Innovations in dedicated PET instrumentation: from the operating room to specimen imaging

This review casts a spotlight on intraoperative positron emission tomography (PET) scanners and the distinctive challenges they confront. Specifically, these systems contend with the necessity of partial coverage geometry, essential for ensuring adequate access to the patient. This inherently leans them towards limited-angle PET imaging, bringing along its array of reconstruction and geometrical sensitivity challenges. Compounding this, the need for real-time imaging in navigation systems mandates rapid acquisition and reconstruction times. For these systems, the emphasis is on dependable PET image reconstruction (without significant artefacts) while rapid processing takes precedence over the spatial resolution of the system. In contrast, specimen PET imagers are unburdened by the geometrical sensitivity challenges, thanks to their ability to leverage full coverage PET imaging geometries. For these devices, the focus shifts: high spatial resolution imaging takes precedence over rapid image reconstruction. This review concurrently probes into the technical complexities of both intraoperative and specimen PET imaging, shedding light on their recent designs, inherent challenges, and technological advancements.


Introduction
Positron emission tomography (PET) imaging plays a crucial role in healthcare delivery by providing valuable information about the physiological and biochemical processes in the human body (Jones and Townsend 2017).It allows for the visualisation and quantification of metabolic activity, blood flow, and receptor binding in various tissues and organs, aiding in the diagnosis, staging, and treatment planning of various diseases, including cancer, cardiovascular diseases, and neurological disorders (Zaidi andKarakatsanis 2018, Meikle et al 2021).Moreover, PET imaging can also be used to monitor the effectiveness of treatment and track disease progression over time.By providing insights into functional and structural changes in the body, PET imaging contributes to the early detection and prevention of diseases.
The predominant application of PET imaging is primarily in the field of oncology which includes clinical diagnosis, staging, and monitoring of various types of cancer, such as lung cancer, breast cancer, and colorectal cancer (Czernin et al 2013).Another prime application of PET imaging is in the field of cardiology, where it is used for assessing myocardial viability and detecting coronary artery disease (Cherry et al 2023).Additionally, PET imaging is commonly employed in neurology for the diagnosis and monitoring of neurodegenerative disorders such as Alzheimer and Parkinson diseases (Hooker and Carson 2019).This has spurred the emergence of organ-specific dedicated PET devices, with many of them being commercially available (Sanaat et al 2024, Zaidi et al 2024).
Apart from conventional applications of PET imaging in oncology, neurology, cardiology, psychiatry, inflammatory and infectious diseases, and paediatrics (Arakawa et al 2020, Takalkar andHernandez Pampaloni 2020) this modality is vastly employed in animal studies, biomedical research, drug development, intra-surgery imaging, range verification in proton radiation therapy, image-guided surgery, laparoscopy, surgery navigation systems, specimen imaging, and organ-on-chip (OOC) studies (Gulec 2007, Futamura et al 2013, Ferrero et al 2018).
The use of intraoperative PET radiotracer or PET imaging in surgery has shown promising potential in improving surgical outcomes and patient care.By combining the precision of PET imaging with the real-time guidance of surgical procedures, intraoperative PET radiotracer or PET imaging enables surgeons to visualise and navigate through complex anatomical structures, identify tumour margins, and assess the extent of disease during surgical interventions (Holland et al 2011).The integration of PET imaging with surgical navigation systems can help surgeons in real-time during operations through enabling surgeons to use radiotracer activity to identify cancerous tissues and guide the surgical procedure.This helps to ensure all diseased tissue is removed, and healthy tissue is preserved as much as possible.PET radiotracers and/or PET data are being employed during surgery in the form of PET-guided navigation surgery systems (Feichtinger et al 2010), PET gamma probes (Gulec 2007, Ay et al 2012, 2013), laparoscopic PET imaging systems (Liyanaarachchi et al 2021), and open-gantry (limited-angle) intraoperative PET systems (Sajedi et al 2019).These systems generally aid the surgeon to better locate the target of surgery or spare healthy tissues.
Intraoperative specimen imaging refers to the use of imaging techniques during surgery to obtain high-resolution images of surgical specimens.This allows for a detailed analysis of the specimen's molecular and histopathological characteristics, thus providing valuable information about tumour heterogeneity and aiding in surgical decision-making and treatment planning.During intraoperative specimen imaging, a PET/computed tomography (CT) scanner is used to scan the tissue specimen immediately after it is resected from the patient's body.This technique provides valuable information about the tumour heterogeneity and can aid in the identification of metastatic locations (Muraglia et al 2023).One of the goals of cancer surgery is to achieve clear margins, meaning that no cancer cells should be found at the outer edge of the tissue that has been resected.Intraoperative PET imaging can help surgeons assess whether the margins of the removed specimen are clear of cancer cells in real-time.If the PET scan indicates the presence of remaining cancer cells, the surgeon may choose to resect additional tissue.Moreover, intraoperative specimen PET imaging is used for lymph node assessment, guidance for biopsy or lesion sampling, and evaluation of treatment effectiveness (Povoski et al 2015).Intraoperative specimen PET imaging has been so far conducted by clinical PET/CT scanners (Povoski et al 2015) or high-resolution small-animal PET imagers (Debacker et al 2021).However, recently a dedicated specimen PET/CT scanner has been introduced for intraoperative examination of tissue samples and tumour margin assessment (Muraglia et al 2023).
In addition to intraoperative specimen PET imagers, high-energy gamma probes and autoradiograph systems are employed intraoperatively to assess the margin of the dissected tumours (Strong et al 2008).Autoradiography is a technique that uses a radioactive tracer to visualise cellular and molecular function.When applied in a surgical setting with a PET radiotracer, it is often used to assess the distribution of the radiotracer within the surgical specimen, which can provide important information to the surgeon.The principal function of autoradiography systems is to detect and visualise the spatial distribution of radioactively labelled substances within a sample.This is accomplished through the process of exposing a photographic film or detector to the emissions from the radioactive substance, creating an image that represents the location and concentration of the substance in the sample.The autoradiograph can then be studied to understand the distribution of the radiolabelled substance to define the margin of the tumour.It can also be overlaid with a microscopic image of the tissue section to give a detailed view of where the radiolabelled substance is located in relation to the tissue structure (Bundy 2001).
Moreover, tissue/OOC technology, also known as microphysiological systems, is an emerging field at the intersection of biology, medicine, and engineering.This revolutionary approach involves the creation of miniaturised, three-dimensional (3D) models of human organs or tissues, structured on a microfluidic chip (S.1 Organ-on-tissue).The application of PET imaging in tissue-on-chip proved to be promising in providing valuable insights into the functionality and behaviour of tissues in a controlled microenvironment.This emerging field combines the advantages of both PET imaging and tissue-on-chip technology, allowing for real-time monitoring and analysis of metabolic processes in living tissues at the microscale level.PET imaging can provide detailed information about the metabolism, oxygenation, and perfusion of tissues.It can be particularly useful in tissue-on-chip applications as it can provide dynamic and quantitative information about the metabolic activity of cells within the tissue (Klontzas and Protonotarios 2021, Clement et al 2022, Kim et al 2012, Marsano et al 2016, Gao et al 2022).
This review provides an in-depth exploration of advancements in dedicated PET imaging systems, a field that has evolved dramatically to include innovative intraoperative PET imaging, intraoperative specimen imaging, PET laparoscopy, and cutting-edge tissue-on-chip PET imaging techniques.These developments have broadened the application of PET imaging, extending its use beyond traditional diagnostic and prognostic roles to active involvement in guiding surgical interventions.Intraoperative PET imaging and specimen imaging provide real-time feedback, allowing for more precise tumour resection and improved margin assessment.Similarly, PET laparoscopy has emerged as a promising tool for minimally invasive surgeries, providing high-resolution, 3D imaging that enhances surgical precision.More recently, the application of tissue-on-chip technology to PET imaging represents a significant breakthrough, offering a novel in vitro platform to study drug pharmacokinetics and disease processes at tissue and organ level.This rapidly expanding field holds immense potential for advancing precision medicine, enhancing patient outcomes, and reshaping the future of medical imaging and surgical practice.

PET-derived information in the surgery room
In recent years, the integration of PET within surgical environments has shown promising potential in enhancing the precision and outcomes of various procedures.The use of PET radiotracers provides surgeons with real-time, high-resolution images of pathological tissues, such as tumour foci, allowing for greater accuracy in targeting and excision.Specifically, these radiotracers can be employed to delineate the boundaries between malignant and healthy tissues, ensuring maximal tumour resection while sparing healthy surrounding structures.Moreover, intraoperative PET imaging can be invaluable in detecting metastatic lesions or nodes that may not have been evident in preoperative imaging, facilitating immediate intervention and reducing the need for subsequent surgeries.This dynamic fusion of radiotracer imaging and surgical practice not only streamlines the surgical process but also augments the potential for more favourable patient prognosis, setting a new standard in personalised surgical care (El Lakis et al 2019, Van Oosterom et al 2019, Valdés Olmos et al 2020).
Building upon the foundational use of PET in the operating room, the advent of PET radio-guided surgery has further revolutionised oncological interventions.In this innovative technique, radiolabelled probes targeting specific tumour markers are administered to patients, enabling surgeons to employ intraoperative probes that detect the emitted radiation, thereby accurately localising tumour tissue.This is especially beneficial in instances where tumours are intimately associated with vital structures or when metastatic lesions are diffuse and scattered.The benefits of PET radio-guided surgery are multifaceted.First, it enhances the surgeon's ability to discern tumour margins in real-time, ensuring complete resection while minimising damage to adjacent healthy tissues.Additionally, it can significantly reduce operative times, given the real-time feedback, and potentially lower post-operative complications.Furthermore, patients benefit from more precise interventions, often translating to reduced hospital stays and improved overall outcomes.As the collection of radiotracers expands and technology evolves, PET radio-guided surgery stands to offer even more transformative changes to the landscape of oncological surgery (Gulec 2007, García-Talavera et al 2014, Pashazadeh and Friebe 2020).

