid PET-imaging technologies like PET/CT and PET/MRI, combined with the more recent theranostic clinical implementations, provides inspiration and a navigator for a bright future of optical and optoacoustic imaging. As has been the case with PET, fundamental challenges lay ahead for optical imaging, but applying and implementing proven strategies from the field of nuclear imaging will contribute to overcome those hurdles. The optical field should benefit from the knowledge gained in PET chemistry, standardisation and quantification and the established pathway to bring imaging techniques into the standard of care. This collaboration may present us with an opportunity to build on the molecular imaging part of ‘nuclear medicine and molecular imaging’ and provide us with a technique that complements our current molecular imaging modalities, such as real-time intraoperative guidance and early-stage cancer detection, to name a few.

The authors of this editorial have a background in nuclear medicine, surgery, gastroenterology, chemistry, optical engineering and physics and support each other in this endeavour at a local, regional level. We appeal to colleagues in our communities to strengthen these synergistic bonds on a broader (inter-)national level at a larger scale and not to miss out on this unique chance for a historical rendezvous of molecular imaging modalities. In short, let us all embrace optical imaging as a growing branch on the clinical molecular imaging tree and a global opportunity to enrich our molecular armamentarium for the benefit of the patient.