PET-guided navigation in surgery
PET guided navigation surgery is an innovative approach that merges the diagnostic precision of PET imaging with surgical navigation technologies.By providing real-time, high-resolution images of metabolic activity, PET can accurately delineate tumours and other pathologies, assisting surgeons in identifying the exact location and extent of the disease.Integrating PET imaging with navigation systems enables surgeons to operate with enhanced precision, ensuring maximal tumour resection while sparing healthy tissue.This integration is especially pivotal in areas with intricate anatomy, such as the brain or near vital organs, where the surgical margins are thin, and the potential risks of damage are high (Feichtinger et al 2010).The fusion of these technologies aids in optimising surgical outcomes, reducing post-operative complications, and potentially improving patient survival rates.
Moreover, PET-guided navigation systems offer potential for real-time feedback during surgeries.This dynamic interactivity allows for immediate adjustments in surgical strategies based on the evolving intraoperative landscape.As surgical tools and instruments advance, the synergy between PET imaging and navigation systems will likely play a fundamental role in the future of precision medicine, bridging the gap between diagnostic imaging and therapeutic interventions.
In this regard, Feichtinger et al presented a novel method for assessing resection margins intraoperatively in patients with advanced head and neck carcinomas, using PET/CT image fusion on a 3D navigation system (2010).Such image-guided surgery allows for more precise navigation of tumour ablation and assessment of resection margins.The technique was tested on six patients, with four indicating unsafe resection margins.Additional image-guided resection achieved local tumour control in three of these cases, while in one, further resection was unachievable due to skull base invasion.The study concluded that this new method may improve local control of advanced head and neck cancer.
A major limitation of the current technology is that it predominantly relies on preoperative PET data.This preoperative imaging data is often fused with structural information from other imaging modalities, such as CT or magnetic resonance imaging, to provide a comprehensive view of the anatomical and metabolic landscape.This fusion enables surgeons or surgical robots to navigate with high precision to the exact location of lesions, maximising the effectiveness of the intervention (Maybody et al 2013).However, a notable limitation in the existing technology is the absence of real-time or online PET data acquisition.This lack of intraoperative PET feedback potentially restricts the surgeon's ability to adapt to dynamic changes during surgery, necessitating reliance on static preoperative imaging.Moreover, a significant dependency exists on advanced registration and tracking algorithms.These computational tools are pivotal in ensuring precision and accuracy during interventions, allowing for the seamless fusion of metabolic and anatomical data.The effectiveness of the surgical navigation process is intrinsically tied to the robustness of these algorithms, underscoring their critical role in contemporary surgical practices.To achieve the pinnacle of surgical precision and adaptability, future innovations in PET-guided navigation systems will need to address this gap and incorporate real-time PET imaging capabilities (Boekestijn et al 2021).
Within the expanding field of PET-guided navigation systems, there exists a predominant reliance on pre-operative PET imaging for surgical guidance.Notably, only a limited number of dedicated intra-operative systems have been developed to date.Given the specific focus of this article's review on intraoperative PET imaging, our analysis will be narrowly tailored to these dedicated intra-operative systems.The emphasis underscores the importance and potential implications of real-time PET imaging in enhancing surgical outcomes and precision.

Biopsy
For optimal needle trajectory guidance during navigated biopsies, some researchers advocate for a navigation method involving thoracic staging through percutaneous biopsy assisted by an automated robotic arm.This technique has been effectively used in targeting lung lesions with 18 F-FDG (Fluorodeoxyglucose) and identifying suspected metastatic lung lesions using 68 Ga-DOTANOC (DOTA,1-Nal3-Octreotide) and 68 Ga-PSMA (Prostate-Specific Membrane Antigen).In this approach, a robotically controlled biopsy needle is directed towards the lesion pinpointed on the diagnostic pre-surgery PET/CT.To ensure precise navigation, the patient is kept still.The precise location is ascertained using two key reference points: one for registration and another for needle insertion.Following the needle's placement, its position is double-checked using an additional intra-procedure PET/CT.The procedure then progresses to tissue extraction, boasting a tissue acquisition rate of 100% and a diagnostic success rate ranging from 96%-100% (Radhakrishnan et al 2017, 2018, Kumar et al 2020, Nath et al 2020).For this application, clinical/conventional PET/CT scanners were employed to verify the accuracy of tissue sampling (Radhakrishnan et al 2017).At present, navigational methods guided by nuclear medicine are predominantly associated with sentinel lymph node procedures, given its well-established nature and widespread application.This provides a straightforward platform for testing new techniques.Nonetheless, navigation approaches utilising PET and single-photon emission computed tomography (SPECT) are anticipated to offer added benefits in procedures specifically targeting tumours (Boekestijn et al 2021).
Conducting biopsy guidance, especially in prostate or breast cancer procedures, can be intricate.Yet, a specialised system can make these tasks more precise and secure.This system would facilitate real-time imaging and 3D reconstruction of tumours while also tracking the needle's trajectory within the patient (Pichler et al 2008).In this regard, West Virginia University's instrumentation group has pioneered the creation of the PEM-PET, which stands for positron emission mammography-tomography imager and biopsy system (Raylman 2009).Using expansive rotating detectors (measuring 15 × 20 cm), this device can craft 3D tomographic images from multiple angles of the breast.Demonstrations have shown the scanner to achieve impressive transaxial and axial resolutions of under 2 mm while maintaining an energy resolution of 25%.The multi-angle approach in tomography allows for precise pinpointing of suspicious areas with radiotracer uptake in three dimensions, negating the need for stereotactic methods.Enhancing the system's capabilities, a computer-aided arm was integrated into the device, thus enabling it to perform image-guided breast biopsies (figure 1).
The PEM-PET system's imaging section is equipped with two sets of radiation detectors.Every detector head houses a 96 × 72 array of 2 × 2 × 15 mm 3 lutetium-yttrium oxyorthosilicate (LYSO) detector elements (with a pitch of 2.1 mm).These elements are linked to a 4 × 3 array of Hamamatsu H8500 position-sensitive photomultiplier tubes (PS-PMT).To reduce the dead space between the detector and the chest wall, the scintillator arrays were custom designed.The gap from the top of the detector head to the scintillator array is specified in millimetres.These four detectors are positioned on a computer-managed rotator, with a separation of 25.6 cm between them.The imaging structure is situated beneath the patient's bed.Part of the bed comprises a 2 mm-thick tungsten alloy, providing shielding for the detectors below against any external radioactivity beyond the scanner's field of view (FOV).The breast is aligned within the imager's FOV through an aperture in the bed.
The system's biopsy feature comprises a device capable of moving in all three dimensions and rotating a full 360 degrees around the scanner's imaging axis (figure 1).The biopsy arm is equipped with three computer-controlled linear sliders, provided by Velmex, Inc., each set along the x, y, and z axes.These sliders are attached to a mount connected through a thrust bearing to the imaging structure.This arrangement permits manual rotation of the arm around the patient's breast to determine the optimal entry point, while the system's interface software consistently tracks this angle.The arm can accommodate a standard Bard Magnum® biopsy gun, supplied by C.R. Bard, Inc., using a custom-made mount.Movement of the needle along the z-axis is managed using a handheld joystick, ensuring that the insertion process remains under the direct oversight of the operator.The needle's position is consistently tracked through linear position encoders situated on all three axes of the biopsy arm.
The utmost positioning discrepancy observed was ±0.27 mm, which is accurate enough to procure samples from the majority of lesions identified by the PEM-PET.

PET gamma probes
Intraoperative PET gamma probes or gamma cameras play a vital role in the field of surgical oncology.One potential application of intraoperative PET gamma probe/camera is in brain cancer surgery.In brain cancer surgery, the use of intraoperative PET gamma probe/camera allows for real-time and specific visualisation of the extent of disease within the brain.This enables the surgeon to assess surgical resection margins and minimise surgical invasiveness, leading to improved patient outcomes.In addition to brain cancer surgery, intraoperative PET gamma probes have shown promise in other areas of surgical oncology (Strong et al 2009, Haiquel et al 2022).For example, the use of intraoperative PET gamma probes has been explored in the detection of lymph node involvement in head and neck, breast, and endometrial cancers.
PET gamma probes are used in lymph surgery to detect and identify lymph node metastases.During the procedure, a gamma-emitting radiotracer, such as 111 In-PSMA or FDG, is injected into the patient prior to surgery.The radiotracer binds to specific receptors on cancer cells, allowing them to be detected by the gamma probe.The surgeon then uses the gamma probe/camera intraoperatively to scan the lymph node area and identify any areas of increased radiotracer uptake, indicating the presence of metastatic lymph nodes (Haiquel et al 2022).
The specific applications of the intraoperative PET gamma probe and/or camera include: Neurosurgery (brain tumours): It can be crucial to differentiate between malignant and surrounding healthy tissues, especially in eloquent areas of the brain where functional preservation is essential (Vilela Filho and Carneiro Filho 2002).
Breast Surgery: For patients undergoing lumpectomy or other breast-conserving surgeries, the probe/camera can help in identifying and removing all malignant tissues, reducing the likelihood of a second surgery (Povoski et al 2009).
Lung Surgery: The probe can be beneficial in thoracic surgeries to ensure that all tumour nodules are removed, especially when dealing with metastatic cancers or multiple nodules (Boni et al 2000).
Colorectal Surgery: For tumours that are not easily palpable or seen, an intraoperative PET probe/camera can help ensure complete removal (Sarikaya et al 2008).
Head and Neck Surgery: Due to the complexity of these regions, an intraoperative PET probe/camera can be beneficial in ensuring all malignant tissue is identified and removed, particularly when dealing with recurrent tumours (Povoski et al 2009).
Melanoma and Skin Cancers: It can be used to detect satellite nodules or in-transit metastasis which might be missed otherwise (Alazraki et al 1997).
Sarcomas: Especially in deep-seated or recurrent sarcomas, the probe can be useful in delineating the tumour's extent (Leong et al 2018).
Gynecologic Surgeries: For ovarian, cervical, and endometrial cancers, the probe/camera might assist in assessing the extent of the disease and ensuring complete resection (van Luijk et al 2004).
Previous studies aimed to evaluate the efficacy of gamma and beta probes in detecting and localising malignant lesions, particularly of varying sizes (Strong et al 2008(Strong et al , 2009)).They also explored the potential advantages of these probes, such as the beta probe's ability to more precisely detect smaller malignant lesions due to less influence from surrounding radiation.In this regard, Strong et al demonstrated a strong correlation between gamma emissions detected through standard PET imaging and beta emissions (2009).The findings suggest that beta counts might be more efficient for on-the-spot pinpointing of tumour locations compared to gamma emissions.
The overall goal seems to be to establish the optimal usage of these probes in surgical procedures, potentially utilising their synergistic effects for granular assessment of malignant tissue and improved accuracy in surgical dissection.Some studies demonstrated that suspicious lesions were identified and subsequently pinpointed during surgery due to their high tumour-to-background ratio (TBR) using the PET probes that were not spotted in preoperative PET/CT examinations (Cohn et al 2008, González et al 2011).These were confirmed to be malignant upon pathological examination.Notably, some smaller tumours that were not visible on the PET scan, were detected during surgery because of their elevated TBR and were also confirmed to be malignant through pathology (González et al 2011).
Intraoperative gamma probes operate differently from traditional PET imaging techniques.While PET imaging capitalises on the detection of annihilation photons, thereby generating an image, intraoperative gamma probes solely detect single gamma emissions without forming any image.Given this fundamental distinction in operational principles, the discussion on the PET gamma probe will not be encompassed in this review, as it deviates from the scope intended for this article.

PET imaging in endoscopy and laparoscopy
The application of PET imaging in endoscopy has the potential to revolutionise the field by providing more accurate and sensitive detection of malignant lesions.Currently, the detection of malignancies during endoscopy relies on nonspecific tumour tools, such as needle localisation, tissue palpation, or direct intraoperative visualisation.This often leads to suboptimal detection rates and missed lesions.The integration of PET imaging in endoscopy could significantly improve tumour detection and localisation.
PET imaging, when combined with endoscopy, allows for real-time in vivo imaging of intraluminal neoplasms, such as rectal cancer, cervical cancer, or head and neck cancer.Unlike traditional endoscopy methods, PET imaging provides a more comprehensive and accurate assessment of tumour response after radiotherapy (Bae et al 2018).This is particularly advantageous in cases where the primary tumour cannot be detected using conventional imaging techniques.Through the use of systemic administration of radiotracers, PET imaging can show tumours and metastases in the whole-body (Zhang et al 2020).Additionally, the use of peptide probes typically in PET imaging during gastroendoscopy has shown promising results in the detection of Barrett's oesophagus, a condition that can progress to oesophageal cancer.The high resolution and real-time performance of PET imaging in endoscopy make it a powerful tool for the early detection and monitoring of gastrointestinal neoplasm (Park et al 2020).
In this regard, EndoTOFPET-US project was initiated to develop a detector unit for endoscopes that is specifically designed to meet the requirements of prostate examinations.This detector was intended to be scalable so that it can be adapted for use with a pancreatic endoscope as well.For reasons of time and cost, only the prostate detector was produced and commissioned for clinical use initially.The detector allows for the calculation of the line of response (LOR) in time-of-flight (TOF) PET (Aubry et al 2013).
The detection device is composed of two primary elements: a PET extension for a standard ultrasound endoscope, and an external PET plate which faces the inner probe.The pancreatic endoscope, along with its PET addition, is roughly half the size of the one designed for the prostate.A typical healthy prostate is an organ with an irregular shape, measuring approximately 4 × 3 × 2 cm 3 .Therefore, a coincidence time resolution (CTR) of 200 ps full width at half-maximum (FWHM), which equates to 3 cm along the response line, is essential to minimise interference from adjacent organs surrounding the region of interest.The design planned to use 15 mm crystals for the external plate and 10 mm crystals for the internal probe.The decision for the 10 mm crystals is driven by the mechanical specifications needed to accommodate the internal probe.
The external plate, measuring 23 × 23 cm 2 , is designed as a detector composed of 256 units.Each unit is defined by a 4 × 4 array of LYSO crystals, each measuring 3 × 3 × 15 mm 3 , achieving an average time resolution of 235 ± 4 ps.These crystals are individually linked to a solid 4 × 4 matrix of Multi-Pixel Photon Counters (MPPC), where each counter has an active region measuring 3 × 3 mm 2 , sourced from Hamamatsu.The external plate's position is controlled by a robotic arm.This arm not only sustains the structure but also adjusts its position for optimal FOV based on the internal probe's alignment (Zvolský et al 2015).Figure 2 presents a 3D drawing of the EndoTOFPET-US external plate and the endoscope PET extension for a pancreatic clinical case.
The prostate-specific internal probe is a combination of a commercial transrectal ultrasound probe (sourced from Hitachi medical systems EUP-U533) and a highly intricate extension.This extension, mirroring the ultrasound probe's cross-section, incorporates a matrix of 18 × 18 LYSO crystals, each measuring 0.71 × 0.71 × 10 mm 3 .Every individual crystal aligns with a digital silicon photomultiplier, encompassing a group of 416 single photon avalanche diodes (SPADs).A SPAD is essentially a Geiger mode avalanche photodiode crafted using a conventional CMOS process, a predominant method in large-scale integration chip creation.Initial evaluations conducted with 0.71 × 0.71 × 10 mm 3 LYSO:Ce crystals linked to an MPPC S10931-050P and a differential amplifier-discriminator revealed an average time resolution of 187 ± 27 ps (Cortinovis 2015).
In addition to endoscopy, the application of PET laparoscopy in modern medicine has revolutionised minimally invasive surgical procedures.With its ability to provide real-time imaging and enhance surgical precision, PET laparoscopy offers numerous benefits for patients and surgeons alike.In the era of minimally invasive surgery, laparoscopy has been suggested as a potential method for detecting minimal peritoneal metastasis because it not only allows the visualisation of abdominal cavities, but also can obtain tissue or peritoneal fluid for pathologic confirmation (Guo et al 2022).
This technique has been particularly useful in cases where CT or PET findings are unreliable for detecting cancer recurrence.The use of PET laparoscopy allows for more accurate diagnosis in patients with advanced gastric cancer and suspected peritoneal metastasis.Recently, staging laparoscopy has been performed in advanced gastric cancer patients with suspected peritoneal metastasis to obtain a more accurate diagnosis and enhance decision-making in treatment planning.Staging laparoscopy has shown promising results in improving diagnostic accuracy and shortening the time until the start of chemotherapy compared to exploratory laparotomy.However, it is important to note that staging laparoscopy may not always detect peritoneal metastasis, as reported in some studies.Therefore, there is still a need for further refinement of the technique to improve its diagnostic accuracy (Sebastian et al 2017).In this regard, Liyanaarachchi et al introduced the concept of a PET-laparoscope system designed to capture 3D PET imagery, specifically targeting the visualisation of lymph node metastasis during gastric cancer surgeries (2020).This innovative approach was subjected to a simulation-based feasibility assessment as detailed in reference (Liyanaarachchi et al 2018).Within this system, two distinct detector modules are employed to identify the 511 keV annihilation photons, wherein the first module consists of a stationary external detector array attached to the patient's surgical bed.Simultaneously, the second component is a dynamic detector probe, carefully inserted into the patient's stomach in tandem with a laparoscope.The conceptual design of the system is presented in figure 3.
The new PET-laparoscope system can be used to locate lymph node metastasis during gastric cancer surgery.The prototype detector, constructed with a pixelated multi-layer movable detector probe with 2 × 2 × 3 mm 3 crystals, was developed to improve the spatial resolution of the system.It successfully reconstructed images of a 22 Na source with 3 mm spatial resolution in the X (coronal) and Z (longitudinal) directions in a 2 min scan.These results suggest that such a system could potentially enhance the accuracy of locating lymph nodes during surgery, thereby improving surgical outcomes for patients with gastric cancer.
The stationary detector is structured as a 7 × 7 array, composed of 10 × 10 × 20 mm 3 Ce:GAGG crystals, which stands for gadolinium aluminium gallium garnet or its chemical representation, Gd3Al2Ga3O12.There's a 1 mm spacing maintained between adjacent crystals.Renowned for its impressive attributes, GAGG boasts a light yield of 46 000 photons/MeV and an energy resolution of 4.9% at 662 keV.Additionally, it possesses a decay time of 88 ns and a density of 6.63 g cm −3 .Notably, GAGG is characterised by its lack of intrinsic radiation and its non-hygroscopic nature.These GAGG crystals were directly paired with PM66 type SiPMs, or Silicon Photomultipliers, which have an active area of 6 × 6 mm 2 .Also recognised as MPPCs, these SiPMs are integrated with a vast array of Geiger-mode avalanche photodiodes, delivering a substantial gain of about 10 6 .Teflon tape served as the chosen light reflector in this setup.
The multi-layer detector served as the transportable detection instrument.Due to the necessity of inserting this detector into the patient's stomach via the laparoscopic instrument incisions, its diameter does not exceed 12 mm.This probe integrates scintillator crystals sized at 2 × 2 × 3 mm 3 , aiming for enhanced spatial resolution.To expedite the scan by enhancing detection efficiency, multiple crystal layers were incorporated into the probe.This includes four layers structured in a cross-shaped pattern with 12 pixels.The scintillator used was the GFAG (gadolinium fine aluminuim gallate) crystal.Given that the detector probe is positioned in close proximity to the lymph node metastasis, superior timing characteristics are paramount to discern each event distinctly under high count rate scenarios.Every 2 × 2 × 3 mm 3 crystal was demarcated by a BaSO4 (Barium Sulfate) reflective layer 0.2 mm in thickness.
Accurate position tracking is critical for such laparoscopy system (Liyanaarachchi et al 2021).The Polaris optical tracking system (Northern Digital Inc., Ontario, Canada), boasting a precision of 0.25 mm, was used to monitor both the position and orientation of the movable detector probe (Wiles et al 2004).The polaris position sensor's illuminators release infrared rays.Utilising the infrared light that gets reflected back, the position sensor discerns the location and orientation of tools equipped with retro-reflective markers.One tracking tool was integrated with the movable detector to trace its movements, while another was affixed to the stationary detector to serve as a point of reference.Given that the spatial relationships of the crystals within the detector probe are pre-established, the location of every individual crystal was deduced in connection to the reference, factoring in the probe's position and orientation quaternion.
This technology aims to pinpoint lymph node metastases previously detected during preoperative PET scans.Surgeons, upon analysing preoperative PET images and with their anatomical understanding of lymph nodes, can deduce which stomach regions need examination for metastasis removal.However, pinpointing specific nodes with metastases amidst surgery proved to be challenging, leading to unwarranted tissue excision.This system is designed to guide surgeons to precise metastasis locations intraoperatively.The system's capability to function without reliance on preoperative PET scans could be assessed in forthcoming clinical trials.During these trials, both preoperative and intraoperative PET imaging sessions will take place.The main objective will be to ascertain if the intraoperative PET setup can capture the metastases noted in the preoperative imagery within a reasonable duration.Following this, lymph nodes will undergo resection based on prevailing surgical protocols.Subsequent pathological assessments will distinguish between benign and malignant nodes.The accuracy and reliability of the intraoperative PET mechanism will then be evaluated.

Surgical PET imaging probes
The primary objective of surgical PET imaging probes is to offer consistent, real-time 3D reconstructed PET imagery throughout surgical procedures.This imagery is dynamically overlayed onto a plane controlled by a tracking apparatus.In essence, the system acts as a portable, clinician-directed imaging device that visualises radiotracer distribution.Consequently, it allows for precise localisation and recognition of tumours during surgery, accommodating any positional alterations due to patient movement.
Tumours that remain undetected during pre-surgical PET scans can pose a risk.The persistent, unseen tumours after surgical procedures, could lead to recurrence.Yet, pinpointing tumours smaller than 1 cm in diameter is an ongoing challenge for traditional whole-body PET scanners.While non-imaging intra-operative probes have shown success in detecting tumours during the surgical removal of lymph nodes, imaging probes offer a more precise option for identifying small tumours and remnants post-tumour removal.These probes can render 2D visuals, reducing the time physicians spend on pinpointing lesions.One approach focuses on directly detecting beta particles from radionuclides that emit these particles.Despite the allure of reducing the interference from 511 keV background from other body parts due to positron-electron annihilation, beta particles' limited penetration depth constrains their utility, especially when attempting to detect deep tumours below the exposed tissue surface.
To identify deeply situated tumours, gamma-ray sensitive imaging probes are more adept than those sensitive to beta particles.However, detecting single photons, typically at 511 keV, results in a dispersed point-spread function, given the extensive attenuation length in almost all materials used for collimation.
To address the limitations of these imaging probes, a compact, high-resolution PET imaging probe was proposed by Huh et al (2010).This tool, which works in tandem with a section of a standard PET scanner, aims to provide surgeons with enhanced visuals.It aids in determining the location and scope of primary tumours during operations and in recognising multifocal diseases.The foundation of this PET imaging probe system draws insights from earlier research on high-resolution imaging techniques.The proposed PET imaging probe setup incorporates a lower-resolution partial ring detector coupled with a high-definition imaging probe, which comes with a position tracking feature.The high-definition imaging probe, along with its close proximity to target lesions, aids in pinpointing smaller lesions.The aim is to continuously offer updated 3D images, which are then relayed in real-time on a plane guided by a tracker.
This PET imaging probe system is made up of a pixelated NaI(Tl) (Sodium Iodide (Thallium-doped)) detector along with a partial ring detector.This partial ring detector includes four bismuth germanate (BGO) detector blocks.We employed a modified version of a one-pass list-mode expectation maximisation (OPL-EM) algorithm, combined with a row-action maximum likelihood (ML) algorithm, for image reconstruction (Reader et al 2002).
The high-resolution imaging probe was designed using a pixelated NaI(Tl) detector.This comprised an arrangement of pixelated NaI(Tl) crystals connected to a 2-inch by 2-inch PS-PMT (Hamamatsu® H8500).A semi-circular detector was assembled using four BGO block detectors.These four BGO detector blocks were set adjacent to each other to craft the semi-circular form (figure 4).
The camera consistently delivered 3D reconstructed images throughout surgical procedures, which were displayed in real time on a plane steered by a tracking mechanism.The intra-operative PET imaging probe system operates like a handheld camera, which, under the guidance of a clinician, can visualise the spread of the radiotracer.
Similar to other portable PET systems, such as the endoscopy PET, the efficacy of the handheld PET scanner system is intrinsically tied to the precision of its tracking system.This is crucial for determining the exact locations of the LORs.Consequently, when evaluating the performance of such scanners, it is imperative to concurrently consider the capabilities and reliability of their tracking systems.

Dedicated intraoperative surgical PET
While traditional whole-body PET systems offer consistent performance, delivering decent spatial resolution, repeatable quantification, and reliable image quality with clinical relevance, their considerable size, restricted patient accessibility (for the purpose of surgery), and significant cost, have constrained their utility in surgical settings, particularly for real-time monitoring of target tissues during operations.Recommendations from the American Association of Clinical Oncology (ASCO) stipulate that surgical procedures should target an identification rate of 85% and maintain false negative rates below 5% (Lyman et al 2005).The prevailing standard of care, utilising gamma probes and near-infrared (NIR) probes (Pesek et al 2012), falls short of these criteria.This underscores a pressing clinical demand for instruments that facilitate a comprehensive and precise intraoperative assessment of nodes prior to their removal.These limitations have driven the advancement of intraoperative tools.By positioning detectors nearer to the target tissue, these tools amplify the solid angle, thereby enhancing detection efficiency.Regarding the current advancement in the field of TOF resolution in PET imaging, Sajedi et al (2022) investigated the potential of TOF-PET for intraoperative imaging, positioning it as a strong rival of intraoperative gamma probes, positron probes, 3D gamma cameras (Bluemel et al 2013), and preoperative specialised PET scanners (Jiang et al 2020).
This system employed two detector panels (a limited angle geometry system), each containing varying numbers of detector modules.The detector panel positioned above the patient is equipped with two modules.Conversely, the panel positioned below the patient bed featured a 3 × 3 module array of identical size.They maintained 27 cm distance between the two panels to accommodate a 25 cm thick phantom/patient, which mimics the average human torso size, and a roughly 2 cm bed structure.Each detector module consists of 12 × 12 array of 4.14 × 4.14 × 20 mm 3 lutetium fine silicate (LFS) crystal pixels, set at a pitch of 4.2 mm, directly coupled to SiPM pixels.The total active surface area of this module spans 51 × 51 mm 2 .When factoring in the light-proof housing, the module's external size measures 53 × 53 mm 2 (figure 5).The measured detector's CTR was 271 ps FWHM, whereas the energy resolution was about 16%.
The scanner was able to provide a FOV of approximately 24.5 × 13.5 × 25 cm 3 .Their empirical results revealed that, with brief data collection, 6 mm diameter spheres were discernible on the generated images when the lesion phantom exhibited a 10:1 activity contrast against the background.They deduced that the deployment of a restricted angle TOF PET detector system indicates significant progress for intraoperative applications.Notably, its capability for lesion detection surpasses traditional gamma-and NIR-based probes.
In the realm of intraoperative surgical PET, the 'Panel PET with Window' (PPW) was introduced by Li et al (2017).A PPW is characterised by a rectangular aperture centrally positioned on one of its detector panels.This design permits medical procedures, including surgical interventions, biopsies, and radiotherapy, to be conducted through the opening.Inheriting the benefits of a traditional panel PET, such as compactness and an adjustable FOV, the PPW utilises fewer detector modules.This not only reduces component requirements but also augments application flexibility and convenience (figure 6).
To investigate the relationship between window size and the imaging performance of the PPW, an experimental study was conducted using a full-size panel PET scanner.In this configuration, each detector panel possessed a detection area of 100.8 × 100.8 mm 2 and was comprised of a 4 × 4 array of detectors.Each individual detector housed a 6 × 6 array of LYSO crystals, each crystal measuring 3.9 × 3.9 × 20 mm 3 in size with a crystal pitch of 4.2 mm.The inter-panel separation distance was fixed at 40 mm.
To investigate various window sizes within the PPW configuration, they systematically removed 6 × 6, 12 × 12, 16 × 16, 18 × 18, and 20 × 20 crystals from the central region of the upper panel.This procedure resulted in the creation of a series of PPWs with differing window sizes.To quantify the window size in relation to the panel area, they introduced the ratio of the window area to the panel area, wherein various PPWs with 6.3%, 25.0%, 44.4%, 56.3%, and 69.4% ratios were examined.
The findings indicate a decline in the PPW system's sensitivity, spatial resolution, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) as the ratio of the window area to the panel area increased.Additionally, the effective FOV shrinks.Interestingly, this decline is not directly proportional to the enlargement of the window size.In an extreme scenario where 69.4% of the detectors were eliminated from a panel, the reductions in SNR, CNR, and spatial resolution were 42.4%, 43.1%, and 25.0%, respectively.Nonetheless, the contrast recovery coefficient of the PPW remains relatively stable (Li et al 2017).
The PPW holds potential for seamless integration into biopsy guidance, as well as combination with other imaging modalities to build multimodal imaging platforms.A significant prospective application of the PPW is its utility in in-beam imaging.It is crucial to note that, unlike the dual-ring OpenPET and partial-ring PET scanners, the PPW does not accommodate an outlet for the flow of light fragments (Enghardt et al 2004, Crespo et al 2006).Given that in-beam PET typically gathers data during treatment breaks and immediately post-treatment, the elevated background noise originating from nuclear reactions has minimal interference with the PPW's operation.
In recent advancements in surgical PET systems, two distinct configurations have been introduced above which facilitate surgical procedures without compromising the imaging capability.The first system employs a dual detector-panel design, wherein one of the panels incorporates a smaller upper detector.This conceptual design ensures that ample space is provided above the patient, thereby facilitating surgical manoeuvres.The second configuration, while also utilising two detector panels, introduces a window within the upper detector, thus granting direct access to the patient.Moving beyond these designs, innovative concepts for surgical PET systems tailored for cranial procedures have emerged.For instance, a model presented in (González et al 2018) depicts a conceptual framework where PET detectors are strategically spaced around the patient's head (figure 7).This arrangement not only permits unobstructed access to the patient but also allows for surgical instruments to be manoeuvred through the gaps between detectors.Furthermore, to optimise imaging capabilities, these detectors are affixed to robotic arms.This enables precise positioning of the detectors, ensuring that critical images can be acquired during the surgical intervention.
In the field of radiation therapy, the introduction of in-beam PET scanners has revolutionised the way we approach treatment planning and monitoring.These innovative scanners allow for real-time imaging and analysis of the dose deposition during radiation treatment, providing valuable information about the effectiveness and accuracy of the delivered therapy.Moreover, in-beam PET scanners have been shown to be particularly useful in particle therapy, such as proton therapy and carbon ion therapy (Parodi 2015).In-beam PET scanners detects positron-emitting isotopes produced during the irradiation of patients in a heavy ion therapy facility.These isotopes are created when the heavy ions interact with patient's tissues, causing nuclear reactions that release positrons.These positrons quickly annihilate with nearby electrons, emitting two annihilation photons with an energy of 511 keV each.The in-beam PET scanner uses detectors, typically comprised of scintillator crystals and PMTs, for coincident detection of these 511 keV photons.The in-beam PET scanner operates by detecting positron-emitting isotopes generated during heavy ion therapy (Bäumer et al 2021).
The unique configuration of in-beam PET scanners, rooted in the principles of partial coverage or limited angle PET imaging, is primarily driven by the need to maintain patient accessibility during the delivery of the prescribed radiation dose, such as carbon or proton beams.Interestingly, the inherent design advantages of these scanners can be extrapolated beyond radiation therapy.Their ability to provide unobstructed access to the patient holds significant promise in surgical contexts.Specifically, these systems can be seamlessly integrated into surgical setups, thus offering direct access for procedures like biopsies.By combining imaging and therapeutic functionalities, these in-beam PET scanners underscore a new era of versatile and patient-centric medical interventions.
In response to the evolving needs of heavy ion radiation therapy, Tashima et al (2011) introduced a notable advancement with the proposal of the OpenPET scanner.This innovative system was specifically designed for online radiation monitoring during therapeutic sessions.The OpenPET system was designed to give patients an open space during PET scans.The inaugural OpenPET design, termed the 'dual-ring OpenPET' features two distinct detector rings (figure 8(A)).This design allows for an expanded axial FOV, capable of imaging both the gap region and the in-ring region.However, tasks such as dose verification via in-beam PET during particle therapy and real-time tumour tracking necessitate a sensitivity concentrated on the gap, rather than a broad FOV.In this study, they introduced an improved single-ring OpenPET (figure 8(B)).This design grants an open, easily observable space while achieving greater sensitivity and requiring fewer detectors than its predecessor (Tashima et al 2011).The updated design resembles a cylinder sliced at an inclined angle, resulting in elliptical ends.They conducted a theoretical sensitivity analysis of their new design and compared it with the dual-ring OpenPET and another design where a conventional PET was tilted against the patient's bed, creating an open space, so called 'slant PET' .The central sensitivity of these designs hinges on their respective solid angles (Tashima et al 2012).The findings revealed that the single-ring OpenPET's sensitivity is 1.2 times superior to the dual-ring OpenPET and 1.3 times better than the slant PET.While the primary design intent of the OpenPET scanner is for in-beam range verification, its functionality extends beyond this.The system, along with other in-beam PET scanners developed for radiation therapy, possesses the potential for application in PET-guided biopsies and surgeries.

Clinical/conventional PET/CT scanners
In the surgical theatre, specimen imaging plays a pivotal role in ensuring the precision and efficacy of interventions.It serves as a bridge between clinical observation and microscopic validation, enabling surgeons to make real-time decisions based on visual assessment of tissues.Specimen imaging can validate the complete excision of a tumour, helping in the determination of tumour margins during oncological procedures.It ensures that no malignant tissue is inadvertently left behind, thereby reducing the risk of recurrence.Additionally, in complex surgeries, it aids in the identification and preservation of vital structures, such as blood vessels and nerves.This real-time feedback is invaluable, not only in guiding the ongoing procedure, but also in planning subsequent steps, making specimen imaging an indispensable tool for modern surgery.Its applications underscore a commitment to enhancing patient outcomes, minimising post-surgical complications, and fostering a more integrated approach to surgical care (Havariyoun et al 2019, Hao et al 2021, Kulkarni et al 2021).
Within the surgical suite, the utilisation of PET radiotracers introduces a transformative dimension to specimen imaging.These radiotracers can be designed to target specific cellular processes, providing surgeons with a functional visualisation of tissue metabolic activity.This is especially critical in oncological surgeries, where differentiating between malignant and benign tissues can be challenging.When a tumour metabolises a PET radiotracer, it 'lights up' on the imaging display, offering clear demarcation of tumour margins and guiding surgical excisions.Beyond tumour identification, PET radiotracers can be instrumental in identifying areas of inflammation, infection, or other metabolic activities.The real-time, functional insights provided by PET radiotracers in the operating room have the potential to enhance surgical precision, reduce the need for re-operations, and improve overall patient outcomes.
In modern surgical suites, the use of PET radiotracers for specimen imaging has seen an innovative convergence of technology.Instruments, such as high-energy gamma probes, beta particle probes, and both clinical/conventional and high-resolution small-animal PET imagers, as well as autoradiographs, have become invaluable tools in the surgeon's arsenal.In post-excision of suspicious tumours, lesions, or lymph nodes, these tools are employed to ascertain the safety margin, ensuring the comprehensive resection of malignant tissues.When relying on clinical or conventional PET scanners, the standard procedure often involves transferring the specimen to a dedicated imaging room to undergo PET/CT imaging.A similar protocol is typically adhered to when utilising small-animal PET scanners for specimen imaging with PET radiotracers.
In this account, Povoski et al (2008) detailed a pioneering approach that integrates perioperative 18 F-FDG PET/CT imaging with intraoperative 18 F-FDG handheld gamma probe detection and intraoperative ultrasound.This multifaceted strategy was employed for the precise localisation of tumours and to confirm the complete removal of all hypermetabolic tumour sites in a case involving concealed recurrent metastatic melanoma.A standout aspect of this research is its application of clinical PET/CT imaging specifically for specimen analysis.This method serves to confirm the precision of surgical procedures, ensuring that the excision is thorough and accurate.This study delves into a unique case of concealed recurrent metastatic melanoma, presenting in three distinct areas within the subcutaneous tissues of the patient's left thigh.On the surgical day, the patient was given an intravenous injection of 437.6 MBq of 18 F-FDG.The trio of excised specimens were subsequently moved to the nuclear medicine department for imaging using a clinical PET/CT scanner (Siemens Biograph 16 PET/CT unit; Knoxville, TN, USA), approximately 210 min post-injection.The specimen PET/CT imaging highlighted three hypermetabolic areas, aligning with the three removed sections of subcutaneous tissue.These zones matched the three hypermetabolic sites in the subcutaneous tissues of the left thigh previously identified in the pre-surgical diagnostic whole-body 18 F-FDG PET/CT scan.The specimen PET/CT played an essential role in affirming that each of the three surgically removed tissue samples encompassed the respective hypermetabolic sites originally identified in the diagnostic whole body 18 F-FDG PET/CT scan.The Siemens Biograph 16 PET/CT bears an axial FOV of 162 mm, a transaxial FOV of 585 mm, and axial and transaxial spatial resolution of 4.6 mm @ 1 cm according to the NEMA (National Electrical Manufacturers Association) standards (Matheoud et al 2011).
To pinpoint hypermetabolic tumours during surgery, determine the extent of the disease, and ensure total removal, Cohn et al (2008) used 18 F-FDG for both perioperative PET/CT scans and real-time gamma probe detection during ovarian cancer surgeries.Three patients showing signs of recurrent ovarian cancer in their lymph nodes were examined.During the surgery, a gamma probe was used to spot tumours.Once removed, the tumours underwent a PET/CT scan to verify metabolic activity.The gamma probe (Neoprobe neo2000 unit, Neoprobe Corporation, Dublin, Ohio) used during the operation and subsequent PET/CT scans post-surgery confirmed the total removal of both apparent and hypermetabolic tumours.The resected specimens underwent a ten-minute PET/CT scan, as outlined in (Hall et al 2007).These images were subsequently analysed and examined for the existence of hypermetabolic foci.Before the abdominal incision, every patient received a preoperative 18 F-FDG injection in the holding area.The dosage was determined by the typical clinical amount used for 18 F-FDG PET/CT imaging (370-740 MBq).This dosage supports both gamma probe detection and postoperative scans for a duration of up to 10 h post-injection.The specimen PET/CT imaging was conducted on a Biograph 16 PET/CT (Siemens Healthineers, Knoxville, TN) scanner.Reference is made to figure 9 for a digital photograph of the resected left breast tissue section, its corresponding fused PET/CT images, and the postoperative PET maximum intensity projection confirming complete resection of the primary breast tumour and axillary metastasis.
In recent studies, clinical PET/CT scanners have been utilised to assess post-surgical specimens, primarily to verify the complete removal of targeted malignant tissue and ensure a safe margin of extraction from adjacent tissues.The imperative nature of intra-or post-operative specimen evaluations necessitates both high sensitivity and precise spatial resolution to thoroughly investigate the excised malignant lesions.However, conventional PET/CT scanners might not offer the requisite granularity for such critical assessments.Given these limitations, there is an emergent need to consider dedicated PET/CT scanners specifically designed for specimen imaging to ensure both accuracy and comprehensive evaluation of the resected tissue.

Autoradiograph
Autoradiograph systems stand distinct from traditional PET imaging techniques in terms of their underlying operational principles.While PET imaging primarily focuses on detecting the annihilation photons resulting from the interaction of emitted positrons with surrounding electrons, autoradiograph systems sidestep this mechanism.Instead, they directly capture the radiation emitted from the radioitracers within a specimen, creating a visualisation of the radiotracer distribution.Despite these fundamental differences, autoradiograph systems have found useful applications in intraoperative settings.Surgeons utilise them for specimen imaging during procedures, offering real-time insights into the exact location and spread of radiotracers (Macaisa et al 2019, Sasaki et al 2019).
Autoradiography is a powerful imaging technique that utilises radioactive isotopes to visualise the distribution and concentration of radioactively labelled compounds, such as those in biological specimens.The basic principle of an autoradiograph system is the detection of radiation emitted from the radionuclides within the specimen, which subsequently exposes a sensitive film or detector, producing an image of the distribution of the radiotracer.For PET radiotracers, the emitted radiation can be captured by a sensitive film placed in close proximity to the tissue or specimen, creating an image.The intensity and pattern of the radiation captured on the film represent the spatial distribution and concentration of the PET radiotracer within the specimen.This information is invaluable for understanding the biodistribution, metabolism, and specificity of PET radiotracers, especially during their development and validation stages (Strong et al 2009, Aguero et al 2019, Li et al 2020).Strong et al (2008) employed handheld PET probes to enable immediate evaluation of the disease and pinpointing tumour locations during procedures.They exploited 124 I-labeled humanised monoclonal antibodies, tailored for colorectal cancer (huA33) and renal tumours (cG250).Patients underwent PET scans roughly a week after the introduction of the tracer, during periods when tumour-to-non-tumour distinctions were pronounced.Handheld beta and gamma PET probes were utilised intraoperatively to examine suspected cancer deposits.They showed that beta probes might present superior precision in real-time pinpointing of tiny cancer deposits, in comparison to its gamma counterpart.Tumour sections from the liver and lymph node, along with a portion of the adjacent normal liver tissue that was removed with the specimen, were collected for both autoradiography and immunohistochemistry to further evaluate tumour targeting.Analysis was conducted through autoradiography, hematoxylin and eosin staining, and using PET probes to measure high-energy gamma and beta counts per second.The autoradiographic analysis revealed the most prominent uptake of the radiolabelled antibody in the liver metastasis with the lymph node showing a marginally lesser uptake.A comprehensive analysis using both autoradiography and immunohistochemistry validated the results observed in preoperative PET imaging and the intraoperative findings from the gamma and beta probes.Given the superior spatial resolution of autoradiography, its outcomes are often considered the benchmark for assessing the findings of gamma and beta probes on specimens.Jurrius et al (2021) reported that among women opting for breast-conserving surgery (BCS), around 20%-25% need a secondary operation due to the partial removal of the tumour.They proposed a flexible autoradiography (FAR) imaging that involved the application of a slim, flexible scintillating film on the specimens.A multi-centric study with a single-arm approach was undertaken to assess the potential of using intraoperative 18 F FDG FAR for tumour margin evaluation in BCS.The LightPath® Imaging System from Lightpoint Medical Ltd (Chesham, UK), a diagnostic tool designed for in vitro use, was employed to map out the position and dispersion of 18 F-FDG through FAR (Ciarrocchi et al 2018).An accessory to the LightPath® System was a 12 µm thick flexible scintillating film.This film had a layered construction: 3 µm of mylar, followed by 6 µm of P43 scintillating phosphor, and another 3 µm of mylar.The image capture was conducted with a 300 s acquisition duration and utilised 8 × 8 pixel binning, resulting in a pixel resolution of 938 µm.A group of eighty-eight BCS patients with invasive breast cancer were administered ⩽ 300 MBq of 18 F FDG between 60 and 180 min before surgery and post-surgical removals were assessed by the LightPath® Imaging System.Out of 385 margins evaluated using FAR during the operation, the results exhibited a sensitivity of 46.2%, specificity of 81.7%, and an overall accuracy rate of 80.5%.They demonstrated that the technique of 18 F-FDG FAR is effective and safe for intraoperative evaluation of tumour margins.
Preparing specimens for autoradiography involves a meticulous process to ensure accurate and clear imaging results.Typically, the sample or specimen is embedded in a suitable medium, often paraffin or resin, and then thinly sectioned using a microtome.These sections are placed on a photographic film or emulsion in a dark environment.Over time, the radioactive decay from the sample exposes the film, creating an image.Once a sufficient exposure time has elapsed, ranging from hours to weeks depending on the radioisotope used and the specific requirements of the study, the film is developed to reveal the distribution of the radioactivity in the specimen.However, despite the unique insights autoradiography offers, it does come with some disadvantages.The technique demands a lengthy exposure time, which might not be ideal for rapidly decaying isotopes or time-sensitive research (Sasaki et al 2019, Naydenov et al 2023).

Small-animal PET scanners for specimen imaging
In the realm of intraoperative specimen imaging, leveraging PET radiotracers becomes paramount to achieve a meticulous assessment of the entire tumour resection process.Ensuring that not only has the tumour been fully excised, but also that an adequate safety margin from the surrounding healthy tissue has been maintained, necessitates a combination of high spatial resolution and substantial sensitivity.In this context, small-animal PET scanners emerged as a particularly promising tool.These devices offer an advantage over traditional clinical PET scanners by delivering notably higher spatial resolution, allowing for detailed imaging of fine tissue structures.Additionally, their heightened sensitivity is attributed to their smaller FOV coupled with a more comprehensive geometrical coverage.This means that these scanners can detect and analyse minute traces of radiotracers more effectively, making them an ideal candidate for ensuring precision in tumour removal and safety margin evaluations (Amirrashedi et al 2020) (figure 10).
In this regard, Debacker et al conducted a preliminary study dealing with evaluating 3D intraoperative margins in head and neck cancers using a high-resolution small-animal PET/CT scanner (2021).Successful surgical intervention for head and neck cancers hinges on the thorough excision of malignant tissues.In cases where surgical margins are insufficient, supplementary treatments become necessary.Notably, postoperative examinations often reveal the deep margins as the main site of positive margins.They explored the potential of high-definition preclinical PET and CT to furnish a clearer understanding of surgical margins in a 3D framework.Before undergoing surgical removal, patients diagnosed with head and neck cancer were given a clinical dose of 4 MBq kg −1 18 F-FDG roughly an hour before the procedure began.After the resection, images of high-definition PET and CT were captured using the ƒ-CUBE and X-CUBE from MOLECUBES (NV, Belgium).An initial 30 min PET scan was conducted, which was then followed by a 5 min high-resolution CT scan protocol.The excised specimen was further evaluated using a subsequent  histopathological analysis (figure 11).The study comprised eight participants, with intraoperative PET/CT imaging carried out on 11 tumour samples and lymph nodes from three patients.The enhanced resolution made it challenging to distinguish between inflamed or dysplastic tissue and cancerous tissue, especially in cases displaying heightened inflammation surrounding the tumour.The adopted method facilitated a 3D Figure 12.Schematic of the PET/CT specimen imager.Displayed on the right is the device, highlighting its primary components: the medical display (marked by a red triangle), the container for the specimen (circled in red), the viewer interface (outlined by a red square), and the dual-modality acquisition hardware for both CT and PET (indicated by a red star).On the left, the imager's viewer showcases images across three orientations: coronal, sagittal, and transversal.The dedicated viewer interface (highlighted in red square) permits users to navigate through each plane, offering a 3D visualisation of the specimen.A 3D rendered image, which users can rotate, is also provided.Reproduced from Muraglia et al (2023).CC BY 4.0.mapping of 18 F-FDG through ultra-fine PET/CT scans.Even though the technique necessitates further refinement and tailored patient selection, its clinical integration had the potential to revolutionise deep margin evaluations in head and neck surgical samples.
The β-CUBE PET scanner relies on a detector design consisting of a 25.4 mm × 25.4 mm × 8 mm 3 thick monolithic LYSO scintillator block, connected to an array of 3.2 mm × 3.2 mm Hamamatsu silicon photomultipliers (MPPC).This compact design not only offers precise light response function measurements but also enables depth of interaction (DOI) measurement.Its enhanced spatial resolution performance is further amplified by a ML clustering algorithm, which determines each interaction's 3D location within the PET detectors.Calibration utilises a precision 3D robot-stage and a collimated 511 keV source, facilitating a 5-layer DOI measurement.The scanner involves 45 PET detectors configured in five rings for an optimum diameter (7.2 cm) and axial length (13 cm), ideal for whole-body imaging of rats and mice (Krishnamoorthy et al 2018).The X-CUBE, a dedicated micro-CT scanner, delivers helical whole-body rat scanning with approximately 80 µm resolution, with a transverse of 6.5 cm and axial FOV of 3.5 cm.A spatial resolution of 1 mm, determined using the filtered backprojection reconstruction technique, is observed at the scanner's centre.Moreover, a peak sensitivity of 12.4% is documented within a 255-765 keV energy window.The scanner's count-rate performance is commendable, showing a peak noise-equivalent-count-rate of 300 kcps and 160 kcps for the NEMA mouse and rat phantoms, respectively, measured at an activity of approximately 33.3 MBq in both NEMA phantoms.

Dedicated PET scanner for specimen imaging
Historically, the methodologies employed for specimen imaging during surgical procedures have predominantly hinged on conventional clinical PET scanners, small-animal scanners, PET gamma probes, or autoradiographs.Notably, these later tools (gamma probes and autoradiographs) do not operate based on the core principles of PET imaging; rather, they identify solely a singular photon resulting from annihilation.Until recently, there had not been a PET imager specifically tailored for specimen imaging.Addressing this gap, Muraglia et al (2023) have pioneered the introduction of the AURA10® PET/CT imager, manufactured by Xeos Medical NV (Ghent, Belgium).This innovative equipment is purposefully crafted for intraoperative specimen PET imaging, signifying a pivotal shift in the approach to surgical specimen evaluation (figure 12).
The Aura 10 PET/CT imager is distinguished by its innovative design and state-of-the-art features.Central to its functionality is a unique specimen container, offering enhanced feedback capabilities and tailored to dimensions of 10 cm in diameter and 4 cm in height.A testament to its advanced engineering is the motorised movement across both the PET and micro-CT imagers, streamlining the scanning process into a seamless, fully automated operation.This efficient setup ensures that imaging results are swiftly available, Included in this figure are macroscopic images of the surgical specimens (B), (F) and (J), high-resolution PET/CT scans of these specimens (C, G, and K), and a histopathological evaluation, where an immunohistochemical analysis using PSMA staining was executed (D and H).Reproduced from Muraglia et al (2023).CC BY 4.0.
typically within a 10 min timeframe.The imager's compactness is attributed to a unique two-stage configuration and pioneering PET/CT detector technology.This not only facilitates a reduced footprint but also delivers an impressive submillimetre spatial resolution for the PET, perfectly complemented by a high-definition CT.For clinicians, a particularly valuable feature is the ability to overlay colour-mapped PET imaging onto the CT scan, providing precise anatomical references for regions exhibiting heightened radiotracer uptake.Furthermore, the system's 3D viewing capability plays a pivotal role in pinpointing malignant cells within the specimen with remarkable accuracy.Muraglia et al (2023) explored the practicality of using the Aura 10 specimen PET/CT imager intra-operatively in a clinical environment.They conducted an initial study involving three patients.Two were administered 68 Ga-PSMA-11 and one with 68 Ga-DOTA-TOC (DOTA-Tyr3-Octreotide.).These patients were given PET radiopharmaceuticals to aid in radioguided surgery using a beta-probe detector.The surgeries involved were radical prostatectomy (RP) for prostate cancer (PCa) and salvage lymphadenectomy for a recurring neuroendocrine tumour (NET) of the ileum.They used immunohistochemistry (PSMA-staining and SSA immunoreactivity) as our reference standard and compared specimen images with baseline PET/CT images and histopathological evaluations.Patients were given either 1 MBq Kg −1 of 68 Ga-PSMA-11 (for PCa) or 1.2 MBq Kg −1 of 68 Ga-DOTA-TOC (for NET) before their surgeries.Specimens were collected and placed in the specialised container for high-resolution PET/CT scanning (figure 13).PET's spatial resolution for the specimen images was notably superior compared to the baseline whole-body PET/CT scans.Importantly, introducing the PET/CT specimen imager into the process did not disrupt the procedures, nor did it extend the duration of the surgeries.The data derived from the specimen PET/CT imaging aligned perfectly with histopathological findings.
In a dual-centre feasibility study (Darr et al 2023), the diagnostic potential of intraoperative ex vivo specimen PET/CT imaging in RP and lymphadenectomy specimens was assessed for ten patients with high-risk PCa.Preoperative PSMA PET/CT was performed, and resected specimens were re-evaluated using AURA10® PET/CT device.The study demonstrated excellent visualisation of index lesions and lymph node metastases, with a high correlation to conventional PET/CT.Importantly, positive or close surgical margins were accurately identified, aligning with histopathology.The findings suggest that specimen PET/CT is valuable for detecting PSMA-avid lesions and merits further investigation to customise RP based on robust correlations with final pathology.Future trials will explore its potential in comparison with frozen section analysis for positive surgical margin detection and assessment of biochemical recurrence-free survival.Overall, specimen PET imaging is deemed feasible and holds promise for enhancing oncological outcomes in PCa patients.

PET imaging of tissue/OOC
The OOC represents a groundbreaking advancement in biochip technology.While a myriad of OOC systems have emerged over the last ten years, playing pivotal roles in areas like drug testing and tailored medicine, their intricate structure-both of the chip itself and the tissue within-has posed significant challenges in terms of imaging and analysis for those in the biomedical field.
A staggering 90% of drugs fail to clear clinical trials despite their success in cell and animal studies.This discrepancy largely arises from the innate differences between animal and human species.Animals often fail to precisely emulate and mirror the disease conditions, their evolution, and the subsequent treatment responses seen in humans (Golebiewska et al 2020).Coupled with this, the constrained throughput of in vivo animal tests elongates the drug development process and escalates associated costs.
The OOC is an amalgamation of disciplines: from cell biology and biomedical engineering to microfabrication and biomaterials.This technology adeptly recreates the biomedical and physical nuances of human organs on microfluidic platforms (Wu et al 2020).Given the compactness of each OOC unit, they enable high-throughput drug screenings, enhancing the efficacy of such processes.As such, OOC stands as a promising alternative to potentially bridge the gaps left by animal testing and might eventually reduce our reliance on the same.
As the applications for OOCs expand, there's a mounting need for advanced measurement techniques.It is essential to monitor metabolic activities and other physiological and pathophysiological processes in OOCs.PET stands out as an optimal choice for OOC imaging due to its capability to glean in vivo insights into metabolism and molecular pathways.Yet, conventional imaging devices employed in clinical trials fall short when applied in OOC imaging.This is largely attributed to their restricted spatial resolution.
In relation to this, Clement et al introduced a conceptual (simulation study) on-chip PET system tailored for functional imaging of OOCs (2022).The development of a scanner comprised of four detectors, each constructed from two seamlessly combined monolithic crystals.They employed Monte Carlo Simulations to optimise the system's design and generate datasets of light pattern images formed on the detector surfaces due to scintillation.Utilising these datasets, they trained and assessed convolutional neural networks to specify the initial interaction locations of gamma rays within the detector.The proposed system can significantly enhance two primary pre-clinical applications of OOCs: disease modelling and precision medicine.Assessing cellular pharmacokinetics provides insights into human diseases by simulating biochemical and genetic interactions.Furthermore, analysing patient-derived organoids aids in identifying the most effective drug tailored to individual patients.
For simulation purposes, they generated a rectangular detector volume, that was repeated four times in a ring pattern around the z-axis.Within each detector volume, a rectangular crystal volume was incorporated and duplicated.Two distinct monolithic crystals (LYSO) having thicknesses of 13 mm and 26 mm were examined.The dimensions of these crystals were selected to ensure arrays of commercially available SiPMs align perfectly on their surfaces (figure 14).
The proposed system demonstrated a sensitivity of 34.81% with 13 mm thick crystals, and maintained consistent prediction accuracy, even near the detector boundaries.They reconstructed an image from a grid consisting of 21-point sources distributed throughout the FOV, achieving a mean spatial resolution of 0.55 mm.

Intraoperative PET scanners
During surgical interventions, high-energy gamma probes and gamma cameras, specifically tailored to work with PET radiotracers, have been increasingly adopted to assist in tumour detection and the removal of metastasised lymph nodes.It is imperative to note that these instruments operate by detecting single photons.This mechanism stands in contrast to the foundational principle of PET imaging, which is predicated on detecting annihilation photons.Given this fundamental difference, the scope of this review did not encompass the application of gamma probes and gamma cameras, even though they are extensively employed in intraoperative contexts.Instead, our focus has primarily revolved around PET imaging systems that have been purposefully designed to provide guidance during surgical procedures.
The design of an intraoperative PET imager, specifically intended for surgical guidance, hinges primarily on its ability to offer real-time imaging of radiotracer distribution.This real-time feedback is crucial, allowing surgeons to make informed decisions during the procedure.Consequently, every component of the imaging process, from data acquisition to preprocessing, from image reconstruction to post-processing, must be optimised for speed.The overarching objective is to ensure that the system can provide instantaneous data that is indispensable for the surgical process.In essence, the seamlessness and rapidity of this imaging chain (including image reconstruction) are paramount to its efficacy in the surgical arena.
An additional consideration in the design of intraoperative PET imagers is ensuring adequate access for the surgeon to the patient.As such, designs with partial coverage or limited-angle PET are often favoured.These configurations offer a balance between imaging capability and practicality, ensuring that the surgeon has both open access and ample space to engage with the patient.Equally crucial is the form factor of these PET scanners.They should not be cumbersome, taking up excessive space in the operating room.Moreover, their design should prioritise mobility and mechanical flexibility.This ensures that the device can be easily adjusted or moved as needed, preventing any disruption or interference with the conventional surgical procedures, allowing the surgical team to work in tandem with the imaging equipment seamlessly.
In the realm of intraoperative PET scanners, minimising radiation exposure to the surgical staff is paramount.A study was conducted by Costa et al to quantitatively characterise the occupational radiation exposure during surgery in the context of Cerenkov luminescence imaging (CLI)-guided robot-assisted RP (2022).A single-injection PET/CT CLI protocol involving the administration of 141.9 ± 57.86 MBq (average ± SD) of 68 Ga-PSMA-11, with preoperative PET/CT imaging, was used.Electronic personal dosimetry (EPD) was employed to assess intraoperative occupational exposure, involving ten participants.Measurements were taken for the first surgical assistant and scrub nurse positioned at the operating table, as well as for the CLI imager/surgeon at the robotic console.The assessments covered the whole duration of the surgery, including CLI image acquisition.EPD readings indicated average personal equivalent doses of 9.0 ± 7.1 µSv, 3.3 ± 3.9 µSv, and 0.7 ± 0.7 µSv for the first surgical assistant, scrub nurse, and CLI imager/surgeon, respectively.
Given that the patient typically receives a radiation dose much smaller than what's administered during conventional diagnostic PET imaging, these scanners must possess high detection sensitivity.Moreover, due to their partial coverage or limited-angle designs, they inherently possess a reduced geometrical sensitivity.This poses a substantial challenge: ensuring that despite these design constraints, the scanner still captures sufficient signals during surgery to reconstruct a reliable PET image.The balancing act between maintaining minimal exposure and achieving precise imaging makes the technological and procedural optimisation of these scanners a formidable task, underscoring the need for innovative techniques to capture dependable data.High-sensitivity PET detectors and the integration of virtual pinhole PET technology (which could be placed in contract with the patient) can potentially address challenges in intraoperative PET scanners, enhancing image quality despite limited signal availability (Tai 2024).
In the context of intraoperative PET imagers, there's a pronounced tilt towards prioritising detection sensitivity over spatial resolution.This prioritisation emerges from several considerations inherent to the operational confines of these imagers.Primarily, the real-time or online nature of PET image formation during surgeries necessitates fast acquisition times.This immediacy, when combined with the challenges of partial detector coverage and a typically reduced activity of radiotracer injections, amplifies the need for heightened detection sensitivity.Further compounding these complexities, different applications of intraoperative PET scanners might contend with challenges like photon attenuation and scatter, both of which have the potential to adversely affect PET image quality.The absence of concurrent anatomical imaging during surgeries aggravates these challenges, making the task of generating accurate and clear images even more daunting.Thus, when weighing sensitivity against spatial resolution, it becomes clear that the former holds the key to the successful implementation of intraoperative PET imagers.
In the realm of intraoperative PET scanners, the integration of TOF PET imaging emerges as a beneficial proposition.Given the typically large FOV in these scanners, TOF technology can be leveraged to its full potential.The unique strength of TOF lies in its ability to significantly aid the reconstruction process of PET imaging, which is inherently challenging for intraoperative PET scanners.Factors, such as partial acquisition coverage (which would lead to image artefacts), reduced sensitivity, rapid acquisition demands, and the presence of attenuated and scattered photons complicate this reconstruction process.However, the advent of artificial intelligence offers a ray of hope in this context.AI-powered algorithms, tailored for fast image reconstruction, quality enhancement, and corrections for both attenuation and scatter and various artefacts, hold significant promise (Haggstrom et al 2019, Arabi and Zaidi 2020, 2021, Mostafapour et al 2021).In effect, they could potentially revolutionise the way image reconstruction challenges are addressed in intraoperative PET scanners, offering clearer images in real-time, a critical requirement for surgical interventions.

PET specimen imaging
While high-energy gamma probes and gamma cameras have gained traction in specimen imaging and are vital tools in confirming the complete removal of malignant tissue during surgeries, they operate on different principles than PET imaging.Due to this distinct operational mechanism, they fall outside the scope of our current discussion and will not be explored in detail in this review.It is crucial to differentiate between the technologies and understand the unique benefits and limitations that each brings to the field of medical imaging and surgical intervention.
PET scanners for specimens are purpose-built for specific applications, which means they do not require a large FOV.In fact, a FOV akin to that of small-animal PET scanners is aptly sufficient for the task.This differs from intraoperative PET scanners which often need broader viewing fields.Another distinction lies in the imaging speed.Whereas intraoperative PET scanners aim for real-time imaging to guide surgical procedures, specimen PET imagers have a more lenient time frame.Verifying excised tissue can be accomplished within a few minutes, eliminating the need for instantaneous image production.Furthermore, given that specimen PET scanners can capture the entirety of the specimen, they are not hampered by challenges that arise from partial coverage.This means there's less emphasis on rapid image reconstruction and fewer concerns regarding potential degradation in image quality.
For the specific task of specimen imaging, many of the existing small-animal PET scanners are perfectly suited.These devices can offer FOV extending up to 10 cm, coupled with sub-millimetre spatial resolution-ideal specifications for PET specimen imaging (Miyaoka and Lehnert 2020).Notably, the imaging process does not necessarily need to be executed in the immediate vicinity of the surgical procedure.PET specimen imaging can indeed be conducted off-site, provided logistical challenges related to sample transportation are not prohibitive.Moreover, it is essential for the design and footprint of PET specimen scanners to remain relatively compact.This ensures that they cannot only fit seamlessly within surgical settings but also maintain a level of portability, allowing for effortless relocation as and when needed.
Specimen PET imagers have a distinct requirement to deliver a high spatial resolution, especially when juxtaposed against intraoperative PET scanners.This relatively high spatial resolution is pivotal for the precise delineation of malignant tissues.In the design hierarchy for these scanners, spatial resolution often takes precedence over detection sensitivity.This is because a mechanical design that encapsulates the sample fully can be adopted, resulting in enhanced geometrical sensitivity.While it is noteworthy that the excised tissue might contain only minimal radiotracer activity, these scanners' compact FOV, combined with the proximity of samples to the detectors, ensures that a favourable SNR is maintained.Consequently, even with relatively low tracer activity, these imagers can still produce clear and informative scans.
It is important to highlight that due to the diminutive size of the sample, photon attenuation and scatter are unlikely to present significant challenges.Conversely, the small sample size renders TOF imaging inefficient for specimen PET scanners.As a result, dedicated specimen PET detectors can prioritise efficiency over timing precision.Detectors with poorer timing resolution, such as those based on BGO, would be aptly suited for this application.
Given the significance of both spatial resolution and high geometrical sensitivity for specimen PET scanners, the principle of virtual pinhole PET imaging could be effectively integrated into their design.A notable example that can be adapted for specimen PET imaging is the small-animal PET scanner design proposed by Perez-Benito et al which emphasises full sample coverage (2019).Traditional PET scanners usually feature detectors arranged in rings, with the specimen positioned axially.However, this configuration does not optimise geometrical sensitivity unless the axial extent significantly exceeds the axial FOV-an approach that would entail substantial cost.The most favourable geometry for a PET scanner is spherical, with the specimen centrally located.Approximating this ideal configuration, the proposed scanner is modelled as an icosahedron, boasting a FOV close to 30 cm.
Furthermore, if an exceptionally high spatial resolution is needed for the specimen, one might contemplate the incorporation of pinhole PET imaging using an actual pinhole collimator.However, it is crucial to recognise that such a strategy would considerably compromise the scanner's sensitivity.As proposed by Van Der Have et al, this design involves positioning an actual pinhole collimator within a stationary triangular detector, and then conducting PET radiotracer imaging in a non-coincidence mode (2009).Due to the magnification factor introduced by the pinhole collimator, the spatial resolution of the resulting PET images becomes notably superior, albeit at a significant sensitivity trade-off.An added advantage of this design is its ability to simultaneously conduct SPECT and PET imaging, achieving an impressive spatial resolution of 400 micrometres.
Enhancing PET performance with DOI capability holds significant promise for improving PET's spatial resolution (Carson et al 2021, Li et al 2023, Zeng et al 2023).By utilising thicker or multi-layer crystals, the detection efficiency of PET imagers can be augmented without compromising the spatial resolution.However, the integration of these techniques into intraoperative PET scanners has been limited, primarily due to considerations regarding cost and the compact size requirements of such scanners, predominantly crucial for PET probes used in endoscopy and laparoscopy.Conversely, the application of DOI PET imaging in specimen and OOC scans presents an opportunity to capitalise on improved spatial resolution and detection efficiency, thereby offering a compelling solution without sacrificing system spatial resolution.
On-chip tissue or organ imaging bears a strong resemblance to specimen imaging.Although it is not utilised within surgical settings, we have chosen to address this concept in the review due to its striking similarity to specimen imaging.The specifications of a dedicated on-chip tissue or organ PET scanner could align closely with those of a specimen PET imager.Such a scanner would prioritise both high spatial resolution and extensive geometric coverage, thereby optimising the sensitivity of PET acquisition.
The summary of the dedicated PET imagers discussed in this work is provided in table 1.The development of intraoperative PET scanners poses several technical challenges, primarily centred around the imperative for compact size to ensure practicality within operating rooms.The compact design, particularly critical for PET probes used in endoscopy and laparoscopy, presents a significant hurdle in their advancement.Furthermore, the limited angle acquisition characteristic of these scanners complicates image reconstruction, as the lack of certain projections can detrimentally impact PET image quality.This limitation not only compromises overall detection efficiency but also challenges image reconstruction in terms of SNR and susceptibility to noise and artefacts.Additionally, the absence of concurrent anatomical imaging in most intraoperative PET scanners complicates the implementation of attenuation and scatter corrections, further exacerbating challenges in image fidelity and quantitative imaging.
Moreover, intraoperative PET scanners relying on tracking systems encounter difficulties in precisely localising PET probes, thereby complicating data acquisition, processing, and image reconstruction procedures.The efficacy of these scanners is heavily contingent upon the accuracy of tracking systems, amplifying the intricacies involved in achieving optimal imaging outcomes.In contrast, specimen and OOC PET scanners aim to deliver high spatial resolution with reasonable detection efficiency.Benefiting from similar design principles and image reconstruction techniques employed in small-animal PET imagers, these scanners face fewer specific challenges, facilitating smoother integration and operation within experimental settings.

Conclusion
In this review, we devoted our attention to intraoperative PET scanners and PET imaging techniques for excised tissues.Given the unique challenges posed by intraoperative PET imaging-such as the need for ample access during surgery, the demand for real-time imaging to support navigation systems, rapid processing and reconstruction times, as well as geometric constraints-our exploration primarily centred around striking a balance between image quality, sensitivity, and imaging coverage.We delved into the latest designs and technological advancements in these specific PET imaging systems.Additionally, we addressed high-resolution specimen PET imaging approaches, underscoring their crucial role in ensuring the complete excision of malignant tissues and verifying safe margins alongside healthy tissues.
Concerning the low activity concentration within patients' bodies during surgery, the prospective design of intraoperative PET systems would prioritise efficient geometric coverage.This entails achieving a balance between high geometrical efficiency and ensuring proper accessibility to patients.Additionally, detectors with high detection efficiency would be essential to guarantee ample count statistics, thereby facilitating reliable image reconstruction.On the other hand, in the context of specimen and tissue-on-chip PET imagers, where there are no limitations on geometric coverage, the emphasis would lean towards high-resolution PET imaging.This approach would be complemented by synergistic anatomical imaging to facilitate a comprehensive assessment of samples.

Figure 2 .
Figure 2. 3D drawing of the EndoTOFPET-US external plate, showing the crystal matrices arranged in a pointing geometry, the aluminium casing, and the attachment for a robotic arm.Sketch of the in vivo configuration for pancreatic clinical case.On the left is the endoscope PET extension for pancreatic clinical case, positioned under the bend of the duodenum opposite to the external plate, enclosing the pancreas in the field-of-view.

Figure 3 .
Figure 3. Schematic sketch of the PET-laparoscope system together with the tracking system.

Figure 4 .
Figure 4. Top-down view of the experimental arrangement for the surgical PET imaging probe.

Figure 6 .
Figure 6.Schematic illustration of the concept of pane1 PET with window.© 2017 IEEE.Adapted, with permission, from Li et al (2017).

Figure 7 .
Figure 7. Illustration of a surgical brain PET with various modules adaptably positioned around the head.

Figure 8 .
Figure 8. Conceptual depictions of PET structures designed with a patient-friendly open space.(A) dual-ring OpenPET providing an open space for medical treatments between the two PET detector rings.(B) A traditional cylindrical PET can similarly offer open space by angling it slantwise relative to the patient's bed (single-ring OpenPET).

Figure 9 .
Figure 9. (A) Digital image showing a cross-section of the excised left breast tissue, excluding the removed left axillary tissue.(B) Cross-sectional fusion of PET/CT images from the excised left breast and axillary tissue, highlighting two areas of increased metabolic activity.(C) Post-surgical PET maximum intensity projection in an anterior chest view, confirming the total removal of the primary breast tumour and the lone axillary metastasis without any remaining hypermetabolic regions.Reproduced from Hall et al (2007).CC BY 2.0.

Figure 10 .
Figure 10.Diagrammatic representation of the implemented methodology.The upper row provides a concise breakdown of the methodology employed during the study, while the lower row depicts detailed imaging results as observed in a representative patient.(A) About an hour before the scheduled surgery, the patient receives a dosage of 4 MBq kg −1 of 18 F-FDG.(B) Routine staging involves a pre-treatment, full-body 18 F-FDG PET/CT, capturing the enhanced 18 F-FDG uptake in the left hemithyroid around 60 min post-administration.(C) Standard surgical procedures are adhered to.(D) A photographic representation of the excised left hemithyroid containing the tumorous growth.(E) The extracted specimen undergoes scanning via high-resolution, preclinical PET and CT imaging devices.(F) Detailed PET/CT imagery of the entire specimen shown in (D).(G) Post-scanning, the specimen advances for standard pathological evaluation, facilitating the assessment of surgical margins.(H) A pathology slide featuring both cancerous and healthy tissues is stained with hematoxylin and eosin, exhibiting pronounced purple (hematoxylin) and pink (eosin) stains, respectively.Reproduced from Debacker et al (2021).CC BY 4.0.

Figure 11 .
Figure 11.Analysis of the 'close margin' observed in PET/CT imaging versus pathological evaluation for a scalp's cutaneous squamous cell carcinoma.The region of interest, highlighted with a blue circle, exhibits a positive margin on PET/CT imaging (A), an unclear margin upon macroscopic pathological evaluation (B), and a 'close' margin with a 0.15 mm distance upon detailed microscopic examination (C).(A) Cross-sectional PET/CT image of the extracted specimen.(B) Sectioned tissue corresponding to the region shown in PET/CT imaging (A).C) Microscopic review of the hematoxylin and eosin-stained section obtained from the macroscopic sample in (B).Reproduced from Debacker et al (2021).CC BY 4.0.

Figure 13 .
Figure 13.Visualisation of a patient with prostate cancer: the 68 Ga-PSMA-11 whole-body PET/CT images reveal abnormal PSMA uptake, signifying high expression (E-PSMA Visual Score = 3) within the prostate (A) and three pelvic lymph nodes (E) and (I).Included in this figure are macroscopic images of the surgical specimens (B), (F) and (J), high-resolution PET/CT scans of these specimens (C, G, and K), and a histopathological evaluation, where an immunohistochemical analysis using PSMA staining was executed (D and H).Reproduced fromMuraglia et al (2023).CC BY 4.0.

Figure 14 .
Figure 14.Visualisation of the simulation setup.Presented from both frontal and lateral perspectives.The crystals are represented by the yellow volumes, the surrounding epoxy layer by the green volumes, and the volume from which the source position is sourced for training dataset creation is depicted in blue.Reproduced from Clement et al (2022).CC BY 4.0.

Table 1 .
Summary of dedicated PET scanners discussed in this survey.